Clinical Pharmacology of the Treatment of Malaria in Papua

Clinical Pharmacology of

the Treatment of Malaria

in Papua New Guinea

Harin Ashley Karunajeewa MBBS FRACP

This thesis is presented for the degree of Doctor of

Philosophy of The University of Western Australia, School of

Medicine and Pharmacology, January 2009

ii

iii

This thesis is dedicated to the memory of my father, Hector

iv

v

i. Abstract Malaria is the most important parasitic disease of man. Of the five species known to

infect humans, Plasmodium falciparum causes most deaths and illness, especially when

it affects children and pregnant women living in highly endemic areas of the rural tropics.

Pharmacological therapies for malaria must be optimised for these groups and must be

practical for administration in critically ill patients in remote settings. The clinical studies

in this thesis evaluated the clinical pharmacology of modern antimalarial treatments in a

Melanesian population exposed to highly endemic malaria. The clinical studies were

conducted between March 2001 and June 2007, with final data analysis completed by

mid-2008. They aimed to evaluate key pharmacokinetic, parasitological, host genetic

and socio-cultural determinants of treatment effectiveness in children with uncomplicated

and severe malaria and in pregnant women.

A multi-centre study of children with uncomplicated malaria evaluated the efficacy of four

treatment regimens, including three artemisinin combination treatments. PCR corrected

recrudescence rates by day 42 were 81.5%, 85.4%, 88.0% and 95.2% for chloroquine +

sulphadoxine-pyrimethamine, artesunate + sulphadoxine-pyrimethamine,

dihydroartemisinin-piperaquine (DHA-PQ) and artemether-lumefantrine (AL),

respectively. Determinants of efficacy in the DHA-PQ group included day 7 piperaquine

(PQ) levels and baseline parasitaemia. Therefore, the worse than expected efficacy in

this group may have been partly due to the high parasitaemias commonly seen in this

population. Based on the results of this trial, the PNG Ministry of Health has decided to

change treatment policy, to adopt AL as the new first-line treatment for uncomplicated

malaria in PNG.

A pharmacokinetic study compared the 4-amino-bisquinoline, PQ with the related

conventional antimalarial chloroquine (CQ). Both drugs showed large volumes of

distribution and multi-phasic elimination kinetics. Combined molar concentrations of CQ

and its active metabolite desethylchloroquine were higher than those of PQ during the

elimination phase. Although PQ had a longer elimination half-life than CQ, its prompt

distribution and lack of active metabolite may undermine its post-treatment suppressive

properties and curative efficacy.

vi

Further studies aimed to define the clinical spectrum of severe malaria and to examine

the role of host genetic factors in influencing its clinical manifestations in Melanesian

children. Preliminary data suggested a protective effect of the erythrocyte polymorphism

caused by the glycophorin C mutation against cerebral malaria. These studies also

evaluated key pharmacokinetic, host genetic and socio-cultural determinants of the likely

effectiveness of a novel pharmaceutical approach using artesunate suppositories for

severe malaria. These demonstrated favourable absorption characteristics, clinical

efficacy, safety and patient/community acceptability. Contrary to previous data, no

evidence was found to suggest that the pharmacokinetic profiles or efficacy of

artemisinin derivatives are likely to be compromised by a high prevalence of �

thalassaemia in this population. However, their highly variable bioavailability raises

questions regarding the consistency of therapeutic response. Given the favourable

efficacy and socio-cultural acceptability of rectal artesunate demonstrated in these

studies, the PNG Ministry of Health has decided to add artesunate suppositories to its

national pharmacopoeia and incorporate them into standard treatment

recommendations.

A final study compared the pharmacokinetics of chloroquine, sulphadoxine and

pyrimethamine in pregnant, versus non-pregnant women. This demonstrated

significantly lower concentrations of all three drugs and active metabolites in the

pregnant group, due to a combination of effects on either volume of distribution,

clearance and elimination half-life. It suggests that significant dosage alterations are

necessary to optimise therapy in pregnant women.

vii

ii. Table of contents i. Abstract.....................................................................................................................v ii. Table of contents....................................................................................................vii iii. Acknowledgements ................................................................................................xv iv. Declaration.............................................................................................................xix v. Statement of ethical approval ...............................................................................xxi vi. Publications and presentations ..........................................................................xxiii

a. Publications...................................................................................................................... xxiii b. Presentations.................................................................................................................. xxvii

vii. Abbreviations ......................................................................................................xxix viii. List of Figures..................................................................................................... xxxv ix. List of Tables ..................................................................................................... xxxix

1. GENERAL INTRODUCTION ......................................................................................1

1.1. Malaria: The global situation ..........................................................................................3 1.2. Malaria: Biology and pathogenesis................................................................................4 1.3. Malaria: Epidemiology....................................................................................................5 1.4. Malaria: Role of host and parasite genetics ..................................................................7 1.5. Clinical manifestations of malaria................................................................................10

1.5.1. Uncomplicated malaria...........................................................................................10 1.5.2. Severe malaria ........................................................................................................11 1.5.3. Malaria in pregnancy ..............................................................................................13

1.6. Antimalarial therapy: Drug development and evaluation............................................13 1.6.1. Historical background............................................................................................13 1.6.2. Pharmacokinetic considerations ...........................................................................16 1.6.3. Pharmacodynamic assessment.............................................................................20 1.6.4. Toxicity and tolerability..........................................................................................23 1.6.5. Operational factors.................................................................................................24

1.7. Antimalarial therapy: Drug classes..............................................................................24 4-aminoquinoline and related drugs ...............................................................................25 1.7.1. Chloroquine ............................................................................................................25

1.7.1.1. Pharmacokinetics ........................................................................................................ 25 1.7.1.2. Safety and tolerability ................................................................................................. 25 1.7.1.3. Efficacy ....................................................................................................................... 25

1.7.2. Piperaquine.............................................................................................................26 1.7.2.1. Pharmacokinetics ........................................................................................................ 26 1.7.2.2. Safety and tolerability ................................................................................................. 26 1.7.2.3. Efficacy ....................................................................................................................... 27

Aryl amino-alcohols: ........................................................................................................27 1.7.3. Lumefantrine...........................................................................................................27


Anti-folates: ......................................................................................................................29 1.7.4. Sulphadoxine-pyrimethamine ................................................................................29


The artemisinin derivatives: ............................................................................................30 1.7.5. Artesunate, dihydroartemisinin and artemether ...................................................30


viii

1.8. Antimalarial therapy: Management of specific malaria syndromes ...........................33 1.8.1. Uncomplicated malaria...........................................................................................33 1.8.2. Severe malaria ........................................................................................................37

1.8.2.1. Rectal administration of artemisinin derivatives......................................................... 40 1.8.3. Malaria in pregnancy ..............................................................................................42

1.9. Malaria in Papua New Guinea and Melanesia..............................................................44 1.9.1. Epidemiology and clinical manifestations.............................................................44 1.9.2. Host genetics: the role of erythrocyte polymorphisms ........................................48 1.9.3. Antimalarial treatment policy and parasite drug resistance.................................49

2. MANAGEMENT OF UNCOMPLICATED MALARIA IN CHILDREN..........................53

2.1. Overview of Clinical trials in this chapter ....................................................................55

2.2. Clinical trial 1: Pharmacokinetics and efficacy of piperaquine and chloroquine in Papua New Guinean children with uncomplicated malaria ...............................................57

2.2.1. SUMMARY...............................................................................................................57 2.2.2. AIMS........................................................................................................................58

2.2.2.1. Primary........................................................................................................................ 58 2.2.2.2. Secondary.................................................................................................................... 58

2.2.3. METHODS ...............................................................................................................58 2.2.3.1. Acknowledgement....................................................................................................... 58 2.2.3.2. Rationale for study design........................................................................................... 59 2.2.3.3. Study site..................................................................................................................... 59 2.2.3.4. Patient eligibility and enrolment ................................................................................. 59 2.2.3.5. Drug administration .................................................................................................... 60 2.2.3.6. Monitoring and outcome assessment .......................................................................... 60 2.2.3.7. Blood sampling ........................................................................................................... 61 2.2.3.8. Laboratory methods .................................................................................................... 61

2.2.3.8.1. Microscopy and parasite genotyping .................................................................. 61 2.2.3.8.2. Drug assays......................................................................................................... 62

2.2.3.9. Pharmacokinetic analysis ............................................................................................ 62 2.2.3.10. Statistical analysis ....................................................................................................... 62

2.2.4. RESULTS ................................................................................................................62 2.2.4.1. Patient characteristics.................................................................................................. 62 2.2.4.2. Response to treatment ................................................................................................. 63 2.2.4.3. Drug assays and pharmacokinetic analysis ................................................................. 65

2.2.5. CONCLUSIONS.......................................................................................................70

2.3. Clinical trial 2: An open-label randomised comparison of four combination therapies for treatment of children with malaria in an area of Papua New Guinea with high transmission of multiple Plasmodium species ..........................................................75

2.3.1. SUMMARY...............................................................................................................75 2.3.2. AIMS........................................................................................................................76

2.3.2.1. Primary........................................................................................................................ 76 2.3.2.2. Secondary.................................................................................................................... 76

2.3.3. METHODS ...............................................................................................................76 2.3.3.1. Acknowledgement....................................................................................................... 76 2.3.3.2. Rationale for study design........................................................................................... 77 2.3.3.3. Study sites and patients ............................................................................................... 78 2.3.3.4. Clinical procedures...................................................................................................... 79 2.3.3.5. Laboratory methods .................................................................................................... 81

2.3.3.5.1. Microscopy ......................................................................................................... 81 2.3.3.5.2. Parasite genotyping............................................................................................. 81

ix

2.3.3.5.3. Drug assays......................................................................................................... 81 2.3.3.6. Efficacy endpoints....................................................................................................... 81 2.3.3.7. Statistical analysis ....................................................................................................... 82

2.3.4. RESULTS ................................................................................................................83 2.3.4.1. Trial profile and patient characteristics ....................................................................... 83 2.3.4.2. Efficacy against P. falciparum .................................................................................... 83 2.3.4.3. Predictors of treatment failure in P. falciparum cases ................................................ 96 2.3.4.4. Efficacy against P. vivax ............................................................................................. 96 2.3.4.5. Predictors of treatment failure in P. vivax cases.......................................................... 99 2.3.4.6. Emergence of P. vivax after treatment for P. falciparum ............................................ 99 2.3.4.7. Relationship between Day 7 drug concentrations and outcome.................................. 99 2.3.4.8. Summary of relative efficacy .................................................................................... 102 2.3.4.9. Safety monitoring...................................................................................................... 102

2.3.5. CONCLUSIONS.....................................................................................................105

2.4. Summary of major findings of this chapter ...............................................................113 2.4.1. Pharmacokinetics of piperaquine and chloroquine ............................................113 2.4.2. Methodological issues in assessing treatment efficacy in a poly-species high transmission environment.............................................................................................113 2.4.3. Efficacy of existing standard treatments for malaria in Papua New Guinea......113 2.4.4. Relative efficacy of two ACTs, dihydroartemisinin-piperaquine and artemether-lumefantrine, in Papua New Guinea ..............................................................................114 2.4.5. Poorer than expected efficacy of dihydroartemisinin-piperaquine against P. falciparum in Papua New Guinea ..................................................................................114

3. MANAGEMENT OF SEVERE MALARIA IN CHILDREN ........................................117

3.1. Overview of Clinical trials in this chapter ..................................................................119

3.2. Clinical trial 3: Safety, efficacy and pharmacokinetics of artesunate suppositories 121

3.2.1. SUMMARY.............................................................................................................121 3.2.2. AIMS......................................................................................................................122

3.2.2.1. Primary...................................................................................................................... 122 3.2.2.2. Secondary.................................................................................................................. 122

3.2.3. METHODS .............................................................................................................122 3.2.3.1. Acknowledgement..................................................................................................... 122 3.2.3.2. Rationale for study design......................................................................................... 123 3.2.3.3. Study site and patients............................................................................................... 123 3.2.3.4. Clinical procedures.................................................................................................... 124 3.2.3.5. Laboratory methods .................................................................................................. 125

3.2.3.5.1. Microscopy, haematocrit and glucose............................................................... 125 3.2.3.5.2. Drug assays....................................................................................................... 125 3.2.3.5.3. Alpha-thalassaemia genotyping ........................................................................ 125

3.2.3.6. Data analysis ............................................................................................................. 126 3.2.3.7. Pharmacodynamic outcomes..................................................................................... 126 3.2.3.8. Pharmacokinetics ...................................................................................................... 126

3.2.4. RESULTS ..............................................................................................................126 3.2.4.1. Patient characteristics................................................................................................ 126 3.2.4.2. Clinical course........................................................................................................... 129 3.2.4.3. Side-effects and complications.................................................................................. 129 3.2.4.4. Pharmacokinetic analysis .......................................................................................... 131 3.2.4.5. Alpha-thalassaemia status ......................................................................................... 137

3.2.5. CONCLUSIONS.....................................................................................................139 3.2.5.1. Efficacy ..................................................................................................................... 139

x

3.2.5.2. Safety and tolerability ............................................................................................... 141 3.2.5.3. Pharmacokinetics ...................................................................................................... 141 3.2.5.4. Implications for practice ........................................................................................... 146

3.3. Clinical trial 4 (a). Severe falciparum malaria in Papua New Guinean children: Clinical features and host genetics...................................................................................147

3.3.1. SUMMARY.............................................................................................................147 3.3.2. AIMS......................................................................................................................147 3.3.3. METHODS .............................................................................................................148

3.3.3.1. Acknowledgement..................................................................................................... 148 3.3.3.2. Rationale for study design......................................................................................... 148 3.3.3.3. Study site and patients............................................................................................... 149 3.3.3.4. Clinical procedures.................................................................................................... 150 3.3.3.5. Laboratory methods .................................................................................................. 150

3.3.3.5.1. Biochemistry, haematology and microscopy.................................................... 150 3.3.3.5.2. Genotyping for erythrocyte polymorphism mutations ...................................... 151

3.3.3.6. Statistical analysis ..................................................................................................... 151 3.3.4. RESULTS ..............................................................................................................151 3.3.5. CONCLUSIONS.....................................................................................................155

3.4. Clinical trial 4 (b). Artesunate suppositories versus intramuscular artemether for severe falciparum malaria .................................................................................................159

3.4.1. SUMMARY.............................................................................................................159 3.4.2. AIMS......................................................................................................................159


3.4.3. METHODS .............................................................................................................160 3.4.3.1. Acknowledgement..................................................................................................... 160 3.4.3.2. Rationale for study design......................................................................................... 160 3.4.3.3. Study site and patients............................................................................................... 161 3.4.3.4. Clinical procedures.................................................................................................... 161 3.4.3.5. Laboratory methods .................................................................................................. 162

3.4.3.5.1. Microscopy and other field laboratory tests...................................................... 162 3.4.3.5.2. Drug assays....................................................................................................... 162

3.4.3.6. Efficacy assessment .................................................................................................. 163 3.4.3.7. Statistical analysis ..................................................................................................... 163

3.4.4. RESULTS ..............................................................................................................163 3.4.4.1. Patient characteristics................................................................................................ 163 3.4.4.2. Clinical course........................................................................................................... 164 3.4.4.3. Side-effects and complications.................................................................................. 169 3.4.4.4. Plasma drug concentrations....................................................................................... 169

3.4.5. CONCLUSIONS.....................................................................................................173

3.5. Clinical trial 5. Social and cultural acceptability of artesunate suppositories in Papua New Guinea ........................................................................................................................179

3.5.1. SUMMARY.............................................................................................................179 3.5.2. AIMS......................................................................................................................180


3.5.3. METHODS .............................................................................................................180 3.5.3.1. Acknowledgement..................................................................................................... 180 3.5.3.2. Rationale for study design......................................................................................... 181 3.5.3.3. Study site and participants ........................................................................................ 182

xi

3.5.3.4. Formative research and questionnaire design............................................................ 183 3.5.3.5. Enrolment procedures ............................................................................................... 183 3.5.3.6. Treatment and clinical monitoring ............................................................................ 183 3.5.3.7. Follow-up questionnaire............................................................................................ 184 3.5.3.8. Data management and analysis ................................................................................. 184

3.5.4. RESULTS ..............................................................................................................185 3.5.4.1. Study participants...................................................................................................... 185 3.5.4.2. Prior knowledge and perceptions of rectal treatment ................................................ 185 3.5.4.3. Acceptance or refusal of suppository treatment ........................................................ 189 3.5.4.4. Self-administration of suppositories.......................................................................... 191 3.5.4.5. Clinical response ....................................................................................................... 191 3.5.4.6. Satisfaction following treatment course .................................................................... 192 3.5.4.7. Preferred mode of administration at home or health-centre ...................................... 194

3.5.5. CONCLUSIONS.....................................................................................................194

3.6. Summary of major findings of this chapter ...............................................................199 3.6.1. Safety and tolerability of rectal artesunate..........................................................199 3.6.2. Bioavailability and pharmacokinetic considerations..........................................199 3.6.3. Role of �-thalassaemia in modulating pharmacokinetics and clinical response to rectally administered artesunate ...................................................................................200 3.6.4. Association between GPC∆∆∆∆ex3 and cerebral malaria.........................................200 3.6.5. Therapeutic efficacy and equivalence of rectal artesunate to other treatments for severe malaria ................................................................................................................200 3.6.6. Operational feasibility of rectal artesunate as a public health intervention ......200

4. PRESUMPTIVE TREATMENT OF MALARIA IN PREGNANCY .............................201

4.1. Overview of Clinical trials in this chapter ..................................................................203

4.2. Clinical trial 6: Pharmacokinetics, safety and efficacy of chloroquine and sulphadoxine-pyrimethamine for presumptive treatment of malaria in pregnancy........205

4.2.1. SUMMARY.............................................................................................................205 4.2.2. AIMS......................................................................................................................207


4.2.3. METHODS .............................................................................................................207 4.2.3.1. Acknowledgement..................................................................................................... 207 4.2.3.2. Rationale for study design......................................................................................... 208 4.2.3.3. Study site and patients............................................................................................... 209 4.2.3.4. Clinical procedures.................................................................................................... 209

4.2.3.4.1. Eligibility criteria and enrolment procedures.................................................... 209 4.2.3.4.2. Drug administration, dosage and dosing times ................................................. 210 4.2.3.4.3. Blood sampling................................................................................................. 210 4.2.3.4.4. Clinical monitoring ........................................................................................... 211

4.2.3.5. Laboratory methods .................................................................................................. 211 4.2.3.5.1. Microscopy , haemoglobin and glucose............................................................ 211 4.2.3.5.2. Drug assays....................................................................................................... 212 4.2.3.5.3. Parasite PCR assays .......................................................................................... 212

4.2.3.6. Pharmacodynamic outcomes..................................................................................... 212 4.2.3.7. Pharmacokinetic outcomes........................................................................................ 212 4.2.3.8. Data analysis ............................................................................................................. 213

4.2.4. RESULTS ..............................................................................................................214 4.2.4.1. Patient characteristics................................................................................................ 214 4.2.4.2. Efficacy outcomes..................................................................................................... 214 4.2.4.3. Safety and tolerability ............................................................................................... 220

xii

4.2.4.3.1. Blood pressure and heart-rate ........................................................................... 220 4.2.4.3.2. Haemoglobin..................................................................................................... 221 4.2.4.3.3. Blood glucose ................................................................................................... 221

4.2.4.4. Pharmacokinetics of chloroquine and desethylchloroquine ...................................... 227 4.2.4.4.1. Drug concentrations.......................................................................................... 227 4.2.4.4.2. Pharmacokinetic analysis.................................................................................. 231

4.2.4.5. Pharmacokinetics of sulphadoxine and N-acetylsulphadoxine ................................. 237 4.2.4.5.1. Drug concentrations.......................................................................................... 237 4.2.4.5.2. Pharmacokinetic analysis.................................................................................. 239

4.2.4.6. Pharmacokinetics of pyrimethamine......................................................................... 244 4.2.4.6.1. Drug concentrations.......................................................................................... 244 4.2.4.6.2. Pharmacokinetic analysis.................................................................................. 246

4.2.5. CONCLUSIONS.....................................................................................................248 4.2.5.1. Pharmacokinetics of chloroquine and desethylchloroquine in pregnancy................. 249 4.2.5.2. Pharmacokinetics of sulphadoxine and N-acetylsulphadoxine in pregnancy............ 253 4.2.5.3. Pharmacokinetics of pyrimethamine in pregnant and non-pregnant women ............ 255 4.2.5.4. Clinical implications ................................................................................................. 257 4.2.5.5. Overall conclusions................................................................................................... 260

4.3. Summary of major findings of this chapter ...............................................................263

4.3.1. Pharmacokinetics of chloroquine and desethylchloroquine in pregnancy .......263 4.3.2. Pharmacokinetics of sulphadoxine and pyrimethamine in pregnancy..............263 4.3.3. Longer than expected elimination half- life of pyrimethamine ...........................263 4.3.4. Use of species-specific PCR diagnostics in pregnancy – the role of non-falciparum species in pregnant women in PNG............................................................263

5. CONCLUSIONS: IMPLICATIONS FOR PUBLIC HEALTH POLICY AND FURTHER RESEARCH.....................................................................................................................265

5.1. Policy implications for the management of uncomplicated malaria in Papua New Guinea ................................................................................................................................267 5.2. The problem of Plasmodium vivax in Papua New Guinea ........................................269 5.3. Need for improved understanding of pharmacokinetics and determinants of therapeutic efficacy of piperaquine in order to guide improved dosing regimens.........269 5.4. Place of rectal artesunate in public health policy in Papua New Guinea .................271 5.5. Further research needs for rectal artemisinins .........................................................272

5.5.1. Safety ....................................................................................................................272 5.5.2. Bioavailability and pharmacokinetic considerations..........................................272 5.5.3. Future comparative trials .....................................................................................273 5.5.4. Drug development: Need for a co-formulated ACT suppository........................274

5.6. Importance of erythrocyte polymorphisms in Papua New Guinea and implications for further research............................................................................................................275 5.7. Improved dosing regimens for intermittent preventive treatment of malaria in pregnancy ..........................................................................................................................276 5.8. Use of species-specific PCR diagnostics in pregnancy – the role of non-falciparum species in pregnant women in PNG..................................................................................278

6. REFERENCES.......................................................................................................281

7. APPENDICES ............................................................................................................1

7.1. Appendix A: Drug assay methodology....................................................................3

7.1.1. PIPERAQUINE...........................................................................................................3 7.1.2. CHLOROQUINE.........................................................................................................3 7.1.3. ARTESUNATE AND DIHYDROARTEMISININ...........................................................4

xiii

7.1.4. ARTEMETHER ..........................................................................................................4 7.1.5. SULPHADOXINE AND PYRIMETHAMINE ................................................................4

7.2. Appendix B: Design of Questionnaire Tool for Use in Clinical Trial 5 ...................7

7.3. Appendix C: Systematic review of pharmacokinetics, safety and efficacy of rectally administered artemisinin derivatives ...................................................................9

7.3.3. METHODS ...............................................................................................................10

7.3.3.2. Rationale for study design........................................................................................... 10 7.3.3.3. Data sources ................................................................................................................ 11 7.3.3.4. Study selection ............................................................................................................ 11 7.3.3.5. Outcome assessment ................................................................................................... 11 7.3.3.6. Data synthesis ............................................................................................................. 12

7.3.4. RESULTS ................................................................................................................12 7.3.4.1. Description of included studies ................................................................................... 12 7.3.4.2. Clinical efficacy .......................................................................................................... 15

7.3.4.2.2. Artemisinin ......................................................................................................... 15 7.3.4.2.3. Dihydroartemisinin ............................................................................................. 15 7.3.4.2.4. Artemether .......................................................................................................... 16

7.3.4.3. Comparative studies .................................................................................................... 22 7.3.4.4. Pharmacokinetics ........................................................................................................ 23 7.3.4.5. Safety and tolerability of rectal artemisinin derivatives.............................................. 24

7.3.5. CONCLUSIONS.......................................................................................................29

xiv

xv

iii. Acknowledgements The clinical studies on which this thesis is based would not have been possible without

the dedicated work and assistance of a number of people who worked either directly in

the field in Papua New Guinea (PNG), helped facilitate the logistics of the clinical studies

or performed laboratory assays either in PNG, Perth or the USA.

In PNG, excellent and attentive clinical supervision of field sites was provided at different

times by Adedayo Kemiki (Modilon General Hospital: MGH), Irwin Law, Sam Salman,

Thomas Schulz, Jovitha Lammey and Michele Senn (Alexishafen Health Centre), Olive

Oa and Penias Suano (Kunjingini Health Centre) and Rachael Hinton (Kunjingini and

Mugil). Day-to-day clinical activities of these trials were performed by a large number of

IMR field staff including Elizah Dabod, Barry Kalissa, Judy Longo (MGH), Wesley

Sikuma, Donald Paino, Sr Maria Goretti, Servina Gomorrai, Francesca Baiwog, Jovitha

Lammey, Kaye Kotab, Susannah Griffin, Ben Maamu, John Kakai (Alexishafen) Rachael

Hinton, Alma Auwun, Grace Pongua, Olive Oa, Peter Maku, Merilyn Uranoli, Jane

Simbrandu, Penias Suano and Petronilla Wapon (Kunjingini). Over 10 000 microscopy

readings have been performed by Kerry Lorry, Nandau Tarongka, Lena Lorry, Kaye

Baea and Moses Lagog.

The collaboration with the PNG IMR has germinated and gone on to flourish due to the

help of its three successive directors, Michael Alpers, John Reeder and Peter Siba, all of

whom have extended me helpful advice and considerable goodwill. Other IMR staff who

have been instrumental in facilitating logistics include Ivo Mueller, Moses Bockarie, John

Taime, Manasseh Baea, Andrew Raiko, Pascal Michon, Will Kastens and Ged Casey.

Further support was provided at MGH by Joseph Amban, Luke Antony, Antonia Demok

and the nursing staff of the children’s outpatient clinic and paediatric ward. I am

particularly indebted to Sr Valsi, Sr Theresetta, Sr Mary-Anthilda, all the health-centre

nursing staff and the Alexishafen Mission community as a whole for being so welcoming

and helpful in allowing us to direct our field activities from their health centre.

In Perth the HPLC and LCMS drug assays were performed by Kitya Duffal, Greg

Chiswell (ARTS and DHA), Juliana Hamzah (ARTS and ARM), Madhu Page-Sharp (PQ,

CQ and SP) and Sam Salman (LUM). Kevin Croft and Lincoln Morton were helpful for

xvi

consultations regarding the LCMS. Genotypic assays were performed by Sheral Patel,

Enmoore Lim and Alice Ura. Logistics and administrative activities in Perth were

coordinated by Michele England, Nicole McCoy and Pat Mumford all of whom always

provided prompt and efficient assistance and made being a PhD student that much

easier.

Financial support was provided to support either myself or the clinical studies by Mepha

Pharmaceuticals, Aesch-Basel, Switzerland, the Rotary Club of Scarborough, Western

Australia, the Royal Australasian College of Physicians, the Australian National Health

and Medical Research Council, the World Health Organization and Rotary Against

Malaria, PNG.

Special thanks go to my supervisors, Tim Davis and Ken Ilett, who have not only

provided me with sound academic guidance and support, but also been warm and

generous friends who have welcomed me into their families. I am particularly indebted to

Helen Ilett and Wendy Davis for providing me with a “home away from home” when

visiting Perth. Similarly Ivo Mueller has been invaluable as a scientific collaborator but

just as valuable as a friend. Ivo has proved to be a limitless source of information on all

matters pertaining to malaria especially in PNG and a major source of inspiration and

ideas – which seem to have been especially fruitful when work-shopped whilst sampling

his wife, Su’s wonderful Thai cooking. As well as Ivo and Su, working in PNG I have

made some lifelong friends, whose companionship made the challenges of performing

these studies so much easier. Nina Veenstra, Ged Casey and Michele Bowe deserve

special thanks in this regard.

I must also give special thanks to my wife, Christine, for enduring the privations of

dealing with a partner who has been absent (both physically and mentally) for much of

the last 4 years. I thank my parents for their support at all times, especially my father,

Hector, who obtained his own PhD in 1966 and instilled in his children the importance of

learning and education. He passed away in 2007, before this thesis was completed. He

would have been very proud.

Most of all, I would like to thank the patients and their parents in PNG who participated in

these studies. Doing so involved enduring various discomforts and a significant

xvii

commitment of their time. This was given generously and cheerfully, in the knowledge

that these studies might lead to knowledge that could benefit others.

xviii

xix

iv. Declaration I declare that, unless otherwise stated, the work contained in this thesis is my own.

The main exceptions to this are the following:

The study conception, design, implementation, analysis and interpretation for

Clinical trial 2 (Section 2.3) were completed by myself in collaboration with Tim

Davis and Ivo Mueller.

Clinical trial 3 (Section 3.2) was initially conceived by Tim Davis and Michael

Alpers in 1996. The population pharmacokinetic modeling was performed by Ken

Ilett, Hugh Barrett and Paolo Vincini.

The design of the questionnaire used in Clinical trial 5 (Section 3.5) was done in

collaboration with Rachael Hinton who also administered the questionnaire to

study participants.

A more detailed acknowledgement of work performed by collaborators is also

included in the preceding acknowledgements and in the methods section of each

Clinical trial.

All co-authors of publications arising from work have given their permission for

inclusion in this thesis.

____________________

Harin Karunajeewa

xx

xxi

v. Statement of ethical approval

The clinical studies detailed in this thesis have all been approved by the Human

Research Ethics Committee (HREC) of the University of Western Australia or the

Medical Research Advisory Committee (MRAC) of PNG. Details are as follows:

Clinical trial 1.

MRAC of PNG approval no. 04.14

University of Western Australia HREC approval no. 1060

Clinical trial 2.


University of Western Australia HREC approval no. 1060

Clinical trial 3.


Clinical trial 4.


Clinical trial 5.


Clinical trial 6.


xxii

xxiii

vi. Publications and presentations

a. Publications 1. Karunajeewa H, Kemiki A, Alpers M, Lorry K, Batty K, Ilett K, Davis T. Safety

and therapeutic efficacy of artesunate suppositories for treatment of malaria

in children in Papua New Guinea. Pediatric Infectious Diseases: 2003;

22:251-5

2. Karunajeewa HA, Ilett KF, Dufall K, , Kemiki A, Bockarie M, Alpers MP,

Barrett PHR, Vicini P, Davis TME. Disposition of artesunate and

dihydroartemisinin after administration of artesunate suppositories in children

from Papua New Guinea with uncomplicated malaria. Antimicrobial Agents

and Chemotherapy. 2004; 48:2966-2972.

3. Davis TM, Hung TY, Sim IK, Karunajeewa HA, Ilett KF. Piperaquine: a

resurgent antimalarial drug. Drugs. 2005; 65(1):75-87. Review.

4. Davis TM, Karunajeewa HA, Ilett KF. Artemisinin-based combination

therapies for uncomplicated malaria. Medical Journal of Australia. 2005 21;

182(4):181-5.

5. Karunajeewa HA, Reeder J, Lorry K, Dabod E, Hamzah J, Page-Sharp M,

Chiswell GM, Ilett KF, Davis TM. Artesunate suppositories versus

intramuscular artemether for treatment of severe malaria in children in Papua

New Guinea. Antimicrobial Agents and Chemotherapy. 2006; 50(3):968-74.

6. Hinton RL, Auwun A, Pongua G, Oa O, Davis TME, Karunajeewa. HA,

Reeder JC. Caregivers’ acceptance of using artesunate suppositories for

treating childhood malaria in Papua New Guinea. American Journal of

Tropical Medicine and Hygiene. 2007; 76:634-640

7. Karunajeewa HA, Manning L, Mueller I, Ilett KF and Davis TME. Rectal

administration of artemisinin derivatives for the treatment of malaria. Journal

of the American Medical Association. 2007; 297 (21):2381-90

8. Karunajeewa HA, Ilett KF, Mueller I, Siba P, Law I, Page-Sharp M, Lin E,

Lammey J, Batty KT , Davis TME. Pharmacokinetics and efficacy of

xxiv

piperaquine and chloroquine in Melanesian children with uncomplicated

malaria. Antimicrobial Agents and Chemotherapy. 2008; 52(1):237-43

9. Gomes M, Ribeiro I, Warsame M, Karunajeewa H and Petzold M. Rectal

artemisinins: a review of efficacy and safety from individual patient data in

clinical studies. BMC Infectious Diseases. 2008; 8:39

10. Law I, Ilett K F, Hackett LP, Page-Sharp M, Baiwog F, Gomorrai S, Mueller I,

Karunajeewa HA, Davis TME. Transfer of chloroquine and

desethylchloroquine across the placenta and into milk in Melanesian mothers.

British Journal of Clinical Pharmacology. 2008; 65 (5):674-679

11. Batty KT, Moore BR, Stirling V, Ilett KF, Page-Sharp M, Shilkin KB., Mueller I,

Karunajeewa HA and Davis TME. Toxicology and pharmacokinetics of

piperaquine in Mice. Toxicology. 2008; 249(1): 55-61

12. Karunajeewa HA, Mueller I, Senn M, Lin E, Law I, Gomorrai S, Oa O, Griffin

S, Kotab K, Suano P, Tarongka N, Ura A, Lautu D, Page-Sharpe M, Wong R,

Salman S, Siba P, Ilett KF, Davis TME. A trial of combination antimalarial

therapies in children from Papua New Guinea. New England Journal of

Medicine. 2008; 359 (24): 2545-2557.

13. Davis TME, Karunajeewa HA and Mueller I. Response to Price et al.:

Antimalarial Therapies in Children from Papua New Guinea. New England

Journal of Medicine. 2009; 360 (12): 1254-5.

14. Karunajeewa HA, Salman S, Mueller I, Baiwog F, Gomorrai S, Law I, Page-

Sharp M, Rogerson S , Siba P, Ilett KF, Davis TME. The pharmacokinetic

properties of sulfadoxine-pyrimethamine in pregnancy. Antimicrobial Agents

and Chemotherapy. 2009; In press

Submitted in mid-2009:

15. In vitro sensitivity of Plasmodium falciparum to conventional and novel

antimalarial drugs. A study from Madang in Papua New Guinea. Wong R,

Laatu, D, Tavul L, Hackett S, Siba P, Karunajeewa H, Ilett K, Mueller I, Davis

TME. (Antimicrobial Agents and Chemotherapy).

xxv

16. Karunajeewa HA, Salman S, Mueller I, Baiwog F, Gomorrai S, Law I, Page-

Sharp M, Rogerson S , Siba P, Ilett KF, Davis TME. Pharmacokinetics of

chloroquine and mono-desethylchloroquine when used for presumptive

treatment of malaria in pregnancy. (Antimicrobial Agents and Chemotherapy)

xxvi

xxvii

b. Presentations 1. Karunajeewa HA, Reeder J, Pakula C, Ilett K F, and Davis TME. Artesunate

suppositories for treatment of childhood malaria in Papua New Guinea.

Australian Society of Infectious Diseases Annual Scientific meeting, Alice

Springs, Australia, May 2004

2. Karunajeewa HA, Reeder J, Lorry K, Dabod E, Hamzah J, Ilett KF, Davis

TME, Artesunate suppositories vs intramuscular artemether for treatment of

severe childhood malaria in Papua New Guinea. XVIth International

Congress of Tropical Medicine and Malaria / IVth European Congress of

Tropical Medicine and International Health / Vth International Congress of the

French Society of Tropical Medicine (“Medicine and Health in the Tropics”).

Marseilles, France, September 2005

3. Karunajeewa HA, Ilett KF, Mueller I, Siba P, Law I, Page-Sharp M, Lin E,

Lammey J, Batty KT , Davis TME. Pharmacokinetics and efficacy of

piperaquine and chloroquine in Melanesian children with uncomplicated

malaria. American Society of Tropical Medicine and Hygiene 56th Annual

Meeting, Philadelphia, USA, November 2007

4. Karunajeewa HA. Country in focus session: “Papua New Guinea”. Asia

Pacific International Conference on Travel Medicine, Melbourne, February

2008 (Invited speaker).

5. Karunajeewa HA, Mueller I, Senn M, Lin E, Law I, Gomorrai S, Oa O, Lautu

D, Page-Sharpe M, Wong R, Salman S, Siba P, Ilett KF, Davis TME.

Pharmacokinetic and parasitological determinants of the efficacy of

dihydroartemisinin-piperaquine for treatment of P. falciparum in Papua New

Guinean children. XVIIth International Congress of Tropical Medicine and

Malaria, Jeju Island, South Korea, October 2008

6. Karunajeewa HA, Mueller I, Page-Sharpe M, Law I, Salman S, Gomorrai S,

Lammey J, Rogerson S, Siba P, Ilett KF, Davis TME. Pharmacokinetic

properties of chloroquine and sulphadoxine-pyrimethamine in pregnancy.

American Society of Tropical Medicine and Hygiene annual meeting, New

Orleans, USA, December 2008

xxviii

xxix

vii. Abbreviations ACPR – Adequate clinical and parasitological response (according to WHO definition)

ACT – Artemisinin combination treatment

AL – Artemether-lumefantrine

ANOVA – Analysis of variance

ARM – Artemether

ARTS – Artesunate

ARTS-SP – Artesunate plus sulphadoxine-pyrimethamine

AUC 0-� - Area under the curve (between zero and infinity)

AUC 0-2 - Area under the curve (between zero and 2 days/ 48 hours)

AUC 0-3 - Area under the curve (between zero and 3 days/ 72 hours)

AUC 0-14 - Area under the curve (between zero and 14 days)



AQ – Amodiaquine

BCS – Blantyre coma score

BMI – Body mass index

CI – Confidence interval

CL – Clearance

CL/F – Clearance (relative to bioavailability)

Cmax – Maximal concentration

CNS – Central nervous system

CQ – Chloroquine

CQ-SP – Chloroquine + sulphadoxine-pyrimethamine

CV% – Coefficient of variation

DECQ – Desethylchloroquine

DHA – Dihydroartemisinin

DHFR – Dihydrofolate reductase gene

DHPS – Dihydropteroate synthetase gene

DNA – Deoxyribonucleic acid

DHA-PQ – Dihydroartemisinin-piperaquine

EIR - Effective inoculation rate

ETF – Early treatment failure

xxx

F – Bioavailability

FCT� – Fever clearance time (Time until axillary temperature is <37.5°C and maintained

below 37.5°C for at least 24 hours).

FDA – Food and Drug Administration (USA)

FRAC – Relative bioavailability term

g – grams

GC-ECD – Gas chromatography – electron capture detection

GIT – Gastrointestinal tract

GLURP – Glutamine rich protein

GMP – Good manufacturing practice

GPC3 – Glycophorin C

GPC∆ex3- glycophorin C exon 3 deletion

h - hours

Hb – haemoglobin

HIV – human immunodeficiency virus

HPLC – High performance liquid chromatography

IC50 – Concentration of drug necessary to provide 50% growth inhibition of in vitro

parasite culture

i.m. – intramuscular

IMCI – Integrated management of childhood illness

IMR – Institute of Medical Research

i.v. – intravenous

i.r. – intrarectal

IPT – intermittent presumptive (or preventive) treatment

IPTi - intermittent presumptive (or preventive) treatment (of infants)

IPTp - intermittent presumptive (or preventive) treatment (in pregnancy)

IRR – incidence rate ratio

IQR – interquartile range

ITN – insecticide-treated bed-net

IUGR – intrauterine growth retardation

k12 – rate constant (compartments 1 to 2)

k23 - rate constant (compartments 2 to 3)

k32 – backwards rate constant (compartments 3 to 2)

k30 – elimination rate constant (from compartment 3)

xxxi

ka – absorption rate constant

L – litres

LCF – Late treatment failure (clinical)

LC-MS-MS - Liquid chromatography-tandem mass spectroscopy

LDH – Lactate dehydrogenase

LDR-FMA – Ligase detection reaction – fluorescent microsphere assay

LOD – Limit of detection

LOQ – Limit of quantification

LPF – Late treatment failure (parasitological)

LTF – Late treatment failure

LUM – lumefantrine

MGH - Modilon General Hospital

MIC – Minimal inhibitory concentration

min – Minutes

mm Hg – millimeters of mercury

mol –moles

MQ – mefloquine

MSP – merozoite surface protein

MUAC – mid-upper arm circumference

n – number

NASDOX – N-acetylsulphadoxine

NAT1 – N-acetyl transferase 1

NAT2 – N-acetyl transferase 2

OR – Odds Ratio

PCR – Polymerase chain reaction

PCT50 – Time to reduction in parasitaemia of 50% from baseline.

PCT90 – Time to reduction in parasitaemia of 90% from baseline.

PC%12 – Parasitaemia at 12 hours, expressed as a percentage of the baseline

parasitaemia.

PC%24 – Parasitaemia at 24 hours, expressed as a percentage of the baseline

parasitaemia.

PCT – Parasite clearance time

PD – Pharmacodynamic

PK – Pharmacokinetic

xxxii

Pf – Plasmodium falciparum

PfCRT – Plasmodium falciparum chloroquine resistance transporter

PfMDR – Plasmodium falciparum multi-drug resistance gene

Pm – Plasmodium malariae

PNG – Papua New Guinea

PNG IMR – Papua New Guinea Institute of Medical Research

Po – Plasmodium ovale

PQ – Piperaquine

Pv – Plasmodium vivax

PYR – Pyrimethamine

QHS – Artemisinin (qinghaosu)

QTc – The electrocardiographic interval between “Q” and “T” waves, corrected for heart-

rate.

RFLP – Restriction fragment length polymorphism

RSD – Relative standard deviation

RSE – Relative standard error

SLC4A1∆27 – 27 base-pair deletion of the band 3 gene

SD – Standard deviation

SDOX - Sulphadoxine

SEAQUAMAT – South East Asian Quinine Artesunate Trial

SP – Sulphadoxine-pyrimethamine

t½ - Half-life

t½abs – Absorption half-life

t½e – Elimination half-life

t½e�1 – Elimination half-life (phase 1) – or “distribution half-life”

t½e�2 – Elimination half-life (phase 2) – or “terminal elimination half-life”

tlag – lag time

tmax – Time to maximal concentration

TGA – Therapeutic Goods Association (Australia)

UWA – University of Western Australia

Vd – Volume of distribution

V/F – Volume of distribution (relative to bioavailability)

V1/F – Volume of distribution of 1st (central) compartment (relative to bioavailability)

V2/F – Volume of distribution of 2nd compartment (relative to bioavailability)

xxxiii

Vss/F – Volume of distribution at steady state (relative to bioavailability)

WAZ – Weight for age Z-score

WHO – World Health Organization

xxxiv

xxxv

viii. List of Figures 1. Chapter 1 1.1. Theoretical relationship between drug concentration over time and clinical

cure of malaria. 1.2. Topographic map of Papua New Guinea.

2. Chapter 2. Clinical trial 1 2.1. Plot of measured and predicted plasma piperaquine concentrations against

time. 2.2. Plots of plasma chloroquine and desethylchloroquine concentrations against

time. A. Chloroquine (measured and predicted). B. Desethylchloroquine.

2.3. Model-predicted plasma concentration-time plot of piperaquine given at a dose of 30mg/kg (base) daily for 3 days compared with piperaquine 12 mg base/kg daily for 3 days and chloroquine 10 mg base/kg daily for 3 days.

Clinical trial 2 2.4. Trial profile

A. Numbers of patients remaining in the study from screening to Day 42 assessment.

B. Reasons for exclusion prior to day 28 and day 42 endpoints (patients with P. falciparum).

C. Reasons for exclusion prior to day 28 and day 42 endpoints (patients with P. vivax).

2.5. Kaplan-Meier plot showing proportion of patients remaining free of parasitaemia.

A. PCR-corrected P. falciparum infection. B. P. vivax.

2.6. Mean changes in parasitaemia from baseline by allocated treatment between Days 0 and 7.

2.7. Percentage of children who were gametocyte slide-positive by allocated treatment between Days 0 and 42.

3. Chapter 3. Clinical trial 3 3.1. Structural model for disposition of artesunate and dihydroartemisinin following

rectal artesunate administration derived from 47 children with uncomplicated malaria treated with artesunate suppositories.

3.2. Plot of population model prediction and measurements for artesunate against time.

3.3. Plot of population model prediction and measurements for dihydroartemisinin against time.

3.4. Plot of weighted residuals for artesunate and dihydroartemisinin versus subject number.

xxxvi

Clinical trial 4(b) 3.5. Parasite clearance kinetics in 77 children with severe malaria treated with

either rectal artesunate or intramuscular artemether. A. Median proportional reduction in parasite density according to treatment

group. B. Individual patient parasite density-time data in 41 children treated with

rectal artesunate. C. Individual patient parasite density-time data in 38 children treated with

intramuscular artemether. 3.6. Plasma drug concentrations in 15 children following 11 mg/kg rectal

artesunate. A. Artesunate. B. Dihydroartemisinin.

3.7. Plasma drug concentrations in 14 children following 3.2 mg/kg i.m artemether.

A. Artemether. B. Dihydroartemisinin.

4. Chapter 4 Clinical trial 6 4.1. Semi-quantitative ligase detection reaction – fluorescent microsphere assay

results showing PCR signal intensity over 42 days following treatment with chloroquine + sulphadoxine-pyrimethamine in both pregnant and non-pregnant groups.

A. P. falciparum (Pf 72). B. P. falciparum (Pf 73). C. P. vivax. D. P. malariae. E. P. ovale.

4.2. Mean blood pressure following treatment with sulphadoxine-pyrimethamine and chloroquine.

A. Pregnant subjects. B. Non-pregnant controls.

4.3. Mean supine and erect mean-arterial pressure following treatment with sulphadoxine-pyrimethamine and chloroquine.


4.4. Mean heart-rate following treatment with sulphadoxine-pyrimethamine and chloroquine.


4.5. Mean haemoglobin concentrations in pregnant and non-pregnant subjects over the 42 days following initiation of treatment.

4.6. Mean blood glucose concentrations in pregnant and non-pregnant subjects in the 3 days following treatment.

4.7. Actual measured plasma chloroquine concentrations over 42 days following administration of three daily doses of 450mg chloroquine base.

A. Pregnant subjects. B. Non-pregnant subjects.

4.8. Comparison of median chloroquine levels in pregnant and non-pregnant groups.

xxxvii

4.9. Actual measured plasma desethylchloroquine concentrations over 42 days following administration of three daily doses of 450mg chloroquine base.


4.10. Comparison of median desethylchloroquine levels in pregnant and non-pregnant groups.

4.11. Simulated concentration-time curves based on median model-derived pharmacokinetic parameters for chloroquine.


4.12. Simulated concentration-time curve using model-derived parameters and based on 50% dosage increase in chloroquine dose in pregnant women.

4.13. Actual measured plasma sulphadoxine concentrations over 42 days following administration of a single dose of 1500/75mg sulphadoxine-pyrimethamine.


4.14. Comparison of median sulphadoxine levels in pregnant and non-pregnant groups.

4.15. Simulated concentration-time curves based on median model-derived pharmacokinetic parameters for sulphadoxine.


4.16. Simulated concentration-time curves of dosage modification using model-derived parameters for sulphadoxine.

A. 33% dosage increase administered as a single dose in pregnant women.

B. 60% total dose increase administered as 3 split doses over 48 hours. 4.17. Actual measured plasma pyrimethamine concentrations over 42 days

following administration of 1500/75mg sulphadoxine-pyrimethamine. A. Pregnant subjects. B. Non-pregnant subjects.

4.18. Comparison of median pyrimethamine levels in pregnant and non-pregnant groups.

xxxviii

xxxix

ix. List of Tables 1. Chapter 1 1.1. Qualities of the “ideal” antimalarial for outpatient treatment of uncomplicated

malaria (e.g. as a partner drug for use in artemisinin combination therapy).

2. Chapter 2. Clinical trial 1 2.1. Details of patients at the time of admission to the study. 2.2. Treatment outcomes by microscopy at days 28 and 42 in children with P.

falciparum mono-infection at baseline. 2.3. Pharmacokinetic parameters for piperaquine, chloroquine and

desethylchloroquine in children with malaria.

Clinical trial 2 2.4. Baseline characteristics of recruited children categorised by parasite species. 2.5. Per-protocol analysis of treatment response in children with falciparum

malaria by treatment allocation. A. PCR uncorrected data. B. PCR corrected data.

2.6. Intention-to-treat analysis of treatment response in children with falciparum malaria by treatment allocation.

A. PCR uncorrected data. B. PCR corrected data.

2.7. Fever and parasite clearance times by Plasmodium species and treatment allocation.

2.8. Per-protocol analysis of treatment response in children with vivax malaria by treatment allocation.

2.9. Intention-to-treat analysis of treatment response in children with vivax malaria by treatment allocation.

2.10. Per-protocol analysis of P. vivax infections occurring after treatment of patients with P. falciparum mono-infections at enrolment.

2.11. Summary table showing relative efficacy of the three artemisinin combination regimens in comparison to chloroquine + sulphadoxine-pyrimethamine.

2.12. Incidence rate of reported or observed signs and symptoms during first 7 days of follow-up, expressed as reports per 100 observations.

3. Chapter 3. Clinical trial 3 3.1. Baseline variables of 47 children with uncomplicated malaria treated with

artesunate suppositories. A. Demographic and clinical features. B. Alpha-thalassaemia genotype.

3.2. Pharmacodynamic outcomes of 47 children with uncomplicated malaria treated with artesunate suppositories.

3.3. Details of the three artesunate suppository patients at the time of a late fever spike.

3.4. Summary of the pharmacokinetic model development process applied to drug concentration-time data.

3.5. Parameters and performance descriptors for the final pharmacokinetic model.

xl

3.6. Pharmacokinetic parameters for the model.

Clinical trial 4 3.7. Demographic and clinical features of 70 children with severe malaria. 3.8. Erythrocyte polymorphism genotype and relationship to severe

manifestations of disease in children with severe malaria. A. Band 3. B. Glycophorin C. C. Alpha-thalassaemia.

3.9. Baseline characteristics of patients in the artesunate suppository and intramuscular artemether groups.

3.10. Clinical efficacy endpoints by treatment group. 3.11. Artemisinin exposure (sum of concentrations of plasma artesunate +

dihydroartemisinin or artemether + dihydroartemisinin) by treatment group at selected times after drug administration.

Clinical trial 5 3.12. Baseline characteristics of 131 subjects.

A. Children. B. Presenting caregiver.

3.13. Presenting caregiver questionnaire responses regarding initial knowledge and perception of rectal administration of medicine.

3.14. Factors associated with declining to use suppositories and therefore refusing to participate.

3.15. Satisfaction/ acceptance following treatment course.

4. Chapter 4 Clinical trial 6 4.1. Baseline variables of 30 pregnant subjects and 30 non-pregnant controls.

A. Anthropometric variables and malaria status. B. Pregnancy status. C. Baseline haemodynamic variables, haemoglobin and glucose.

4.2. Efficacy outcomes. A. Conventional WHO-defined criteria (based on clinical and microscopy

follow-up to 28 days) for patients with microscopy-proven P. falciparum at baseline.

B. Microscopy and parasite PCR-positivity at day 28 and 42. 4.3. Compartmentally derived pharmacokinetic parameters for chloroquine in

pregnant versus non-pregnant subjects. 4.4. Non-compartmentally derived pharmacokinetic parameters for chloroquine

and desethylchloroquine in pregnant versus non-pregnant subjects. 4.5. Compartmentally derived pharmacokinetic parameters for sulphadoxine in

pregnant versus non-pregnant subjects. 4.6. Non-compartmentally derived pharmacokinetic parameters for sulphadoxine

in pregnant versus non-pregnant subjects. 4.7. Non-compartmentally derived pharmacokinetic parameters for pyrimethamine

in pregnant versus non-pregnant subjects.

xli

7. Appendix 7.1. Pharmaceutical preparations of artemisinin derivatives for rectal

administration. 7.2. Clinical efficacy of rectally administered artemisinin derivatives.

A. Artesunate. i. Low-dose regimens. ii. High dose regimens.

B. Artemisinin. C. Dihydroartemisinin. D. Artemether.

7.3. Pharmacokinetics of rectally administered artemisinin derivatives. A. Artesunate.

i. Low-dose regimens. ii. High dose regimens.

B. Artemisinin. C. Dihydroartemisinin. D. Artemether.

xlii

1

1. GENERAL INTRODUCTION

2

3

1.1. Malaria: The global situation A comprehensive analysis of global population health data from 2001 suggested that

malaria was the world’s eighth leading cause of premature death, being responsible for

almost 4% of total years of life lost to disease. [1] This burden affected predominantly

low and middle-income countries, where malaria accounted for almost 40 million

disability adjusted life years. Only three infectious disease classifications - lower

respiratory infection, HIV/AIDS and diarrhoeal disease - had greater impact in these

countries. The same study demonstrated a rise in global malaria mortality throughout the

1990’s, due predominantly to the proportions of deaths from malaria in African children

(rising from 15% in 1990 to 22% in 2001). It has been suggested that this mortality

increase may be attributable to worsening parasite drug -resistance. [2]

The overwhelming burden of malarial disease falls on children under the age of 5 and

pregnant women. [3, 4] One analysis based on empirical methods estimated that of 1

824 000 malaria-attributable childhood deaths per year in Africa, most were due to

severe anaemia (53%), low birth weight (20%), hypoglycemia (15%), respiratory disease

(6%) and cerebral malaria (6%). [3]

The five plasmodia species known to cause human malaria include P. falciparum, P.

vivax, P. ovale, P. malariae and P. knowlesi. The distribution of each varies

geographically. Plasmodium ovale and P. malariae occur widely throughout the tropical

world but their overall contribution to the burden of malaria disease is relatively small. [5]

Plasmodium knowlesi has only recently been recognised as a human pathogen in

Malaysian Borneo where its primary definitive host is thought to be old-world primates.

[6] Plasmodium falciparum is the most important cause of malaria, being responsible for

the bulk of its global mortality burden. [7] This reflects its high prevalence, wide

geographic distribution (including South America, Asia, Oceania and Africa), and greater

pathogenicity. [8, 9] Plasmodium vivax is rare in Africa and has always been considered

of secondary importance when compared with P. falciparum. However, there is

accumulating evidence that this Plasmodium species can be associated with acute

complications and death [10-13] and recurrent infections are likely to contribute to

chronic anaemia and its deleterious effects in high-transmission areas. [14, 15] It

remains prevalent throughout Asia, South America and Oceania. [16] For these reasons,

4

the overall contribution of P. vivax to the global burden of disease may have been

seriously underestimated. [12, 16]

Along with environmental control of the mosquito vector and the provision of insecticide

treated bed-nets (ITNs), early effective drug treatment of malarial infections remains a

cornerstone of malaria control measures. [17] Chloroquine (CQ) has been an important

mainstay of treatment for uncomplicated malaria since its development in the 1940s.

However, P. falciparum parasite resistance to this cheap and readily available drug is

now virtually ubiquitous throughout the malaria-endemic world. [18] Effective, practical

and affordable alternative therapies are urgently required to avert what many see as an

imminent global health disaster. [19]

1.2. Malaria: Biology and pathogenesis Human malarias are transmitted almost exclusively by the anopheline mosquito. The life-

cycle of P. falciparum in the human host begins when an inoculum containing

approximately 8-15 infecting sporozoites is acquired from the salivary glands of a biting

female mosquito. [20] After entering the systemic circulation, sporozoites invade

hepatocytes and begin a process of asexual multiplication lasting approximately 6 days,

before the hepatic schizont ruptures to release many thousand merozoites. [9]

Merozoites then rapidly invade erythrocytes by a complex interaction with host cell

ligands that facilitates endocytosis of the merozoite. [21] The intracellular parasite

multiplication that ensues constitutes the 48-hour period of the parasite’s intra-

erythrocytic life-cycle. [22] This includes the development of early trophozoites (small

ring forms visible by microscopy in the first 12 h), maturation of trophozoites (larger ring

forms that may display pigment on microscopy after 12 h), nuclear division (leading to

the segmented appearance of the schizont visible after 36 h) and finally rupture of the

erythrocyte schizont (at 48 h). [9] Schizont rupture releases between 16 and 32

merozoites, each of which may infect another erythrocyte and re-commence the intra-

erythrocytic cycle. [21] In this way blood-stage infection is characterized by an

exponential expansion in parasite numbers with increments occurring every 48 h in

synchronous infections. [22]

The efficiency of merozoite invasion may be attenuated by cellular and humoral immune

processes so that individuals with pre-existing immunity have a slower rate of

5

erythrocytic parasite expansion. However, in general, each 48-hour erythrocytic life-

cycle produces an approximately 8-fold increase in total parasite numbers. [22] Even in

non-immune hosts, the infection is unlikely to become symptomatic until the whole-body

parasite burden is greater than 108 (close to the lower limit of detection by light

microscopy). [9] Therefore, a pre-patent phase or incubation period includes the pre-

erythrocytic phase and the first few cycles of erythrocytic multiplication before the host

develops symptoms. In total this is typically 13 days in non-immune hosts with P.

falciparum but longer in immune individuals. [9]

After 12-14 h of development, mature P. falciparum trophozoites express a variety of

strain-specific variant antigens on the erythrocyte surface that may mediate attachment

of the erythrocyte to vascular endothelium, a phenomenon described as

“cytoadherence.” [23] This phenomenon, which is unique to P. falciparum, causes the

sequestration of infected cells within the microcirculation that may underlie the

pathogenesis of severe malaria by impairing blood flow and oxygen exchange in the

tissues of vital organs. [24] Consequent anaerobic metabolism contributes to metabolic

acidosis and hyperlactatemia which are key indicators of disease severity. [25-27]

Sequestration in a variety of tissues also leads to organ-specific dysfunction including

cerebral malaria (manifesting as impairment of conscious state), renal failure, pulmonary

disease, jaundice, retinal haemorrhage and placental dysfunction (leading to foetal

growth retardation). [24]

In addition to asexual multiplication during pre-erythrocytic and erythrocytic stages, the

development of sexual forms occurs during erythrocytic development over a period of 7-

10 days. [9] This produces gametocytes which enable transmission back to the

anopheline vector when it ingests a blood meal. Sexual reproduction within the mosquito

alimentary system then results in sporozoites which migrate to salivary glands where

they can be inoculated into the next human host.

1.3. Malaria: Epidemiology The epidemiology of malaria is dependent on the environmental tropism, breeding

activity and biting habits of its anopheline vector. [28] Geographic variations in ambient

temperature, humidity, rainfall and the proximity and density of human populations may

determine how well mosquitoes thrive. [28, 29] In addition, as many as 80 separate

6

anopheline species have been identified worldwide as being capable of transmitting

malaria to humans. [28, 30] Each species may have differing breeding requirements, life-

spans, feeding habits and vulnerability to environmental stressors or insecticides. Taken

together, these factors explain why malaria endemicity can vary so widely within

relatively small geographic areas. [28, 29] Some of the highest rates of transmission

(and therefore malaria endemicity) have been described in Sub-Saharan Africa and

Oceania. [31, 32] This reflects both the environmental conditions in these areas

(including high rainfall and humidity) that are highly favourable for the vector and the

relative robustness of the anopheline species endemic to these areas. [29]

The clinical epidemiology of P. falciparum malaria in any human population is the

product of a complex interplay between the factors of age, acquired immunity and the

level of endemicity of P. falciparum infection in the area where a population resides. [33,

34] Acquisition of protective immunity is a gradual process that appears to occur in a

stepwise fashion. Initial protection is limited to “disease modifying” immunity, that does

not protect the individual from becoming ill, but ameliorates disease severity so that the

risk of dying is lower. [35] With time and following further infective episodes, infections

may become asymptomatic. Eventually, the host may be protected from even

developing detectable parasitaemia following inoculation. The primary determinant of the

rate of acquisition of immunity is the frequency that infections occur over time in each

individual. [34, 35] This is obviously dependent on the intensity of mosquito-borne

transmission in the community. In high-endemicity regions, mosquito-borne transmission

(measured using a standard entomological index such as the effective inoculation rate:

EIR) will be high, resulting in frequent infections occurring in early childhood. Protective

immunity develops relatively early in life. [34] The level of endemicity will thus determine

the population’s age-specific disease distribution. Because clinical manifestations of

disease are different in children and adults, the level of disease endemicity will therefore

also determine the pattern of clinical syndromes seen in the community as a whole. [33,

34]

Traditionally, malaria endemicity has been classified according to the rates of clinically

detectable splenomegaly or microscopically detectable parasitaemia in children aged 2-9

years. [30] Areas with spleen or parasite rates <10% are classified as hypo-endemic,

7

those 10-50% as meso-endemic, 50-75% as hyper-endemic (adult spleen rate is high)

and >75% as holo-endemic (adult spleen rate is low and parasitaemia in infancy is high).

In areas of the highest (holoendemic) transmission, such on the North coast of Papua

New Guinea (PNG) or parts of Sub-Saharan Africa, long-term residents who survive to

adulthood have experienced frequent and regular infections throughout childhood so that

as adults, they rarely experience symptomatic infections. Infants born in these areas are

also relatively protected from severe infection by passively acquired maternal immunity

and by foetal haemoglobin which is less favourable to parasite growth. [36] The disease

has its greatest impact in the 1-3 year age group and severe anaemia is its primary

manifestation. [24, 27] Where transmission is less intense (hyper-endemic), disease

extends to older children and cerebral malaria is the most prominent manifestation of

severe disease. [24] By contrast, where transmission is low, erratic, highly seasonal or

epidemic, children and adults of all ages are susceptible to severe disease. [33, 34] The

full range of manifestations of severity is seen in adults in these areas. [24]

Pregnant women are more susceptible to malarial infection and therefore contribute a

disproportionate number of all cases of severe malaria. [24, 37] In hyper- and holo-

endemic areas, asymptomatic infection is common, contributing to a high prevalence of

maternal anaemia and intrauterine growth retardation (IUGR) secondary to placental

malaria. [38] Because birth-weight is the single most important determinant of risk of

perinatal mortality, even relatively small increases in the number of low-birth weight

neonates (<2.5kg) will add significantly to the population’s overall mortality. [38] In

addition, more recently the prevalence and impact of malaria in pregnancy have been

compounded by the effects of HIV infection in areas where HIV prevalence is high. [39]

HIV-positive pregnant women in Africa have been shown to be almost twice as likely to

have P. falciparum malaria during pregnancy [40] and those with co-infection with

malaria and HIV twice as likely to be anaemic as those with malaria alone [41], leading

to dual additive effects on maternal, perinatal and infant mortality. [39]

1.4. Malaria: Role of host and parasite genetics It is estimated that human populations have been living with malaria for as long as they

have existed. [5] During this time, evolution of parasite species has been driven by

selection pressures related to the host’s ability to suppress or self-cure infections.

8

Conversely, evolution of the human host has been driven by the selection pressure

exerted by malaria mortality. [5, 42] Therefore, a complex dynamic evolutionary

equilibrium exists between the host’s attempts to protect itself from infection and the

parasite’s attempts to thrive. Here, the parasite has a relative advantage due to its very

much shorter life-cycle and therefore greater capacity for rapid genetic adaptation. This

is reflected, for example, by the ability of parasites to develop drug-resistance within a

fraction of a human lifetime.

Not surprisingly, the most pathogenic species of malaria, P. falciparum, is thought to

have the shortest history of all human malaria parasites, possibly having afflicted

humans for as little as 5 000 years. [5] Its relatively greater pathogenicity may therefore

reflect the mere <400 generations that humans have had to adapt to this parasite.

Nonetheless, it is clear that a number of genetic adaptations have arisen in human

populations that protect the host from P. falciparum malaria. These appear to have

arisen de novo in populations exposed to a heavy burden of disease where a heavy

selection pressure exists for adaptations that increase survival from malaria. [5, 42]

Many such adaptations involve polymorphisms of the erythrocyte that facilitate host-

protection by interfering with parasite invasion or by altering the capacity of parasitized

erythrocytes to sequester. [43, 44] These have arisen independently in many different

parts of the world and include sickle cell anaemia, the thalassaemias, hereditary

ovalocytoses, Gerbich blood group and glucose-6 phosphatase deficiency (G6PD) (for

P. falciparum) and polymorphisms of the Duffy-binding protein (for P. vivax). [5] In the

case of some of these, such as alpha thalassaemia, there is good evidence the

mutations protect against morbidity and mortality associated with malaria infections. [45]

However, for others such as the GYPC�ex3 mutation causing the Gerbich blood group,

the evidence currently remains circumstantial. In this case a protective effect is

hypothesized based firstly on the close association between the geographic distribution

of P. falciparum malaria and of the GYPC�ex3 mutation in human populations in PNG.

[46] Secondly, biological plausibility exists in that erythrocyte surface components

affected by the mutation have been shown to be binding ligands for the P. falciparum

merozoite, utilised during the invasion process. Erythrocytes affected by the mutant

gene cannot use this invasion pathway and are therefore less prone to parasite invasion

in vitro. [47] Nonetheless, there is still no clinico-epidemiological evidence to support the

hypothesis of a protective effect from malaria. In particular, no relationship between

9

carriage of the mutation and the prevalence of asymptomatic parasitaemia has been

demonstrated. [46] However, it is possible that a protective effect leading to lower

morbidity and mortality is mediated not through susceptibility to infection but by

ameliorating the severity of disease itself. Therefore, ideally, confirmation of this

hypothesis will require epidemiological studies with case-control design, comparing the

incidence of the mutation in healthy controls, with those with clinically symptomatic

malaria, as has been performed previously for other mutations with putative protective

benefit. [45] Cases in such a study should be those with severe malarial disease or

subjects who have died from malaria.

In modern times, the key manifestation of the malaria parasite’s propensity to genetic

adaptation is the rapidity with which drug-resistance can arise de novo and then spread

throughout parasite populations. [48] De novo resistance may arise when a single

parasite has a spontaneous mutation that enables it to survive drug concentrations that

would usually be lethal. [49] The mutant parasite replicates and transfers a lower

susceptibility to the drug to its progeny. This survival advantage means that, in areas of

heavy drug-use, the mutation is rapidly selected and spreads throughout the population.

[49, 50] A number of specific parasite genetic polymorphisms associated with drug

resistance have been identified. These include mutations of the PfCRT and PfMDR1

genes (CQ and mefloquine resistance) the DHPS genes (sulphonamide resistance) and

the DHFR gene (pyrimethamine resistance). [51] When mutations affect enzymatic

pathways critical to the drug’s activation or mechanism of action, high-level resistance

can be induced by a single mutation. This is the case with the drug atovaquone, which is

rendered ineffective by a single base pair mutation of the gene encoding the cyt b

enzyme. [52]

For other drugs, the development of resistance is more complex and takes place as a

multi-step process. [53] The widely used antimalarial combination of sulphadoxine-

pyrimethamine (SP) is usually only subject to clinically significant drug resistance once

there is a high population prevalence of a number of mutations involving both the DHFR

and DHPS genes. [54-56] Genetic mechanisms underlying parasite resistance to

quinoline antimalarials are even more complex and less-well understood. [57, 58]

Mutations of the PfCRT gene are often fixed within P. falciparum populations throughout

the world but clinical response to CQ may remain relatively intact in endemic areas,

10

reflecting the contribution of pre-existing immunity to parasite clearance. [59-61] The

impact of the PfMDR gene on drug resistance may be mediated largely through the

number of gene-copies the parasite expresses as well as in specific mutations of this

gene. [62, 63]

The propensity of malaria parasites to de novo selection of drug-resistant mutants

underlies the theoretical justification for combination antimalarial therapy. [49, 64] This

supposes that simultaneously administering at least two antimalarial compounds that

have differing mechanisms of actions will prevent or slow the development of resistance

to either drug. Parasite mutants that arise with resistance to one drug will still be

susceptible to the second, and therefore will not survive to pass on their resistance

characteristics to the next generation. [49] A similar approach has underpinned the

treatment of tuberculosis, HIV, Hepatitis B infections and most cancers. However, in

terms of the impact of this strategy on the population genetics of antimalarial drug

resistance, it remains a theory which is difficult to test and to prove.

1.5. Clinical manifestations of malaria

1.5.1. Uncomplicated malaria

Symptomatic malaria that is not complicated by features of severe disease is described

as being “uncomplicated”, “simple” or “non-severe.” [9] An estimated 500 million such

episodes occur world-wide every year. [8] The cardinal feature is fever, which is almost

always present. Other clinical symptoms may include chills, rigors, lethargy, anorexia,

vomiting and headache, and children may become restless and irritable. [9] However, no

clinical feature (either alone or in combination) is sufficiently sensitive or specific to either

confirm or exclude malaria as a cause of fever. Attempts to design clinical algorithms to

do this have all demonstrated poor positive and negative predictive values. [65]

Therefore, given the potential serious sequelae of untreated P. falciparum infection,

microscopy of an appropriately stained thick blood film is necessary if malaria is to be

reliably excluded. [9, 65] Rapid diagnostic tests employing antigen detection of the P.

falciparum histidine-rich protein2 (PfHRP2) or parasite LDH may be useful alternatives

or adjuncts to microscopy, but lack sensitivity for low P. falciparum parasitaemias and for

non-P. falciparum infections. [66]

11

1.5.2. Severe malaria

The most serious clinical manifestations of P. falciparum infection carry a high risk of

death. If this risk is to be minimized, it is important that those clinical features of greatest

prognostic significance are well-defined and easily recognised by health-workers. In this

way treatment modalities most suitable for severe disease can be instituted with

appropriate expediency.

Non-immune adults are prone to a wide range of serious manifestations including

cerebral malaria, renal failure, jaundice, haemolysis, anaemia, pulmonary insufficiency

and hypoglycemia. However, the greater burden of severe malaria affects children living

in endemic areas, for whom the manifestations of severe malaria are more protean. In

particular, involvement of the kidneys, liver and lungs is uncommon in children. [24]

Instead, most children at risk of death have manifestations of one or more of just three

major clinical syndromes: cerebral malaria, respiratory distress/metabolic acidosis or

severe anaemia. [24, 25, 27, 35, 67, 68] It is likely, therefore, that together these three

syndromes constitute a significant portion of malaria’s global mortality burden. [3, 4]

Determining the relative contribution of each syndrome to overall mortality is, however,

difficult because most deaths occur in the community, often in remote rural areas. [3] In

Africa, an estimated 80% of all malaria-related deaths occur without the individual having

reached a health-centre or hospital. [4] Therefore, data derived from hospitalized

patients may not be representative of the overall pattern of disease in the community. [3,

4, 8] Nonetheless, hospital-based [25, 27, 67-70] and broader epidemiological studies [3,

8, 34] have done much to define the clinical spectrum of disease in different contexts, to

validate key prognostic markers and to identify factors that may modulate the clinical

manifestations, severity and mortality risk of P. falciparum infection.

Descriptive studies of severe paediatric malaria from Africa have consistently identified

impairment of consciousness and respiratory distress or metabolic acidosis as the key

prognostic indicators for risk of death. [27, 67-72] A study by Marsh et al. of 1 844

children hospitalized with malaria in Kenya demonstrated that, of the 64 (3.5%) who

died, 84.4% either had impaired consciousness (which was associated with a case

fatality rate of 11.9%), respiratory distress (case fatality rate of 13.9%) or both. [27]

12

Although severe anaemia has long been defined as a common manifestation of severe

malaria, its acute mortality was low in this (case fatality rate of 1.3%) and other studies.

[27, 67, 68] However, the very high prevalence of malaria-related anaemia in the

community may mean that hospital case fatality rates may underestimate its contribution

to the overall malaria-related mortality burden [3]

Impaired consciousness, the cardinal manifestation of cerebral malaria, is defined using

the standardised Blantyre coma score (BCS) which grades motor, verbal responses and

eye movements. [3, 24, 27] Scores of �4 and �2 out of a total of 5 indicate impaired

consciousness and coma, respectively, and have been validated as powerful indicators

of the risk of death. [3, 24, 25, 27] Autopsy studies [73, 74] and more recently, retinal

angiography, [75] suggest that the pathogenesis of cerebral malaria relates

predominantly to impaired microvascular circulation and perfusion, probably secondary

to erythrocyte sequestration and likely to exert effects via ischaemia and hypoxia in

cerebral tissue. Retinal examination provides a means by which the clinician can

effectively directly visualize changes related to the intracerebral microcirculation. It has

been proposed that retinal appearances could provide a means by which severity and

prognosis can be further stratified in severe malaria. [76]

Whilst impaired consciousness has long been recognized as a marker of life-threatening

disease, understanding of the pathological significance and mechanisms underlying

respiratory distress has been a more recent development. [25, 26, 77] This simple

clinical marker now appears to have similar prognostic significance to impaired

consciousness and its pathogenic antecedents have been the subject of intense interest.

[26, 78, 79] In children, respiratory distress appears closely linked to the development of

metabolic acidosis and hyperlactatemia (which in themselves may be even more

sensitive indicators of an adverse prognosis). [25, 68, 79] Their aetiology and

pathogenesis are complex and poorly understood. However, they may reflect tissue

hypoxia resulting from microvascular sequestration of parasitized erythrocytes,

hypovolaemia, products of parasite metabolism and/or neuroendocrine dysregulation of

glucose metabolism. [25, 27, 80, 81] Both are easily defined using conventional

biochemical assays.

13

1.5.3. Malaria in pregnancy

The increased susceptibility of pregnant women to malarial infection may reflect a strain-

specific loss of immunity secondary to pregnancy-related immunosuppression. [82] In

low transmission areas pregnant women are at higher risk of rapidly developing severe

manifestations of disease, especially hypoglycemia, and face risks of an associated high

foetal and maternal mortality. [24]

In hyper- and holo-endemic areas women are less likely to develop severe malaria and

even symptomatic infection may be rare. However, asymptomatic placental infection is

common. [38] Plasmodium falciparum parasites exhibit a predilection for intense

sequestration in the placenta, leading to placental insufficiency and IUGR. [82] In areas

of the most intense transmission this results in attributable reductions in birth-weight of

approximately 170g. [9] However, the effect appears to be limited to primigravidaeae

only. [38] P. vivax may also contribute to reductions in birth weight although this is less

well-characterised and the effect is probably not as pronounced as that with P.

falciparum. [38]

Asymptomatic parasitaemia in pregnant women also contributes to maternal anaemia.

This may increase the risks of maternal mortality by 35% when moderate anaemia (Hb

40-80g/L) is present and by 350% when there is severe anaemia (Hb <40g/L) [9]

Plasmodium vivax also contributes to maternal anaemia though the magnitude of its

attributable effect is less clear.

1.6. Antimalarial therapy: Drug development and evaluation

1.6.1. Historical background

Until recently, the origins of modern antimalarial pharmacology could be traced directly

to use of the juice of Cinchona bark amongst native populations of South America,

following which quinine was first recognized by Europeans as a viable treatment for

malaria in the 17th century. [9] Subsequently, pharmaceutical grade preparations of

quinine were developed formulated as a variety of salts for oral or parenteral

administration. However the use of synthetic compounds only dates back to 1891, when

Paul Ehrlich, having noted the affinity of methylene blue dye for structural components of

14

the malaria parasite, attempted using it therapeutically in patients with malaria infection.

[83] This compound had generally weak antimalarial activity and was soon superceded

by various synthetic quinolines, designed with chemical structures based on the quinine

prototype. [84] These included numerous aminoquinoline drugs developed in Germany

during the first half of the 20th century. Many of these were abandoned early in their

development due to their poor toxicity profiles. Others, including mepacrine and

sontoquine were used extensively, especially during World War II in spite of significant

problems with side-effects. [84] However, others, including CQ and later amodiaquine

(AQ), proved highly effective and to have acceptable safety profiles. [84] They also had

the advantage of being able to be cheaply manufactured in large quantities.

Chloroquine, first developed in 1934 and licensed in 1936, proved particularly

successful, becoming the most widely used antimalarial in the post-war period. [84]

While the first aminoquinoline drugs were being developed, the anti-plasmodial activity

of the sulphonamide drug-class was recognized in 1937. However, they did not achieve

widespread use until the development of sulphadoxine (SDOX), which had a longer half-

life and improved toxicity profile compared to its predecessors. The sulphonomides were

used in combination with dihydrofolate reductase (DHFR) inhibitors including

pyrimethamine (Daraprim® and Fansidar®) and proguanil (Paludrine®). [84] In China,

other 4-aminoquinoline-like compounds including piperaquine (PQ) and naphthoquine

and the closely-related Mannich Base derivative, pyronaridine were developed in the

1960’s. Of these, PQ (which had also been simultaneously discovered and

manufactured by Roche pharmaceuticals in France) was used very widely in mass-

treatment malaria eradication campaigns. [85]

The advent of CQ resistance in South-East Asia, and American military involvement in

the Vietnam War added impetus to the development of alternative antimalarial

compounds. The re-establishment of the US Army Malaria Research Program in 1963

lead to the aryl amino alcohol, mefloquine (MQ), becoming its drug of choice for CQ-

resistant malaria. [84] However, its high manufacturing cost made it prohibitively

expensive for much of the tropical world. Although still now used extensively in some

parts of Asia and South America (in combination with artesunate) where it maintains

high efficacy, its use is compromised by issues of toxicity and tolerability (vomiting in

children and neurological side-effects in adults) and significant resistance has developed

15

very rapidly following its introduction in South-East Asia. [86] Chemically related to MQ,

halofantrine was widely used particularly in Africa but its poor toxicity profile (including

significant cardiotoxicity) has curtailed its use. [87] A relatively recently developed

combination of the antibiotic atovaquone with the biguanide, proguanil (Malarone ®) is

too expensive to have found use outside residents or travelers from developed

countries. Other antibiotic drugs (including tetracycline/doxycycline, trimethoprim,

chloramphenicol and clindamycin) have generally weak antimalarial activity, limiting their

use to prophylaxis or adjunctive therapy alongside another agent. [9, 84] The demise of

CQ from widespread resistance saw the anti-folate combination SP become the most

practical alternative to CQ for routine antimalarial treatment in many areas of the

developing world. More recently an alternative cheap anti-folate combination comprised

of dapsone and chlorproguanil (“Lapdap”) has been developed [88], but it was recently

withdrawn because of haematological toxicity. [89, 90]

It is now clear that a number of plant-derived traditional medicines have antimalarial

properties. Their use may have pre-dated that of Cinchona bark in some parts of the

world. In particular, ancient Chinese texts clearly detail the use of an infusion of the

medicinal herb Qing Hao Su (Milk wormwood or Artemisia annua) for the treatment of

periodic fevers as early as 340 AD. [91] A push to re-investigate traditional therapies

initiated in China during Mao Tse Dung’s Cultural Revolution lead to the discovery of the

potent anti-plasmodial activity of the plant extract in vitro. Identification of its active

component (artemisinin) lead to the first empirical human trials in China during the 1970s

[92], followed by development of a range of pharmaceutical preparations (including

artemether, artesunate and dihydroartemisinin) by synthetic derivatisation of raw

artemisinin (QHS). [84] The first English-language publications regarding clinical efficacy

appeared in 1982 [93, 94] and clinical trials demonstrating efficacy against multi-drug

resistant strains of P. falciparum in Thailand were published in 1991. [95] In particular,

the efficacy of the combination of artesunate (ARTS) and MQ for uncomplicated malaria

[96] appeared to be excellent, with the introduction of this combination coinciding with a

reversal in the trend towards worsening MQ resistance in the area. [97] Susceptibility to

MQ has subsequently remained high in this area despite use of the ARTS-MQ

combination for many years. [97]

16

Current proponents of combination therapy now favour the use of an artemisinin in

combination with a second longer-acting drug (artemisinin combination therapy: ACT),

as a strategy to limit the development of parasite resistance. [19] This has now been

endorsed by the World Health Organization (WHO) as the strategy of choice for

treatment of uncomplicated P. falciparum malaria. [64] However, it should be

remembered that the arrest of worsening MQ resistance observed at the time the ARTS-

MQ combination was introduced effectively represents ecological data. [97] Therefore, it

constitutes a relatively low level of evidence and such an effect has not been

demonstrated when artemisinins have been combined with other conventional

antimalarials (such as CQ and SP). [60] Also, because of issues of cost, intrinsic

efficacy, tolerability/toxicity and dosing convenience, significant uncertainty still exists

regarding the optimal partner drug to use alongside the artemisinin. [98, 99] At present,

only one co-formulated ACT (artemether-lumefantrine) is endorsed by WHO for use in

areas of multi-drug resistance and in Africa. [64] Other promising combinations available

or being developed as co-formulations include dihydroartemisinin-piperaquine (DHA-

PQ), ARTS-MQ, ARTS-AQ and artemisinin-naphthoquine. [99]

1.6.2. Pharmacokinetic considerations

A sound understanding of the pharmacokinetic (PK) disposition of antimalarial drugs is

desirable to predict and optimise each drug’s therapeutic potential. Drug absorption,

metabolism and excretion, may depend on a number of factors including age, sex,

ethnicity, host genetics, the effects of pregnancy and of acute disease itself. [9, 100-102]

Therefore, PK studies should be conducted in suitably representative human

populations. In the case of malaria, studies should ideally be performed in children in

endemic areas (who constitute the bulk of those affected by the disease) and have

malarial illness (which might impact on PK disposition). [103] This presents a number of

technical and logistical challenges. Firstly, in areas of the developing world,

infrastructure necessary to conduct clinical trials and to properly process and store

relevant blood samples may be poor. [103] Secondly, the number and volume of blood

samples taken is subject to ethical constraints, particularly in small children. Therefore,

generating “rich” concentration-time datasets is often difficult. An approach to the latter

problem is to conduct analyses according to a population PK approach, whereby a

relatively small number of samples is taken from a large number of subjects and the

pooled population data analysed. [103]

17

When considering the pharmacokinetics of anti-microbial drugs, the essential

relationship between drug concentration and clinical effect exists at the level of the body

tissues in which the pathogen resides. Fortunately, in the case of malaria, the infected

tissue is blood, which is easily sampled in human subjects. Drugs will partition between

the erythrocytes and plasma. Most, however, are likely to rapidly attain a steady-state

equilibrium, meaning that measuring drug in plasma demonstrates a concentration-time

relationship parallel to that in whole blood or erythrocytes. Therefore, assays of drug

concentration in plasma are considered appropriate, also having technical and logistical

advantages for samples that must be processed, stored and transported from remote

tropical locations. An additional consideration is that for extensively protein-bound

antimalarial drugs such as quinine, it is the free (non-protein-bound) drug fraction that is

primarily responsible for clinical potency. [104-106] Therefore, situations that can alter

the proportion of protein-bound: non-bound drug may also impact on pharmacodynamic

outcomes including toxicity. [105]

In antimalarial pharmacology, the clinically important PK parameters include measures

of the rate of absorption, rates of elimination, volume of distribution (Vd) and clearance

(CL). Compartmental modeling can be applied to concentration-time data and a number

of computer software packages are available to do this. [107] These apply a model

containing a predefined number of compartments analogous to the physiological

processes of drug disposition. [107] For instance drugs can be considered as being

absorbed into an initial (or central) compartment (e.g. the intravascular space) before

being redistributed to a larger compartment (e.g. tissues). Fitting models to

concentration-time data assigns rate constants for the time-course of movement of drug

from one compartment to another (including absorption of drug into the first

compartment, and both “forward” and “backward” rate constants for passage between

subsequent compartments).

A simple 2-compartment model, will consist of a central compartment and a second,

larger (usually), distribution compartment. The rate of absorption of an orally

administered drug from the gut into the central compartment will be defined by an

absorption rate constant ka from which can be derived an absorption half-life (t½abs).

Another constant will define the rate of movement of drug from the first to the second

18

compartment (k12) and back again (k21). Elimination constants will be derived for the

initial distribution phase of the concentration time profile (�1) and the terminal elimination

phase (�2). From these are derived the distribution and elimination half-lives t½e �1 and

t½ e�2. The model will also derive volumes of distribution (Vd) for the central (V1) and

distribution compartments (V2) and the rate of clearance (the volume of blood that is

cleared of the drug per unit time). For drugs that are not injected directly into the

vascular space (oral or rectal administration) Vd and CL may be adjusted according to a

bioavailability function, F (i.e. Vd/F and CL/F). For sufficiently rich datasets, a two-step

approach can be used such that a model is fitted to each individual patient

concentration-time dataset and individual PK parameters derived. The derived

parameters from each patient are then pooled and measures of centrality and variance

derived for each parameter in the population sample. These parameters can then be put

back into the model to simulate hypothetical concentration-time curves with a variety of

dosing regimens. [107]

In addition to compartmental methods, non-compartmental analysis uses simple

mathematical analysis of the individual concentration-time plot to derive parameters

such as the maximal concentration achieved (Cmax), the time to maximal concentration

(tmax), the area under the curve (AUC) and a derivation of terminal elimination half life

(t½e) using linear regression of the log-linear portion of the terminal phase of the

concentration-time curve to determine its slope. [107]

Because assessment of outcomes can extend well beyond the time over which

treatment is given, the relationship between PK parameters and clinical response is

difficult to define for many antimalarial drugs. However, in general, consistent and rapid

absorption that achieves optimal parasiticidal drug concentrations is desirable, especially

for treatments of severe malaria. Radical cure of P. falciparum infections probably

requires that effective therapeutic concentrations of drug are maintained long enough to

kill the entire body parasite burden, including the residuum that remains long after the

patient has defervesced and parasites are no longer detectable on the periperhal blood

film (see Figure 1.1). [9, 108] For rapidly parasitocidal agents like artemisinin derivatives,

this is probably means a minimum of 7 days from the initiation of treatment. However, for

slower acting drugs, where the slope of the parasite clearance curve (see Figure 1.1) is

lesser, much longer durations (many weeks) may be necessary. Therefore, the t½e�2 of

19

the drug is important in determining the drug’s likely therapeutic efficacy when given as a

short course. In this example, the t½e�2 will also be important in determining the “post-

treatment prophylactic” effect: the duration of protection from subsequent re-infections

following a course of therapy. This may be important in highly epidemic regions,

especially for interventions such as intermittent presumptive treatment (IPT). [109]

It has been observed that drugs with the slowest elimination profiles are also those

prone to the most rapid development of resistance. [50] For example MQ, which has a

t½e�2 of approximately 14 days, suffered widespread clinically significant resistance

within 3 years of its introduction in the Thai-Burmese border area. [86, 97] By contrast,

quinine (t½e�2 8 h) has been spared significant resistance despite over 300 years of use.

This difference may reflect the higher mathematical probability of parasites being

exposed to intermediate concentrations of drug (i.e. those most likely to select resistant

parasites), if the drug has a slow elimination profile. [50] When widely deployed, a large

proportion of the population will have intermediate blood concentrations of drug at any

one time. [49] New sporozoite inoculations have a high likelihood of exposing parasites

to intermediate concentrations of residual drug remaining from a previous antimalarial

treatment. [49] For these reasons a drug’s t½e�2 has been recognized as an important

determinant of the propensity to resistance. [50] Most partner drugs used in artemisinin

combination therapy have a very much longer t½e�2 than the artemisinin derivative.

Therefore, although they may offer protection from resistance to the artemisinin

component, they may not themselves receive much protection from the artemisinin. [50]

The case of quinine represents an example of how a detailed analysis of PK disposition

has enabled design of more rational dosing schedules. These aim to rapidly achieve

optimal therapeutic concentrations whilst minimizing the risk of serious toxicity.

Knowledge of the disposition of quinine in severe malaria and in renal failure has

enabled suitable dose adjustments to be recommended in these circumstances. [101,

110, 111] However, quinine is distinguished by a reasonably well-defined therapeutic

margin, where a threshold for toxic concentrations has been well-characterised, enabling

limits to guide rational dosing design based on PK data. [9] The situation is somewhat

different for other antimalarial drug classes such as the artemisinin derivatives. These

have a very wide therapeutic margin; routine administration achieves drug

concentrations many orders of magnitude higher than the parasite inhibitory

20

concentrations but without resulting in toxicity. [112] Whether a toxic threshold for

artemisinin concentrations exists in humans is unclear as is the nature of the drug

concentration-therapeutic response relationship. Therefore, it is difficult to rationally

design optimal dosing schedules for these drugs based on PK data alone. In many

cases dosing schedules have been determined empirically.

1.6.3. Pharmacodynamic assessment

The most important pharmacodynamic (PD) outcome for treatments of uncomplicated

malaria is definitive cure of the infection. This requires a complete eradication of asexual

parasites from the body. Initial treatment may rapidly reduce parasitaemia below the

threshold for detection by microscopy but if a residuum of viable parasites persists below

this threshold, recrudescence of infection is likely to occur (see Figure 1.1). [9, 108]

Therefore, recrudescence usually follows a period of resolution of fever and detectable

parasitaemia, often some weeks after the initial treatment. For this reason, standard in

vivo assessments of cure require a lengthy follow-up in order to detect recrudescent

infections. [108]

Standardised methodology for in vivo assessment of therapeutic response has been

defined by WHO. [113] This requires a follow-up period of 28 days, including weekly

assessments for the presence of fever and parasitaemia. However, longer follow-up

periods may be required for drugs with longer half-lives to allow for the suppressive

effect of residual drug concentrations that may delay recrudescence. [108] In addition,

recrudescent infections must be distinguished from re-infections. Only the former can be

considered genuine treatment failures. Distinguishing the two can be attempted using

genotypic characterization of parasites from the initial infection and the relapsed

infection, using PCR and restriction fragment length polymorphism directed at the

parasite merozoite surface proteins 1 and 2 (MSP1 and MSP2) and glutamine-rich

protein (GLURP) gene loci according to standard methods. [59] Definitions of therapeutic

outcomes for P. falciparum infections according to WHO methodology [113] can be

summarised as follows:

1. Early treatment failure (ETF): clinical deterioration or inadequate parasitological

response in the first 4 days

21

2. Late treatment failure-clinical (LCF): Parasitaemia with fever occurring between

days 4 and 28.

3. Late treatment failure-parasitological (LPF): Parasitaemia without fever occurring

between days 4 and 28.

4. Adequate clinical and parasitological response (ACPR): Absence of parasitaemia

occurring between days 4 and 28 (irrespective of axillary temperature) without

previously meeting any of the criteria of ETF, LCF or LPF.

The methodologies and definitions above replaced a previous WHO classification where

treatment failure was classified according to three grades: RI (Low-grade: following initial

resolution of symptoms and parasite clearance, recrudescence occurs between day 7

and 28), RII (High-grade: Reduction of parasitaemia by >75% but failure to clear

parasites by 7 days) and RIII (Complete: Parasitaemia does not fall by >75% within 48

h). [9]

The rate of reduction in parasite density has also been used as a PD outcome,

particularly in studies of severe malaria. [108] This usually requires taking a thick blood

film relatively frequently (typically every 4-6 h) and counting the number of asexual

parasites per leucocyte at each time point. The number of parasites per µl of blood (the

parasite density) is calculated by multiplying this figure by the known or assumed

absolute leucocyte count. A plot of parasite density versus time can thus be constructed

to represent the kinetics of parasite clearance and the following parameters derived:

1. Parasite clearance time (PCT): Time taken until parasites become undetectable

on the peripheral blood thick blood film. Usually, an additional requirement is that

subsequent blood films are negative for parasitaemia for at least the next 24 h

(also called the PCT�). [108]

2. The time taken until 90% and 50% reductions in parasitaemia have been

achieved (PCT90 and PCT50). These can be derived by simple linear

interpolation of the parasite density-time plot. [108]

3. The parasite density at 12 and 24 h expressed as a percentage of the baseline

parasitaemia (PC%12 and PC%24). [108]

22

Parasite clearance has limitations as a PD marker. Firstly, microscopy of a thick blood

film is relatively insensitive and subject to error, and ceases to define the decline in

parasitaemia once the parasite burden falls below the limits of detection (10-50 parasites

per µL: equivalent to a total parasite burden of approximately 108). Therefore, the

majority of the decline in parasitaemia is not defined using this technique. Secondly,

microscopy detects ring and early trophozoite forms that predominate in the peripheral

circulation. The more pathogenic mature forms that sequester in tissues will not be

detected. Therefore, a patient with a high burden of sequestered parasitaemia may have

a falsely reassuring reduction in peripheral parasitaemia if treatment clears peripheral

parasitaemia faster than the sequestered parasite load. [9, 108] Markers of parasite

clearance therefore represent only surrogate markers of clinical response in severe

malaria and it is unclear to what extent markers of parasite clearance correlate with the

most clinically relevant endpoint (mortality). This is especially so because different drug

classes may have differential activity throughout the parasite life cycle. Nonetheless

markers of parasite clearance remain a useful and widely-used surrogate marker for

assessing early parasite clearance in therapies for severe malaria. [108] When parasite

clearance times for a particular agent become longer over time, they may also be useful

surrogate markers of emerging parasite resistance. [114]

Mortality and the incidence of permanent neurological sequelae represent the most

robust clinical endpoints when assessing response to severe malaria. [115] However, in

practice, comparative drug trials using mortality as an endpoint requires huge sample

sizes for adequate statistical power. Although one large multi-centre study has managed

to do this successfully [116], historically, few studies of severe malaria have been

adequately powered to assess mortality as an endpoint, and of necessity the best

available surrogate markers (those relating to parasite clearance) have been used. [117]

The use of parasite clearance as an endpoint in comparative trials is probably more

robust when comparing either the same drug (e.g. given by different routes or regimens)

or at least drugs belonging to the same drug class (which would be expected to have the

same differential activity by stage of the parasite life-cycle). In this case, differences in

parasite clearance between the two regimens are likely to reflect a genuine therapeutic

advantage (e.g. due to PK factors rather than differences in intrinsic activity of the drug).

23

Fever clearance time (FCT) is also used in a similar manner to parasite clearance as a

PD marker. [108] Unfortunately FCT has been poorly standardised; variability may occur

due to the mode of body temperature measurement (axillary, oral, rectal or tympanic),

the frequency of measurements, the use of antipyretic medications, the cut-off definition

of “normal temperature” and the duration of temperature monitoring. However, intra-trial

differences in fever clearance time between drug treatments may be of interest if fever

clearance has been rigorously defined. A common methodology involves determining

axillary temperature every 4-6 h. FCT is defined as the time in hours to achieve and

maintain a temperature below 37.5 °C for a full 24 h of monitoring (FCT�). [108]

1.6.4. Toxicity and tolerability

Assessment of the safety, toxicity and tolerability of antimalarial drugs is an important

aspect of their clinical evaluation. However, this is also subject to important limitations.

Distinguishing reported side-effects from the clinical manifestations of malaria itself

(including gastrointestinal, neurological, haematological, hepatic or nephrologic) is often

difficult. Assessment of toxicity in clinical trials is generally determined by focused side-

effect questionnaires and laboratory investigations (that may be limited in availability in

the developing world). Data from animal studies, human volunteer studies and clinical

trials of chemically related drug compounds may help guide this process. For example

drugs from the quinoline class have long been recognised as having cell-membrane

destabilizing properties that may predispose to cardiac arrhythmias. [118] This has been

demonstrated particularly with halofantrine, but also MQ, quinine and CQ (in overdose).

[118] Therefore, assessment of new quinoline drugs should include monitoring for sub-

clinical cardiac toxicity by use of electrocardiographic assessment and assessment of

QT prolongation.

Clinical investigations focused on mechanistically predictable drug toxicity are, however,

less well-suited to detecting idiosyncratic events that may occur as rare but serious

events. These are usually not related to predictable physiological processes. [119] An

example would be anaphylaxis occurring as an isolated event in occasional susceptible

individuals. Detecting such events requires systems of post-marketing surveillance that

are generally absent in developing countries. [102] Clinical trial data collected during the

early stages of drug evaluation (Phase-I through to Phase-III) is limited to a modest

number of patients may not detect these less common but potentially serious side-

24

effects. [102] This emphasizes the importance of post-marketing pharmacovigilance

systems and of large Phase-IV studies that include appropriate designs for safety

monitoring for severe adverse events, including the use of autonomous drug safety

review panels to independently assess possible adverse events and their likely

relationship to the study drug. [102]

1.6.5. Operational factors

A drug’s efficacy can be defined as its intrinsic therapeutic benefit when administered

under ideal conditions such as during a formal clinical trial. The extent to which efficacy

translates to true effectiveness (when administered routinely through the health system)

are dependent on numerous operational factors. [120] These require careful

consideration during the development of antimalarial drugs and in planning the nature of

their deployment in the health system. In particular, compliance with therapy is a vital

consideration. [121] Factors that may impact on compliance include the duration of

therapy, complexity of the dosing regimen, the incidence of unpleasant side-effects,

patient preferences and socio-cultural attitudes towards different routes of

administration. Clearly drugs that can be administered using simple short-course

regimens are most appropriate. However, some studies have shown that compliance

with even a 3-day course can be as low as 14%. [120] There is a relative paucity of data

that examines the determinants of compliance and the manner in which socio-cultural

factors determine preferences for different treatment modalities. [121]

1.7. Antimalarial therapy: Drug classes The basic clinical pharmacology of the drugs evaluated in studies contained in this thesis

is detailed briefly in this section. Most currently used antimalarial drugs can be classified

as belonging to one of seven drug classes. These include the 4-aminoquinolines, 8-

aminquinolines, aryl amino-alcohols, artemisinin derivatives, anti-folates, inhibitors of the

respiratory chain and antibiotics. [122] Drugs from the 8-aminoquinoline, respiratory

chain inhibitor and antibiotic classes were not evaluated in any of the studies in this

thesis. Therefore, they are not discussed here.

25

4-aminoquinoline and related drugs

1.7.1. Chloroquine

Chloroquine has probably been the most-used of any antimalarial drug. [84] It has the

advantages of very low cost, wide availability, relatively good tolerability and feasible

administration as short course oral therapy.

1.7.1.1. Pharmacokinetics

Despite extensive use over the last 60 years, the PK of this prototypical antimalarial

remain poorly defined. In particular, a valid measure of its t½e has been problematic.

[123] Chloroquine has a very large Vd and appears to exhibit multi-exponential

elimination kinetics that probably reflect ongoing redistribution from peripheral tissues

rather than genuine elimination processes. [123, 124] Early PK studies were limited by

poor assay sensitivity and sampling durations have varied widely between studies, such

that t½es have ranged from 2 to 60 days. [123, 124] Most have been conducted in small

numbers of adults and have shown substantial inter-individual variability in calculated PK

parameters. [123, 124]

Chloroquine has an active metabolite, desethylchloroquine (DECQ), which is not

considered an important contributor to its initial therapeutic effect. [123]

1.7.1.2. Safety and tolerability

Chloroquine is relatively non-toxic and well-tolerated at usual therapeutic doses.

However, it commonly causes an unpleasant side-effect of pruritus (usually in dark-

skinned individuals) that may compromise compliance. [125] It is also bitter-tasting

making it difficult to administer to young children. Overdosing may lead to potentially

fatal ventricular arrhythmia characterized by a sine wave electrocardiogram. Deliberate

overdose of CQ therefore constitutes a common cause of death in some developing

countries. [126] It can be administered orally as either tablet or syrups. Although a

parenteral preparation was developed, the high initial drug concentrations associated

with i.v. or i.m. injection were associated with acute hypotension. [127]

1.7.1.3. Efficacy

Decreased in vitro susceptibility of P. falciparum is now virtually ubiquitous throughout all

endemic areas of the world. [64] However, in highly endemic regions, in vivo response

26

may remain acceptable, due to the additive benefits of pre-existing immunity on

therapeutic response. [61] It remains the drug of choice for treatment of P. vivax where

transmission of this parasite occurs, although CQ-resistant P. vivax is now a problem in

some parts of the world, including the island of Papua. [128, 129]

1.7.2. Piperaquine

Piperaquine is a 4-amino-bisquinoline currently undergoing a resurgence. [85] Originally

used as mono-therapy in the 1960’s and 1970’s for mass-treatment and prophylaxis in

China, more recently it has generated interest as a component of ACT. Piperaquine co-

formulated with DHA as Duo-cotecxin™ and Eurartekin™ (formerly Artekin™ or

Artekin2™) may be an attractive alternative ACT. The combination has now been shown

to be safe and highly effective in a number of studies. [130-134] In PNG, DHA-PQ could

replace the combination of CQ and SP which is increasingly associated with treatment

failure. [59]


Despite its long and relatively heavy use [85], the PK properties of PQ have been

investigated in only small numbers of healthy Caucasian or South-East Asian volunteers.

[135-137] and in one study of patients with malaria. [138] There is a need to define its

disposition in other ethnic groups and to generate more data from children who are sick

with malaria.


Piperaquine has not undergone systematic and rigorous safety evaluation of the sort

usual with modern approaches to drug development. Safety data at all levels (including

in vitro, animal and human data) are lacking. [85] The drug has not been registered with

regulatory bodies such as the FDA nor endorsed by WHO. However, it is notable that no

significant adverse events were reported during extensive use in China (including an

estimated 1.2 million treatment doses). [85] Also, available in vitro and animal data

suggests that its therapeutic margin is wider than that of the chemically related drug,

CQ. [85, 139] A reasonable amount of safety data has now been accrued from clinical

trials. [130-134] In human trials, rates of reported side-effects have ranged from 2% to

21%. [131, 132, 140] These have all been reported as minor and mostly related to the

gastrointestinal tract (nausea, abdominal pain and diarrhoea), with transient nausea

reported in 2%-5% of subjects. Prolongation of the QTc interval and hypotension do not

27

occur with usual therapeutic doses suggesting that significant cardiotoxicity is unlikely.

[130] One child enrolled in a study of uncomplicated malaria became drowsy and

eventually died 2 days after completing a course of DHA-PQ. [131] Although the authors

felt this was likely to be due to a concomitant non-malarial systemic infection, an adverse

reaction to the drug could not be excluded.

1.7.2.3. Efficacy

Published in vivo clinical evaluations of PQ have evaluated co-formulations with DHA

(DHA-PQ: Artekin™, Eurartekin™ or Duo-cotecxin™) or DHA and trimethoprim plus

primaquine (“CV8™”). Most have been performed in South-East Asian patients [130-

134] but data has also been derived from Africa and West Papua province of Indonesia.

[140, 141] All studies have demonstrated 28-day in vivo cure rates of >95%. Although

PQ resistance was well-described following mass-drug administration in China, parasite

populations elsewhere in the world have not had significant exposure to PQ. Although

the possibility of cross-resistance existing between different 4-aminoquinoline

compounds has not been closely examined, pre-existing CQ resistance does not appear

to impact on the clinical efficacy of PQ. In one study of P. falciparum parasites collected

in Cameroon, PQ IC50s in CQ-resistant parasite lines were not significantly higher than

those for CQ-sensitive strains. [142] However more recent data from parasite strains

collected collected in West Papua, Indonesia have demonstrated a positive correlation

between IC50s for CQ and PQ in both P. falciparum and P. vivax. [143] It is reassuring,

therefore, that in vivo cure rates remain high in Thailand, Vietnam, Laos and Cambodia

where multi-drug resistant strains of P. falciparum are prevalent. [131-134, 138] The

DHA-PQ combination also appears to have excellent activity against P. vivax, including

in areas where CQ-resistant P. vivax is prevalent. [141]

Aryl amino-alcohols:

1.7.3. Lumefantrine

Lumefantrine (also previously known as benflumetol) was also originally developed in

China. It is available only as a fixed co-formulation with artemether (marketed as

“Coartem ™” or “Riamet™”) which is the only co-formulated ACT currently approved by

regulatory bodies such as the FDA and TGA and endorsed by WHO. [64] However, its

complex dosing regimen may be problematic in some settings. [144] Although marketed

in the developed world at a relatively high cost, the manufacturer has recently made it

28

available to developing countries at cost-price and additional subsidies have made it

more affordable at approximately US$1 per course.


Lumefantrine (LUM) is a highly lipophillic compound and its absorption is poor and

variable. [145, 146] It should, therefore, be administered with dietary fat (such as full-

cream milk) in order to improve its absorption, though only relatively small quantities (30

mls of full-cream milk or equivalent) are necessary. [147] Its t½eß2 is thought to be

approximately 60 h but it has not been well-defined in children. [145, 146] This relatively

short half-life means that most regimens employ twice daily dosing in order to achieve

and maintain therapeutic concentrations over the first few days. [148] It also means that

this drug is unlikely to have a significant period of post-treatment parasite suppression

and would therefore seems less suited to IPT or use as prophylaxis. [109] Most PK

studies of LUM have been conducted in South-East Asian or Caucasian adult subjects.

[145, 146, 148-150] Recently acquired PK data suggest that its elimination may be even

more rapid during pregnancy. [151]


Lumefantine is chemically related to MQ and halofantrine. However, extensive

experience with LUM suggests that it is very well-tolerated and electrocardiographic

studies have shown no lengthening of the QTc interval to suggest sub-clinical

cardiotoxicity with therapeutic doses. [152]

1.7.3.3. Efficacy

Artemether-lumefantrine (AL) was originally used as a twice daily 2-day regimen.

However, cure rates of <80% were recorded in non-immune subjects. [153] This

probably reflected the short half-lives of both artemether and LUM causing drug

concentrations insufficient to eradicate the parasite residuum. Clinical cure correlates

well with either the measured concentration of LUM at day 7 or the calculated LUM AUC.

[63] A 6-dose 3-day regimen is now recommended. [64] This has demonstrated cure

rates of >95% in Africa, including in paediatric populations. [153] However, a study in

Cambodia showed a 70% cure-rate even with the 3-day regimen. [154] Because LUM

AUC is increased significantly by administration with fatty food, it is probable that cure

rates have been compromised when this has not been the case. [154] Poor clinical

29

response has also been linked to an increase in the copy number of the PfMDR1 gene.

[63] Also, an increased prevalence of the PfMDR1 86N genotype has been observed in

recurrent P. falciparum infections following treatment with AL, suggesting selection of

less susceptible parasite strains. [155] It is possible that even minor, sequential

decreases in drug susceptibility may seriously compromise the efficacy of this drug

combination in the future. [156] Although studies in Africa have indicated adequate

adherence to the 6-dose regimen [157, 158], concerns remain that the need to

administer each dose with fatty food, will prove impractical and compromise compliance

in many parts of the developing world.

Anti-folates:

1.7.4. Sulphadoxine-pyrimethamine

Sulphadoxine-pyrimethamine (SP) is a fixed combination of two dihydrofolate reductase

inhibitors. It is cheap to manufacture and is conveniently administered as a single dose.

It was the most affordable alternative to CQ and was adopted as first-line treatment for

uncomplicated malaria in much of Africa following the emergence of CQ resistance. [84,

122]


Both sulphadoxine (SDOX) and pyrimethamine (PYR) are well-absorbed following oral

administration and have first order elimination kinetics. [159] However, t½e of SDOX

(approximately 8 days) is longer than that of PYR (4 days). [159, 160]


Sulphadoxine-pyrimethamine is generally very well tolerated. However, rare,

idiosyncratic reactions to the sulphadoxine component may be very serious. These have

been reported when SP was used for prophylaxis (rather than case-management of

acute disease) and include severe, even fatal exfoliative dermatitis (estimated to have a

frequency of 1:7000) hepatitis, blood dyscrasias, methaemoglobinaemia and other

allergic reactions. [161] Pyrimethamine may also cause blood dyscrasias. The SP

combination is one of very few antimalarials for which there is adequate experience

regarding its safety in pregnancy. [102]

30

1.7.4.3. Efficacy

Although previously a highly effective drug-combination, resistance to SP has arisen

more rapidly than for CQ in most parts of the world. [64, 162, 163] A number of

mutations affecting the dihydrofolate reductase (DHFR) and dihydropteroate synthase

(DHPS) genes have been identified that, in concert, are associated with in vivo

resistance to PYR and SDOX, respectively. [54, 163] Combining ARTS with SP (ARTS-

SP) hastens asexual and gametocyte clearance and improves cure rates. [98] In some

settings this combination may be at least as effective as other ACTs (such as AL) even

in areas where parasite resistance to SP is increasing. [164-166] However, in general,

cure rates with the ARTS-SP combination remain relatively poor where SP resistance is

well-established. [98] Activity against P. vivax is generally weak and resistance is

common. [167] Sulphadoxine-pryimethamine has proved remarkably effective when

used as IPT for pregnant women (IPTp) and infants (IPTi) in Africa, despite high rates of

resistant P. falciparum in the areas where these studies were undertaken. [168-170]

Sulphadoxine-pyrimethamine has also been used successfully in combination with 4-

aminoquinolines (CQ and AQ) often with good cure rates in spite of established

resistance to either drug. [171-175]

The artemisinin derivatives:

1.7.5. Artesunate, dihydroartemisinin and artemether

All currently available artemisinin derivatives are manufactured from artemisinin which is

extracted from Artemisia annua. As well as artemisinin itself (QHS), synthetic

derivitisation of QHS produces artemether (ARM), dihdyroartemisinin (DHA) and

artesunate (ARTS). These all share the same basic chemical structure of artemisinin

with different chemical groups at the 2-keto position giving ARTS (hemisuccinate), DHA

(hydroxy) or ARM (methyl). [122]


The 2-keto position radical determines the solubility of each artemisinin derivative and

therefore influences its diffusion across mucosal membranes. The order of lipid solubility

(log10 oil:water partition coefficient at pH 7) is ARM (3.07)> DHA (2.6)> QHS (2.27) >

ARTS (0.25). [176] Artemether and ARTS are both metabolized to DHA by rapid

esteratic hydrolysis of ARTS (median t½e 3 min), or slower cytochrome P450-mediated

31

demethylation for ARM (median t½e 3.6 h). Dihydroartemisinin and QHS have a median

t½e of 0.9 h, and 2.6 h, respectively. [177] All four artemisinin drugs can be administered

orally or rectally. Artemisinin, ARM and ARTS can also be given i.m., but ARTS is the

only drug that can be administered i.v. [177] Oral bioavailability (F) is higher for the more

water-soluble ARTS (88%) [178] than for the lipid soluble QHS (32%) [179], DHA (16%)

[180] and ARM (43%). [181] Lipid solubility is also likely to influence both F and

absorption profile after rectal administration. The PK of artemisinin derivatives such as

ARTS are not influenced significantly by severity of infection [182] and dose adjustment

in patients with hepatic or renal impairment is not required.


The artemisinin derivatives appear to have a very wide therapeutic margin, being

generally extremely well-tolerated at usual therapeutic doses in humans. [183-185]

Reported adverse effects have included nausea, vomiting, bowel disturbance, abdominal

pain, headache and dizziness, symptoms that can result from malaria infection itself and

are usually mild and self-limiting. [186] Mild and reversible haematological and

electrocardiographic abnormalities, such as neutropenia and first-degree heart block, are

observed infrequently. [186] Animal toxicity studies using these compounds in supra-

therapeutic doses in rats and dogs demonstrated a characteristic fatal neurotoxicity

affecting predominantly the brainstem of animals receiving high doses over long periods.

[187, 188] This toxicity now appears to be dose-related and to be restricted to

compounds and dosing regimens that result in sustained high plasma concentrations of

the primary drug or its active metabolite DHA. [189] Therefore, it is likely that the depot-

like slow release of lipophillic artemisinin derivatives when administered by i.m. injection

facilitates this toxicity. [190] This theory is supported by data showing that mice fed

comparable oral doses of more rapidly cleared water soluble artemisinin derivatives do

not develop neurotoxicity. [191] Neurological side-effects such as ataxia, slurred speech

and hearing loss have also been reported in a small numbers of humans [192, 193] but

ascribing these to a drug side-effect rather than malarial disease itself is difficult. [194,

195] The limited penetration of artemisinin derivatives into the CSF also makes the

possibility of neurotoxicity at therapeutic doses less likely. [196] Nonetheless, concerns

still exist that neurotoxicity may manifest in certain vulnerable individuals, including

children, for whom the developing CNS may be at greater risk. [197] Ongoing safety

32

monitoring and attention to PK disposition are warranted in order that optimally safe and

effective regimens can be developed.

A large body of data from animal models (including rats, rabbits and monkeys) now

raises significant concerns about the reproductive toxicity of the artemisinins,

demonstrating toxicities including embryonic death, foetal resorption, limb dysgenesis

and cardiac defects following treatment with therapeutic doses in the first trimester. [198]

By contrast, birth outcomes have now been documented in >500 women exposed to

artemisinins during their pregnancy, and adverse events have not been reported. [199]

Nonetheless, given the strength of the animal data, artemisinin derivatives are

considered contraindicated in the first trimester of pregnancy. [198]

1.7.5.3. Efficacy

Each of the artemisinin derivatives is highly active against asexual forms of all the

Plasmodia that infect humans. The in vitro P. falciparum IC50 is similar for all 4 drugs

[177], and is a number of orders of magnitude less than plasma drug concentrations

achieved through routine therapeutic administration. The initial reduction in parasitaemia

is the most rapid of all available antimalarial drugs, and seems to be a consistent

feature, regardless of the derivative or dosing regimen used. [183-185] Given that they

have very short plasma half-lives, and that parasite clearance persists for many hours

after drug is no longer detectable in the circulation, it is likely that the equivalent of a

“post-antibiotic effect” exists, where there is persistent suppression of microbial growth

following limited exposure to an antimicrobial agent. The artemisinin derivatives are also

active against some stages of the sexual form of the parasites (stage 1 and 2

gametocytes) and can therefore reduce transmission rates. [200]

Their exact mechanism of action is unclear. An endoperoxide moiety, which is essential

for antimalarial activity, may cause destructive free-radical generation within the parasite

and, through the formation of covalent bonds, alter the function of key parasite proteins

including membrane transporters. [201] Recently, specific inhibition of an enzyme

essential for oxidative metabolism in the parasite, PfATPase6 has been postulated as a

mechanism of action. [202] Short courses (3–5 days) are associated with

recrudescence rates that are typically greater than 25%. [183, 184] Although such

33

recrudescences can be re-treated successfully with an artemisinin drug, this

phenomenon highlights the limitations of monotherapy.

It is possible that their extremely short half-lives will protect them from the development

of parasite resistance in the long term. [49] However, parasite resistance has been

successfully induced in the laboratory [203] and recently, a trend towards slower

parasite clearance times has been observed in Cambodia, raising concerns that this

might represent an early sign of the development of artemisinin resistance. [114]

However, it is not clear if this phenomenon represents genuine parasite resistance.

1.8. Antimalarial therapy: Management of specific malaria syndromes

1.8.1. Uncomplicated malaria

Ideally, treatment of malaria should be based on a microscopically-confirmed diagnosis

and parasite speciation. [9] However, in practice, neither microscopy nor rapid diagnostic

tests are available in many malaria-endemic countries. Febrile illness is usually treated

empirically with a “first-line” antimalarial treatment. [204] Given the lack of any

sufficiently sensitive or specific clinical features of malaria, and the relative safety and

tolerability of many antimalarial drugs, this is usually justified. [205] Therefore, “over-

treatment” of malaria is the rule and huge quantities of antimalarials are dispensed

worldwide. Although improved diagnostic services are desirable to improve this situation,

achieving this will be difficult in resource-poor settings in the short-term. In this context,

where routine empirical treatment is the norm, the ideal antimalarial should be cheap,

safe and efficacious against all Plasmodial species endemic to the area.

Initial assessment of patients to be treated for malaria should include careful

assessment for signs of severity to exclude the need for parenteral treatment and in-

patient monitoring. [9, 24] Provided this is not the case, uncomplicated malaria is usually

managed on an outpatient basis in most of the tropical world. Unfortunately, compliance

cannot be assured, an important factor that may significantly undermine treatment

effectiveness. [206] Ideally the treatment course should be as short as possible (3 days

or less), employ a simple dosing regimen (e.g. once daily), not require additional dietary

instructions (such as the need to consume fatty food at the time of each dose) and be

34

free of unpleasant side-effects. Even relatively minor “annoying” side-effects (such as

pruritus with CQ or nausea and dizziness with MQ) may substantially reduce compliance

with the full course of therapy.

Because most symptomatic malaria occurs in children living in highly endemic areas, it is

important that antimalarial treatment is available in forms appropriate for children.

Paediatric formulations (including syrups, granules and suppositories) may be

particularly useful in children for whom it is often difficult to administer oral medicines,

especially those that are unpleasant tasting.

The burden of antimalarial use in highly endemic areas also raises the issue of parasite

resistance. The advantages that longer half-life drugs carry in terms of suitability for

simpler short-course regimens must be balanced against a possibly higher susceptibility

to parasite resistance [50] which makes them less desirable from a public health

perspective.

Combination treatment may, in theory, reduce the risks of drug resistance developing

and this underlies the rationale for ACT, where the use of a second partner drug may at

least “protect” the artemisinin component from the development of resistance. [207]

However, substantial challenges exist in finding and developing suitable partner drugs.

For practical reasons antimalarials must be administered as short-course therapy.

Therefore, sustained antimalarial activity is necessary for many days after the treatment

course has been completed in order to eliminate residual parasitaemia and effect

definitive cure. This requires a good understanding of the drug’s elimination profile

(particularly the t½e�2) and the likely in vivo susceptibility of local parasite strains to the

drug (see Figure 1.1 below).

35

Time (days)

0 2 4 6 8 10 12 14

Dru

g co

ncne

ntra

tion

/ tot

al b

ody

para

site

bur

den

Figure 1.1 Theoretical relationship between drug concentration and clinical cure

(logarithmic scale). The solid line represents the concentration-time plot of an

antimalarial drug given as 3 daily doses. The diagonal dashed line represents the

decline in total body parasite burden during treatment. The horizontal dotted line

represents the in vivo parasite MIC. Definitive cure will be achieved if the length of

time that drug concentrations are maintained above the MIC (X in the Figure)

exceeds the time taken to eradicate the entire parasite burden (Z). In the example

above, drug treatment will have been effective.

X

MIC

Z

k�2

k�1

k�

P

36

It can be deduced from Figure1.1 that the chances of affecting a definitive cure will

depend on a number of factors including:

1. Parasitological factors:

a. The initial parasite burden at the time treatment is initiated (P)

b. The rapidity with which parasites are cleared by the drug i.e. the slope of

the parasite clearance curve (k�)

c. The inherent susceptibility of the parasite to the drug (represented by the

in vivo MIC)

2. Pharmacokinetic factors:

a. The bioavailability of the drug

b. The extent and rate of drug distribution i.e. the slope of the concentration

time plot during the early distribution phase (k�1)

c. The rate of drug elimination i.e. the slope of the concentration time plot

during the elimination phase (k�2)

3. The dosage regimen:

a. Individual dosage

b. Frequency of dosing

c. Duration of treatment

A rational approach to designing optimal antimalarial therapies (including those for use

as partner drugs for ACT) should take all of the above factors into account. In particular,

accurately defining the PK of an antimalarial, especially its elimination kinetics, is vital to

predicting its utility and designing optimal dosage regimens that can practicably be

administered as short courses in developing countries.

In summary, the qualities of the “ideal” antimalarial to partner with an artemisinin are

listed in the Table 1.1 (below).

37

Table 1.1 Qualities of the “ideal” antimalarial for outpatient treatment of

uncomplicated malaria (e.g. as a partner drug for use in ACT)

1. Affordable

2. Available in form easily administered to children

3. Efficacious against all local parasite species and strains

4. Without serious toxicity

5. Free of unpleasant side-effects that may compromise compliance

6. Safe in pregnancy

7. Effective when given as a short, simple regimen

8. Free of drug interactions (especially with artemisinin derivatives)

9. Less prone to development of drug resistance

10. Co-formulation to enhance compliance

In addition, the choice of ACT should be based on local parasite resistance to the non-

artemisinin partner drug. However, the cost and logistical difficulties of changing

established therapies can be substantial, and there may be insufficient contemporary

parasite resistance data to enable a rational choice between available ACTs.

Artemether-lumefantrine is currently the only ACT considered suitable by WHO for areas

of multi-drug resistance in South-East Asia as well as for Africa. [64] However, recent

head-to-head studies in Asia and Africa have suggested that DHA-PQ is more effective

than AL. [141, 175, 208]

1.8.2. Severe malaria

In contrast to the management of uncomplicated malaria, a different set of

pharmacological properties are desirable in drugs used for treating severe malaria. In

particular, the oral route of administration is usually impractical, dangerous or impossible

due to impairment of the patient’s conscious state, frequent vomiting or disordered GIT

function. [209] Also, because of the risk of death in patients with severe disease, there is

a far greater imperative to kill parasites rapidly in order to prevent the progression of

microvascular parasite sequestration and the consequent metabolic derangements that

can lead ultimately to death. [108] Therefore, for practical purposes, the ideal

antimalarial drug should not be administered orally but in a manner that rapidly achieves

optimal parasiticidal concentrations without causing dangerous side-effects.

38

Until recently, parenterally administered quinine had been the mainstay of therapy for

severe malaria. The parenteral formulation is a highly water-soluble dihydrochloride salt,

enabling injection by either i.m. or i.v. routes. However, it suffers from a relatively narrow

therapeutic margin, meaning that therapeutic concentrations result in tinnitus, dizziness

and nausea (“cinchonism”) [210] and high drug concentrations may lead to life-

threatening haemodynamic instability and cardiac arrhythmias. [124, 211] Often a

loading dose is administered in order to more rapidly achieve therapeutic concentrations.

[106, 212] This requires rigorous supervision and monitoring to prevent inadvertent over-

dosing (with significant risks of toxicity) or under-dosing (and therefore inadequate

therapeutic response). This may be impractical in much of the developing world,

particularly at the periphery of the health system, where trained staff and equipment

necessary to administer such infusions are often not available. In addition, the CL and Vd

of quinine exhibit marked inter-individual variability, being partially dependent on the

severity of the patient’s illness (which causes a contraction of the Vd and therefore higher

plasma concentrations). Therefore, it is difficult to predict the steady state concentrations

that will be achieved in the individual patient with standard dosing. [213] Furthermore

quinine treatment is associated with additional toxicities including hyperinsulinaemic

hypoglycemia and (rarely) drug-induced haemolytic anaemia [124, 214], complications

that can be difficult to diagnose and manage in resource-poor settings. Quinine’s onset

of action, even following intravenous loading dose is also relatively slow, with a “lag” of

up to 24 h occurring before decline in peripheral parasitaemias are seen. [215] This may

reflect the relatively poor activity against young, ring-form trophozoites. [215] These

forms (which are often predominant on the peripheral blood smear) may not be killed by

the drug until they have matured into later trophozoite forms. [215]

Whilst difficult and even risky to administer in some settings, there were few alternatives

to parenteral quinine for severe malaria until recently. Parenteral preparations of CQ and

halofantrine were compromised by even more problematic acute, dose-related toxicity

[124, 216-218] and development of parasite resistance. [64] Only the development of the

artemisinin derivatives as pharmaceutical-grade preparations for parenteral

administration in the 1990’s has offered a viable alternative. These include the lipophillic,

poorly water-soluble compounds ARM and arteether (manufactured dissolved in lipid

such as peanut oil for intramuscular injection), and the water-soluble artemisinin

39

derivative, ARTS (manufactured as a powder for dissolution in sodium bicarbonate

immediately prior to i.v. or i.m. injection). When compared with quinine, all artemisinin

derivatives share the advantages of consistently more rapid parasite clearance [184,

185] and a more favourable toxicity profile [186, 215, 218], and they can be more

conveniently administered as single bolus injections (rather than infusions), making them

attractive for resource-poor settings.

Most of the early studies of artemisinin derivatives for severe malaria evaluated ARM

given by i.m. injection. [219-228] A meta-analysis of trials comparing ARM with quinine

showed an overall lower mortality in ARM-treated patients but the difference did not

reach statistical significance. [115] Rates of parasite clearance were superior with i.m.

ARM. [115] A number of countries, including PNG, adopted i.m. ARM as standard

recommended treatment for severe malaria. Most studies of ARM for severe malaria had

been conducted at a time when PK data were limited. It has since become clear that

ARM has poor and highly variable absorption into the systemic circulation from the i.m.

injection site, raising concerns that a proportion of patients treated might be at risk of

sub-optimal therapeutic response. [229]

The PK profile of the water-soluble artemisinins appears to be more favourable for

treatments of severe malaria. In contrast to ARM, both i.v. and i.m. injections of ARTS

result in consistent and rapid absorption, with peak levels occurring within minutes of

administration. [182] The depot effect seen with ARM does not occur so that the drug’s

apparent plasma t½es (5 min for ARTS and 40 min for DHA) are not distorted by ongoing

absorption from the injection site. Its superior PK properties to those of ARM suggested

ARTS as a more promising replacement for parenteral quinine. Recently a large multi-

centre comparison of parenteral quinine with i.v. ARTS, the South-East Asian Quinine

Artesunate (SEAQUAMAT) trial, was published. [116] This showed a clear mortality

benefit (relative risk reduction of 32%, absolute risk reduction of 7%) for ARTS over

quinine. Because the majority of participants in this study were South-East Asian non-

immune adults, it is not yet clear whether this mortality benefit can be extrapolated to

children living in malaria endemic areas of the world such as Africa and PNG. However,

it seems likely that, in future, i.v. ARTS will be considered the gold standard by which

other treatments for severe malaria will be judged.

40

1.8.2.1. Rectal administration of artemisinin derivatives

Because the majority of deaths from P. falciparum infection occur in children living in

remote rural areas with limited access to medical care, a rapidly acting antimalarial drug

in suppository form could be of particular value. [230] In contrast to oral therapy, rectal

administration is not precluded by vomiting, prostration or impaired consciousness, all

common features of malaria in children. Although parenteral therapy can also be used

when oral dosing is problematic, equipment and trained personnel necessary to safely

administer i.m. or i.v. injections are often unavailable in these areas. An effective rectally

administered treatment could mean that children too sick to take medication orally could

be treated in the community, either as emergency “pre-referral” treatment (before

transportation of the child to a health facility for higher-level treatment) or, if referral is

not possible, as part of a complete treatment course given at home. [64, 231]

Artemisinin derivatives that have been formulated for rectal administration have included

ARM, QHS, DHA, and ARTS. Of these, ARTS has the most favourable characteristics,

because its high water solubility should facilitate absorption across the rectal mucosa.

[176] At present, most of the data accrued regarding rectally administered artemisinin

derivatives comes from studies of ARTS suppositories. These have been produced by

two manufacturers and are usually formulated as 50mg or 200mg suppositories

(Plasmotrim Rectocaps®, Mepha Pharmaceuticals, Aesch-Basel, Switzerland, Scanfarm

A/E Denmark). [117, 232-244] Rectal administration of artemisinin drugs has been

identified by WHO as having great promise, especially for the treatment of childhood

malaria. [245]

It may be feasible for suppository administration and subsequent monitoring for

extrusion to be performed by a village health worker, or even by a sick child’s mother,

with minimal training. Recently, home-based treatment strategies, in which mothers were

trained to administer oral CQ, were shown to dramatically reduce childhood mortality in

areas of poor health-care infrastructure in Africa. [246] However, the sickest children will

be unable to take oral therapy so an effective suppository might additionally enable

community-based treatment. An artemisinin-based therapy would have additional

advantages over oral CQ due to its activity where CQ-resistant P. falciparum is

prevalent. Because the artemisinin drugs have an excellent safety profile and a wide

therapeutic margin (meaning that the likelihood of drug-related toxicity would seem low

41

even if recommended dosages were exceeded), rectal administration could be safely

and effectively performed by individuals with little or no medical training.

Rectally administered artemisinins are likely to have their greatest value in critically ill

patients residing in geographically remote settings in whom definitive therapy may be

many hours or days away. Therefore, the primary imperative is for early and rapid

parasite clearance. This should prevent or delay the progression of microvascular

parasite sequestration and the consequent metabolic derangements that can lead to

death, [108] and buy time while the patient is being transported to a health facility. For

this reason, the most useful indicators of clinical efficacy are those that describe the

rapidity of parasite clearance, particularly in the first 12-24 h. [108] These include the

PC%12, PC%24, PCT50 and PCT90. [108] Definitive clinical and parasitological cure is

determined largely by either the duration of therapy or the efficacy of the longer acting

partner drug used as part of artemisinin-based combination therapy (ACT) and is

therefore of secondary interest in evaluation of rectal artemisinin drugs.

A recent large multi-centre placebo-controlled trial has shown that rectal artesunate

given as pre-referral therapy to children with suspected severe malaria results in a

significant reduction in overall death and disability, especially when arrival at the health

clinic is delayed > 6h from the time of administration. [230] Therefore, artemisinin

derivatives formulated for rectal administration have the potential to be amongst the

most widely used of all drugs in tropical countries. In common with oral and parenteral

formulations of artemisinin derivatives, their introduction into clinical use has not followed

classical pathways of rational drug development and dosage regimens have largely

been derived empirically. Although likely to share the intrinsic efficacy of orally or

parenterally administered artemisinins, rectal preparations must also be quickly and

consistently absorbed if they are to have a reliably rapid parasiticidal effect. Generating

the necessary PK and PD data has, however, lagged well behind the widespread clinical

introduction of a large number of rectal formulations. The reasons for this include the

practical challenges of performing PK studies in suitably representative populations in

the rural tropics, and the relative lack of stringent control of drug-licensing that exists in

many developing countries. The situation is complicated by the number of rectal

artemisinin derivatives on the market with the potential for multiple manufacturers of

single formulations. Because pharmaceutical factors may also influence PK disposition,

42

preparations from different manufacturers cannot necessarily be assumed to be

bioequivalent.

There are four published studies reporting PK data from children with falciparum malaria

treated with rectal ARTS. Two are from Africa [232, 238] and one from Thailand [242]

Two of these [238, 242] reported only basic PK data based on non-compartmental

methods. A larger crossover study, using higher rectal and i.v. doses in moderately

severe falciparum malaria [232], generated absolute bioavailability data but only single-

dose analysis of plasma DHA concentrations was possible. An additional study used a

combination of rich and sparse sampling, a population approach and a one-compartment

model to derive PK parameters for DHA in a more heterogeneous population including

adults and children from South-East Asia and Africa with moderately severe malaria.

[247] There is evidence that the PK of artemisinin derivatives change significantly with

repeated dosing [248-250], but this has not been assessed in the case of rectal ARTS.

Optimal strategies for deploying i.r. artemisinin drugs at community level are yet to be

determined. In particular, it is not clear to what extent this strategy is socially or culturally

acceptable and feasible across the range of societies in which this intervention could be

deployed. Information on existing attitudes to rectal administration in specific cultural

contexts should help to inform the development of the most appropriate deployment

strategies.

1.8.3. Malaria in pregnancy

The pharmacological management of malaria poses additional challenges in pregnancy.

In particular, physiological changes in pregnancy can lead to marked alterations in drug

disposition (especially through an increase in Vd and CL). However, little has been done

to define the PK of antimalarial drugs in pregnancy so that dose regimens can be

optimised. Almost all of the small number of studies that have attempted this have been

performed in South-East Asian populations and have evaluated drugs which are not

appropriate for use in much of the developing world due to issues of either cost (e.g.

MQ), safety (e.g. AL) and practicality (short-acting drugs such as quinine are not useful

for preventive therapeutic approaches to malaria in pregnancy). [151, 251-253] In

addition, these studies have had methodological limitations, such as the lack of

appropriately matched non-pregnant comparator groups. [151, 251-253] The ideal study

43

design would administer drug to both pregnant women and a control group of non-

pregnant women carefully matched for age, ethnicity and malarial parasitaemia, before

comparing PK disposition.

An additional challenge to pharmacological approaches to malaria in pregnancy is the

issue of drug safety including reproductive toxicity. No antimalarial drug has been

subject to formal systematic evaluation to exclude the possibility teratogenicity or foetal

toxicity. [102] Although CQ and SP are now considered safe for use in pregnancy, this

has come about not through structured evaluation but many decades of field-experience

with use of these drugs. [102] This unstructured approach to safety evaluation is no

longer considered appropriate. [102] Formal evaluations including pre-clinical animal

reproductive toxicity studies of alternatives to CQ and SP are lacking, thus limiting the

range of antimalarial drugs available for use in pregnancy.

In non-immune women, with increased susceptibility to severe disease, there is an even

greater need to optimise case management of acute infections, whether severe or non-

severe. [37] However, in immune women living in highly endemic areas, the need is to

limit the effects of placental parasitaemia throughout the duration of the pregnancy.

Because regular prophylaxis is difficult to implement in the developing world, an

alternative, simpler strategy of administering a small number of treatment courses during

pregnancy has been employed. [170, 254, 255] This strategy, known as intermittent

presumptive treatment of pregnancy (IPTp) using either 4-aminoquinoline drugs (CQ or

AQ) or SP has been shown to be have benefit in improving both maternal and neonatal

outcomes, that almost certainly outweighs the known risks of therapy with these drugs.

[170, 254, 255] However, because of the lack of safety and PK data for most

antimalarials, the range of drugs that can be utilised is currently limited to CQ, AQ and

SP. [102] It is important that PK data in pregnancy be obtained in order to optimise

dosing regimens of CQ, AQ and SP for IPTp.

44

1.9. Malaria in Papua New Guinea and Melanesia

1.9.1. Epidemiology and clinical manifestations

The Melanesian peoples populate the Western South Pacific islands that include New

Guinea, the Solomon Islands, Vanuatu, New Caledonia and Western Fiji. Human genetic

and linguistic analyses suggest that they arrived during two major migratory movements,

the first approximately 40 000 years ago (the non-Austronesian peoples) and the second

6 000 years ago (the Austronesian peoples). [256] Anopheline mosquito species and

malaria transmission are only present in four countries; the island of New Guinea

including both Indonesian West Papua and PNG, the Solomon Islands and Vanuatu. [29]

Anopheline species diversity diminishes moving westward through the Melanesian

islands, suggesting that this was the direction of their historical spread through the area.

[29] In PNG at least 8 species belonging to the Anopheles punctulatus group exist

(including Anopheles punctulatus itself, Anopheles koliensis and a number of species

formerly collectively identified as Anopheles farauti). [29]

Papua New Guinea is the largest of the Melanesian countries with a current population

of approximately 5.2 million. Its total land-mass is 450 000 km2 (roughly the size of

Germany) including many widely dispersed islands, separated in some cases by

hundereds of kilometers (see Fig 1.2). Papua New Guinea is distinguished by its

profound environmental diversity with topography that includes offshore coastal atolls,

lowland coastal swamps, dense tropical rainforests and a rugged central mountain

range. The many geographic barriers to the movements of peoples within the country

have, historically, contributed to maintaining a high degree of cultural and linguistic

diversity and in the modern era have impeded the delivery of services (including primary

health-care) to the population. Over 800 separate languages are spoken in the entire

country. [256]

Because of the environmental trophism of the anopheline vector, the main climatic

determinant of malaria endemicity in PNG is ambient temperature. [29] All areas of the

country are situated close to the equator which means that there is little seasonal

variation in temperature and the major source of variability is altitude. Therefore, malaria

transmission intensity in different areas of the country relates predominantly to its

45

altitudinal zones. [29] Hyper- and holo-endemic transmission exists throughout many

areas of the country lying below 600m. Transmission in “highland” areas above 1700-

1800m does not occur. At intermediate altitudes, transmission may be low, unstable or

epidemic. Interestingly, human populations are concentrated mostly in either the lowland

(<600 m) or highland (>1300 m) areas leading some to hypothesise that the relative

sparseness of human population density at intermediate altitudes has arisen due to the

higher mortality associated with unstable and epidemic transmission in these areas. [29]

All areas of PNG experience high annual rainfall (1300-7000 mm/year) and in most

lowland areas transmission is perennial. However, peaks and troughs in transmission

intensity may still occur during wet and dry seasons, respectively. [29]

Four of the five species of malaria known to infect humans occur in PNG. [29]

Plasmodium falciparum now predominates and is highly endemic throughout most

coastal regions of the country. Plasmodium vivax is also highly endemic, and may

previously have been more prevalent than P. falciparum. [29] The shift in predominance

from P. vivax to P. falciparum seems to have occurred during intensive malaria control

efforts that included mass drug administration (mostly with CQ). [29, 257] These

measures were initiated from 1957 onwards but abandoned in the 1970s when it had

become clear that eradication was not feasible. [29] It is suspected that mass CQ

administration may have given P. falciparum (which seemed to develop CQ resistance

very early) a selective advantage over P. vivax (in which CQ resistance appears to be a

later and less extensive phenomenon) and therefore enabled it to re-emerge quickly and

occupy the niche left by P. vivax, following the abandonment of eradication efforts. [29]

Nonetheless, PNG probably still has the highest prevalence of P. vivax in the world and

P. vivax remains the dominant species in altitudes >1600 m (where P. falciparum

transmission cannot be supported). [29] Plasmodium malariae also occurs in a patchy

distribution, with prevalence of up to 13% in East Sepik province. [29] However, it has

been little studied. Both P. vivax and P. malariae may have additional relevance to the

clinical epidemiology of P. falciparum due to the possibility that some cross-protective

immunity exists between these species. [258] The prevalence of P. vivax and P.

malariae might therefore ameliorate the severity of P. falciparum infections in PNG.

Plasmodium ovale also occurs uncommonly. Plasmodium knowlesi has not been

described and its presence in PNG would be unlikely, given the absence of a suitable

primate host.

46

According to PNG Ministry of Health statistics, malaria is the leading outpatient

diagnosis (annual incidence of 303/1000 population at risk), the second-leading cause of

hospital admissions (834/100 000 population) and the second-leading cause of death

nationwide. [259] However, these statistics are based on passive case reporting and

diagnostic facilities necessary to confirm diagnoses are lacking, so their accuracy is

uncertain. Most malaria-related deaths are likely to occur in children under the age of 5

years. [31, 32, 260] Eighty-five percent of PNG’s population lives in rural areas. [259]

Here microscopy services are poor or non-existent and most treatment is administered

based on presumptive diagnosis. [29, 204] Primary health-care services have

deteriorated recently [261] and supplies of essential medicines to remaining health

centres have been erratic, adding to the challenges for rural populations in accessing

antimalarial treatment.

The clinical studies that will be detailed in this thesis were all conducted in Madang and

East-Sepik (Maprik area) provinces which are situated on the North coast of PNG (see

Figure 1.2). The epidemiology of malaria has been extensively studied in both locations,

including detailed studies of anopheline entomology and ecology, field-based

malariometric surveys, clinical epidemiology and clinical studies of severe malaria. [31,

32, 260, 262, 263] Both areas have high transmission intensities (EIRs range from 100

in the East Sepik to 400 in the Madang area) that occur on a year-round basis but with

some seasonal variation (in Madang this is highest during the wet season peaking in

December and January and lowest in the dry season from June to October). [31, 32]

Malariometric surveys suggest that both areas are hyper-endemic for P. falciparum

(population prevalence of parasitaemia in under-five children 60-80%). [31, 32] In

pregnant women in Madang, peripheral parasitaemia prevalence ranges from 19% in

multigravidae to 34% in primigravidae women. [264] Rates of placental parasitaemia are

similar and 17% of pregnant women have Hb <70g/L. [265] Plasmodium vivax, P. ovale

and P. malariae are also endemic in both areas. Aside from the high prevalence of non-

falciparum malaria in this setting, the level of P. falciparum transmission intensity means

that age-specific malaria morbidity and mortality mirrors that of much of sub-Saharan

Africa. [31, 32, 260, 262, 263] Verbal autopsy series have suggested that malaria

contributes 17% of childhood deaths in Madang [263] and 4% in the Maprik area of East

Sepik. [260]

47

Figure 1.2 Topographic map of Papua New Guinea

A large study of clinical manifestations of severe malaria in children in Madang in 1997

demonstrated an overall mortality lower than that generally seen in comparable African

studies that have used similar definitions. [25, 27, 70, 266, 267] It also demonstrated that

the three strongest predictors of death were impaired consciousness, hyperlactatemia

(venous plasma lactate >5 mmol/L) and metabolic acidosis (venous plasma bicarbonate

<15 mmol/L) which were associated with mortalities of 8%, 17% and 15%, respectively.

[25, 27, 80, 81] These rates accord well with those from Africa, with the notable

exception that mortality associated with impaired conscious state or coma was

significantly lower than that seen in comparable African studies. [27, 81, 266, 267] This

lower than expected mortality raises interesting questions about possible protective

mechanisms in this population. Putative protective factors include cross-protection due

to P. vivax infections and local host genetic factors.

48

1.9.2. Host genetics: the role of erythrocyte polymorphisms

The Melanesian populations inhabiting coastal PNG are likely to have been exposed to a

high burden of malarial disease for many thousands of years. [42] This is thought to

have lead to a number of de novo genetic adaptations to malaria. [256] Although

different to those seen in other malaria-endemic parts of the world, they may have

similar mechanisms of protection and therefore be examples of “parallel evolution”. The

manner in which these host genetic mutations may have arisen as evolutionary

adaptations may provide insights into malarial pathogenesis and mechanisms of innate

protection from severe disease. If they have become prevalent through natural selection

then it follows that, in order to confer a survival advantage, they must protect the host

from dying from malaria. [42, 43] The characteristic clinical spectrum of severe childhood

malaria seen in PNG [25, 27, 80, 81] may be of additional interest, given the unique

pattern of erythrocyte polymorphisms that co-exist here. [43] These include South-East

Asian ovalocytosis (SAO), Gerbich antigen blood group and �-thalassaemia. [43] South-

East Asian ovalocytosis is associated with heterozygosity for a 27 base-pair deletion in

the band 3 gene (SLC4A1∆27) [268] and Gerbich blood group with the glycophorin C

exon 3 deletion (GPC∆ex3). [46] All are postulated to protect relatively isolated PNG

populations from mortality associated with P. falciparum infection, having been selected

as evolutionary adaptations to a heavy malaria burden. [42, 47] In Madang, for example,

the prevalence of �-thalassaemia is >90%, SLC4A1∆27 15-35% and GPC∆ex3 32-71%.

[42, 45, 46, 269, 270] Each polymorphism can be characterized by well-validated PCR-

based assays, [270-272] with the respective genes lying on separate chromosomes and

exhibiting independent assortment within the population. [46]

In areas of coastal PNG such as Madang, clinical disease in children with malaria

encompasses a spectrum through asymptomatic parasitaemia, simple fever, illness

warranting hospitalization, and life-threatening multi-organ failure. However,

susceptibility to asymptomatic or mild infections does not appear to be influenced by any

of the prevalent erythrocyte polymorphism mutations. [45, 46] Rather, they appear to

provide a survival advantage by ameliorating the severe end of the disease spectrum.

[45, 269] Case-control studies have demonstrated that �-thalassaemia is associated with

reduced risk of all forms of severe malaria, and that SLC4A1∆27 is protective against

cerebral malaria. [45, 269] However, the underlying mechanisms remain unclear and

49

GPC∆ex3 has not been studied in severe disease. This mutation is therefore of

particular interest. Because it does not appear to prevent falciparum malaria infection,

and so if it reduces the overall risk of death, its survival benefit must occur by an effect

on the incidence or severity of the most lethal manifestations of the infection. This

requires further studies of children with severe malaria in areas of PNG where the

mutation is endemic.

1.9.3. Antimalarial treatment policy and parasite drug resistance

Until very recently, antimalarial drugs used in PNG have been restricted to those from

the quinoline and anti-folate drug classes. Quinine has been used at least since the time

Robert Koch performed some of the first formal clinical therapeutic trials in malaria, in

the Madang area in 1899-1900. [273] In the post-war era, CQ became the mainstay of

treatment for uncomplicated malaria, with quinine being reserved for severe malaria. [29,

274] Chloroquine-resistant P. falciparum was first described in an Australian returning

from PNG in 1976. [275] However, monotherapy with CQ (or the related 4-

aminoquinoline, AQ) continued to be the standard treatment recommended by PNG

Ministry of Health guidelines until 1997. Rates of resistance rose steadily during this

time such that RII/RIII in vivo failure rates were greater than 30% by the mid 1990’s. [29]

Molecular studies from 1997 onwards have shown that both PfCRT and PfMDR

mutations associated with CQ resistance are fixed in local parasite populations, at rates

of >90%. [59, 276] The molecular epidemiology based on sequencing of the PfCRT

gene suggests that CQ resistance seen in PNG is likely to have arisen de novo, rather

than by introduction from Asian parasite populations. [277] Case reports documenting

the presence of chloroquine-resistant P. vivax in PNG first emerged in 1989 [278, 279]

However since this time there has been very little prospective data generated to

document in vivo efficacy of P. vivax in PNG. One multi-centre study evaluated WHO-

defined day 28 ACPR in 190 children with either P. vivax monoinfection or mixed P.

vivax/P. falciparum infections treated with either AQ-SP or CQ-SP. [280] However most

(92%) received AQ-SP. Of the 16 children who received CQ-SP, 4 (25%) were classed

as treatment failures. For AQ-SP, depending on the study site, treatment failure rates

ranged from 0 (East Sepik province) to 27% (Madang area). More data is available from

neighbouring West Papua province of Indonesia, where two studies have demonstrated

much higher day 28 treatment failure rates (65-84%) with chloroquine monotherapy.

[281, 282]

50

Anti-folate drugs were not used extensively as treatment of symptomatic malaria in

PNG until standard treatment policy was changed in 1997. The recommended first-

line treatment then became the combination of a 4-aminoquinoline (either CQ or AQ)

with SP. [204] This recommendation did not take effect until well after 2000. However,

one component of this combination, PYR, had been extensively used as part of mass-

treatment programs in the 1950’s-1970’s. [257] This probably explains the high pre-

existing rates of DHFR mutations seen in P. falciparum strains in many areas of PNG.

[29] Sulphadoxine-pyrimethamine has also been used since the 1970’s in combination

with quinine as “second-line” therapy (for uncomplicated infections failing treatment

with CQ) and for the treatment of severe malaria. [224, 274, 283] In vivo resistance to

SP was first described in 1980. [284] The currently recommended combination of CQ

(or AQ) and SP has been shown to have parasitological failure rates of approximately

20% (using the more recent WHO 28-day in vivo test with PCR correction for re-

infection/recrudescence) in P. falciparum in two areas of PNG. [280] Data from these

areas also indicated that 20% of P. falciparum parasites from cross-sectional

population surveys had already acquired the PfDHPS 437 mutation (associated with

sulphadoxine resistance) and that 31% carried the triple DHFR mutation (affecting the

108, 59 and 51 position base pairs and associated with high level PYR resistance).

Five percent had both the triple DHFR combined with a single DHPS mutation.

Generally, a combination of triple DHFR mutations and at least two DHPS mutations

(the so-called “quintuple mutation”) is associated with a high level of in vivo resistance

to the SP combination. [162] Therefore, whilst in vivo efficacy of the 4-

aminoquinoline+SP combination appears to be acceptable, in some parts of PNG only

one additional mutation (of the DHPS gene) is required before widespread failure of

the current first line treatment is expected to occur.

At the same time that the recommended first-line treatment for uncomplicated malaria

was changed, the treatment policy for severe malaria was also revised such that an

artemisinin derivative (ARM, given by i.m. injection) became first-line treatment. [204]

Therefore, artemisinin derivatives have only been used in PNG for less than a decade

and their official use is restricted to the treatment of severe malaria (ARM) and as

second-line treatment for treatment-failure (ARTS). In both instances, it is

recommended that artemisinin treatment be combined with SP. [204]

51

Because the majority of PNG’s population live in rural areas where access to medical

care is severely limited by poor health-care infrastructure and the extremely challenging

nature of the country’s terrain, an effective antimalarial suppository may be particularly

beneficial in this setting. However, the applicability of dose regimens developed in non-

Melanesians is uncertain. There have been no studies of rectally administered

artemisinin from the Oceania region. In fact, only one previous study has evaluated

artemisinin drugs in any form in a Melanesian population. [224] In addition, Hb variants

such as �- �thalassaemia that may alter the disposition and efficacy of artemisinin drugs

[285-289] have a high prevalence in some Melanesian populations. [45] There have

been no previous PK studies of rectal ARTS in patients with haemoglobinopathies.

Unlike many areas of Asia and Africa, access to antimalarials through the private and

informal sector has not been widespread in most areas of PNG. A variety of

artemisinin monopreparations (including ARTS, ARM and DHA) have been available

for sale from private pharmacies in PNG for some time. However, in practice these are

only accessed by a small segment of the urban population. Other conventional

antimalarials including those incorporating alternative anti-folate drugs (dapsone-

chlorproguanil, atovaquone-proguanil) or the aryl amino-alcohols or related

compounds (halofantrine, MQ and LUM) have historically not been part of the national

formulary, nor used through the private sector to any significant extent in PNG.

The nature of the health system in PNG means that most malaria treatment is

administered empirically rather than being based on microscopic confirmation and

speciation. [204] As in other countries in the Oceania and Asian regions P. vivax is an

additional significant cause of morbidity. [12] Therefore, there is a need for deployment

of antimalarial drug regimens that are effective against both falciparum and vivax malaria

in PNG. In view of increasing evidence that current standard treatment regimens (CQ or

AQ combined with SP) are losing effectiveness in PNG, [280] the results of recent

comparative trials of available ACTs [141, 164-166, 174, 175, 208] and current WHO

recommendations [64] there is a need to evaluate the efficacy of candidate replacement

ACT regimens directly in comparison with CQ-SP for both falciparum and vivax malaria.

52

53

2. MANAGEMENT OF UNCOMPLICATED MALARIA IN

CHILDREN

54

55

2.1. Overview of Clinical trials in this chapter This chapter includes two clinical trials that evaluate treatments for uncomplicated

malaria in Melanesian children. These investigate alternative treatments to the currently

recommended first-line treatment for malaria in PNG, CQ-SP, which appears to be at

imminent risk of failure due to worsening drug-resistance. The second study compares

three artemisinin combination therapies (ARTS-SP, AL and DHA-PQ) directly with

existing standard therapy. Of these, DHA-PQ is the most novel and little is known of its

PK or other determinants of efficacy. Therefore, the first study provided preliminary data

that included characterisation of PQ’s PK profile in a representative population.

The two studies aimed to address the following research questions:

1. What are the PK properties of PQ in Melanesian children and how do these

compare with those of the structurally related antimalarial, CQ?

2. Based on efficacy and safety profile, should either of the fixed ACT co-

formulations DHA-PQ or AL replace CQ-SP as standard treatment for

uncomplicated malaria in PNG?

3. Are there clinically-significant differences between the efficacy of the three ACTs

against either P. falciparum or P. vivax in an area of high multi-species

transmission?

4. What host-, parasite- and drug-specific factors are important determinants of

treatment response?

56

57

2.2. Clinical trial 1: Pharmacokinetics and efficacy of piperaquine and chloroquine in Papua New Guinean children with uncomplicated malaria

2.2.1. SUMMARY

Aims: To analyse and compare the pharmacokinetic (PK) disposition of chloroquine

(CQ) and the related 4-aminoquinoline, piperaquine (PQ), in Papua New Guinean

children with uncomplicated malaria.

Methods: Twenty-two children were randomised to three-days of PQ phosphate 20

mg/kg/day (12 mg PQ base/kg/day) co-formulated with dihydroartemisinin (DHA-PQ:

Duo-cotecxin™) and 20 children to three days of CQ 10 mg base/kg/day with a single

dose of sulphadoxine-pyrimethamine (SP: 25/1.25 mg/kg). Following a 42-day intensive

sampling protocol, PQ, CQ and its active metabolite mono-desethylchloroquine (DECQ)

were assayed in plasma using high performance liquid chromatography. A two-

compartment model with first-order absorption was fitted to the PQ and CQ data.

Results: There were no significant differences in age, gender, body-weight or admission

parasitaemia between the two groups. The PCR-corrected 42-day adequate clinical and

parasitological response was 100% for DHA-PQ and 94% for CQ-SP but P. falciparum

re-infections during follow-up were common (33% and 18%, respectively). For PQ, the

median (interquartile range) volume of distribution at steady state allowing for

bioavailability (Vss/F) was 431 (283-588) L/kg, clearance (CL/F) was 0.85 (0.67-1.06)

L/h/kg, distribution half-life (t½e�1) 0.12 (0.05-0.66) h and elimination half-life (t½e�2) 413

(318-516) h. For CQ, Vss/F was 154 (101-210) L/kg, CL/F 0.80 (0.52-0.96) L/h/kg, t½e�1

0.43 (0.05-1.82) h and t½e�2 233 (206-298) h. The non-compartmentally derived DECQ

elimination half-life was 290 (236-368) h. Combined molar concentrations of DECQ and

CQ were higher than those of PQ during the elimination phase.

Conclusions: Although PQ has a longer elimination half-life than CQ, its prompt

distribution may limit its post-treatment malaria suppressive properties.

58

2.2.2. AIMS

2.2.2.1. Primary

To characterise the PK properties of PQ and in Melanesian children with uncomplicated

malaria. Because the nature of PK-PD relationships are not well characterized for PQ, in

order to provide a comparative context, a simultaneous PK study of the better-known

chemically-related 4-aminoquinoline, CQ was also performed.

2.2.2.2. Secondary

To provide preliminary comparative efficacy data for DHA-PQ and CQ-SP in Melanesian

children with uncomplicated malaria.

2.2.3. METHODS

2.2.3.1. Acknowledgement

The following collaborators contributed work to this study as follows:

Irwin Law (School of Medicine and Pharmacology, UWA) supervised some of the field

work and performed some of the clinical procedures. The PNG IMR field staff who

performed the clinical activities (including patient enrolment, evaluation, clinical

procedures and data collection) were Jovitha Lammey, Servina Gommorai, Sr Maria

Goretti, Bernard (“Ben”) Maamu, Wesley Sikuma and Donald Paiva. Enmoore Lin and

Alice Ura performed the MSP1/MSP2/GLURP genotyping. Nandau Tarongka, Lena

Lorry and Kaye Baea performed the microscopy.

The HPLC assays for PQ, CQ and DECQ were developed and validated by Madhu

Page-Sharp (School of Medicine and Pharmacology, UWA) under the supervision of Ken

Ilett (School of Medicine and Pharmacology, UWA). Madhu Page-Sharp performed all of

the drug assays.

I took primary responsibility for the original study concept, design and ethical approval. I

supervised the clinical aspects of the study, including data collection, clinical evaluations

and procedures. I performed the PK modeling, analysed and interpreted the clinical and

59

PK data and prepared the initial manuscript (publication 8) and presentation

(presentation 3).

2.2.3.2. Rationale for study design

An intensive sampling PK-PD study of PQ was designed in order to provide the richest

possible dataset, whilst bearing in mind the logistical, practical and ethical constraints of

blood sampling in children. Children aged 5-10 years with uncomplicated malaria were

deemed the best group to evaluate by providing a compromise between the need for the

study to be generalisable to the most relevant patient group (young children with

malaria) and the practical and logistical constraints (ease of taking blood from children

who were old enough to not be unduly distressed by i.v. cannulation or venesection). It

was decided that no more than 15 blood samples should be drawn over the 42-day

study period in each individual subject. A sample size of 20 was considered appropriate

for adequate characterisation of PK parameters. The sampling protocol was designed

assuming a t½e�2 of 13.5 days (based on a population PK analysis of PQ in Cambodian

children with malaria by Hung et al. [138]), aiming to have at least 3-4 sampling points

during the elimination phase. The PK characteristics of CQ were evaluated in a

matched cohort of 20 children, but the comparison was intended to be exploratory and

not subject to any a priori statistical power estimation. Because it was expected that CQ

would have a similarly long elimination to profile to PQ, an identical sampling protocol

could be used in the control group. To enable the most robust comparison between the

two groups, treatment allocation was performed in a randomised controlled fashion,

though for practical reasons, the study was open-label.

2.2.3.3. Study site

The study was conducted at Alexishafen Health Centre, Madang Province from August

2005 to March 2006 in an area hyper-endemic for P. falciparum where most clinical

disease is in children [31] and P. vivax, P. ovale and P. malariae are also endemic.

2.2.3.4. Patient eligibility and enrolment

Children aged 5-10 years who presented with an axillary temperature >37.5°C or a

history of fever in the previous 24 h were screened for malaria with a Giemsa-stained

thick blood film read by a trained microscopist. Those with mono-infections of either P.

falciparum (at >1000 asexual parasites/�L whole blood) or P. vivax, P. ovale or P.

malariae (>250 parasites/�L) were eligible for recruitment provided that i) they had no

60

features of severe malaria [24], ii) they had not taken one of the study drugs in the

previous 14 days, iii) there was no clinical evidence of another infection as a possible

cause of fever, iv) there were no signs of malnutrition or significant other co-morbidity,

and v) the child’s parents gave informed consent. A sequential study code was

allocated, an initial detailed clinical assessment was performed, and a 3 mL venous

blood sample taken. Baseline tests included Hb (HaemoCue™, Angelholm, Sweden)

and blood glucose (HaemoCue™, Angelholm, Sweden). Separated plasma was stored

at below -20oC for subsequent drug assay and a red-cell pellet was retained for parasite

genotyping.

2.2.3.5. Drug administration

Following enrolment, an opaque envelope labeled with a unique sequential code was

opened and a computer-generated randomised treatment read and assigned. This

process was conducted independently of the children, their parents/guardians and study

staff. Children in treatment arm A were allocated Duo-cotecxin™ (Beijing Holley-Cotec,

Beijing, China) tablets containing DHA and PQ in a mass-based ratio of 1:8 (40 mg DHA

and 320 mg PQ phosphate per tablet). A dose of approximately 2.5 mg/kg DHA and 20

mg/kg PQ phosphate (equivalent to 11.5 mg/kg PQ base) was given daily for three days

(at 0, 24 and 48 h: total dose of PQ phosphate 60mg/kg). Children in treatment arm B

received CQ (Chlorquin™, Aspen Healthcare Australia Pty Ltd, St Leonards, Australia)

10mg base/kg daily for three days (0, 24 and 48 h: total CQ dose CQ base 30 mg/kg)

and a single dose of SP (Fansidar™, Roche, Basel, Switzerland; 25 mg/kg SDOX and

1.25 mg/kg PYR) co-administered with the first dose of CQ. Drugs were administered as

a combination of whole, half or quarter tablets (divided using a tablet cutter) that best

approximated the target mg/kg dose and were swallowed whole with water. All doses

were administered under direct supervision and the exact time of administration was

recorded. Children were observed for 30 min post-dose and those who vomited were re-

administered the same dose. Subjects were not required to fast. Because the

manufacture of Duo-cotecxin™ is not compliant with Good Manufacturing Practice

(GMP) standards, tablets from each batch used were assayed for content (see below).

2.2.3.6. Monitoring and outcome assessment

All patients were admitted for 48-72 h to enable supervised administration of all drug

doses, blood sampling and clinical monitoring. All children had a daily assessment which

61

included a symptom checklist, measurement of axillary temperature, respiratory rate,

capillary blood glucose and a Giemsa-stained thick blood film. Children who developed

signs of severe malaria or early treatment failure according to WHO criteria [24] were to

be withdrawn from the study and given intramuscular quinine, as recommended for

complicated malaria in PNG. [204] Following discharge, children were reviewed as

outpatients on days 7, 14, 28 and 42 at which times axillary temperature and Hb were

measured and a thick blood film taken. Treatment response was defined according to

current WHO definitions of adequate clinical and parasitological response (ACPR). [113]

This analysis was conducted per protocol for both P. falciparum and P. vivax. Children

who developed recurrent infection with P. vivax were not censored or removed from the

P. falciparum outcome analysis (ie they remained in both the numerator and

denominator when proportions were calculated). Children experiencing late treatment

failure during the follow-up period were re-treated with quinine if symptomatic.

2.2.3.7. Blood sampling

Each child had a heparinised intravenous cannula inserted at the time of enrolment. In

addition to a baseline sample, 3 mL samples were drawn at 1,2, 4, 6, 12, 18, 24, 30, 48

and 72 h and by venesection at 7, 14, 28 and 42 days. The sampling protocol was

identical regardless of the allocated treatment arm. The exact time of each sample was

recorded. Blood samples were collected into lithium heparin tubes, centrifuged at 1800

x g for 5 min and separated plasma stored at -80oC until analysed.

2.2.3.8. Laboratory methods

2.2.3.8.1. Microscopy and parasite genotyping

Blood smears were examined by two skilled microscopists who were blinded to

treatment allocation and the other microscopists’ result. At least 100 fields at 1,000x

magnification were viewed before a slide was considered negative. Parasite density was

determined by counting asexual parasites/1000 leucocytes and assuming a total

leucocyte count of 8,000/�L. Any slide discrepant for positivity/negativity, speciation or

parasite density (>10x difference) was read by a third microscopist whose result was

final. In cases of microscopically-confirmed recrudescence/re-infection with P.

falciparum, parasite DNA from the follow-up blood sample was compared with that of the

baseline sample for polymorphisms of the MSP1, MSP2 and GLURP genes as

described previously. [59]

62

2.2.3.8.2. Drug assays

Drug assay methodology for PQ and CQ (as well as DECQ) are described in Appendix

1, sections 7.11 and 7.12, respectively.

2.2.3.9. Pharmacokinetic analysis

For both PQ and CQ, individual 2 or 3-compartment models with a lag time and using

weighting of 1/y or 1/y2 were fitted to individual patient concentration-time data using

Topfit 2.0 . [290] Discrimination between models was decided on the basis of the

overall model goodness of fit criterion (B), visual inspection of the distribution of

residuals, and the Akaike Information Criterion value. Parameters of interest derived

from the model included t½abs, t ½e�1 t½e�2, Vss/F, CL/F, AUC0-42 and AUC0-�. For

DECQ, AUC0-42 (log-linear trapezoidal rule), AUC42-� (C42/terminal elimination rate

constant), AUC0-� (AUC0-42 + AUC42-�) and t½e (log-linear regression of last 3-4 data

pairs) were estimated by non-compartmental analysis using Topfit 2.0 . [291] In order

to compare absolute concentrations of CQ, DECQ and PQ, drug concentrations and

AUC measurements are here expressed as molar concentrations.

2.2.3.10. Statistical analysis

Student’s t-test (for normally distributed data) or the Mann-Whitney U Test (for non-

normally distributed data) was used for between-group comparisons of admission

characteristics, PK parameters and efficacy outcomes. Categorical data were compared

using either Pearson Chi-squared or Fisher’s exact test where appropriate. A two-tailed

level of significance of P<0.05 was used for all tests (SPSS v 9.0, Chicago, IL, USA).

2.2.4. RESULTS

2.2.4.1. Patient characteristics

A total of 101 febrile children aged 5-10 years were screened, of whom 35 were

ineligible due to a negative blood slide or parasitaemia <1000 asexual parasites/�L. A

further 12 declined consent and 12 were excluded due to recent treatment (n=2),

inability to attend for follow-up (n=9), concomitant illness as possible cause of fever

(n=3) and/or signs of severe malaria (n=2). Of the remaining 42 children, 36 had

falciparum (including 3 with mixed infections with P. vivax), 3 vivax and 3 malariae

malaria. Their baseline characteristics by allocated treatment are summarised in Table

63

2.1. There were no statistically significant differences between the two treatment groups,

except that children treated with CQ had a higher mean axillary temperature. Only one

child vomited the first treatment dose within the 30 minute observation period. This dose

(of CQ and SP) was re-administered promptly without further incident. No children were

known to have vomited subsequent to the 30-min observation period.

2.2.4.2. Response to treatment

Two children in the CQ-SP group with P. falciparum failed to complete follow-up.

Treatment outcomes in the remaining 34 falciparum patients are summarised in Table

2.2. Parasite clearance (the number of days taken to achieve a negative blood slide)

was faster in the DHA-PQ group (median 2 vs 3 days in CQ-SP group: P=0.001). The

PC%24 was significantly higher in the CQ-SP group (median 13%: range 0-256%) than

the DHA-PQ group (median 0.2%: range 0-5.1%: P for difference with CQ-SP <0.001). A

higher number of CQ-SP patients (8/15) were gametocytaemic at day 14 compared with

DHA-PQ treated patients (1/19: P=0.007). According to WHO definitions, all 19 (100%)

DHA-PQ-treated patients achieved ACPR at 28 days. For CQ-SP treated patients, only

one recrudescence occurred (PCR-corrected ACPR 93%). By day 42, a further 6 DHA-

PQ-treated patients and 2 other CQ-SP-treated patients had developed re-infections

with P. falciparum but there were no further recrudescences. Thus the PCR-corrected

ACPR at day 42 was unchanged (100% for DHA-PQ and 93% for CQ-SP). There were

no statistically significant differences between treatment arms for P. falciparum re-

infections, recrudescences or ACPR at days 28 or 42.

A significantly higher number of those in the CQ-SP group developed P. vivax infections

during follow-up (Table 2.2). Of 3 patients with P. vivax, 1 treated with CQ-SP was lost to

follow up, 1 treated with DHA-PQ had ACPR, and 1 treated with DHA-PQ had

reappearance of P. vivax on day 42. All 3 patients with P. malariae (2 treated with CQ-

SP and 1 with DHA-PQ) made a full recovery without relapse during the 42-day follow-

up period.

One child in the DHA-PQ group developed a painful foot, reluctance to walk and general

malaise 7 days following treatment in association with a painful, indurated intravenous

cannula insertion site. Her symptoms were consistent with disseminated bacterial sepsis

secondary to thrombophlebitis. She made an uneventful recovery with a course of

64

antibacterial therapy (flucloxacillin). No other significant adverse effects were observed

in either treatment arm.

Table 2.1 Details of 42 Papua New Guinean children at the time of admission to Clinical trial 1. Data are number (%), mean ± SD or median (range) unless otherwise specified.

DHA-PQ CQ-SP

Number of patients 22 20

Age (years) 6.9 ± 1.4 6.7 ± 1.5

Male sex 17 (77%) 13 (65%)

Body weight (kg) 19.1 ± 3.8 18.8 ± 3.1

Axillary temperature (°°°°C) 37.2 ± 1.2 38.1 ± 1.51

Splenomegaly 15 (68%) 10 (50%)

Respiratory rate (/min) 27 ± 6 30 ± 12

Haemoglobin (g/L) 86 ± 18 78 ± 12

Number with P. falciparum2 19 (86%) 17 (85%)

P. falciparum parasitaemia (/µµµµL)

20 400

(3520–191 000)

34 500

(1640–322 000)

Plasma glucose (mmol/L) 6.2 ± 1.2 5.5 ± 1.9

Drug dose (mg base/kg/day) 11.5 (10-14.5)3 10.2 (7.2-11.7)

1 P= 0.03 vs DHA-PQ 2 A further 3 patients treated with DHA-PQ had mono-infections with either P. vivax (2) or

P. malariae (1); 1 P. vivax and 2 P. malariae received CQ-SP 3 Equivalent to 19.9 (17.3-25.0) mg/kg/day of PQ phosphate

65

Table 2.2 Treatment outcomes by microscopy at days 28 and 42 in Papua New Guinean

children aged 5-10 with P. falciparum mono-infection at baseline (n=34 after 2 children in

chloroquine+sulphadoxine-pyrimethamine arm excluded due to loss to follow-up) treated

with 3-day courses of Duo-cotecxin (DHA-PQ: 11.5mg/kg PQ-2.5mg/kg DHA daily) or

chloroquine+sulphadoxine-pyrimethamine (CQ-SP: 10mg/kg daily) Recurrent infections

during follow-up were differentiated as either re-infection or recrudescence by

genotyping (see text). Data are number (%).

1 P<0.05 for difference between treatment groups

2.2.4.3. Drug assays and pharmacokinetic analysis

Mean drug content of 24 tablets of Duo-cotecxin™, was 109 ± 7% for PQ and 106 ± 4%

for DHA.

In the development of PK models for both the CQ and PQ datasets, a 2-compartment

model, with 1/y2 weighting gave the best fit. Pharmacokinetic parameters for CQ and PQ

are summarised in Table 2.3. Non-compartmental analysis derived similar t½e�2 values

for PQ and CQ to those derived by compartmental methods (non-compartmental t½e�2

values were 431 [322-531] h for PQ and 235 [212-298] h for CQ; P<0.001). Plasma

concentration profiles of PQ, CQ and DECQ are presented as spaghetti plots in Figures

2.1 and 2.2. The similarity of t½e�2 values for DECQ (290 [236-368] h) to those for CQ

suggests that DECQ t½e�2 may be formation rate-limited.

Day 28 Day 42

DHA-PQ (n=19)

CQ-SP (n=15)

DHA-PQ (n=19)

CQ-SP (n=15)

P. falciparum parasitaemia

Re-infection

Recrudescence

0

0

0

2 (13)

1 (7)

1 (7)

6 (32)

6 (32)

0 (0)

4 (27)

3 (20)

1(7)

P. vivax parasitaemia 0 6 (40)1 1 (5) 6 (40)1

66

Table 2.3 Pharmacokinetic parameters for piperaquine (PQ), chloroquine (CQ) and

desethylchloroquine (DECQ) in 42 Papua New Guinean children with malaria following

3-day courses of Duo-cotecxin (11.5mg/kg PQ-2.5mg/kg DHA daily) or chloroquine

(10mg/kg daily). Data are median (interquartile range).

Parameter PQ

(n=22)

CQ

(n=20)

DECQ

(n=20)

t½e�1 (h) 0.12 (0.05-0.66) 0.43 (0.05-1.82) -

t½e�2 (h)1 413 (328-515)

(17.2 d)

233 (206-298)

(9.7 d)

290 (236-368)

(12.1 d)

t½abs (h) 8.3 (6.2-11.8) 12.8 (7.1-19.4) -

tlag (h) 0.89 (0.03-1.02) 0.55 (0.01-0.86) -

Vss/F (L/kg)1 431 (283-588) 154 (101-210) -

CL/F (L/h/kg) 0.85 (0.67-1.06) 0.80 (0.52-0.96) -

AUC0-42days (µmol*h/L)

76.4 (51.9-84.7) 117.2 (98.7-153.0) 83.6 (58.1-127.4)

AUC0-� (µmol*h/L) 87.4 (63.3-113.1) 122.5 (101.1-165.8) 98.5 (69.5-133.8

1 P<0.001 for piperaquine vs chloroquine comparison

67

Time (h)

0 200 400 600 800 1000

Con

cent

ratio

n (n

M)

1

10

100

1000

10000

Figure 2.1. Plot of measured (circles and grey lines) and predicted (solid line) plasma

piperaquine concentrations against time in 22 Papua New Guinean children treated with

with Duo-cotecxin (approximately 20mg/kg piperaquine phosphate daily for 3 days). The

predicted concentrations were derived from geometric mean parameters for the

pharmacokinetic model.

68

Time (h)

0 200 400 600 800 1000 1200

Con

cent

ratio

n (n

M)

1

10

100

1000

10000

Time (h)

0 200 400 600 800 1000

Con

cent

ratio

n (n

M)

1

10

100

1000

10000

Figure 2.2

A) Plot of measured (circles and grey lines) and predicted (solid black line) plasma

chloroquine concentrations against time in 20 Papua New Guinean children treated with

approximately 10mg/kg chloroquine daily for 3 days. The predicted concentrations were

derived from geometric mean parameters for the pharmacokinetic model.

B) Plot of plasma desethylchloroquine concentrations (circles and grey lines) against

time. The solid black line represents the terminal elimination phase, based on linear

regression of the final four data points (at 7, 14, 28 and 42 days).

A

B

69

Using the model parameters for PQ, a simulation based on a hypothetical dose of

30mg/kg/day for 3 days, is shown in Figure 2.3. Modeled curves based on the actual

doses of PQ and CQ used in the current study are shown for comparison.

Time (h)

0 200 400 600 800 1000 1200

Con

cent

ratio

n (n

M)

1

10

100

1000

10000 Piperaquine 12mg/kg/day x 3 (current study)Chloroquine 10mg/kg/day x 3 (current study)Piperaquine 30mg/kg/day x 3 (simulated)

Figure 2.3 Model-predicted plasma concentration-time plot of piperaquine given at a

dose of 30mg/kg (base) daily for 3 days (solid line; assuming linear kinetics). Plots

showing modeled curves for piperaquine and chloroquine given at the doses used in the

present study are shown for comparison (piperaquine 12 mg base/kg daily for 3 days:

short dashed line, chloroquine 10 mg base/kg daily for 3 days: long dashed line).

70

In DHA-PQ-treated patients on day 7, median concentrations of PQ (78 nM) were lower

than median concentrations of CQ (223 nM) and DECQ (167 nM) seen in CQ-SP-treated

patients, consistent with the significantly larger Vss/F for PQ. This was also the case at

day 28 (median PQ concentration 30 nM, CQ 28 nM and DECQ 35 nM).

The association between drug exposure and post-treatment suppression of infection was

assessed by comparing the PQ and CQ AUC0-42 in patients who experienced a

recurrence of P. falciparum parasitaemia (including all re-infections and one

recrudescence in the CQ-SP group) during the 42-day follow-up period to those in whom

there was no evidence of relapse. There was a lower PQ AUC0-42 (median 59.3 [43.1-

76.4] µmol*h/L) for patients treated with DHA-PQ who experienced relapse than in those

who did not (79.0 [36.7-147.0] µmol*h/L; P=0.036). However, there was no such

relationship in the case of CQ-SP whether the AUC0-42 was that for CQ alone or for the

composite of CQ and DECQ (P>0.6 in each case). Day 7 PQ concentrations were

significantly correlated with PQ AUC0-42 (Spearman correlation coefficient of 0.79,

P<0.01). However there was no statistically significant difference in day 7 PQ

concentration for those who did versus those who did not experience relapse (median 38

vs 43 nmol respectively: P=0.69).

2.2.5. CONCLUSIONS

The present study adds to the limited knowledge of the PK properties of two widely-used

and chemically-related long half-life antimalarial compounds. One (PQ) is likely to see

increased used as part of ACT and the other (CQ) is subject to declining effectiveness in

many areas of the tropics despite co-administration with other drugs such as SP. The

disposition of PQ and CQ in the Melanesian children with uncomplicated malaria in this

study is broadly consistent with limited data from other studies in different contexts. By

suggesting no important ethnically determined differences in PK disposition (that could

compromise its safety or efficacy) it therefore provides preliminary data supportive of

further evaluation for use in treatment of uncomplicated malaria in Melanesian children.

However, the prompt distribution phase of PQ may limit its post-treatment suppressive

properties in areas of high transmission such as PNG and it may be less well suited to

use as IPT. By contrast, the production of DECQ, an active metabolite with a long

terminal elimination phase, may explain why the decline of CQ effectiveness as mono-

therapy has been slower than other quinoline drugs such as mefloquine.

71

The PQ PK parameters from the 22 Melanesian children (who were co-administered

DHA as Duo-cotecxin™) in the present study are consistent with those of the few other

similar published studies. A population PK analysis of Cambodian patients with malaria

by Hung et al. [138] demonstrated similar median values for t½abs (9.3 and 9.1 h for

children and adults, respectively), t½e�2 (13.5 and 22.7 days), Vss/F (614 and 574 L/kg)

and CL/F (1.85 and 0.9 L/h/kg) to those in our patients. Although there was more rapid

clearance in Cambodian children than adults, the median CL/F from the present study

(0.85 L/kg) was closer to that seen in Cambodian adults. This is unlikely to reflect a

difference in age (6.9 ± 1.4 vs 7 ± 2 years in the Cambodian children [138]), but the PNG

children from the present study were heavier (19.1± 3.8 vs 16 ± 4 kg), more likely to be

male (77% vs 55%), had higher parasitaemias (range 3,520–191,000 vs 3,100–33,342

/µL) and were not required to fast like their Cambodian counterparts. These

demographic, anthropometric, infection and design-related differences may also have

contributed to the much shorter calculated median t½e�1 in the present study (0.12 h)

than that reported in Cambodian children and adults (1.7 and 2.6 h, respectively) [138],

but this discrepancy might also reflect the relatively rich sampling schedule and

consequently better characterization of the distribution phase in the PNG children.

A study of PQ by Sim et al. [137] was limited to non-compartmental analysis but derived

similar mean values for t½e�2 (20.3 and 20.9 days), Vss/F (716 and 365 l/kg) and CL/F

(1.14 and 0.6 L/h/kg) in Caucasian adults who were either fasting or fed a high fat meal,

respectively, to those of the present study. Röshamaar et al. [136] also demonstrated a

similar t½e�2 (11.7 days using a 2-compartment and 19 days using a 3-compartment

model) and CL/F (approximately 0.97 L/h/kg) in healthy Vietnamese adults but a much

lower Vss/F (approximately 103 L/kg). This study used a different preparation of PQ,

“CV8” (a co-formulation including trimethoprim and primaquine in addition to DHA).

Therefore, it is possible this lower Vss/F may have reflected drug interactions or other

pharmaceutical factors. In addition, Röshamaar et al. [136] demonstrated multiple

absorption peaks and used a dual pathway first-order absorption PK model in order to

improve the goodness-of-fit, which may also have contributed to the lower calculated

Vss/F.

Although the median t½e�2 of approximately 17 days derived in the current study was

within the range of those from the 3 previous PQ PK studies [136, 137, 138], a study of a

72

single Caucasian volunteer by Tarning et al. [135] employed a 93-day sampling duration

that appeared to suggest prolonged multi-phasic elimination kinetics. The authors

calculated a t½e of 33 days but, based on linear regression of the final three sampling

points, suggested that the real t½e may have even been as long as 80 days. [135]

Because of the difficulties in following up patients in remote rural settings, and the ethical

constraints to performing excessive blood sampling in children, the current study was of

insufficient duration to confirm these findings. However, in the current study, a two-

compartment model generated a curve whose log-linear elimination phase fitted well

with the last 4 data points (from 7 to 42 days: see Figure 2.1). Also, whilst a longer

elimination phase following the last sampling point cannot be excluded, it seems doubtful

that this would be of clinical significance, given the PD data suggesting that

concentrations of drug are likely to be well below the P. falciparum minimal inhibitory

concentration (MIC) by 42 days. Such concentrations should not influence

recrudescence but may have implications for the selection of resistant parasites if re-

infection occurs. [50, 85]

This study sought to compare the PK disposition of PQ with that of the related 4-

aminoquinoline compound CQ (co-administered with SP). Although used extensively for

more than 50 years, CQ pharmacokinetics remain poorly defined and have been

determined mainly in adult patients. [123, 124] However, consistent with the present

study, all have shown large Vd (ranging from 59 to 882 L/kg) [123, 124, 293, 294] and,

when reported, a rapid distribution phase with initial half-lives ranging from 0.16 to 1.07

h. [293, 295, 296] A consistent finding has been the demonstration of multi-exponential

disposition kinetics. However, our median t½e�2 of 9.7 d is consistent with values of 5.8-

20.0 d generated from studies with sampling performed over 28-56 days. [294, 297-303]

One study conducted over 6 months calculated a t½e�2 of 42 days. [293] However, in the

same manner as PQ, the presence of a prolonged elimination phase for CQ is unlikely to

be of clinical significance if this occurs after plasma concentrations have fallen to well

below the parasite MIC. Values for CL/F for CQ calculated in the current study (0.8

L/h/kg) were similar to those previously reported (range 0.13-1.1 L/h/kg). [294, 297-303]

However, many previous studies have relied on non-compartmental methods and none

has documented a value for CQ clearance in children, nor an absorption rate constant or

half-life in adults or children.

73

Although the present study was not powered to assess therapeutic efficacy, it is

encouraging that the 42-day PCR corrected ACPR was 100% for the 19 P. falciparum

patients treated with DHA-PQ. However, ACPR was also 93% in the CQ-SP comparator

group. This suggests that CQ-SP retains some efficacy in this semi-immune pediatric

population despite a high prevalence of genotypic resistance markers for CQ and SP in

the study area. [59] The rate of P. falciparum re-infection seen in both groups confirms

local endemicity but also suggests that, for both drug combinations, plasma

concentrations had fallen to below the parasite MIC well before 42 days post-treatment.

The preliminary observation that a higher PQ AUC was associated with a better

response to treatment at 42 days is consistent with a Cambodian study in which the

patients with recrudescent infections were those given the lowest mg/kg doses. [122]

However, larger-scale PK-PD studies are needed to explore this association further.

In the case of PQ, the apparently short-duration post-treatment chemoprophylactic effect

is perhaps surprising, given its long t½e�2. However, it is notable that its short t½e�1

means that plasma PQ concentrations have already fallen precipitously by day 7. At the

doses used, combined plasma concentrations of CQ and DECQ were almost 5 times

higher at day 7 and twice as high at day 28, than those of PQ. Modeling suggests that,

because of rapid distribution kinetics, even as much as a 2.5-fold increase in daily PQ

dosage in a 3-day regimen is likely to result in only modest increases in drug

concentrations during the elimination phase (see Figure 2.3). Higher acute plasma

concentrations may increase the risk of side-effects such as nausea and vomiting.

The curative and suppressive effects of CQ and PQ will depend not only on their

elimination kinetics but also on the intrinsic susceptibility of local parasite strains. An in

vitro susceptibility of Cameroonian P. falciparum field isolates demonstrated that for CQ-

sensitive strains, mean IC50s for CQ (41.6 nM), DECQ (17.8 nM) and PQ (35.5 nM) were

similar. [142] For CQ-resistant strains (mean CQ IC50 201 nM), PQ and DECQ sensitivity

were preserved (40.7 nM and 37.4 nM, respectively). In vitro drug susceptibility assays

are limited in their capacity to predict therapeutic response and IC50s cannot necessarily

be considered accurate markers of the in vivo parasite MIC. In addition, CQ and DECQ

exist as enantiomers, with stereoselective differences in protein binding, PK and intrinsic

potency. [104] Nonetheless, IC50s can be useful for comparing geographic and temporal

trends in drug resistance and characterising local resistance patterns. Therefore, the

74

similar plasma DECQ concentrations to those of PQ and CQ seen from day 7 to 42 in

our patients may be clinically significant with DECQ contributing more than CQ to post-

treatment prophylaxis. The greater parasite susceptibility of PQ may have been

counterbalanced by the higher levels of CQ and its more potent active metabolite during

the elimination phase. No comparable active metabolite of PQ has yet been described

and seems unlikely based on the present PD data.

Statistically significant differences in the rates of emergent P. vivax infections during the

follow-up period and resolution in gametocytaemia were observed between the two

treatment groups. The former phenomenon is well-described [304] even in patients

treated with ACT [305], representing either suppression of P. vivax parasitaemia to a

sub-microscopic level by a dominant P. falciparum co-infection at the time treatment is

initiated, relapses from hypnozoites, or newly acquired infections. If the drug being used

is more effective against P. falciparum than P. vivax, eradication of P. falciparum will

allow P. vivax parasitaemia to emerge. The pattern seen in the present study could

reflect the known high level of P. vivax resistance to CQ in PNG [128, 278] and suggests

that PQ has greater activity than CQ against local strains. This has recently been

confirmed in nearby West Papua province of Indonesia. [281] The more rapid reduction

in P. falciparum seen in the DHA-PQ arm reflects the well described potent gametocidal

activity of the artemisinin component. [200]

The advent of new malaria control strategies such as ACT and IPT in pregnancy and

infancy mean that characterizing the elimination kinetics of long-acting antimalarial drugs

is of particular importance. Short-course ACT regimens require that the non-artemisinin

partner drug achieve parasiticidal concentrations for several parasite lifecycles in order

to eliminate the sub-microscopic parasite residuum and effect definitive cure. [144] The

effectiveness of IPT strategies may also rest on the duration of the post-treatment

prophylactic phase. [109] Elimination kinetics may have implications for future

development and spread of parasite drug resistance. [50] However, the present study

demonstrates that drug elimination must be considered in association with distribution

kinetics, parasite susceptibility and role of active metabolites. Because drug disposition

can vary significantly according to age and with pregnancy, planned strategies for

deploying drugs such as PQ in pregnant women and infants should include careful PK

evaluation.

75

2.3. Clinical trial 2: An open-label randomised comparison of four combination therapies for treatment of children with malaria in an area of Papua New Guinea with high transmission of multiple Plasmodium species

2.3.1. SUMMARY

Aims: To determine whether, based on efficacy and safety, an artemisinin combination

therapy (ACT) should replace chloroquine+sulphadoxine-pyrimethamine (CQ-SP) as the

currently recommended treatment of uncomplicated malaria in Papua New Guinea

(PNG). Secondary aims were to determine whether clinically-significant differences exist

between the efficacy of the three ACTs against either P. falciparum or P. vivax and to

explore host-, parasite- and drug-specific determinants of treatment response.

Methods: Between April 2005 and July 2007, an open-label, randomised, parallel-group

study of CQ-SP, artesunate-SP (ARTS-SP), dihydroartemisinin-piperaquine (DHA-PQ)

and artemether-lumefantrine (AL) was conducted in children aged 0.5 to 5 years from

Madang and East Sepik Provinces with falciparum and vivax malaria. The primary

endpoint was the 42-day P. falciparum PCR-corrected adequate clinical and

parasitological response (ACPR). Secondary endpoints included the P. vivax PCR-

uncorrected ACPR.

Results: Of 2802 febrile children screened, 482 children with P. falciparum and 195 with

P. vivax were included. All treatments were well-tolerated. The P. falciparum ACPR was

highest in the AL group (95·2%) and 81.5% in CQ-SP, 85.4% in ARTS-SP and 88% in

DHA-PQ treated patients (P=0.001, P=0.02 and P=0.06 vs AL, respectively). Treatment

failure was associated with low Day 7 plasma PQ concentrations and high baseline

parasitaemias. The P. vivax ACPR was highest in the DHA-PQ group (69.4%), and

13.3% in CQ-SP, 33.3% in ARTS-SP and 30.3% in AL treated patients (P<0.001,

P=0.002 and P=0.001 vs DHA-PQ, respectively).

Conclusions: Although AL was the most efficacious regimen against P. falciparum, in

this poly-species high transmission setting it was little better than CQ-SP against P.

vivax. Conversely, whilst DHA-PQ was the most efficacious regimen against P. vivax,

the treatment-failure rate for P. falciparum was unexpectedly high, possibly reflecting PK

76

factors, cross-resistance of local parasite strains between PQ and CQ and relatively high

baseline parasitaemias in this population.

2.3.2. AIMS

2.3.2.1. Primary

To determine whether, based on efficacy and safety, ACT could replace CQ-SP as

standard treatment for uncomplicated malaria in PNG.

2.3.2.2. Secondary

i. To determine whether there are clinically-significant differences between the efficacy

of the three ACTs against either P. falciparum or P. vivax in an area where

transmission of both species is high.

ii. To explore host-, parasite- and drug-specific determinants of treatment response.

2.3.3. METHODS



Michele Senn, Olive Oa, Penias Suano (all of PNG IMR), Irwin Law, Thomas Schulz and

Sam Salman (School of Medicine and Pharmacology, UWA) supervised the field work

and performed some of the clinical procedures. The PNG IMR field staff who performed

the clinical activities (including patient enrolment, evaluation, clinical procedures and

data collection) were Jovitha Lammey, Servina Gommorai, Sr Maria Goretti, Bernard

(“Ben”) Maamu, Wesley Sikuma, Petronilla Wapon, Jane Simbrandu, Merilyn Uranoli,

Peter Maku, and Donald Paiva. Enmoore Lin and Alice Ura performed the

MSP1/MSP2/GLURP genotyping. Nandau Tarongka, Lena Lorry, Moses Lagog and

Kaye Baea performed the microscopy under the supervision of Nandau Tarongka.

The HPLC assays for PQ, CQ and DECQ were developed and validated by Madhu

Page-Sharp (School of Medicine and Pharmacology, UWA) and Ken Ilett (School of

Medicine and Pharmacology, UWA). Madhu Page-Sharp performed the PQ, CQ and

DECQ drug assays. The lumefantrine (LUM) assays were performed by Sam Salman

who also developed and validated this assay.

77

In collaboration with Ivo Mueller and Tim Davis, I was responsible for the original study

concept and design. I prepared the funding application and obtained the ethical

approval. I supervised the setting up of the Clinical trial and, in its initial stages, directly

supervised the data collection, clinical evaluations and procedures. I continued to

supervise the data collection and clinical aspects of the study over the remaining 2 years

of enrolments, by direct liaison with Michele Senn, Olive Oa, Penias Suano, Irwin Law,

Sam Salman and Thomas Schulz. In collaboration with Tim Davis and Ivo Mueller, I

analysed and interpreted the data and prepared final manuscripts for publication

(publication 11). Statistical analysis was provided by Ivo Mueller and Wendy Davis.


This study needed to fulfill operational requirements, necessary to inform public health

policy in PNG, specifically the need or otherwise to change antimalarial treatment policy

to incorporate ACTs as first–line therapy and, if so, the most appropriate ACT.

Therefore, national health planners (Leo Makita, National Malaria Program, PNG

Ministry of Health) were consulted at the study design stage, in association with WHO

(Dapeng Luo, WHO country malaria advisor). Both organizations expressed a high level

of interest in the results of this study.

In order to provide the most robust comparison of treatment efficacy, a prospective,

randomised, parallel-group comparative trial design was used. For practical reasons

(due to total number of individual drugs used in the four arms), placebo-blinding was not

feasible so the study was conducted in an open-label fashion. The primary endpoint was

the reappearance of malarial parasitaemia within 42 days after correction for PCR-

determined re-infection. This was based on standardised WHO methodology [113] but

with a follow-up period extended to 42 days to account for the possible suppressive

effects of long half-life drugs (such as PQ) that might mask the presence of recrudescent

parasitaemia if a shorter follow-up period was used. [108]

Sample size calculations were based on the assumption that the P. falciparum Day 42

PCR-corrected ACPR for each ACT would be �95%, within WHO-recommended range

for adoption of new antimalarial therapy. [64] Because there were no robust local CQ-SP

efficacy data when the study was designed, a sample size of 100 children/treatment arm

was chosen to detect a >5% treatment failure rate in any arm with 5% precision and 95%

78

confidence after allowing for 20% loss to follow-up. [113] Early in 2006, data became

available from Madang and East Sepik Provinces from a 2004 efficacy study that

revealed a CQ-SP P. falciparum Day 28 PCR-corrected ACPR of 76.7%. [280] Under

the conservative assumption that this ACPR persisted to Day 42 and allowing for 20%

attrition, 452 children (113/treatment arm) were required to detect a significant difference

in the primary endpoint between each ACT and CQ-SP with 80% power and a two-tailed

1% Type 1 error. [306] Secondary endpoints were i) reappearance of PCR-uncorrected

P. falciparum parasitaemia within 42 days, ii) reappearance of PCR-corrected and -

uncorrected P. falciparum within 28 days, iii) appearance of any P. vivax parasitaemia

within 28 and 42 days after treatment for vivax malaria and for falciparum malaria, iv)

initial fever and parasite clearance times, v) gametocyte persistence, and v) laboratory

and clinical indices of drug safety during follow-up. These secondary endpoints have all

been commonly used and well-standardised endpoints in a large number of previous

comparative trials of antimalarial efficacy. The primary efficacy analysis was conducted

according to a modified intention-to-treat analysis as well as a time to event analysis,

consistent with current WHO recommendations. [113] Intention-to-treat analysis

assumes patients lost to follow-up to be treatment failures, thus removing the possibility

of results being biased by “missed” treatment failures. However, WHO recommendations

for national policy makers utilise cut-points for absolute efficacy (including a

recommendation to consider changing policy if an existing drug-treatment has an

efficacy <90% and that an efficacy of >95% is acceptable for any new therapy being

considered for national treatment policy) that are based on “per-protocol”, analysis. [64]

There fore the analysis was also performed per protocol so that the results could be

used to guide treatment policy in PNG (see Statistical analysis, 2.3.3.7, below for further

detail).

2.3.3.3. Study sites and patients

The study was conducted at Alexishafen Health Centre, Madang Province between April

2005 and June 2007, and at Kunjingini Health Centre, East Sepik Province between May

2005 and April 2006. Both areas served by these clinics are populated almost

exclusively by Melanesian people and are considered hyper-endemic for P. falciparum

[31, 32] Plasmodium vivax, P. ovale and P. malariae are also endemic in both areas.

79

Children aged 6 months to 5 years who presented with fever (axillary temperature

>37·5°C) or a history of fever in the previous 24 h were screened with a Giemsa-stained

thick blood film. Those with mono-infections of either P. falciparum (density >1000

asexual parasites/�L), or P. vivax, P. ovale or P. malariae (>250 parasites/�L) were

eligible provided that they had i) no features of severe malaria, [24] ii) not taken one of

the study drugs in the previous 14 days, iii) no clinical evidence of another infection, and

iv) no signs of malnutrition or other co-morbidity. Informed consent was obtained from

parents or legal guardians prior to recruitment and all children assented to study

procedures.

2.3.3.4. Clinical procedures

Once consent had been obtained, an initial detailed clinical assessment was performed

and a 3 mL venous blood sample taken. Physical examination included measurement of

height, weight and mid-upper arm circumference (MUAC), and calculation of body mass

index (BMI) and weight-for-age nutrition Z-score (WAZ).[307] Venous blood was taken

for measurement of Hb (HaemoCue™, Angelholm, Sweden) and blood glucose

(HaemoCue™). The remainder of the sample was centrifuged and separated plasma

stored frozen. The cell pellet was retained for parasite genotyping. A repeat blood smear

was prepared for subsequent confirmation of the diagnosis and quantitation of

parasitaemia.

An envelope labeled with a unique sequential code was opened and a computer-

generated randomised treatment read and assigned. This process did not involve the

children, their parents/guardians or clinic staff. A single randomization list was used for

all children (regardless of parasite species on initial microscopy) so that post hoc

assignation of either P. falciparum or P. vivax could be made on the basis of definitive

expert microscopy. Children were allocated i) CQ-SP as CQ (Chlorquin™, Aspen

Healthcare Australia Pty Ltd, St Leonards, Australia) 10 mg base/kg daily for three days

(30 mg base/kg total) and single-dose SP (Fansidar™, Roche, Basel, Switzerland; 25

mg/kg S and 1·25 mg/kg P) given with the first CQ dose, ii) ARTS-SP as single-dose SP

(Roche) as in i) plus ARTS (Sanofi-Aventis, Paris, France) 4 mg/kg daily for three days,

iii) DHA-PQ (Duo-cotecxin™, Beijing Holley-Cotec, Beijing, China) containing DHA and

PQ (40 mg DHA and 320 mg PQ phosphate per tablet) at a dose of 2·5 mg/kg DHA and

20 mg/kg PQ phosphate (equivalent to 11·5 mg base/kg) daily for three days, or iv) AL

80

(Coartem™, Novartis Pharma, Basel, Switzerland) containing ARM 20 mg and LUM 120

mg per tablet at a dose of 1·7 mg/kg A and 10 mg/kg L twice daily for three days (total

dose of A 10·2 mg/kg and L 60 mg/kg).

Doses were given on Days 0, 1 and 2 as combinations of whole, half or quarter tablets

(prepared using a tablet cutter) that best approximated the target dose. Tablets were

swallowed whole with water or, for younger children, crushed lightly and mixed with

water. Children were observed for 30 min afterwards and a repeat dose given to those

who vomited during this time. Subjects were not required to fast. However, since LUM

bioavailability is increased by co-administration of fat, [146] children allocated AL were

also given at least 30 mL cow’s milk or breast-fed at the time of drug administration.

Dosing was directly supervised apart from the three evening AL doses which were

administered at home by the mothers after appropriate instruction and provision of cow’s

milk for dosing in non-breast-fed children. Since Duo-cotecxin tablets were not produced

under Good Manufacturing Practice standards, tablets from each batch were assayed

for content (see below).

All children were asked to re-attend for standardised review on Days 3, 7, 14, 28 and 42.

As well as a detailed assessment including a symptom questionnaire on Days 1, 2, 3

and 7, an axillary temperature, a Giemsa-stained thick blood film and a Hb concentration

were taken at each visit. Blood glucose was measured on Days 1, 2 and 3. Children

who, at any time, developed signs of severe [24] or uncomplicated malaria were given

i.m. quinine or a course of oral quinine together with single-dose SP, respectively,

according to PNG guidelines. [204] Children who were hospitalized or who had an

unscheduled clinic review were considered as having a likely adverse event and a

standard form was completed with details of the clinical assessment and management.

The relationship of each such event to study medication was coded independently by at

least two study physicians as definite, probable, possible or unlikely. Repeated attempts

were made over a week to assess children who did not attend a scheduled visit. If such

attempts were unsuccessful, the child was withdrawn by a parent/guardian or there was

any other breach of protocol, the child was excluded from the study.

81


2.3.3.5.1. Microscopy

All thick blood smears were examined independently by two skilled microscopists who

were blinded to treatment allocation. At least 100 fields were viewed at 1,000x

magnification before a slide was considered negative. Parasite density was determined

by counting asexual parasites per 1,000 leucocytes and assuming a total leucocyte

count of 8,000/�L. Any slide discrepant for positivity/negativity, speciation or parasite

density (with a difference of more than a factor of 3) was referred to a third microscopist.

2.3.3.5.2. Parasite genotyping

Re-infection and recrudescence were distinguished by comparing PCR-restriction

fragment length polymorphism (RFLP)-generated genotype patterns of MSP2 in pairs of

samples obtained at enrollment and on the day of reappearance of parasitaemia. [308,

309] Previous work in Clinical trial 1 and other studies performed by PNGIMR had

shown that additional use of MSP1 and GLURP added little or no discriminative benefit

over and above use of a single allele (MSP2). (Personal correspondence: Ivo Mueller

and Enmoore Lim, PNGIMR). Therefore, genotyping was restricted to use of MSP2 in

this study.


Assay of PQ, CQ and DECQ plasma drug concentrations on Day 7 and PQ tablet

content were identical to those used in Clinical trial 1 (Section 2.2) are as described in

Appendix 1 (Sections 7.11 and 7.12). Lumefantrine assays were performed according to

a HPLC methodology previously described. [146]

2.3.3.6. Efficacy endpoints

Early treatment failure (ETF) was defined as i) danger signs or severe malaria on Days

1–3 in the presence of parasitaemia, ii) Day 2 parasitaemia higher than on Day 0

irrespective of axillary temperature, iii) parasitaemia on Day 3 with axillary temperature

�37·5°C, or iv) parasitaemia on Day 3 �25% of baseline. Any child developing

parasitaemia between Days 4 and 42 was classified as late parasitological failure (LPF)

or, if this was associated with an axillary temperature >37·5oC, late clinical failure (LCF).

Remaining patients were classified as having an adequate parasitological and clinical

response (ACPR).

82

Parasite clearance time (PCT) and fever clearance time (FCT) were taken as the

number of days to the first of at least two follow-up assessments in which the child was

afebrile (axillary temperature <37·5oC) and slide-negative, respectively.


Statistical analysis was according to an a priori plan. Data were double-entered with

discrepancies resolved by a trained data manager. Per-protocol analyses included

children with complete follow-up (up to and including Days 28 and 42) or a confirmed

treatment failure, and excluded those treated for malaria without confirmatory

microscopy or who defaulted from follow-up despite repeated attempts at contact. These

excluded patients were retained in modified intention-to-treat analyses utilizing i) a

worst-case approach (ETF assumed for Day 3 exclusions, LPF/LCF otherwise) and ii) a

best-case approach (all missing follow-up blood films assumed parasite-negative). In

PCR-corrected per-protocol analyses, only children with confirmed recrudescences were

considered failures; those with LCF caused by a different Plasmodium species, a PCR-

confirmed new infection or who had been treated outside the study were excluded

unless they had experienced a prior PCR-corrected failure. Between-group differences

in baseline characteristics were assessed by chi-squared test and ANOVA. All

comparisons of efficacy between treatments were by chi-squared and Fisher’s exact

tests. Kaplan-Meier estimates were computed for each endpoint defined by parasite

species. Treatments were compared by log-rank test, including post hoc comparisons

between ACT arms. No interim efficacy analyses were performed. Cox regression using

backward-stepwise modeling was used to determine predictors of treatment failure

(based on the “worst-case” definitions used in the modified intention to treat analysis)

among age, sex, measures of growth/nutrition and baseline parasitaemia (pre-specified),

and Day 7 drug concentrations (exploratory). Safety and tolerability were assessed by

comparison of the incidence of symptoms by treatment during the first 7 days in all

children who completed a full treatment course. Common symptoms were compared

using Poisson regression with the individual exposure defined as number of

observations per child. Infrequent symptoms were compared using Fisher’s exact test,

and Hb and blood glucose levels using ANOVA.

83

2.3.4. RESULTS

2.3.4.1. Trial profile and patient characteristics

Of 2,802 children screened (see Figure 2.4A), 742 met the eligibility criteria but 83

(11.2%) of these were excluded post hoc because they received incorrect treatment

(n=24) or ineligibility was detected after treatment (n=59). In the majority of this latter

group, confirmatory microscopy revealed a parasite density below the inclusion

threshold (n=41). Of the remaining 659 children, 482 (73·6%) were slide-positive for P.

falciparum and 195 (29·6%) for P. vivax. There were 21 patients (3·2%) with mixed P.

falciparum/P. vivax infections who were included in both arms because their parasite

densities were above each species-specific threshold, and a further 3 children with P.

malariae mono-infections who were not included in the formal efficacy analysis. Among

the 482 children with P. falciparum infections, there were 24 with a second species (P.

vivax or malariae) at densities below the inclusion threshold. Similarly, 4 patients with P.

vivax had a low-level P. falciparum co-infection.

For the P. falciparum cases, a higher percentage of children were lost to follow-up in the

CQ-SP group at Day 28 compared to the other three groups combined (22·7% vs 9·7%,

P<0·001; see Figure 2.4B), a difference that was also evident at Day 42 (P<0·001). For

the P. vivax treatment arm, there was no difference in attrition rates between treatments

(P=0·47; see Figure 2.4C). The P. vivax cases were younger than those with P.

falciparum (25 ± 14 vs 33 ± 14 months) and had lower parasite densities (geometric

mean [SD range]; 4,863 [1,155-20,474] vs 49,804 [14,744-168,226] /µL; P<0·001 in both

cases), but there were no other differences in demographic, clinical or parasitological

variables between groups categorised by parasite species or treatment allocation (Table

2.4). Total mg/kg dose varied according to where each child sat within the dosage

banding schedule. Expressed as a percentage of the target dose, actual mg/kg doses

ranged from 78-150% for CQ-SP, 51-154% for ARTS-SP, 51-160% for DHA-PQ and

54%-120% for AL.

2.3.4.2. Efficacy against P. falciparum

Outcomes (both PCR-uncorrected and PCR-corrected) by treatment at Days 28 and 42

are summarised in Table 2.5. A high proportion of children in all groups re-developed P.

falciparum parasitaemia (overall 19·4% by Day 28 and 34·1% by Day 42), the majority of

whom (>80%) were LPF. There were no between-group differences in PCR-uncorrected

84

treatment outcomes at Day 28 (Day 28 ACPR 80.0%, 77.7%, 77.5% and 82.3% for CQ-

SP, ARTS-SP, DHA-PQ and AL, respectively, P=0.79) and Day 42 (Day 42 ACPR

67.9%, 63.9%, 62.6% and 64.2% for CQ-SP, ARTS-SP, DHA-PQ and AL, respectively,

P=0.79) However, the AL group had a higher PCR-corrected Day 28 ACPR (97·3%) than

all other treatments (�90·1%, p�0·027; see Table 2.5B and Figure 2.5) which was also

evident at Day 42 (95·2% vs �88·0%, p�0·063). There was a similar result in the best-

case intention-to-treat analysis while, in the worst case model, treatment failure was

greatest with CQ-SP (see Table 2.6).

The mean [95%CI] PCT in the CQ-SP group (4.3 [3.5-5.1] days) was slower than with

the other three regimens (means �3.1 days, P<0·001; Figure 2.6 and Table 2.7) and

children treated with AL showed a slower PCT (P=0·012; Table 2.7) and higher parasite

densities on Days 1 and 2 than those allocated ARTS-SP and DHA-PQ (P<0·001; Figure

2.6). FCT was also longer in the CQ-SP group than in the other three (P=0·048; Table

2.7). The prevalence of gametocytaemia in the CQ-SP group was higher that that of the

three ACT regimens between Days 3 and 28 (Figure 2.7).

85

Figure 2.4A: Clinical trial 2 trial profile showing numbers of patients remaining in the study from screening to Day 42 assessment.

2,802 children with fever/history of fever

742 enrolled and randomised

1968 did not meet eligibility criteria 92 refused consent

83 excluded post hoc because they did not meet eligibility criteria (59) or violated protocol (24)

482 falciparum malaria (including 40 mixed vivax, 5 mixed malariae)

195 vivax malaria (including 25 mixed falciparum)

Chloroquine- sulfadoxine-

pyrimethamine n=110

Artesunate- sulfadoxine-

pyrimethamine n=122

Dihydro- artemisinin- piperaquine

n=123

Artemether- lumefantrine

n=127


pyrimethamine n=61


pyrimethamine n=51


n=44


N=39

Day 28 n=85 (all)

n=85 (PCR-corrected)

Day 28 n=112 (all)


Day 28 n=111 (all)


Day 28 n=113 (all)


Day 28 n=51 (all)

Day 28 n=39 (all)

Day 28 n=38 (all)

Day 28 n=33 (all)

Day 42 n=81 (all)


Day 42 n=108 (all)


Day 42 n=107 (all)


Day 42 n=109 (all)


Day 42 n=46 (all)

Day 42 n=39 (all)

Day 42 n=36 (all)

Day 42 n=33 (all)

659 eligible and followed protocol

3 malariae malaria (not analysed)

86

Figure 2.4B: Trial profile showing reasons for exclusion prior to day 28 and day 42 endpoints (patients with P. falciparum).


pyrimethaminen=110

Day 28 n=85 (all)


Day 42 n=81 (all)



pyrimethamine n=122

Day 28 n=112 (all)


Day 42 n=108 (all)



n=123

Day 28 n=111 (all)


Day 42 n=107 (all)



n=127

Day 28 n=113 (all)


Day 42 n=109 (all)


All: 23 lost, 1 protocol

violation, 1 withdrawn PCR-corrected: 23 lost, 1 protocol

violation, 1 withdrawn


violations, 1 withdrawn PCR-corrected: 10 lost, 1 protocol

violation, 1 withdrawn


violations PCR-corrected: 10 lost, 2 protocol

violations


violations, 1 withdrawn PCR-corrected:

11 lost, 4 protocol violations, 1 withdrawn



violations


violation PCR-corrected: 3 lost, 4 protocol

violations



violations


violation PCR-corrected: 4 lost, 3 protocol

violations

87

Figure 2.4C: Trial profile showing reasons for exclusion prior to day 28 and day 42 endpoints (patients with P. vivax).


pyrimethaminen=61

Day 28 n=51 (all)

Day 42 n=46 (all)


pyrimethamine n=51

Day 28 n=39 (all)

Day 42 n=39 (all)


n=44

Day 28 n=38 (all)

Day 42 n=36 (all)


n=39

Day 28 n=33 (all)

Day 42 n=33 (all)

7 lost, 3 protocol violations

11 lost, 1 protocol violation







88

Table 2.4: Baseline characteristics of the 659 Papua New Guinean children with uncomplicated malaria enrolled in Clinical trial 2 categorised by parasite species. Data are, unless otherwise stated, percentages or means.

* Weight range was 0.5-5.0 in all treatment groups

Plasmodium falciparum (n = 482) Plasmodium vivax (n = 195)

CQ-SP ARTS-SP

DHA-PQ AL

P-value

CQ-SP ARTS-SP

DHA-PQ AL

P-value

Gender (% female) 52 39 49 43 0.18 54 45 59 51 0.58

Age (months) 36 35 37 38 0.38 27 24 24 25 0.49

Weight (kg)* 11.4 11.3 11.2 11.7 0.39 10.0 9.6 9.5 9.4 0.60

WAZ -1.68 -1.75 -1.79 -1.74 0.89 -1.66 -1.66 -1.75 -1.87 0.78

MUAC (cm) 14.1 14.0 14.5 14.0 0.56 13.5 13.9 13.5 14.0 0.73

BMI (kg/m2) 14.5 14.6 14.6 14.6 0.94 15.5 15.2 15.4 15.2 0.91

Parasitaemia (/µµµµL) (median and range)

43,900 [1,160,

467,000]

51,000 [1,600,

366,000]

56,000 [1320,

610,000]

48,500 [2300,

450,000]

0.49 4,070 [320,

55,700]

5,720 [320,

86,9000]

5,890 [320,

93,800]

4,190 [280,

223,000]

0.43

Axillary temperature (°C)

38.0 38.0 37.9 38.0 0.95 37.3 37.7 37.2 37.1 0.18

Enlarged spleen (%) 52 51 52 59 0.56 40 37 36 44 0.88

Hemoglobin (g/L) 85 86 83 84 0.58 87 90 89 90 0.71

Blood glucose (mmol/L)

7.2 7.0 7.0 7.0 0.95 7.2 6.9 7.1 7.2 0.86

Heart rate (/min) 119 120 125 118 0.23 119 113 125 111 0.17

Respiratory rate (/min) 33 33 33 35 0.49 34 37 35 36 0.81

89

Table 2.5A: Per-protocol analysis of treatment response in children with falciparum malaria by treatment allocation – PCR

uncorrected data (see text). The 95% confidence intervals (CI) for %ACPR are given in each case.

CQ-SP ARTS-SP DHA-PQ AL Total

Total 85 112 111 113 421

ETF 3 (3.5%) 2 (1.8%) 0 (0.0%) 0 (0.0%) 5 (1.2%)

LCF 1 (1.2%) 3 (2.7%) 2 (1.8%) 3 (2.7%) 9 (2.1%)

LPF 13 (15.3%) 20 (17.9%) 23 (20.7%) 17 (15.0%) 73 (17.3%)

ACPR 68 (80.0%) 87 (77.7%) 86 (77.5%) 93 (82.3%) 334 (79.3%)

Day 28

PCR-uncorrected

P=0.79

[95% CI]

[69.9-87.9] [68.8-85.0] [68.6-84.9] [74.0-88.8] [75.1-83.1]

Total 81 108 107 109 405

ETF 3 (3.7%) 2 (1.9%) 0 (0.0%) 0 (0.0%) 5 (1.2%)

LCF 2 (2.5%) 6 (5.6%) 7 (6.5%) 6 (5.5%) 21 (5.2%)

LPF 21 (25.9%) 31 (28.7%) 33 (30.8%) 33 (30.3%) 118 (29.1%)

ACPR 55 (67.9%) 69 (63.9%) 67 (62.6%) 70 (64.2%) 261 (64.4%)

Day 42

PCR-uncorrected

P=0.90

[95% CI]

[56.6-77.8] [54.1-72.9] [52.7-71.8] [55.5-73.2] [59.6-69.1]

90

Table 2.5B: Per-protocol analysis of treatment response in children with falciparum malaria by treatment allocation –PCR corrected

data (see text). The 95% confidence intervals (CI) for %ACPR are given in each case.


Total 85 110 111 111 417

ETF 3 (3.5%) 2 (1.8%) 0 (0.0%) 0 (0.0%) 5 (1.2%)

LCF 1 (1.2%) 0 (0.0%) 1 (0.9%) 1 (0.9%) 3 (0.7%)

LPF 9 (10.6%) 9 (8.2%) 10 (9.0%) 2 (1.8%) 30 (7.2%)

ACPR 72 (84.7%) 99 (90.0%) 100 (90.1%) 108 (97.3%) 379 (90.9%)

Day 28

PCR-corrected

P=0.022

[95% CI]

[75.3-91.6] [82.8-94.9] [83.0-94.9] [92.3-99.4] [87.7-93.5]

Total 81 103 100 104 388

ETF 3 (3.7%) 2 (1.9%) 0 (0.0%) 0 (0.0%) 5 (1.3%)

LCF 1 (1.1%) 0 (0.0%) 1 (1.0%) 2 (1.9%) 4 (1.0%)

LPF 11 (13.6%) 13 (12.6%) 11 (11.0%) 3 (2.9%) 38 (9.8%)

ACPR 66 (81.5%) 88 (85.4%) 88 (88.0%) 99 (95.2%) 341 (87.6%)

Day 42

PCR-corrected

P=0.030

[95% CI]

[71.3-89.2] [77.1-91.6] [80.0-93.6] [89.1-98.4] [83.9-90.7]

91

Table 2.6A: Modified intention-to-treat analysis of treatment response in children with falciparum malaria by treatment allocation –

PCR uncorrected data (see text). The 95% confidence intervals (CI) for %ACPR are given in each case.


Total 110 122 123 127 482

ETF 5 (4.5%) 5 (4.1%) 4 (3.3%) 3 (2.4%) 17 (3.5%)

LCF 26 (23.6%) 16 (13.0%) 13 (10.6%) 19 (15.0%) 74 (15.4%)

LPF 12 (10.9%) 19 (15.6%) 23 (18.7%) 15 (11.8%) 69 (14.3%)

ACPR 67 (60.9%) 82 (67.2%) 83 (67.5%) 90 (70.9%) 322 (66.8%)

Day 28

PCR-uncorrected

P=0.033

[95% CI]

[51.2-70.1] [58.1-75.4] [58.4-75.6] [61.1-78.6] [62.4-71.0]

Total 110 122 123 127 482

ETF 5 (4.5%) 5 (4.1%) 4 (3.3%) 3 (2.4%) 17 (3.5%)

LCF 29 (26.4%) 23 (18.9%) 22 (17.9%) 26 (20.5%) 100 (20.7%)

LPF 20 (18.2%) 29 (23.8%) 33 (26.8%) 31 (24.4%) 113 (23.4%)

ACPR 56 (50.9%) 65 (53.3%) 64 (52.0%) 67 (52.8%) 252 (52.3%)

Day 42

PCR-uncorrected

P=0.99

[95% CI]

[41.2-60.6] [44.0-62.4] [42.8-61.1] [42.7-61.7] [47.7-56.8]

92

Table 2.6B: Modified intention-to-treat analysis of treatment response in children with falciparum malaria by treatment allocation –

PCR corrected data (see text). The 95% confidence intervals (CI) for %ACPR are given in each case.


Total 110 123 123 127 482

ETF 5 (4.5%) 5 (4.1%) 4 (3.3%) 3 (2.4%) 17 (3.5%)

LCF 26 (23.6%) 13 (10.7%) 12 (9.8%) 17 (13.4%) 68 (14.1%)

LPF 8 (7.3%) 9 (7.4%) 10 (8.1%) 1 (0.8%) 28 (5.8%)

ACPR 71 (64.5%) 95 (77.9%) 97 (78.9%) 106 (83.5%) 369 (76.6%)

Day 28

PCR-corrected

P=0.001

[95% CI]

[54.9-73.4] [69.5-84.9] [70.6-85.7] [75.8-89.5] [72.5-80.3]

Total 110 122 123 127 484

ETF 5 (4.5%) 5 (4.1%) 4 (3.3%) 3 (2.4%) 17 (3.5%)

LCF 29 (26.4%) 17 (13.9%) 17 (13.8%) 22 (17.3%) 85 (17.6%)

LPF 10 (9.1%) 13 (10.7%) 11 (9.8%) 2 (1.6%) 37 (7.7%)

ACPR 66 (60.0%) 87 (71.3%) 90 (73.2%) 100 (78.7%) 343 (71.2%)

Day 42

PCR-corrected

P=0.015

[95% CI]

[50.2-69.2] [62.4-79.1] [64.4-80.8] [70.6-85.5] [66.9-75.2]

93

Table 2.7: Fever and parasite clearance times by Plasmodium species and treatment allocation. Data are number (%) and median

[interquartile range].

Plasmodium falciparum (n = 482) Plasmodium vivax (n = 195)

CQ-SP

(n = 110)

ARTS-SP

(n = 122)

DHA-PQ

(n = 123)

AL

(n = 127)

P-value

CQ-SP

(n = 61)

ARTS-SP

(n = 51)

DHA-PQ

(n = 44)

AL

(n = 39)

P-value

Febrile at day 1 (approx. 24 hours)

12 (11) 8 (7) 8 (7) 16 (12) <0.001 10

(16)

2

(4)

4

(10)

4

(10)

<0.001

PC%24 47.1 [13.6-116.1]

0.8

[0.2-2.5]

1.6

[0.3-3.8]

3.5

[1.0-15.4]

<0.001 3.5

[0-16.1]

0 [0-0] 0 [0-0] 0 [0-0] <0.001

94

Figure 2.5 Kaplan-Meier plots showing proportion of patients remaining free of infection

A: PCR-corrected P. falciparum. B: P. vivax.

B

A

95

Figure 2.6 Mean changes in parasitaemia from baseline by allocated treatment between

Days 0 and 7.

Figure 2.7 Percentage of children who were gametocyte slide-positive by allocated

treatment between Days 0 and 42.

0

20

40

60

80

100

0 1 2 3 4 5 6 7

Time (days)

Par

asit

aem

ia a

s p

erce

nta

ge

of

bas

elin

eCQ-SPARTS-SPDHA-PQAL

0

20

40

60

80

100

0 7 14 21 28 35 42

Time (days)

Per

cent

age

of p

atie

nts

gam

etoc

ytae

mic

CQ-SPARTS-SPDHA-PQAL

96

2.3.4.3. Predictors of treatment failure in P. falciparum cases

In Cox proportional hazards modeling of pooled P. falciparum outcome data, the overall

risk of treatment failure (ETF, LCF or LPF) with AL at Day 42 was 75% less than with

CQ-SP (Hazard ratio (HR) [95% CI] 0·25 [0·09, 0·69], P=0·007), 69% less than ARTS-

SP (0·31 [0·11, 0·85], P=0·022) and 61% less than DHA-PQ (0·39 [0·14, 1·10], P=0·074).

Risk of treatment failure was also independently associated with BMI (1·22 [1·09, 1·38]

for each 1·0 kg/m2 increase, P=0·001) and axillary temperature (2·33 [1·18, 4·61] for an

increase of 1·0oC, P=0·015). There was no difference in treatment failure between the

two study sites (P=0·365).

In separate Cox models for each treatment arm, treatment failure in the ARTS-SP group

was independently associated with age (1·12 [1·05, 1·20] for each 1 year increase,

P=0·001) and BMI (1·43 [1·16, 1·76] for each 1·0 kg/m2 increase, P=0·001). In the DHA-

PQ group, independent predictors of treatment failure were baseline parasitaemia (1·07

[1·03, 1·11] for each 10,000/�L increase, P=0·001) and WAZ (1·22 [1·09, 1·38], for each

1·0 unit increase, P=0·001). Risk of treatment failure in the DHA-PQ group increased

substantially for a baseline parasitaemia >200,000/�L (5·59 [1·77, 17·69], P=0·003). No

variables were associated with treatment failure in the other two treatment groups.

2.3.4.4. Efficacy against P. vivax

Outcomes by treatment for the 154 children with P. vivax infections are shown in Table

2.8. Almost two-thirds of all patients re-developed P. vivax parasitaemia during the 42-

day follow-up period and 23% of these (15% of the total) were clinical failures. The

highest ACPR at both Day 28 and Day 42 was in the DHA-PQ group (P �0·002 vs the

other three regimens; Table 2.8 and Figure 2.5B). The lowest rates of LCF were also

seen with DHA-PQ (2·8% at Day 42 vs �15·2%, P �0·02; Table 2.8). The least

efficacious treatment was CQ-SP, a group in which only 13·0% of children remained free

of P. vivax for 42 days (Table 2.8). Parasite clearance time was also slower in the CQ-

SP group compared to the other regimens (log rank-test, P<0·001). Intention-to-treat

analyses were consistent with per-protocol results at Days 28 and 42 (see Table 2.9).

97

Table 2.8 Per-protocol analysis of treatment response in children with vivax malaria by treatment allocation (see text). The 95%

confidence intervals (CI) for %ACPR are given in each case. Data are not PCR-corrected.


Total 51 39 38 33 161

ETF 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%)

LCF 4 (7.8%) 1 (2.6%) 0 (0.0%) 2 (6.1%) 7 (4.3%)

LPF 21 (41.2%) 18 (46.2%) 6 (15.8%) 15 (45.5%) 60 (37.3%)

ACPR 26 (51.0%) 20 (51.3%) 32 (84.2%) 16 (48.5%) 94 (58.4%)

Day 28

P= 0.003

[95% CI]

[36.6-65.2] [34.8-67.6] [68.7-94.0] [30.8-66.5] [50.4-66.1]

Total 46 39 36 33 154

ETF 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%)

LCF 10 (21.7%) 7 (17.9%) 1 (2.8%) 5 (15.2%) 23 (14.9%)

LPF 30 (65.3%) 19 (48.7%) 10 (27.8%) 18 (54.5%) 77 (50.0%)

ACPR 6 (13.0%) 13 (33.3%) 25 (69.4%) 10 (30.3%) 54 (35.1%)

Day 42

P< 0.001

[95% CI]

[4.9-26.3] [19.1-50.2] [51.9-83.7] [15.6-48.7] [27.6-43.2]

98

Table 2.9 Modified intention-to-treat analysis of treatment response in children with vivax malaria by treatment allocation (see text).

The 95% confidence intervals (CI) for %ACPR are given in each case. Data are not PCR-corrected.


Total 61 51 44 39 195

ETF 2 (3.3%) 4 (7.8%) 0 (0.0%) 1 (2.6%) 7 (3.6%)

LCF 12 (19.7%) 12 (23.5%) 6 (13.6%) 7 (17.9%) 37 (19.0%)

LPF 19 (31.1%) 15 (29.4%) 6 (13.6%) 15 (38.5%) 55 (28.2%)

ACPR 28 (45.9%) 20 (39.2%) 32 (72.7%) 16 (41.0%) 96 (49.2%)

Day 28

P=0.033

[95% CI] [33.1-59.2] [25.8-53.9] [57.2-85.0] [25.6-57.9] [42.0-56.5]

Total 61 51 44 39 195

ETF 2 (3.3%) 4 (7.8%) 0 (0.0%) 1 (2.6%) 7 (3.6%)

LCF 20 (32.8%) 16 (31.4%) 8 (18.2%) 10 (25.6%) 54 (27.7%)

LPF 28 (45.9%) 17 (33.3%) 10 (22.7%) 18 (46.2%) 73 (37.4%)

ACPR 11 (18.0%) 14 (27.5%) 26 (59.1%) 10 (25.6%) 61 (31.3%)

Day 42

P=0.99

[95% CI] [9.4-30.0] [15.9-41.7] [43.2-73.7] [13.0-42.1] [24.8-38.3]

99

2.3.4.5. Predictors of treatment failure in P. vivax cases

In Cox regression analysis, the risk of treatment failure (ETF, LCF and LPF) with DHA-

PQ was 76% less than for CQ-SP (0·24 [0·12, 0·78], P<0·001), 66% less than ARTS-SP

(0·34 [0·17, 0·70], P=0·003) and 70% less than AL (0·30 [0·15, 0·62], P=0·001). Despite

the higher total number of treatment failures at Day 42 in the CQ-SP group (P=0·03), the

overall failure rates of CQ-SP, ARTS-SP and AL were similar (log rank test, P=0·54).

Age, sex, BMI, WAZ and baseline parasitaemia were not independently associated with

the risk of treatment failure.

2.3.4.6. Emergence of P. vivax after treatment for P. falciparum

Approximately half (52%) of the 371 children with P. falciparum mono-infections by

microscopy developed P. vivax parasitaemias during the 42-day follow-up period (see

Table 2.10). Of these only 4% were symptomatic infections. Consistent with the P. vivax

mono-infection treatment data, these emergent infections occurred least in DHA-PQ

treated children (P<0·001), with CQ-SP, ARTS-SP and AL having similar ACPR rates

(P>0·4; Table 2.10). The percentage of children with an emergent P. vivax infection

following treatment for P. falciparum was lower than that of recurrent infections in P.

vivax cases in children treated with CQ-SP (60·0% vs 87·0%, P<0·001) but not with the

three ACTs (P>0·4).

2.3.4.7. Relationship between Day 7 drug concentrations and outcome

In children with P. falciparum in the DHA-PQ and AL groups, there was a trend towards

a lower risk of any treatment failure (PCR-uncorrected) with higher Day 7 plasma PQ

(HR 0·86 [0·73, 1·01] for a 10 µg/L increase, P=0·06) and LUM (0·87 [0·74, 1·02] for a

100 µg/L increase, P=0·09) concentrations, respectively, in univariate analyses. Two of

the three AL treatment failures had a plasma LUM <175 µg/L, and seven of 10 DHA-PQ

treatment failures had a plasma PQ <30 �g/L. There was no association between Day 7

plasma CQ and DECQ concentrations and treatment failure (P>0·6). Plasma

concentrations of both CQ (0·97 [0·95, 1·00], P=0·043) and DECQ (0·97 [0·94, 0·99],

P=0·012) were negatively associated with treatment failure in P. vivax cases. In P.

falciparum cases with plasma LUM <175 µg/L and plasma PQ <20 µg/L, there was an

increased risk of emergent P. vivax infections during follow-up (HR 2·65 [1·43, 4·91] and

2·42 [1·03, 5·70], respectively, P�0·043).

100

Day 7 concentrations did not correlate with the actual mg/kg dose given for either PQ

(Spearman correlation coefficient, R= 0.06, P=0.5) or LUM (R=0.09, P=0.3). Children in

the lowest quartiles for mg/kg dosing were no more likely to have day 7 concentrations

below the 30 µg/L cut-off for PQ (10 of 21 vs 54 of 104: P=0.7) or the 175ug/L cut-off for

LUM (8 of 18 vs 30 of 115: P=0.1) than those in the upper 3 quartiles.

101

Table 2.10: Per-protocol analysis of P. vivax infections occurring after treatment of patients with P. falciparum mono-infections at

enrolment (see text). The 95% confidence intervals (CI) for %ACPR are given in each case.


Total 75 94 100 102 371

ETF 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%)

LCF 2 (2.7%) 1 (1.1%) 1 (1.0%) 4 (3.9%) 8 (2.2%)

LPF 43 (57.3%) 55 (58.5%) 26 (26.0%) 62 (60.8%) 186 (50.1%)

ACPR 30 (40.0%) 38 (40.4%) 73 (73.0%) 36 (35.3%) 177 (47.7%)

Day 42

P< 0.001

[95% CI]

[28.9-52.0] [30.4-51.0] [63.2-81.4] [26.1-45.4] [42.5-52.9]

102

2.3.4.8. Summary of relative efficacy

A summary of relative efficacy of the three ACT regimens against CQ-SP is shown in

Table 2.11. In Cox proportional hazards modeling of all re-treatments (both inside and

outside the study), clinical failures and any (clinical or parasitological) treatment failure,

DHA-PQ was better than CQ-SP in all categories apart from single-species clinical

failure and any (clinical or parasitological) failure of the same P. falciparum genotype.

Artemether-lumefantrine was better than CQ-SP for the latter category with a 75% risk

reduction.

2.3.4.9. Safety monitoring

All four antimalarial regimens were well tolerated and there were no withdrawals

attributable to drug-related adverse effects. Data from side-effect questionnaires

administered between Day 0 and Day 7 showed a lower incidence rate ratio (IRR) for

fever (P=0·35) and vomiting (P=0·012) with the three ACT regimens than CQ-SP (see

Table 2.12) with trends to fewer chills (P=0·062) and less pruritus (P=0·091), but there

was no between-treatment difference for other symptoms including headache,

weakness, myalgia, abdominal pain, diarrhoea, dyspnoea and cough. The results of

physical examination findings to Day 7 showed a higher IRR for skin rash with ARTS-SP

and DHA-PQ than CQ-SP (P=0·004) but, when combined with all skin lesions observed,

this difference was no longer significant (see Table 2.12). The presence of a palpable

spleen was least in the DHA-PQ group (P=0·006).

Haemoglobin concentrations were similar across the four treatment groups on each day

of follow-up, with mean levels on Day 42 �17 g/L more than at study entry in each case

(data not shown). Although blood glucose concentrations were also similar on Days 1

and 2, the mean concentration in the CQ-SP treated children on Day 3 (6·5 ± 1·7

mmol/L) was higher than in those treated with ARTS-SP, DHA-PQ or AL (6·0 ± 1·4

mmol/L, 6·2 ± 1·4 mmol/L and 6·1 ± 1·4 mmol/L, respectively, P=0·037). No child

developed clinically significant hypoglycaemia during the study.

103

Table 2.11: Relative efficacy (day 42) of the three ACT regimens in comparison to chloroquine+sulphadoxine-pyrimethamine.

Hazard ratios (HR) and [95% confidence intervals] are shown.

ARTS-SP DHA-PQ AL

HR [95% CI] HR [95% CI] HR [95% CI.] P-value

Re-treatments (including outside the study)

0.57 [0.32,1.00] 0.39 [0.21,0.74] 0.88 [0.53,1.46] 0.009

Clinical failure (any species) 0.80 [0.42,1.55] 0.42 [0.29,0.93] 0.88 [0.46,1.68] 0.13

Clinical failure (same species) 0.85 [0.59,1.22] 0.48 [0.20,1.16] 0.62 [0.28,1.40] 0.33

Any failure (all species) 0.98 [0.75,1.28] 0.53 [0.39,0.71] 1.00 [0.77,1.31] < 0.001

Any failure (same species) 0.71 [0.48,1.07] 0.28 [0.17,0.46] 0.77 [0.52,1.14] < 0.001

Any failure (same genotype,

P. falciparum only)

0.83 [0.40,1.69] 0.66 [0.31,1.40] 0.25 [0.09,0.70] 0.024

104

Table 2.12: Incidence rate (IR) of reported or observed signs and symptoms during first 7 days of follow-up, expressed as reports per 100 observations. 95% confidence intervals (CI) for Poisson analysis are given for the difference between treatment groups in the case of signs or symptoms with P�0.1. CQ-SP ARTS-SP DHA-PQ AL

Number of observations 644 620 639 641

IR 95% CI IR 95% CI IR 95% CI IR 95% CI P-value

Fever 44.3 [39.3,49.7] 42.9 [37.9,48.4] 37.9 [33.3,43.0] 37.0 [32.4,42.0] 0.004

Chills 2.3 [1.3,3.8] 0.6 [0.2,1.7] 1.6 [0.8,2.9] 1.9 [1.0,3.3] 0.06

Headaches 2.0 [1.1,3.5] 0.6 [0.2,1.7] 0.8 [0.3,1.8] 0.8 [0.3,1.8] 0.07

Child irritable/frequent crying 5.3 [3.7,7.4] 5.6 [3.9,7.9] 3.8 [2.4,5.6] 2.8 [1.7,4.4] 0.06

Trouble sleeping 3.0 [1.8,4.6] 2.9 [1.7,4.6] 1.6 [0.8,2.9] 1.7 [0.9,3.1] 0.02

Problems eating/breast feeding 5.0 5.6 4.2 3.4 0.36

Vomiting 4.0 [2.6,5.9] 1.9 [1.0,3.4] 1.6 [0.8,2.9] 2.2 [1.2,3.7] 0.01

Diarrhoea 2.0 2.7 3.3 2.7 0.58

Abdominal pain 1.6 2.4 1.7 1.6 0.71

Cough 19.9 24.0 22.7 24.0 0.43

Difficulty breathing 1.6 1.6 1.6 0.8 0.50

Rhinorrhoea 21.1 23.5 20.5 22.9 0.77

Skin rash (reported) 0.5 [0.1,1.4 1.3 [0.6,2.5] 1.3 [0.5,259] 0.2 [0.0,0.9] 0.03

Skin rash (observed) 0.2 [0.0,0.9] 1.0 [0.4,2.1] 1.4 [0.6,2.7] 0.0 [0.0,0.6] 0.004

Skin lesion (observed) 3.1 [1.9,4.8] 1.6 [0.8,3.0] 3.6 [2.3,5.4] 2.3 [1.3,3.9] 0.10

Pallor 14.1 12.3 13.9 15.4 0.21

Crackles at lung bases 2.3 2.6 3.4 2.0 0.44

Presence of enlarged spleen 40.8 [36.1,46.1] 39.4 [34.6,44.6] 31.3 [27.1,36.0] 42.4 [37.5,47.8] 0.006

105

2.3.5. CONCLUSIONS

The present study provides valuable comparative data on different ACTs in an area

in which prolonged use of CQ and SP has resulted in increasing treatment failure.

[280] The range of data collected allows a unique insight into host-, parasite- and

drug-specific determinants of the efficacy of the four different treatment regimens

under study. In addition, the findings highlight the difficulties in performing and

evaluating an efficacy study in an area with intense transmission of multiple

Plasmodium species. Based on per-protocol PCR-corrected treatment failure rates

for P. falciparum, AL is the treatment of choice for PNG children with uncomplicated

infections but this combination is significantly less efficacious than DHA-PQ against

P. vivax, both as treatment for a primary infection and in suppressing its emergence

after treatment for P. falciparum. The overall efficacy of CQ-SP and ARTS-SP was

significantly less than the other two regimens.

There is strong evidence that CQ-SP is inappropriate treatment in PNG children

with uncomplicated malaria. The treatment failure rate for P. falciparum at Day 28,

whether PCR-corrected or uncorrected, was >10%, the level at which the WHO

recommends a change in therapy.[64] Approximately half of the CQ-SP treated

children with P. vivax malaria had redeveloped parasitaemia by Day 28, and the

greater level of Day 42 failures compared with that for emergent P. vivax in

similarly-treated patients with P. falciparum implies significant P. vivax resistance to

this regimen. Parasite clearance time and FCT were the slowest and gametocyte

carriage the highest of the four treatments. The higher blood glucose concentrations

on Day 3 probably reflect persistent insulin resistance associated with acute

infection. [310] Consistent with the association between CQ and pruritus in dark-

skinned people [125], itch was more common in CQ-SP treated children. The

relatively large number of children who were lost to follow-up from this group may,

because of the open-label design, have reflected dissatisfaction that novel

treatment had not been allocated and/or a relatively slow initial symptomatic

response. However, even if all these children had finished the study and were

assessed as ACPR, the overall CQ-SP treatment failure rate would still have been

unacceptable (data not shown).

106

The addition of ARTS to established therapy such as SP can greatly improve cure

rates but there are concerns regarding the use of failing drugs as partners for the

artemisinins. [98] The present response of children with P. falciparum or P. vivax to

ARTS-SP treatment was better than CQ-SP overall but not as good as AL for P.

falciparum or DHA-PQ against P. vivax. Despite being an endorsed candidate ACT

(depending on local parasite resistance) [64], the 28-day PCR-corrected failure rate

for P. falciparum in this study was right on the WHO’s 10% cut-point, indicating that

it is not preferable to either DHA-PQ or AL for PNG children.

The PCR-corrected ACPR for P. falciparum in the AL group was >95% at Days 28

and 42. This is within the range that the WHO recommends for a potential new

antimalarial therapy. [64] However, as in the ARTS-SP group, >50% of AL-treated

patients developed recurrent P. vivax parasitaemia by Day 28 and almost 70% had

become slide positive by Day 42. These rates were significantly less in DHA-PQ-

treated children. Consistent with these findings, the rate of P. vivax emergence after

AL treatment of P. falciparum mono-infections was similar to that in the ARTS-SP

group and double that in DHA-PQ treated children. The relatively short t½e of LUM

compared with that of PQ [138, 146] rather than P. vivax resistance is the most

likely explanation for our findings, especially since there were comparable rates of

emergent or recurrent P. vivax after AL treatment of P. falciparum and P. vivax,

respectively, and an increased likelihood of P. vivax infections in children with low

Day 7 plasma LUM concentrations.

Although it was by far the most efficacious combination against P. vivax, the PCR-

corrected 28-day failure rate of DHA-PQ in P. falciparum cases was barely below

the 10% WHO cut-point [64] and was 12·0% by Day 42. This result is at odds with a

variety of studies in Asia, Africa and South America where efficacies of >95% have

been consistently achieved using follow-up periods of up to 63 days. [311] In vitro

data from P. falciparum isolates from this study sample (Rina Wong, see publication

12) show a strong association between IC50 values for CQ and PQ, suggesting that

cross-resistance of local parasite strains may have contributed to the unexpectedly

high DHA-PQ failure rates. Similar correlations have also been demonstrated in

West Papua. [143] However in P. falciparum isolates from this study sample there

was no such relationship between in vitro parasite sensitivity to CQ and LUM. A

107

similar albeit weaker positive correlation between IC50 values for CQ and PQ has

been found in isolates from Cameroon. [142] In vitro studies also suggest that there

is antagonism between DHA and PQ in CQ-resistant P. falciparum [312, 313] and

this may also have contributed to the lower than expected efficacy of DHA-PQ.

Consistent with the results of studies from West Papua [314] and Myanmar, [134]

we found that low Day 7 plasma PQ concentrations were associated with treatment

failure. Although bioavailability of PQ is improved by co-administration of fat in

healthy volunteers, [137] this is not part of the current manufacturers’ dosing

recommendations In a study in which DHA-PQ was administered with food [141],

38% of children <15 years of age had a plasma PQ <30 �g/L [314] compared with

52% in our younger subjects. Although this might reflect lower bioavailability in our

patients, it could also result from age-specific differences in the PK properties of

PQ. [138] Indeed, Price et al. have called for consideration of increased DHA-PQ

doses in children [314], and have suggested that the poorer efficacy of DHA-PQ

compared with AL in the current study reflected a bias in that only children in the AL

arm were co-administered food. [315] However, drug administration in this study

was performed in accordance with the manufacturer’s instructions for both drugs.

Indeed the non-requirement for co-administering DHA-PQ with food has often been

cited as an operational advantage of DHA-PQ over AL. [208, 316] Notably, until

now, co-administration of food with DHA-PQ has not been routine practice in most

clinical trials, the majority of which have demonstrated excellent clinical efficacy.

[131-134, 138, 140, 208, 316-321] The safety and toxicity implications of higher

peak PQ concentrations likely to result from food co-administration are uncertain.

Additionally, in resource-poor settings, a need to co-administer food would

represent an operational disadvantage due to issues of cost and compliance.

The independent positive association between WAZ and treatment failure in the

present study suggests that nutritional status as well as age might influence PQ

disposition and thus treatment response. The few treatment failures in our AL group

were associated with low plasma LUM concentrations. This relationship has been

described previously in an African study, [322] which also found that unsupervised

AL treatment was associated with lower Day 7 plasma LUM concentrations than

supervised. If AL is to be introduced into PNG, measures to ensure adherence to

108

the 3-day, 6-dose schedule are essential.

We found that baseline P. falciparum parasitaemia was positively and

independently associated with treatment failure in the DHA-PQ group. This has not

been reported in other DHA-PQ studies with high cure rates that included significant

numbers of children [131, 134, 141], perhaps because parasite densities were lower

than in the present study (geometric means <10,000/µL vs approximately

50,000/µL; see Table 2.4). Parasite burden is a well-recognised predictor of

treatment failure across a range of antimalarial treatments. [323] It is not clear why

the regimen in the present study utilizing PQ, the drug with the longest t½e [85],

showed this relationship, as exposure to therapeutic drug concentrations over many

parasite lifecycles is also an important determinant of response. [108] However, PK

data from the PNG children described in section 2.2 suggests that, in comparison

with long half-life antimalarial drugs, PQ has a rapid distribution phase and a

consequently early post-treatment fall in plasma concentrations. This phenomenon,

in concert with relatively high densities of PQ-resistant strains of P. falciparum,

could mean that a proportion of our DHA-PQ treated children had sub-therapeutic

plasma PQ concentrations at an early stage.

The differences between PCR-uncorrected and PCR-corrected ACPR imply that

approximately 10% of patients were re-infected with P. falciparum by Day 28 and

approximately one quarter by Day 42. This confirms the intense transmission in the

two study areas. The emergence of P. vivax parasitaemia after treatment of P.

falciparum mono-infections is well recognised even with ACT [305], but not to the

same extent as in the present study. Over 50% of our children with falciparum

malaria developed P. vivax during the 42-day follow-up period. This could be due to

recrudescence of sub-microscopic P. vivax parasitaemia present at baseline,

activation of liver-stage hypnozoites or a new infection. It was not possible to

assess the relative contributions of these (nor their contribution to treatment failure

in P. vivax monoinfections) in the present study. Nevertheless, emergent vivax

malaria could contribute to complications such as anaemia [14, 15] even if, as in the

present study, it remains mostly asymptomatic.

109

The present study highlights important issues involved in performing efficacy

studies in hyper- or holo-endemic areas with multiple Plasmodium species. As well

as the need for expert microscopy, PCR assessment of treatment failures is

essential. A high P. falciparum re-infection rate could mask the development of

significant parasite resistance unless such assays are available. This would be of

particular importance for treatment of children moving from areas of less intense

transmission and without immunity to local strains of P. falciparum. [60] The large

percentage of children developing vivax malaria during follow-up suggests that this

parasite may contribute substantially to morbidity and perhaps also mortality. [324-

326] It is of relevance that, although Hb concentrations increased after each of the

four present treatment regimens, the Day 42 levels were still less than in a study in

West Papua which also showed a greater improvement with DHA-PQ than AL.

[141] Also, changes in Hb may be underestimated in a study like this, because

children who are withdrawn due to clinical treatment failure (and need for re-

treatment) are missed and because the drop in Hb is likely to occur many days after

the second infection (and therefore be missed due to the limited follow-up duration

in this study).

When considering the poorer than expected efficacy of DHA-PQ for P. falciparum, a

number of potential limitations of the current study merit careful consideration.

Firstly, the statistical power of this study was moderate, with relatively wide

confidence intervals for the absolute estimates of treatment failure (6.4-20% for the

day 42 corrected per-protocol analysis of DHA-PQ efficacy). [315] However, even

with this wide confidence interval, the lower limit remains above that (<5%)

recommended by WHO for adoption of any new therapy.

Secondly, per-protocol analysis may over-estimate the risks of treatment failure

(assuming that those lost to follow-up were actually less likely to have failed

treatment respectively). [315, 327] This issue was addressed by performing an

additional modified intention to treat analysis, using a “best” case scenario (where it

was assumed that all losses to follow-up had not failed treatment). This showed

similar results to the per-protocol analysis (see Table 2.5B), though in both cases

statistical significance was marginal (P=0.027 at day 28 and P=0.67 at day 42 for

difference between DHA-PQ and AL). Due to the relatively high number of losses to

110

follow-up, differences between the treatment arms were no longer statistically

significant when conventional modified intention to treat analysis (which assumes

that all losses to follow up are treatment failures) was performed (see Table 2.6B).

Thirdly, it is possible that differential efficacy of drugs on P. vivax may have biased

the P. falciparum analysis. For example, if P. vivax infections occurring during the

follow-up period act to suppress recrudescent P. falciparum infections, then the

superior efficacy of DHA-PQ against P. vivax would make this treatment group

more susceptible to relapse of P. falciparum. Some have suggested therefore that

follow-up data from patients relapsing with P. vivax should be censored (removed

from the analysis) from the time of the relapse in order to remove this potential

source of bias. [327] However, doing this would have meant that approximately

50% of subjects were censored prior to the study’s 42-day endpoint. Combined with

losses to follow-up, this would result in a dramatically reduced sample size and poor

statistical power.

Fourthly, PCR methods for distinguishing recrudescence from reinfection are likely

to significantly overestimate the risk of treatment failure in high transmission

settings. Where multiple parasite strains exist there is an increased probability that

at least one strain in the pre-treatment and recurrent parasitaemia samples has the

same genotype purely by chance. [328, 329] Although previous studies at PNGIMR

have shown that use of a single MSP2 allele has equal discriminating power to use

of three markers (MSP1, MSP2 and GLURP), it is believed that greater numbers of

alleles reduce misclassification of reinfections as recrudescences. [330]

Nonetheless, others have estimated that use of even up to 6 markers may

overestimate recrudescence by up to 45%. [328]

The ultimate choice of an antimalarial regimen in a country such as PNG will

depend on factors additional to efficacy and safety. Cost, availability, adherence to

unsupervised treatment and appropriate education of health workers involved in

diagnosis and management are crucial to successful deployment and effective use

of antimalarial drugs. For these reasons, ideally policy decisions will ultimately be

based, not only on the results of this study, but also on larger, Phase IV studies,

that are designed to evaluate effectiveness (rather than efficacy) outside the setting

111

of closely regulated supervised treatment and powered to detect less common

adverse effects. Policy decisions must also consider health-economic factors. A

health-economic analysis, utilizing data from this study is planned (Wendy Davis,

UWA). However, the present study demonstrates that local data can provide

important insights, especially in the case of DHA-PQ which had the lowest efficacy

of any large-scale trial published to date. The present in vitro data facilitate an

understanding of this unexpected result and, as implied by WHO guidelines [64],

underscore the need for regular monitoring of parasite sensitivity and in vivo

response. Such data can identify failing conventional therapy at an early stage and

inform a pro-active choice of replacement regimens.

112

113

2.4. Summary of major findings of this chapter

2.4.1. Pharmacokinetics of piperaquine and chloroquine

• Although PQ has a long t½e, its much more rapid and extensive distribution

phase than CQ causes concentrations of drug to fall precipitously within 4

days of treatment completion, possibly limiting its post-treatment

suppressive properties.

• Unlike PQ, CQ’s post-treatment suppressive properties may be augmented

by those of its active metabolite, DECQ which is present at levels in a similar

range to those of the parent drug in the weeks following treatment.

Combined molar concentrations of CQ and its active metabolite were many

times higher than those of PQ from day 7 onwards.

2.4.2. Methodological issues in assessing treatment efficacy in a poly-species high transmission environment

• Clinical trial 2 emphasized the unique challenges of efficacy evaluations

where intense transmission of multiple Plasmodium species occurs. In

addition to genuine treatment failures leading to recrudescence,

approximately one quarter of all patients became re-infected with P.

falciparum and more than half developed or re-developed P. vivax

parasitaemia within the 42-days. This demonstrated the importance of PCR

assessment of treatment failures in this setting.

2.4.3. Efficacy of existing standard treatments for malaria in Papua New Guinea

• Efficacy of existing first and second-line treatments (CQ-SP and ARTS-SP,

respectively) for uncomplicated P. falciparum in Clinical trial 2 was

inadequate (corrected day 28 treatment failure rates of �10%) and

significantly worse than that of the two ACT regimens evaluated, thereby

confirming existing suspicions of worsening P. falciparum resistance to CQ

and SP in PNG.

114

• High level P. vivax resistance to CQ now appears to exist in PNG.

Approximately half of the CQ-SP treated children with P. vivax malaria in

Clinical trial 2 re-developed P. vivax parasitaemia by Day 28. Chloroquine-

based therapy for P. vivax in PNG can no longer be considered effective,

consistent with high levels of in vivo CQ resistance reported from

neighbouring West Papua. [281, 282]

• The efficacy of ARTS-SP was superior to that of CQ-SP for both P.

falciparum and P. vivax, but not as good as AL for P. falciparum or DHA-PQ

against P. vivax suggesting that its efficacy is compromised by pre-existing

resistance to its SDOX and PYR components.

2.4.4. Relative efficacy of two ACTs, dihydroartemisinin-piperaquine and artemether-lumefantrine, in Papua New Guinea

• Whilst AL was highly efficacious against P. falciparum (based on PCR-

corrected ACPR of >95%) in PNG, its efficacy in Clinical trial 2 was

compromised by high rates of re-emergent P. vivax (>50% at day 28 and

almost 70% at day 42). Conversely, whilst DHA-PQ was the most

efficacious regimen against P. vivax, the treatment failure rate for P.

falciparum was unacceptably high.

• The few treatment failures in the AL group of Clinical trial 2 were associated

with low plasma LUM concentrations, as has been previously described.

[322]

2.4.5. Poorer than expected efficacy of dihydroartemisinin-piperaquine against P. falciparum in Papua New Guinea

• The PCR-corrected treatment failure rate of DHA-PQ for P. falciparum in

Clinical trial 2 was barely below the 10% WHO cut-point [64] at Day 28 and

was 12·0% by Day 42. These rates were higher than those seen in any

other large-scale trial published to date and were, therefore, altogether

unexpected.

• Possible reasons for DHA-PQ’s unexpectedly poor efficacy include PK

factors, cross-resistance of local parasite strains with CQ and very high

115

baseline parasitaemias in this population. Treatment failure in Clinical trial 2

was associated with baseline P. falciparum parasite density. The ability to

cure high parasitaemia infections might be further compromised by PQ’s

rapid, early post-treatment fall in plasma concentrations (shown in Clinical

trial 1). An association between low day 7 plasma PQ concentrations and

treatment failure demonstrated in Clinical trial 2 and by others [314] further

support the importance of drug concentrations early in the post-treatment

phase.

116

117

3. MANAGEMENT OF SEVERE

MALARIA IN CHILDREN

118

119

3.1. Overview of Clinical trials in this chapter This chapter includes four Clinical trials. These aimed to address the following

research questions, which relate specifically to severe paediatric malaria in PNG

and the potential of a new treatment modality (rectally administered artesunate

suppositories) to be used in this setting:

1. How safe, tolerable and efficacious are artesunate (ARTS) suppositories in

Melanesian children with malaria (including P. falciparum and non P. falciparum

infections, uncomplicated and severe malaria and in children of all ages)?

2. What is the PK disposition of rectally administered ARTS? In particular, how

rapidly does absorption occur and how consistent or variable is this? Is PK

disposition or efficacy influenced by erythrocyte polymorphisms prevalent in

Melanesian populations?

3. What is the clinical spectrum of severe malaria in children in coastal PNG and

how might this be modified by the prevalence of putative genetic adaptations to

malaria? In particular what is the role of the GPC∆ex3 mutation which has yet to

be studied in severe malaria?

4. Are ARTS suppositories at least as effective as current standard hospital-based

treatments for severe malaria?

5. Are ARTS suppositories a feasible health intervention in PNG? In particular is

rectal administration socially and culturally acceptable in this population and will

mothers be amenable to self-administering suppositories to their children? What

pre-existing perceptions and attitudes regarding rectal administration need to be

considered for health education and deployment strategies?

All four clinical trials were undertaken at sites on the North Coast of PNG and are

summarised below:

120

1. Clinical trial 3 was designed to investigate the safety, efficacy and PK of

ARTS and DHA in Melanesian children who i) had uncomplicated falciparum

or vivax malaria, ii) were from a high �-thalassaemia prevalence area

(research questions 1 and 2 above).

2. Clinical trial 4a) was designed to define the clinical spectrum of severe

paediatric malaria in Madang and to examine the influence of common

erythrocyte polymorphisms (including the GPC∆ex3 mutation: research

question 3, above).

3. Clinical trial 4b) aimed to determine the therapeutic equivalence of ARTS

suppositories to a standard hospital-based injectable artemisinin treatment

for severe malaria (i.m. ARM which is standard treatment for severe malaria

in PNG: research question 4, above).

4. Clinical trial 5 aimed to address the issue of socio-cultural acceptability by

examining the pre-existing attitudes of Papua New Guinean caregivers

towards rectal drug administration to children (research question 5, above).

121

3.2. Clinical trial 3: Safety, efficacy and pharmacokinetics of artesunate suppositories

3.2.1. SUMMARY

Aims: To assess the safety, efficacy and pharmacokinetic (PK) disposition of rectal

artesunate (ARTS) in Melanesian children with uncomplicated malaria and to

examine the influence of �-thalassaemia genotype on PK disposition and efficacy.

Methods: Forty-seven children aged 5 to 10 years with uncomplicated malaria (42

P. falciparum and 5 P. vivax) received rectal ARTS (Rectocaps) approximately 13

mg/kg given in two doses 12 h apart. Following an intensive sampling protocol,

ARTS and its primary active metabolite, dihydroartemisinin (DHA) were assayed

using liquid chromatography-mass spectrometry and a population PK model

developed to describe the data.

Results: Artesunate suppositories were well tolerated. After 24 h, only one child

(with P. falciparum) had persistent parasitaemia and only one (with P. vivax) had

not defervesced. Three children redeveloped fever and tachycardia at 24 h but each

responded to simple supportive measures and remained aparasitemic. Following

the first dose, the mean Cmax of ARTS and DHA were 1085 nmol/L at 0.9 h and

2525 nmol/L at 2.3 h, respectively. Absorption half-life for ARTS was 2.3 h and

conversion half-life (ARTS to DHA) 0.27 h, while the elimination half-life of DHA was

0.71 h. The mean common V/F for ARTS and DHA was 42.9 L (2.15 L/kg), and

mean CL/F values were 121.2 L/h (6.07 L/h/kg) for ARTS and 44.9 L/h (2.25 L/h/kg)

for DHA, respectively. Substantial inter-patient variability was observed and the

bioavailability of the second dose relative to the first was estimated at 0.72. The

covariates age, sex and �-thalassaemia genotype were not influential in the PK

model development, but the inclusion of weight as a covariate significantly improved

model performance. Seven children had undetectable levels of ARTS and DHA

after one of the two doses suggesting either unrecognized passage of the

suppository soon after administration or negligible absorption.

Conclusions: Although absorption is subject to substantial inter-individual

variability, overall this was rapid and sufficient to effect a rapid parasitological and

clinical response in all patients. Although a transient fever spike sometimes

occurred after parasite clearance, drug concentrations achieved in this study were

122

not associated with any adverse events. Intrarectal ARTS is safe, effective initial

treatment for uncomplicated malaria in children. A treatment regimen of two doses

of ≥10 mg/kg rectal ARTS within the first 24 h is appropriate for children with

uncomplicated malaria.

3.2.2. AIMS

3.2.2.1. Primary

To generate detailed concentration-time data for ARTS and its primary metabolite

DHA in order to perform compartmental modeling and a robust derivation of

standard PK parameters (V/F, CL/F, t½e, Cmax and t½abs)

3.2.2.2. Secondary

i. To define safety and tolerability of ARTS suppositories with respect to reported

adverse events and findings on clinical examination

ii. To evaluate efficacy in children with malaria with respect to standard PD

measures of parasitological (parasite clearance times) and clinical (fever

clearance time) response.

iii. To evaluate �-thalassaemia genotype as a covariate with respect to PK

disposition and clinical response.

3.2.3. METHODS



Tim Davis (School of Medicine and Pharmacology, UWA), Inoni Betuela (PNG IMR,

Goroka, PNG) and Michael Alpers (PNG IMR, Goroka, PNG) were responsible for

the initial study concept, design and ethical approval. Dr Luke Antony (MGH,

Madang, PNG) performed some of the clinical procedures. Assistance was also

provided by other nursing and medical staff at MGH. Dr Adedayo Kemiki (MGH,

Madang, PNG) also assisted in the clinical supervision of the study. Kerry Lorry

(PNG IMR, Madang, PNG) performed all the microscopy. The LC-MS-MS assays

for ARTS and DHA were developed and validated by Kittya Duffal (School of

Medicine and Pharmacology, UWA) and Ken Ilett (School of Medicine and

Pharmacology, UWA) with input and advice from Kevin Croft (School of Medicine

123

and Pharmacology, UWA) and Lincoln Morton (School of Medicine and

Pharmacology, UWA). Kittya Duffal performed most of the drug assays. Ken Ilett,

Hugh Barrett (School of Medicine and Pharmacology, UWA) and Paolo Vinci

(Resource Facility for Population Kinetics, Department of Bioengineering, University

of Washington, Seattle, USA) devised the population PK model. Dr John Beilby

(Path West Laboratory Medicine, Nedlands, WA) performed the �-thalassaemia

genotyping assays.

I supervised the clinical aspects of the study, performed most of the data collection,

clinical evaluations and procedures, and took primary responsibility for analysis and

interpretation of the clinical and PK data and the preparation of manuscripts for

publication (publications 1 and 2).


The study design chosen was of a single arm clinical intervention in Melanesian

children with uncomplicated malaria. Because PK disposition, treatment efficacy

and drug safety and tolerability may vary widely according to age and ethnicity

[103], a study population of Melanesian children was considered most appropriate,

in order that the study results could be generalized to the setting in which the drug

is likely to be most used in future. For practical and ethical reasons, this was limited

to older children (in whom more frequent blood sampling was feasible). Because PK

disposition of antimalarial drugs has also been found to vary according to the

presence and severity of acute malaria infection, [101, 331] only subjects with

documented malaria infection (including P. vivax and P. falciparum infection with a

range of disease severity) were eligible. This also enabled assessment of

conventional PD markers of response (fever and parasite clearance). Given the

lack of existing data relating to this preparation at the time of this study, it was not

considered ethical to include children with severe malaria. Therefore, subject

selection was limited to children with uncomplicated malaria. The study design

chosen thus met definitions of a Phase 2a clinical trial. [332]

3.2.3.3. Study site and patients

The study was conducted in Madang during a 3-month period between March and

May 2001. Subjects were recruited by screening children presenting to the

124

ambulatory children’s outpatient clinic of MGH (passive case detection) or by

visiting nearby villages and taking a blood film from children with fever (active case

detection). Children presenting to the MGH clinic have either been referred from

surrounding rural health centres or come directly from peri-urban settlements and

are likely to have been exposed to a hyper-endemic level of P. falciparum

transmission. [31] Children aged 5-10 years with a blood film positive for either P.

falciparum or P. vivax were eligible provided that i) there were no features of severe

malaria [24] ii) there was no history of antimalarial drug use in the previous 28 days,

iii) there was no clinical evidence of another infection as a possible cause of fever,

iv) there was no history of diarrhoea, and v) the child’s parents gave informed

consent.


All recruited children were admitted to MGH for at least 24 h. An initial clinical

assessment was performed, and blood was taken for baseline tests including full

blood count and plasma glucose. The diagnosis was confirmed by microscopy and

the parasite density quantified.

Artesunate was then given per rectum in a dose of 10 to 15 mg/kg body weight to

the closest whole suppository. A combination of between one and four 50 mg and

200 mg thermostable suppositories (Plasmotrim Rectocaps�, Mepha Ltd, Aesch-

Basel, Switzerland) were used to make up the total dose for each child.

Suppositories were kept at room temperature, removed from their airtight foil

packaging immediately before use and administered base first without lubricant. A

second dose of 10–15 mg/kg was administered per rectum 12 h after the initial

dose. If the child passed a suppository within 1 hour of administration, a repeat

dose was administered. If the suppository was passed after 1 hour, the child was

withdrawn from the study and a single dose of SP (25 mg/kg SDOX / 1.25 mg/kg

PYR) given followed by CQ 15 mg base/kg for 3 days in accordance with standard

treatment recommendations in PNG. [204]

Each child had a heparinised i.v. cannula inserted at the time of admission. A

baseline sample of 3 mL of venous blood was taken and further samples of 3 mL

drawn through the cannula at 1, 2, 3, 4, 6, 8, 12 h after administration of the first

dose and 2, 4, 8 and 12 h after the second. Blood was collected into

125

fluoride/oxalate blood tubes kept on ice for < 30 min prior to centrifugation, and the

separated plasma stored and transported at <–20oC until analysed.

Patients were assessed 4-hourly with measurement of oral temperature, pulse rate,

blood pressure, respiratory rate and plasma glucose. Urine output and conscious

state were also assessed, and a thick blood film was also taken at these times.

Children with an oral temperature >38.0 °C were administered paracetamol and

tepid sponging and fanning were commenced. Children who developed signs of

complicated malaria at any time were to be withdrawn from the study and given i.m.

quinine, as recommended for complicated malaria in PNG. [204] Subjects were kept

in hospital beyond 24 h if they remained febrile or the treating physician considered

that their clinical condition warranted continued inpatient care. Each child received a

standard course of SP and CQ [204] at 24 h after administration of the first dose of

ARTS and was subsequently discharged when aparasitaemic and afebrile.


3.2.3.5.1. Microscopy, haematocrit and glucose

Giemsa-stained thick blood smears were examined by a skilled microscopist (Kerry

Lorry, PNG IMR). At least 100 fields were examined at 1,000x magnification before

a slide was considered negative. For positive slides, parasite density was calculated

from the number of asexual forms/1,000 leucocytes and the whole blood leucocyte

count. Venous haematocrit was measured using a microcentrifuge and plasma

glucose was measured at the bedside (Exactech, Medisense, Abingdon, UK).


Methodology for assays of ARTS and DHA is described in Appendix 1, section 7.13.

3.2.3.5.3. Alpha-thalassaemia genotyping

Whole blood was collected in EDTA tubes and assayed for the 6 common deletional

variants of the two α-thalassaemia genes using a single tube multiplex PCR using

the method described by Chong et al. [271] On the basis of the assay, each patient

was classified as being either normal (homozygous wild type), heterozygous (one

abnormal gene), homozygous (two copies of a single gene deletion type) or dual

heterozygous (one copy each of two different gene deletion types).

126

3.2.3.6. Data analysis

Data were recorded in a Microsoft Excel spreadsheet. Statistical analysis was by

Epi Info 2000. [333] Data are presented as mean ± SD or, for non-normally

distributed variables, median and (range).

3.2.3.7. Pharmacodynamic outcomes

Pharmacodynamic outcomes measured included fever clearance time (from

admission to the first oral temperature <37.5 °C; FCT), parasite clearance time

(from admission to the first of two negative thick blood films; PCT) and time taken

for parasite density to fall by more than 50% and 90% of the baseline value (PCT50

and PCT90).


Plasma ARTS and DHA concentrations were analysed using population PK

analysis using NONMEM Version V. [334] Four ARTS data points between the

limits of detection and quantitation were allowed to remain in the analysis to assist

with model stability. For covariate inclusion in the final population PK model, an

increase of at least 6.63 points (P<0.01) in the value of the approximate first order

objective function, as reported by NONMEM, was required.

3.2.4. RESULTS


A total of 74 children were screened of whom 26 (35%) were excluded because of

recent antimalarial treatment (16), parental refusal of consent (7) or disappearance

before recruitment (3). One further patient was excluded during the trial because of

passage of a suppository more than 1 hour after the dose.

The remaining 47 enrolled patients were divided into three groups. Group 1

comprised 30 children with i) an axillary temperature >37.5°C or parental

confirmation of fever during the previous 24 h, and ii) P. falciparum parasitaemia

>2000/µL. These criteria were chosen so that Group 1 patients were similar to

those included in other pediatric studies of ARTS suppositories. [232, 238, 242]

Group 2 consisted of 12 patients with P. falciparum but either i) no fever at

presentation or within the previous 24 h, or ii) parasite density <2000/µL. Five

127

children had P. vivax infections (Group 3). The percentages of patients recruited

from local villages compared to the clinic were 33% of Group 1, 100% of Group 2

and 40% of Group 3. The characteristics of patients in the three groups including �-

thalassaemia genotype, are summarised in Table 3.1, below.

128

Table 3.1. Details of 47 Papua New Guinean children with uncomplicated malaria

at the time of admission to Clinical trial 3. Data are mean ± SD or median (range).

A. Demographic and clinical features

Group 1 Group 2 Group 3 TOTAL

Number of patients 30 12 5 47

Age (years) 7.1 ± 1.5 7.9 ± 1.7 6 (5-6) 7.2 ± 1.6

(range: 5-10)

Sex (%males) 60 58 60 60

Body weight (kg) 20.4 ± 4.6 21.5 ± 3.4 17.5 (15 – 23)

20.5 ± 4.2

(range: 12.5-31.0)

Pulse rate (/min) 107 ± 151 90 ± 15 108 ± 13 102 ± 16

Respiratory rate (/min)

25 ± 10 22.2 ± 4.0 25 ± 8.5 24.3 ± 8.6

Systolic blood pressure (mm Hg)

90 ± 12 94 ± 10 96 ± 11 92 ± 11

Diastolic blood pressure (mm Hg)

56 ± 12 55 ± 5 58 ± 8 56 ± 10

Oral temperature (°°°°C)

38.3 ± 1.0 37.6 ± 0.3 38.2 (37.6-40.8)

38.1 ± 0.9

Splenomegaly (%) 70 83 40 70

Venous haematocrit (%)

33.7 ± 5.5 35.0 ± 4.0 33.3 ± 5.9 34.0 ± 5.0

Parasite density (/µµµµL)

17,800 (2,070–

326,000)1,2

536 (119–2,830)2

7,480 (630–13,300)

11, 600 (119-326,000)

Plasma glucose (mmol/L)

6.5 ± 1.4 5.9 ± 0.7 5.2 ± 1.2 6.2 ± 1.3

Artesunate dose (mg/kg)

12.9 ± 0.9 12.5 ± 0.7 12.8 ± 0.9 12.9 ± 0.8

(range: 11.3-16.0)

1 P<0.05 vs Group 2, 2 P<0.05 vs Group 3 by Kruskal-Wallis test

129

B. Alpha-thalassaemia genotype. Data are number (%)

�-thalassaemia genotype

Group 1 Group 2 Group 3 TOTAL

Wild type (2 normal genes)

7 (23) 2 (17) 1 (20) 10 (21)

Heterozygous (1 gene deletion)

11 (37) 3 (25) 1 (20) 15 (32)

Homozygous (2 gene deletions)

11 (37) 6 (50) 3 (60) 20 (43)

Dual heterozygous (2 different gene deletions)

1 (3) 1(8) 0 2 (4)

3.2.4.2. Clinical course

The PD outcome variables are summarised in Table 3.2, below.

Table 3.2. Pharmacodynamic outcomes in 47 Papua New Guinean children with

uncomplicated malaria administered approximately 13mg rectal artesunate at 0 and

12h.

Group 1 Group 2 Group 3

FCT (h) 8 (4-20) - 4 (4-36)

PCT50 (h) 8 (4-20) 4 (2-8) 4 (4-16)

PCT90 (h) 12 (4-24) 6 (2-8) 6 (4-16)

PCT (h) 20 (4->24)* 14 (2-24) 12 (6-20)

* One patient still had parasitaemia (480/µL) after 24 h

3.2.4.3. Side-effects and complications

Four children, three with P. falciparum and one with P. vivax, vomited during the 24-

hour observation period. Two of these children, one with P. falciparum and the other

with P. vivax malaria, were still vomiting, febrile and/or unwell at 24 h despite each

130

becoming blood-slide negative within 20 h. Both were treated with i.m. quinine at 24

h and made an uneventful recovery with fever and vomiting settling over a

subsequent day of inpatient observation.

One child developed hyperpyrexia (oral temperature 43.0°C) 8 h after the first

suppository. The high fever resolved quickly after administration of paracetamol and

tepid sponging. A further three patients, all from Group 2, developed high

temperature, tachycardia and vomiting 24 h after having received their first dose of

ARTS. The characteristics of the episodes in these 3 children are summarised in

Table 3.3, below. None had a history of significant fever or chills in the days before

administration of ARTS. Repeat thick and thin blood films at the time of fever were

all negative. All remained normotensive and none had rash or wheeze to suggest

an allergic reaction. Each of these children was observed for a further 24 h in

hospital with 4-hourly temperature and thick and thin blood films. No further fever or

relapse of peripheral blood parasitaemia was observed and all three children made

an uneventful recovery.

No other adverse events or complications were observed during the study. In

particular, hypoglycemia was not observed and no alterations in conscious state or

neurological sequelae were seen.

131

Table 3.3. Details of the three Group 2 patients at the time of the late fever spike.

Patient number

Baseline oral

temp (°°°° C)

Baseline parasite density

(/µµµµL)

PCT (h) Peak oral

temp (°°°° C)

Time of peak

temp (h)

Heart rate

(/min)

24 37.7 440 6 39.0 24h 100

36 37.7 220 4 39.3 24h 132

37 37.7 120 2 40.6 28h 148

3.2.4.4. Pharmacokinetic analysis

In spite of the high sensitivity, particularly of the DHA assay, seven patients (3 from

group 1, 3 from group 2 and 1 from group 3) had undetectable levels of ARTS and

DHA following one of their doses (4 after the first dose, and 3 after the second). All

showed good absorption of the other dose. In these 7 patients, PCT90 was 16.3 vs

9.5 h in the remaining patients (P=0.008) suggesting that little or no active drug had

been absorbed in these 7 instances. Given the difficulties in closely monitoring

patients, it was suspected that these children extruded the suppository after

administration without the knowledge of the investigators. Pharmacokinetic data

following these 7 doses have been excluded from the analysis. In 27 datasets,

ARTS concentrations were not available due to difficulties with the LC-MS-MS

instrument setup and sensitivity, and the fact that there was insufficient sample to

repeat these analyses. Overall, 43 first and 44 second DHA datasets, and 19 first

and 19 second ARTS datasets were available for PK analysis.

Selection of an initial structural model (Figure 3.1, below) was guided by previous

studies of the metabolic disposition [335, 336] and PK [180, 337, 338] of ARTS and

DHA in adults.

In keeping with knowledge of the metabolism of ARTS and DHA, conversion of

ARTS to DHA (rate constant = k23) was assumed quantitative [336] and elimination

of DHA was assumed to occur from compartment 3 (rate constant = k30). [180]

132

Since each patient received two equal doses of ARTS 12 h apart, a “relative

bioavailability” term was incorporated into the model which was set at 100% for the

first administration and estimated as a separate parameter for the second dose

(FRAC). Dose was entered in the model in nmol/patient. In addition, since the

dose was administered rectally, model-derived estimates of both CL and Vd were

estimated relative to bioavailability (F). There were 78 data points available for

ARTS and 291 for DHA. These included ten DHA and four ARTS data points

between the limits of detection and quantitation.

The initial values for the fit were obtained by performing a naïve pooled estimate of

all data against the model using SAAM II. (SAAM II Program and User Guide,

SAAM Institute Inc, University of Washington, Seattle, WA, USA, Version 1.2,

2000), with average figures for the dose and the time of the second dose. The

steps in model development are summarised in Table 3.4.

Figure 3.1. Structural model for disposition of artesunate and dihydroartemisinin

following rectal artesunate administration. Compartment 1 = rectum with application

of dose, compartment 2 = artesunate plasma concentration-time dataset with

volume V1/F, compartment 3 = dihydroartemisinin plasma concentration-time

dataset with volume V2/F, k12 = absorption rate constant for artesunate, k23 = rate

constant for metabolic conversion of artesunate to dihydroartemisinin, k30 =

elimination rate constant for dihydroartemisinin. Bioavailability (F) of the first dose

was set at 100% and estimated as a variable for the second dose (FRAC, relative to

F for the first dose).

1 2 3

k12 k23

k30

133

Table 3.4. Summary of the model development process

Model Comment Objective function

Base 1; k12, k23, k30, V1/F, V2/F, FRAC

V1/F, V2/F and FRAC were fixed effects in the population.

5792

Base 2; k12, k23, k30, V1/F, V2/F, FRAC

V1/F and V2/F were different but had random effects in the population. FRAC is a fixed effect.

5790.6

Base 3; k12, k23, k30, V1/F =V2/F, FRAC

FRAC is a fixed effect. V1/F set to be same asV2/F. Proportional measurement error and constant coefficient of variation for between-subject variability.

5789.6

Final; k12, k23, k30, V1/F =V2/F, FRAC

As for base 3, but including weight as a covariate.

5721.2

Base model 1 consistently underestimated the predicted drug concentrations and,

despite alterations such as the inclusion of a lag time, failed to adequately represent

the data on standard criteria. Base model 2 marginally improved the overall

objective function but there was still a problem with model identifiability, particularly

at higher concentrations, and estimates of V2/F and k30 were poorly defined.

Because of the latter, V1/F and V2/F were set as equal in model 3, and this resulted

in improved model identifiability, more robust parameter estimates and a marginal

improvement in the objective function.

In the final model age, sex, �-thalassaemia genotype and weight were investigated

as covariates. Only weight was influential, and its inclusion in the volume term (as

V/F * weight/median weight) resulted in a substantial improvement in objective

function and overall model performance. The mean parameter estimates for the

final model, and the individual and residual variability factors for the model are

summarised in Table 3.5.

134

Table 3.5. Parameters and performance descriptors for the final pharmacokinetic

model.

Theta (parameters)

Final estimate

% RSE4 Lower 95% CI

Upper 95% CI

CV (%)1

1 (k12) 0.253 18.1 0.163 0.343 -

2 (k23) 2.8 22.8 1.55 4.05 -

3 (k30) 1.03 19.4 0.638 1.42 -

4 (V/F) 41.8 20.4 25.2 58.4 -

5 (FRAC) 0.72 9.68 0.583 0.857 -

2Omega (CV2) - Between subject variability 3

k12 0.412 31.3 0.158 0.665 64.2

k23 0.303 126.0 -0.448 1.05 5 55.0

k30 0.162 37.3 0.0434 0.281 40.2

FRAC 0.179 61.2 -0.0338 0.374 5 41.2

2Sigma (CV2) – Random unexplained variability

ARTS 3.02 53.3 -0.136 6.18 5 174

DHA 0.625 15.7 0.433 0.817 79.1

1 CV = coefficient of variation = SD *100/mean 2 Omega and Sigma are maximum likelihood estimates of variances of the between

subject, and random unexplained variability, respectively 3 Note that V/F has no random effect since its variation is only modeled by the

covariate weight 4 %RSE is percent relative standard error (100 * SE/estimate) 5 95% confidence interval includes zero

Plots of model predicted and observed artesunate and dihydroartemisinin

concentration-time data are shown in Figures 3.2 and 3.3, respectively, while Figure

3.4 shows a plot of the weighted residuals for the model by subject.

135

Figure 3.2. Plot of population model prediction (PRED) and measurements (DV)

for artesunate against time (semi log scale) in 47 Papua New Guinean children

treated with approximately 13mg/kg rectal artesunate at 0 and 12h. The solid line

shows the plasma artesunate concentration simulated using the mean parameters

for the model, a 12.75 mg/kg dose at zero time and 72% of that dose at 12 h.

Time (h)

0 4 8 12 16 20 24

Art

esun

ate

(nm

ol/l)

10

100

1000

10000

136

Figure 3.3. Plot of population model prediction (PRED) and measurements (DV) for

dihydroartemisinin against time (semi log scale) in 47 Papua New Guinean children

treated with approximately 13mg/kg rectal artesunate at 0 and 12h. . The solid line

shows the plasma dihydroartemisinin concentration simulated using the mean

parameters for the model, a 12.75 mg/kg dose at zero time and 72% of that dose at

12 h.

Time (h)

0 4 8 12 16 20 24

Dih

ydro

arte

mis

inin

(nm

ol/l)

10

100

1000

10000

137

Subject number0 10 20 30 40 50

Wei

ghte

d re

sidu

al

-8

-6

-4

-2

0

2

4

6

8

Figure 3.4. Plot of weighted residuals for artesunate () and dihydroartemisinin (∆)

versus subject number (linear scale).

Table 3.6 summarizes the individual PK parameters derived from post-hoc

Bayesian predictions and/or calculated from these using standard PK equations.

[290]

Mean Cmax and tmax values for ARTS and DHA were simulated using the kinetic data

from Table 3.5. For ARTS a Cmax of 1085 nmol/L occurred at 0.9h and for DHA Cmax

was 2525 nmol/Lat 2.3 h. Four patients with P. vivax were initially analysed

separately but no statistically significant differences with patients with P. falciparum

were seen in any derived PK parameters and they have been included together in

the pooled analysis.

3.2.4.5. Alpha-thalassaemia status

Alpha-thalassaemia genotype was determined in 47 patients and was wild type (two

normal genes) in 22%, heterozygous for one abnormal gene in 33% and either

homozygous or dual heterozygous in 46% (see table 3.1 B). Comparison of

138

pharmacodynamic outcomes (FCT, PCT, PCT50 and PCT90) for each of these three

groups by ANOVA revealed no significant between group variability.

Table 3.6. Pharmacokinetic parameters for the model

Parameter (mean ±±±± SD)

ARTS

F (dose 2) 0.72 ± 0.21

V1/F (L) 42.9 ± 8.1

(2.15 L/kg)

k12 (/h) 0.27 ± 0.18

t½ absorption (h) 2.3 ± 1.0

k23 (/h) 2.81 ± 0.65

t½ conversion (h) 0.27 ± 0.13

CL/F conversion (L/h) 121.2 ± 35.4

(6.07 L/h/kg)

DHA

V2/F (L) 42.9 ± 8.1

(2.15 L/kg)

k30 (/h) 1.06 ± 0.28

t½ elimination (h) 0.71 ± 0.22

CL/F elimination (L/h) 44.9 ± 13.0

(2.25 L/h/kg)

139

3.2.5. CONCLUSIONS

3.2.5.1. Efficacy

This is the first study to have evaluated the efficacy of an artemisinin derivative in

Melanesian children. Although a comparator group was not used, parasitological

measures of outcome were impressive. All but one of 47 patients cleared their

parasitaemia within 24 h and most had achieved >90% reductions in parasite

density within the first 12 h. These results are consistent with those of studies from

South-East Asia and Africa that demonstrate more rapid parasite clearance with

artemisinin derivatives than other antimalarial drugs. [184, 185] Parasite clearance

reflects factors other than the type of treatment, including pre-existing immunity, the

initial parasitaemia and infection severity. [339] Because of this, comparisons

between treatment responses in different populations are difficult. Nevertheless, in

children with uncomplicated malaria treated with conventional drugs such as CQ

and quinine, the PCT is generally >48 h. [184, 340]

The three other published studies of ARTS suppositories in children with

uncomplicated falciparum malaria have used different dosage regimens. In 26 Thai

children who received Plasmotrim Rectocaps 15 mg/kg daily for three days [242],

PCT and FCT (means 42 and 43 h, respectively) were longer than in patients in this

study. Given that the initial parasitaemias were similar to those in Group 1 patients

from the present study, this difference might reflect the lower background immunity

in the Thai children. In a second study [238], 12 Gabonese children received

Plasmotrim Rectocaps of an average dosage of only 1.8 mg/kg at 0 and 4 h. Eight

of 11 evaluable patients (73%) were still blood slide-positive after 24 h. Because the

median initial parasitaemia (21,450/µL) and likely level of malarial immunity in these

children were similar to those in Group 1 patients in the present study, the relatively

low dose may have been the major reason for their slower parasite clearance.

Alternatively, parasite strains in PNG are inherently more sensitive to the

artemisinin derivatives than those in Africa. The third study by Gomez et al.,

published after the current study was performed, compared 3 different regimens

(3.9-13.8mg/kg administered 1-3 times per day) in a total of 150 children in

Ecuador. [239] Mean PCTs of 8.3-9.2 h in each of the three treatment arms were

even shorter than those demonstrated in the current study, possibly reflecting either

140

the co-administration of another antimalarial (MQ) in 2/3 of patients or the much

lower baseline parasitaemia (mean parasite densities were 2838-4408/µl) than

those of our Group 1 children.

Two other published studies have now evaluated parasite clearance in children with

moderately severe [117, 232] and one in severe malaria. [341] Although one of

these, a study by Krishna et al [232], showed mean PCT values (20-24 h) similar to

those in the present study, the others [117, 341] showed higher mean/median PCTs

(24-48 h) possibly reflecting the greater disease severity and higher baseline

parasitaemias than our PNG children. However, subsequently reported median

values for PCT50 (8-15 h) and PCT90 (16-18h) [247] from the same patients studied

by Krishna et al., Barnes et al. [117, 232] and a further 44 children from Thailand

(all of whom were treated with a dose of 10mg/kg) all showed consistency both

between study sites and when compared with the present study. [247]

Some studies of rectal ARTS in adults with severe malaria have employed an

intensive initial regimen, administering doses every 4 h. [240, 241, 342, 343]

However, such schedules may be too complex for use in remote rural settings. The

regimen used in the present study replicates the situation in which a sick child is

given a single rectal dose at a primary health care facility before a long journey to a

centre where more intensive therapy can be administered. The critical determinant

of such an approach is the extent to which parasite burden is reduced in the first 12

h. A first dose of 10-15 mg/kg proved effective in this study. A more comprehensive

review of efficacy with different dosing regimens and other rectally administered

artemisinin derivatives is described in the Appendix.

The high rate of recrudescence after short-course mono-therapy with an artemisinin

derivative can be reduced but not eliminated by longer duration therapy [183, 344],

and compliance becomes more difficult to maintain. The combination of an

artemisinin derivative with a partner drug such as SP or MQ can circumvent these

problems, albeit adding to the complexity and perhaps cost of treatment. This study

was not designed to assess definitive cure as an endpoint (this would require

follow-up over at least a 28-day period). Children in this study received standard

oral therapy with CQ-SP at 24 h. At the time of this study, recent data from PNG

141

had demonstrated acceptable cure-rates with CQ-SP. However, more recent data

(see 2.3) suggests that even in combination with ARTS, the CQ-SP combination is

unlikely to provide adequate definitive treatment in PNG.

In contrast to other previous studies of ARTS suppositories in children [232, 238,

242], patients with uncomplicated malaria were recruited who had no symptoms,

low parasitaemias or concomitant vivax malaria. This allowed an assessment of

subjects with a range of clinical and parasitological features representative of those

in the field. Each group responded rapidly to rectal ARTS in a regimen that could be

easily used in remote areas. Nevertheless, larger and longer-term studies are

needed to confirm the acceptability, safety and efficacy of ARTS suppositories as

part of combination therapy in both uncomplicated and severe malaria in children

from PNG and other tropical countries.


Artesunate suppositories were well tolerated and no serious side-effects were

observed in this study. The FCT was short in each of the three patient groups, but

the observation of a late temperature spike in three Group 2 children with mild

falciparum malaria was unexpected. All had cleared their parasites rapidly, and had

a negative thick blood film at the time of the fever and subsequently. Because of

this, the late fever seems unlikely to have been a product of active parasite

replication or of the parasiticidal effect of ARTS. Its timing raises the possibility of an

idiosyncratic drug or metabolite effect. Although there is little published evidence of

this phenomenon, drug fever was reported as a common finding in early healthy

volunteer trials of artemisinin derivatives. [345] Given that the artemisinin drugs are

widely regarded as safe in humans, this phenomenon may be uncommon but merits

further evaluation in future clinical trials.


The present study has also provided novel data relating to the population PK of

both ARTS and DHA in Melanesian children with a high prevalence of α-

thalassaemia mutations and in whom blood sampling was carried out after both

doses of rectal ARTS given 12 h apart. In one out of seven children, there was

evidence that one of the doses was expelled soon after administration but this did

142

not result in any cases of early treatment failure. It is notable in this respect that

expulsion of suppositories has been noted in up to 19% of Malawian children. [247]

In those with valid data over the 24-hour sampling period, the bioavailability of DHA

was an average of 28% lower after the second compared with the first dose. Alpha-

thalassaemia had no influence on the PK properties of ARTS and DHA, nor on

measures of parasite and fever clearance.

The PK parameters derived from this study are difficult to compare with those of the

previous studies in children with falciparum malaria performed by Halpaap et al.

[238] and Sabchareon et al. [242] The former authors used a dose of rectal ARTS

of only 1.2-2.2 mg/kg given at 0 and 4 h and determined Cmax and tmax after the first

dose. [238] Mean plasma ARTS and DHA Cmax values were 317 and 634 nmol/L,

respectively and they occurred close to 1 h after drug administration. Given the

proportionately lower dose used these data are consistent with those observed in

the present study. Out of 12 subjects, one had very low concentrations after the first

dose, implying that, as in the present study, the suppository had not been retained.

In the Thai study by Sabchareon et al. [242], PK analysis was performed in an

undefined subset of 19 out of 52 children who received 10-19 mg/kg rectal ARTS

once a day for three days, and Cmax, t½e and AUC0-12 for DHA were derived. The

mean Cmax, presumably after the first dose, was 2382 nmol/L and the mean

apparent t½e 0.7 h. These values are similar to those observed in our PNG children

and, consistent with the findings of Halpaap et al. [238], show that peak plasma

DHA concentrations are achieved promptly after drug administration.

Krishna et al. [232] used doses of approximately 10 mg/kg or 20 mg/kg in Ghanaian

children with moderately severe malaria. The mean Cmax values for DHA were 2400

and 3100 nmol/L, respectively in these two groups, with tmax values just under 2 h in

each case. As with the first two studies [238, 242], these results are in accord with

those in our PNG children. Krishna et al. also reported an inverse relationship

between dose and bioavailability. [232] Krishna et al. also incorporated a lag time

(means 0.6 and 0.4 h in low and high-dose groups, respectively) into their one-

compartment model [232], and the mean DHA t½abs estimates in their patients were

0.7 and 1.1 h, respectively. These parameters describe the same processes as our

143

ARTS t½ abs (mean 2.3 h) and t½ ARTS conversion to DHA (mean 0.3 h), with the

result that mean tmax for DHA in our patients (2.3 h) and in theirs (1.8 h) were

similar. Values from our population model for V/F and CL/F (2.15 L/kg and 2.25

L/h/kg, respectively) were within the range of values for these two variables from

the 26 subjects studied by Krishna et al. [232] (1.1-14.4 L/kg and 1.0-22.3 L/h/kg)

albeit towards the lower limit in each case. This most likely reflects the fact that our

patients were older (mean age 7.2 vs 4.1 years).

A small study of Thai children with severe malaria by Pengsaa et al. [341] using

single compartment modeling was limited to just 2 subjects, but also demonstrated

similar DHA Cmax (3890 nmol) and tmax (1.2 h) to the present study.

A study recently published by Simpson et al. [247] applied population PK methods

to a mixture of sparse and rich sampling data from a heterogeneous sample of 141

children with moderately severe malaria from Thailand and Africa (including some

children in the study by Krishna et al. [232]). Consistent with our findings, this

demonstrated that inclusion of body-weight as a covariate improved objective

function of the model, (higher body weight was associated with higher V/F). Notably

however, other covariates, including age, geographic location (Thailand, Ghana,

Malawi or South Africa), parasitaemia and other indicators of disease severity

(lactate, glucose and packed cell volume) were not influential. The values for DHA

V/F calculated (1.81 L/kg for 15kg children and 3.04 L/kg for 30kg children) were

comparable to our overall value of 2.15 L/kg in PNG children (who had a mean

body weight of 20.5 kg). Simpson et al. [247] also calculated similar values for CL/F

(2.64 L/kg/h) and t½e (0.72 h) although their model predicted a slightly later DHA

tmax (approximately 3 h) than our data (2.3 h). However, theirs was a single

compartment model which, due to the sparseness of the dataset, employed both a

fixed lag time and a fixed appearance rate constant.

A consistent and important finding in the studies by Simpson et al. [247], Krishna et

al. [232] and the present has been the degree of inter-individual variability in PK

parameters. The final model used by Simpson et al. calculated inter-patient

variability of 96% and 66% for V/F and CL/F, respectively, whilst Krishna et al.

demonstrated that relative bioavailability of DHA after rectal vs i.v. ARTS ranged

144

between 6% and 131%. [232] It is likely that the patients from the present study

would show the same magnitude of variation. This has implications for treatment.

Although a clear relationship between drug concentrations and clinical response

has not been demonstrated in any study of ARTS suppositories, Simpson et al.

[247] found an association between a higher apparent V/F and the probability of

requiring rescue treatment. Because of the study design used it was not possible to

determine whether this effect was due to a genuinely larger Vd in these patients or

to reduced bioavailability (F). This raises the question as to whether patients on the

lower end of the very wide spectrum of bioavailability may be at risk of sub-

therapeutic response. There is evidence that the therapeutic index of the

artemisinin derivatives is large [184, 185] and so high rectal doses of ARTS (e.g. 20

mg/kg) may improve the probability of achieving therapeutic plasma concentrations

without necessarily raising the risk of unacceptable side-effects in children with

malaria.

Relative to the first dose of ARTS in the present patients (F = 1.0 or 100%), there

was a significant decrease in F after the second rectal dose in our children to a

mean of 0.73, and only three of the 41 patient datasets (7.3%) had a second-dose F

>1.0 (values 1.08, 1.23 and 1.52, respectively). Although the data were not

presented, Krishna et al. [232] stated that there were no differences in PK

parameters for DHA suppositories given 12 h apart in high dose. However, these

same authors found a borderline (P=0.06) difference between DHA AUC values for

i.v. doses at 0 and 12 h. [232] In addition, a variety of studies have suggested a

time-dependent reduction in the bioavailability of artemisinin derivatives during

treatment [248, 250, 346, 347] including during the first few days of ARTS therapy

[249]

The reasons for the change in bioavailability between the first and second doses

are unclear but may relate to changes in physiological factors. In the present study,

a relatively high core body temperature at the time of the first dose may have lead

to greater rectal blood flow and better absorption of drug compared to the second

dose. Given the less frequent sampling schedule used following the second dose of

ARTS, it was not practicable to analyse the PK of both doses separately. However,

the induction of the metabolic clearance of ARTS or DHA seems less likely because

145

their metabolic pathways (hydrolysis [336] and glucuronidation [335], respectively)

are unlikely to change acutely. Nevertheless, there is clear evidence for induction

of artemisinin metabolism during repeated oral or rectal dose administration [248,

348], largely as a result of CYP2B6 induction. [349]

The high prevalence of �-thalassaemia in this region of PNG was confirmed with

78% of our sample having at least one abnormal �-thalassaemia gene and 46%

being either homozygous or dual heterozygous for abnormal genes. In non-malaria-

infected volunteers with thalassaemia, Ittarat et al. found that the mean AUC of

ARTS plus DHA measured by bioassay after i.v. ARTS was 9-fold higher, and the

Vss 15-fold lower, than in normal subjects. [286] These findings were attributed to

accumulation of artemisinin derivatives in thalassemic erythrocytes [287], or

perhaps to differences in ARTS metabolism in thalassemic subjects. During the

model building process, we attempted to improve the description by using �-

thalassaemia genotype as a covariate of V/F. However, this did not alter the V/F

significantly, or decrease its variability, while between-subject variability also was

unchanged. There may be other factors relating to malaria infection itself that

explain the differences between our results and those of Ittarat et al. [286], but their

investigation was beyond the scope of the present study.

Studies in vitro have demonstrated markedly increased IC50s of artemisinin drugs in

parasite cultures using thalassaemic red blood cells. [285, 287-289, 350] The

determinants of reduced parasite susceptibility to artemisinin drugs in thalassaemic

cells remain unclear but may relate to competitive uptake and inactivation of drug

by thalassaemic host cell components, the abnormal Hb itself [289], to erythrocyte

membrane-bound haem [350] or alterations in oxidative stress within parasitized

thalassaemic erythrocytes. [288] Whether these findings have relevance in vivo is

unclear as we found no influence of thalassaemia mutations on PCT or FCT. The

present findings, coupled with the lack of observed toxicity and excellent clinical

efficacy in this population, are reassuring for the future use of artemisinin

derivatives in Melanesian and other populations with a high prevalence of this

common Hb variant.

146

3.2.5.4. Implications for practice

Analysis of patients treated with oral ARTS seems to suggest that a dose-parasite

clearance response relationship may exist for DHA at concentrations of up to 1400

nmol/L. [351] Current manufacturer and WHO guidelines recommend a rectal dose

of 5 mg/kg. [352] A higher dose of 10-20 mg/kg may be justified based on PK and

PD data from the present and other studies, current knowledge of dose-response

relationships of artemisinin drugs, wide variability in bioavailability of ARTS

suppositories and the fact that the toxic threshold is very unlikely to be exceeded.

Pharmacokinetic data published by Simpson et al. subsequent to this study [247],

that evaluated a larger more heterogeneous population from various geographic

locations, suggest that factors such as age, ethnicity and disease severity are

unlikely to have profound effects on PK disposition. Therefore, the findings of the

present study are probably applicable to a wide range of patients in various

settings. However, findings in this and the present study of a positive association

between body weight and V/F may have relevance to the treatment of larger

children and adults who may be at greater risk of achieving sub-optimal drug

concentrations.

147

3.3. Clinical trial 4 (a). Severe falciparum malaria in Papua New Guinean children: Clinical features and host genetics

3.3.1. SUMMARY

Aims: To define the clinical features of severe malaria in Melanesian children and

examine the influence of common erythrocyte polymorphisms on clinical

presentation and markers of disease severity

Methods: As part of an interventional drug study, clinical features of disease

severity (including impaired consciousness, metabolic acidosis and

hyperlactatemia) and genotypes related to the three common erythrocyte

polymorphisms prevalent in coastal PNG ( �-thalassaemia, GPC∆ex3 and

SLC4A1∆27) were determined in 70 children with severe malaria in Madang, PNG.

Univariate analysis was used to examine associations between genotype and

clinical markers of severity.

Results: Children with impaired consciousness were significantly less likely to carry

the GPC∆ex3 mutation than those with other features of severity, suggesting

protection from cerebral malaria. Consistent with previous studies, no children with

the SLC4A1∆27 mutation presented with cerebral malaria. However, children with

metabolic acidosis were more likely to carry the SLC4A1∆27 mutation.

Conclusions: Survival benefit from GPC∆ex3 and SLC4A1∆27 mutations may

relate specifically to protection from cerebral malaria but this may be offset by

increased susceptibility to other complications.

3.3.2. AIMS

1. To define the clinical features of severe malaria according to current (year

2000) WHO definitions of severe malaria in a cohort of Papua New Guinean

children.

2. To define the genotypic prevalence of the three most common erythrocyte

polymorphisms including �-thalassaemia, Gerbich blood group (GPC∆ex3

mutation) and South Asian Ovalocytosis due to band 3 mutation

(SLC4A1�27) in a group of children with WHO-defined severe malaria and

then to:

148

a. Compare these rates with published prevalence rates from healthy

populations in the area

b. Examine their relationships to the primary markers of disease

severity (impaired consciousness, metabolic acidosis and

hyperlactatemia) within this group of children with severe malaria.

3.3.3. METHODS



Clinical assessments, procedures, data collection and sample preparation were

performed with the help of field workers Elizah Dabod, Barry Kalissa and Judy

Longo (PNG IMR, Madang). Microscopy was performed by Kerry Lorry (PNG IMR,

Madang). DNA extraction and PCR-detection of the �-thalassaemia, SLC4A1�27

and GYPC�ex3 mutations were performed by Sheral Patel (Case Western Reserve

University, Cleveland, Ohio). Statistical analysis was performed in collaboration with

Ivo Mueller (PNG IMR, Goroka).

I conceived and designed the study, prepared the funding application, analysed and

interpreted the data and took primary responsibility for preparation of a manuscript

for publication (submitted).


The study design chosen was of a cross-sectional descriptive study. The patient

population required for an interventional trial (described in 3.4 of this chapter)

enabled the opportunity to evaluate children aged 1-10 with strictly WHO-defined

severe malaria.

For description of features of severity, emphasis was placed on three clinical

features which have been shown to have the greatest prognostic significance as

predictors of mortality (in this and other settings), namely impairment of conscious

state, metabolic acidosis and hyperlactatemia. [25] This enabled definition of three

major sub-classifications of disease severity within the sample, which could then be

used to examine their associations with each of three erythrocyte polymorphisms.

149

This sub-analysis effectively constituted a case-control design intended to generate

hypotheses regarding protective effects of these polymorphisms. Because �-

thalassaemia and band 3 mutations had already been well-studied in this regard,

the erythrocyte polymorphism of most interest was GPC∆ex3.

The study of erythrocyte polymorphisms was opportunistic and the data analysis

exploratory. Therefore no a priori estimation of required sample size was

performed.


The study was conducted at MGH which serves an area hyper-endemic for P.

falciparum as described previously. [263] Severely ill children are referred from

surrounding rural health centres or come directly from peri-urban settlements.

Those from peri-urban settlements are often trans-migrants from the Sepik area of

PNG. Children aged 1-10 years with clinical features of severe malaria [352] were

eligible for recruitment. Screening was based on the World Health Organization

(WHO) Integrated Management of Childhood Illness (IMCI) procedures and

inclusion criteria included drowsiness or unconsciousness (Blantyre coma score <5,

assessed >1 h after a convulsion), respiratory distress (respiratory rate >40/min

and/or chest in-drawing), severe anaemia (Hb <50 g/L), prostration (inability to sit

unaided in a child usually able to do so), convulsions within the previous 24 h, and

nausea and vomiting precluding oral therapy. Children with at least one of these

features and a blood slide with P. falciparum parasite density >500/µL at entry were

eligible. Mixed infections with P. vivax, P. ovale or P. malariae were allowed,

provided that the P. falciparum parasitaemia was >500/µL. Patients with features of

another infection, malnutrition, diarrhoea or other significant co-morbidity, treatment

with an artemisinin derivative within the previous 24 h or whose parent/guardian

declined to give consent, were excluded and treated according to PNG standard

treatment guidelines. All children evaluated in this study were subsequently

enrolled in the antimalarial treatment study detailed in the following section of this

chapter (3.4). Enrolments were conducted between August 2003 and March 2004.

150


Detailed baseline clinical assessment included Blantyre Coma Score (BCS), which

was delayed for at least one hour following a convulsion, [27] respiratory rate and

signs of respiratory distress (in-drawing, intercostal recession and nasal flaring),

and prostration (an inability to sit or stand in a child who would ordinarily be able to

do so). A history of multiple convulsions (two or more in the previous 24 h), inability

to eat, drink or breast feed and vomiting “everything” (indicating an inability to

tolerate oral therapy) was also taken in line with the latest WHO criteria for defining

severe malaria. A thorough history and physical examination was conducted to

exclude other potential causes of fever including especially pneumonia, meningo-

encephalitis and measles. Chest X-ray and lumbar puncture were performed as

clinically indicated to exclude these significant co-morbidities.


3.3.3.5.1. Biochemistry, haematology and microscopy

Haemoglobin was assayed in whole venous blood using a portable

spectrophotometer (Haemaccue®, Angelholm, Sweden) and glucose was assayed

at the bedside from capillary blood using a portable glucometer (Exactech,

Medisense, Abingdon, UK).

Three mL of venous blood was taken into lithium heparin tubes and centrifuged

immediately at 1500 g for 5 min. Plasma was then used for biochemical assays

within the following 24 h or stored frozen at -20 °C and assayed at a later time

following transport to Australia. Biochemistry, including plasma lactate and

bicarbonate were performed using a Kodak Ektachem DT60 autoanalyser,

Eastman-Kodak, Rochester, NY.

Giemsa-stained thick blood smears were examined by a skilled microscopist (Kerry

Lorry, PNG IMR). At least 100 fields were examined at 1,000x magnification before

a slide was considered negative. For positive slides, parasite density was calculated

from the number of asexual forms/1,000 leucocytes and the whole blood leucocyte

count.

151

3.3.3.5.2. Genotyping for erythrocyte polymorphism mutations

EDTA-anticoagulated whole blood was collected at the time of enrolment. DNA

extraction and PCR-detection of the SLC4A1�27 and GYPC�ex3 mutations were

performed as described previously. [268, 272]


Statistical analysis was performed using STATA8 (STATA Corp, College Station,

USA). Data are presented as mean ± SD or median (interquartile range (IQR)).

Categorical data were compared using Fisher’s exact test.

3.3.4. RESULTS

Of 79 children enrolled in the initial treatment study (see 3.4), DNA was available in

70 (89%). The baseline characteristics of the children who provided a sample are

summarised in Table 3.7. Due to missing or unusable specimens, only 60 had

complete plasma lactate and bicarbonate results. All 70 children responded to

allocated therapy and recovered uneventfully.

Relationships between band 3, GPC and �-thalassaemia genotypes and the three

disease manifestations of interest are shown in Table 3.8 A, B and C, respectively.

SLC4A1�27 appeared to be associated with an increased risk of metabolic acidosis

and possibly hyperlactatemia (Table 3.8 A). No child with impaired consciousness

carried SLC4A1�27. GPC∆ex3 was significantly less common in children with

impaired consciousness (1 in 14, or 7%). No significant association was seen with

�-thalassaemia carrier state which was present in 86% of the total sample.

152

Table 3.7. Demographic and clinical features of Papua New Guinean children

enrolled with WHO-defined severe malaria enrolled in Clinical trial 4. Data are

number (%), mean ± SD or median (interquartile range). N=70 unless stated

otherwise.

Characteristic

Age 3.1 (2.1-5.0)

Sex (male) 38 (54)

Duration of symptoms (days) 3 (1.3-5.0)

Axillary temperature (°C) 38.7 ± 1.3

Respiratory rate (/min) 43.1± 7.5

Parasite density (/µL) 50 080 (16 920 – 100 440)

Mixed infection: P. falciparum and P. vivax 4 (5.7)

*Impaired consciousness: BCS �4 14 (20.0)

Coma: BCS �2 3 (4.3)

Hyperlactatemia: Plasma lactate >5 mmol/L (n=60)

19 (31.7)

Acidosis: Plasma bicarbonate �15 mmol/L (n=60)

19 (31.7)

*Hypoglycaemia: <2.2mmol/L 0

*Anaemia: Hb<5 (n=66) 9 (14)

*Multiple convulsions 29 (41.4)

*Frequent vomiting 22 (31.4)

*Unable to drink or breastfeed 25 (35.7)

*Prostration 22 (31.4)

*WHO criteria for severe malaria [24]

153

Table 3.8. Band 3, GPC3 and �-thalassaemia genotype and relationship to severe

manifestations of disease

A. Band 3

Band 3 genotype Manifestation of severity

Wild type SLC4A1�27 heterozygote

Fisher’s exact test

Odd-ratio (95%CI) hetero-

zygote vs wild type

Impaired consciousness (n=70)

BCS ≥4 50 6

BCS <4 14 0

P=0.25

0 (0-2.4)

Metabolic acidosis (n=60)

Plasma bicarbonate >15 mmol/L

40 1

Plasma bicarbonate �15 mmol/L

14 5

P=0.01

14.3 (1.4-693)

Hyperlactatemia (n=60)

Plasma lactate ≤5 mmol/L

39 2

Plasma lactate >5 mmol/L

15 4

P=0.07

5.6 (0.7-65.7)

154

B. GPC3

GPC genotype Manifestation of severity

Wild type

GPC∆∆∆∆ex3 hetero-zygote

GPC∆∆∆∆ex3 homo-zygote



zygote vs wild type


BCS ≥4 27 25 4

BCS <4 13 1 0

P= 0.008

0.08 [0-0.7]



25 12 4


8 11 0

P= 0.06

2.9 [0.8-10.5]



23 14 4


10 9 0

P= 0.38

1.5 [0.4-5.2]

155

C. Alpha-thalassaemia

�-thalassaemia genotype Manifestation of severity

Wild type Carrier



zygote vs wild type


BCS ≥4 7 49

BCS <4 3 11

P= 0.31

0.52 (0.1-3.1)



6 35


4 15

P=0.39

0.64 (0.13-3.23)



6 35


4 15

P=0.39

0.64 (0.13-3.23)

3.3.5. CONCLUSIONS

The 7% prevalence of GPC∆ex3 in children with impaired consciousness is much

lower than published prevalence rates from the community (32-71%). [46] Within

the present sample, children with impaired consciousness were significantly less

likely to carry the mutation than those with other manifestations of severe malaria.

Taken together, this may suggest a specific protection against cerebral malaria by

the GPC∆ex3 mutation.

No child with impaired consciousness carried the SLC4A1�27i mutation. Although

not statistically significant, the absence of the SLC4A1�27i mutation in any of the

156

children with impaired consciousness is consistent with earlier findings of complete

protection by SLC4A1�27 against cerebral malaria in this population. [269, 353]

Paradoxically, SLC4A1�27 appeared to be associated with an increased risk of

metabolic acidosis and possibly hyperlactatemia (Table 3.8 A). The reasons for this

are unclear as no such associations were found in a previous study. [353] It is of

interest that the band 3 erythrocyte membrane protein functions as the principal

exchanger of bicarbonate in the process of removing carbon dioxide from tissues.

[354] Therefore, SLC4A1�27 might influence systemic acid-base status. It is also

possible that, if SLC4A1�27 delays or prevents the worrying symptom of impaired

consciousness, the child’s parents may not seek prompt treatment, allowing greater

time for metabolic complications to develop. This also may explain the weak, non-

significant association of GPC∆ex3 with metabolic acidosis.

Genetic factors could influence disease progression at any point from asymptomatic

malaria to death. However, available data suggest that the protection associated

with SLC4A1�27 or GPC∆ex3 does not manifest early in this spectrum. [46] The

observation that hospitalized children carrying GPC∆ex3 are less likely to present

with impaired consciousness is the first time that this GPC mutation has been

associated with protection against any form of severe malaria. In addition to the

geographic relationship between the mutation and prevalence of P. falciparum and

in vitro data showing an effect of the mutation on parasite invasion (see section

1.5), it supports the hypothesis of a naturally selected protective effect against

malaria-related morbidity and mortality. However, the design of the current study,

limited as it was only to subjects with severe malaria, was imperfect and subject to

selection bias and confounding. An adequately designed case-control study

comparing GPC∆ex3 mutation prevalence in children with cerebral malaria versus

carefully matched healthy community controls would provide the best possible level

of evidence to further test the hypothesis that GPC∆ex3 mutation is protective

against cerebral malaria.

Impaired consciousness, acidosis and hyperlactatemia have been validated as

powerful predictors of mortality both in PNG and African countries. [27, 80, 81] If

SLC4A1�27 or GPC∆ex3 do confer an overall survival advantage to populations

157

exposed to malaria, these data suggest that this relationship may be more complex

than previously envisaged. In particular, the association of SLC4A1�27 with

acidosis and hyperlactatemia indicates that protection against cerebral malaria may

be offset by an increased risk of metabolic complications.

The present findings are from a relatively small group of children and the design of

this study was limited by the lack of a suitably matched control group of healthy

children. There is a need for larger case-control studies exploring the effects of

SLC4A1�27 or GPC∆ex3 on cerebral malaria and acid-base status. Such studies

could help to identify metabolic processes leading to severe, life-threatening

pediatric malaria and facilitate the development of effective preventive and/or

therapeutic interventions.

158

159

3.4. Clinical trial 4 (b). Artesunate suppositories versus intramuscular artemether for severe falciparum malaria

3.4.1. SUMMARY

Aims: To compare the therapeutic efficacy of rectally administered artesunate

(ARTS) with intramuscularly (i.m.) administered artemether (ARM).

Methods: In an open-label randomised trial, children with WHO-defined severe

Plasmodium falciparum malaria were randomised to receive either ARTS

suppositories (n=41; 8-16 mg/kg at 0 and 12 h then daily) or i.m. ARM (n=38; 3.2

mg/kg at 0 h, then 1.6 mg/kg daily). Parasite density and temperature were

measured 6-hourly for ≥72 h. Primary end-points included PCT50, PCT90 and the

time to per os status. In a subset of 29 patients, plasma levels of ARM, ARTS and

their common active metabolite DHA were measured during the first 12 h.

Results: One suppository-treated patient with multiple complications died within 2 h

of admission but the remaining 78 recovered uneventfully. Compared to the ARM-

treated children, those receiving ARTS suppositories had a significantly lower mean

PCT50 (9.1 vs 13.8 h, P=0.008) and PCT90 (15.6 vs 20.4 h: P=0.011). Mean time to

per os status was similar in each group. Plasma concentrations of primary drug

plus active metabolite were significantly higher in the ARTS suppository group at 2

h post-dose.

Conclusions: The earlier initial fall in parasitaemia with ARTS is clinically

advantageous and mirrors higher initial plasma concentrations of active

drug/metabolite. In severely ill children with malaria in PNG, ARTS suppositories

were at least as efficacious as i.m. ARM and may, therefore, be useful in settings

where parenteral therapy cannot be given.

3.4.2. AIMS

3.4.2.1. Primary

To compare the therapeutic efficacy of rectally administered ARTS with i.m.

administered ARM using the rate of clearance of parasites from the peripheral blood

and time taken for the child to return to per os status as the primary PD

measurements of response.

160

3.4.2.2. Secondary

i. To compare efficacy based on other PD indicators of clinical response including

rate of fever resolution

ii. To compare the bioavailability of the two treatment regimens by examining drug

concentrations in plasma during the first 12 h following drug administration in a

subset of patients.

3.4.3. METHODS



Clinical assessments, procedures, data collection and sample preparation were

performed with the help of field workers Elizah Dabod, Barry Kalissa and Judy

Longo (PNG IMR, Madang, PNG). Microscopy was performed by Kerry Lorry (PNG

IMR, Madang, PNG). The HPLC-MS and gas chromatography-mass spectrometry

assays for ARTS, ARM and DHA were developed and validated by Ken Ilett,

Juliana Hamzah and Madhu Page-Sharp (all of School of Medicine and

Pharmacology, UWA) and performed by Juliana Hamzah, Madhu Page-Sharp and

Gregory Chiswell.

I conceived and designed the study, prepared the funding application, obtained

ethical approval, and took primary responsibility for analysis and interpretation of

the clinical and PK data and preparation of a manuscript for publication (publication

5).


Key health policy-makers (Professor John Vince, senior paediatrician Port Moresby

Hospital and University of PNG) were consulted at the study design stage, in

association with WHO (Steve Bjorge, WHO country malaria advisor for PNG). Both

expressed a high level of interest in the results of this study.

161

The study design was an open-label randomised controlled trial. Because of the

nature of the two preparations (i.m. and i.r. administration), blinding or the use of

placebo were neither feasible nor ethical. A comparator group of i.m. ARM was

chosen because this is officially recommended standard first line treatment for

severe malaria in PNG and therefore represents the best available standard of care

in this setting. Because of recent data suggesting poor and erratic absorption of i.m.

ARM [229, 347, 355, 356] and the present data showing good absorption and

excellent parasite clearance with ARTS suppositories (Clinical trial 3 – see 3.2), it

was felt that the intervention (rectal ARTS) was likely to be at least as safe and as

effective as the current standard of care, and that therefore a direct comparison in

children with severe malaria was ethically justifiable. The intended clinical use of the

intervention is in children with severe malaria, so this was chosen as the most

appropriate study group, with current WHO definitions of disease severity being

used as inclusion criteria. Primary outcomes related to early parasite clearance

(PC%12, PC%24, PCT50 and PCT90) were chosen as the best available surrogate

markers likely to have discriminatory power in a relatively small sample and as

being most appropriate given the intended clinical use of the intervention as an

emergency pre-referral treatment where early rapid reduction in parasite biomass is

the primary therapeutic goal.


The present study enrolled children with WHO-defined severe malaria who

presented to MGH, Madang during a 9-month period between August 2003 and

March 2004 at Madang. The catchment population, screening procedure and

eligibility criteria are as previously described in section 3.3.


At recruitment, an initial clinical assessment was performed, and blood taken for full

blood count and biochemistry as detailed in Section 3.3. A computer-generated

randomisation schedule was used to assign children to ARTS suppositories or i.m.

ARM. The subjects in the ARTS suppository group received a combination of one or

two 50 mg and/or 200 mg thermostable suppositories (Rectocaps®, Mepha

Pharmaceuticals, Aesch-Basel, Switzerland) to a total dose of 8-17 mg/kg. The

suppositories were kept at room temperature, removed from their air-tight foil

162

packaging immediately before use and administered base first without lubricant.

Children were observed for 1 h after dosing and, if the suppositories were expelled,

the same dose was readministered and the observation continued. Staffing ratios

to supervise and detect expulsion ranged from 1:1 to 1:3. A second dose was given

12 h later and then daily until the child was able to tolerate oral medication.

Children in the i.m. ARM group were administered ARM (Kunming Pharmaceuticals,

Kunming, China) at a dose of 3.2 mg/kg as an i.m. injection into the anterior thigh or

buttock. A daily i.m. dose of 1.6 mg/kg was given subsequently until the child was

able to tolerate oral medication. In both treatment groups, continuation oral therapy

comprised ARTS 2 mg/kg daily to complete 7 days of treatment and a single dose

of SP (25 mg/kg SDOX/ 1.0 mg/kg PYR) was given by mouth at 72 h.

All children were assessed comprehensively every 6 h, including their ability to

tolerate oral medication, to eat, drink and mobilize normally for age, axillary

temperature, pulse and respiratory rate, capillary blood glucose, BCS and parasite

density. Those with an axillary temperature >38.0 °C were given paracetamol, and

tepid sponging and fanning were started. Due to the possibility of serious bacterial

infections in children with severe malaria [357], parenteral antibiotics (i.m.

chloramphenicol under national treatment guidelines) were also administered if

considered appropriate by the treating physician. Other concomitant treatment,

including anticonvulsants, iron and folate, were recorded, as were blood

transfusions. All children were kept in hospital for at least 72 h and were

discharged when they had been afebrile for 24 h, they were eating, drinking and

mobilizing normally, and serial blood slides had been negative for >24 h.


3.4.3.5.1. Microscopy and other field laboratory tests

Microscopy, biochemistry and haematology was performed as previously described

in Section 3.3.3.5.1


A subset of the children aged >2 years, who were not anaemic and whose

parent/guardian provided informed consent had 4 or 5 additional blood samples

taken during the first 12 h. Blood was taken at predetermined intervals of 1, 2, 4, 6

163

and 12 h after the first dose of study drug, centrifuged within 30 min and separated

plasma stored at –20oC prior to transport to Australia on dry ice for drug assay.

Assay methodology is described in Appendix 1 (sections 7.13 and 7.14).

3.4.3.6. Efficacy assessment

The primary efficacy endpoints were the time for parasite density to fall, as

assessed from linear interpolation from the parasitaemia-time curve, by 50% and

90% of the baseline value (PCT50 and PCT90), and the equivalent percentage

reduction (relative to baseline) in parasite density at 12 and 24 h (PC%12 and

PC%24). An additional primary efficacy endpoint was the time to return to per os

status (from admission to first dose of oral medication). Secondary endpoints

included the parasite clearance time (from admission to the first of two consecutive

negative thick blood films; PCT), fever clearance time (time from first dose to the

first axillary temperature of a 24-hour period <37.5°C, FCT�), death, neurological

sequelae, suspected adverse drug reactions, time to full consciousness in children

admitted with impaired conscious state, and the time to normal mobility.


Based on parasite clearance data from Clinical trial 4, a sample size of 35 patients

per group was required to detect a 30% difference in PCT50 or PCT90 with power

≥80% and α=5%. Statistical analysis was performed using SPSS version 10 (SPSS

Inc., Chicago, IL, USA). Data are presented as mean ± SD, mean and 95%

confidence intervals (CI), or median and (range or interquartile range (IQR)).

Normally distributed continuous variables were compared by Student’s t test and

non-normally distributed data by the Mann-Whitney U test. Categorical data were

compared using Fisher’s exact and Chi-squared tests.

3.4.4. RESULTS


One-hundred and twenty-four children were identified with features of severe

malaria and an initial blood slide positive for P. falciparum. After excluding ineligible

subjects and those found to have a parasite density <500/µL by microscopic

quantification subsequent to randomization, 79 children (41 in the ARTS

suppository group and 38 in the i.m. ARM group) were included in the final analysis

164

(Table 3.9). This group included the 70 children previously described in Section 3.3

and an additional 9 children who did not have genotyping performed. Clinical

features in the two treatment groups are compared in Table 3.9. The features of

severity warranting inclusion in the trial for the group as a whole included multiple

convulsions (41%), impaired conscious state (28%), respiratory distress (76%),

anaemia (23%), prostration (34%), frequent vomiting (32%) and inability to tolerate

oral medication (35%). There were no significant differences in the prevalence of

these features or in the results of baseline laboratory tests by treatment allocation

except that the admission parasitaemia was higher in the ARTS suppository group

than the i.m. ARM group.

3.4.4.2. Clinical course

After excluding subjects with missing or unreadable slides, primary parasitological

endpoints were evaluable in 37 ARTS suppository and 35 i.m. ARM patients (Table

3.10). Median parasitaemia-time curves for the two treatment groups are shown in

Figure 3.5.A and individual patient plots in Figure 3.5. B and C. There was a more

rapid initial parasite clearance with rectal ARTS. Eight children in the i.m. ARM

group and 3 children in the ARTS suppository group had an increase in parasite

density during the first 6 h (odds ratio 3.8, 95% CI 0.7-17.8, P=0.08).

One child in the ARTS suppository group, a 5 year old girl, died during the study.

She presented with impaired conscious state (BCS 3), multiple convulsions,

frequent vomiting, hyperpyrexia (axillary temperature 39.6°C) and borderline

hypoglycemia (2.3 mmol/L). In addition to rectal ARTS (13 mg/kg) at study entry,

she received an i.v. bolus of 50% glucose followed by a 5% dextrose infusion and

an i.m. dose of chloramphenicol. Shortly after this, the child vomited and aspirated

before progressing to respiratory arrest. Resuscitation, including intubation and

ventilation, was unsuccessful and the child died 2 h after the first rectal ARTS dose

which, on post mortem examination, had been retained.

Twenty-two children had impaired consciousness (BCS <5) at enrolment. Mean ±

SD time to recovery of normal conscious state (BCS =5) in the ARTS suppository

group (15 ± 13 h; n=13) was similar to that in the i.m. ARM group (20 ± 12 h; n=9;

165

P>0.2). No child had a detectable residual neurological deficit on clinical

examination at discharge.

Table 3.9. Characteristics of children with severe malaria enrolled in Clinical trial 4

according to treatment arm (artesunate suppository or intramuscular artemether) at

the time of admission to the study. Data are mean ± SD or median (range) unless

otherwise specified.

Artesunate suppositories

Intramuscular artemether

Number of patients 41 38

Age (years) 3.6 ± 1.6 3.8 ± 2.2

Sex (% males) 20 (50) 21 (55)

Initial drug dose (mg/kg) 11.2 ± 2.2 3.6 ± 0.9

Body weight (kg) 12.6 ± 3.4 13.4 ± 4.8

Axillary temperature (°°°°C) 38.7 ± 1.1 38.7 ± 1.5

Respiratory rate (/min) 43 43

Pulse rate (/min) 122 ± 18 122 ± 16

Blantyre coma score 4.5 ± 1.0 4.7 ± 0.7

Parasite density (/µµµµL) 69,000 (770-458,000)* 40,100 (775-473,000)

Mixed infection – P. vivax (%) 1 (2) 3 (8)

Haemoglobin (g/L) 91± 22 84 ± 30

Blood glucose (mmol/L) 7.1± 2.7 6.6 ± 2.4

Plasma bicarbonate (mmol/L) 18.5 ± 5.3 17.1 ± 4.8

Plasma lactate (mmol/L) 4.4 ± 2.7 5.5 ± 4.3

Plasma urea (mmol/L) 4.1± 1.7 3.4 ± 1.2

Plasma bilirubin (µµµµmol/L) 28 ± 22 28 ± 25

* Mann-Whitney U test, P<0.05

166

Table 3.10. Clinical efficacy endpoints by treatment group. Data are mean ± SD or

median (range). Differences between groups () and their 95% confidence intervals

(CI) are also shown.

Outcome Artesunate suppositories

(n=41)


(n=38)

� (95% CI)

PCT50 (h) 9.1± 4.9 13.5 ± 7.9 -4.3 (-7.5 to -1.2)1

PCT90 (h) 15.6 ± 7.4 20.4 ± 8.0 -4.8 (-8.5 to -1.1)2

PC%12 17 (0-190)1 56 (0-3020) -

PC%24 0.5 (0-40) 1.2 (0-580) -

Time to per os status (days)

1.35 ± 0.76 1.69 ± 0.82 -0.25 (-0.62 to +0.11)

PCT (h) 30.3 ± 14.2 32.8 ± 12.9 -2.5 (-8.8 to +3.8)

FCTb (h) 18.3 ± 10.7 15.6 ± 9.2 +2.8 (-1.9 to +7.5)

1 P<0.01, 2 P<0.02

167

A

Time (h)

0 6 12 18 24 30 36

Par

asite

den

sity

com

pare

d to

bas

elin

e

0.0

0.2

0.4

0.6

0.8

1.0

Artemether i.m.Artesunate suppositories

Figure 3.5

A: Median (interquartile range) proportional reduction in parasite density following

treatment of Papua New Guinean children with severe malaria with either rectal

artesunate (open circles: n=37) or intramuscular artemether (closed circles :n=35)

in Papua New Guinean children aged 5-10 with severe malaria.

168

B

Time (h)

0 6 12 18 24 30 36

Par

asite

den

sity

com

pare

d to

bas

elin

e

0.001

0.01

0.1

1

10

100

C

Time (h)

0 6 12 18 24 30 36

Par

asite

den

sity

com

pare

d to

bas

elin

e

0.001

0.01

0.1

1

10

100

Figure 3.5

B: Individual patient parasite density-time data (various symbols) in 41 Papua New

Guinean children with severe malaria treated with rectal artesunate.

C: Individual patient parasite density-time data (various symbols) in 38 Papua New

Guinean children with severe malaria treated with intramuscular artemether.

169

3.4.4.3. Side-effects and complications

Five children passed suppositories within 1 h of the first dose. The suppositories

were retained following repeat dosing in all of these cases. Five children (2 in the

ARTS suppository group and 3 in the i.m. ARM group) developed constipation

and/or abdominal distension during early convalescence. The distension could be

marked and usually occurred around days 2-4. One child in the i.m. ARM group

developed abdominal pain, diarrhoea, nausea and anorexia on days 2-4, and

another ARM-treated child developed late fever on day 3. In all these cases,

symptoms were mild, their onset was after parasite and fever clearance, resolution

occurred spontaneously within 1-3 days, and oral ARTS therapy did not have to be

discontinued.

3.4.4.4. Plasma drug concentrations

A total of 29 children (15 in the ARTS suppository group and 14 in the i.m. ARM

group) had plasma samples taken for drug assay. Data from the ARTS suppository

group are shown in Figure 3.6 A and B, and for the i.m. ARM group are shown in

Figure 3.7 A and B.

Median plasma levels of both ARTS and DHA in this small group of severely ill

children were similar to those seen in Clinical trial 3. Simulated concentration-time

curves based on mean PK parameters from Clinical trial 3 are shown in Figure 3.6A

and B for comparison purposes. Because of the small patient numbers, the sparse

sampling protocol employed and the difficulty in generating valid PK data in the

ARM-treated patients [229, 347, 355, 356], formal PK modeling was not performed

in either treatment group.

170

Time (h)

0 2 4 6 8 10 12

Con

cent

ratio

n (n

mol

/L)

10

100

1000

10000

Time (h)

0 2 4 6 8 10 12

Con

cent

ratio

n (n

mol

/L)

10

100

1000

10000

Figure 3.6. Plasma artesunate and dihydroartemisinin concentrations in 15 Papua

New Guinean children with severe malaria following 11 mg/kg dose rectal

artesunate. The closed circles are individual data points, while the horizontal bars

show median concentrations at the different times. The solid lines show simulated

(using mean pharmacokinetic parameters) concentration-time profiles from the

previous study of children with uncomplicated malaria (Clinical trial 3). [358]

A: Artesunate B: Dihdyroartemisinin

B

A

171

Time after dose (h)

0 2 4 6 8 10 12

Con

cent

ratio

n (n

mol

/L)

10

100

1000

10000

Time after dose (h)

0 2 4 6 8 10 12

Con

cent

ratio

n (n

mol

/L)

10

100

1000

10000

Figure 3.7. Plasma artemether and dihydroartemisinin concentrations in 14 Papua

New Guinean children with severe malaria following a 3.2 mg/kg dose of

intramuscular artemether. The closed circles are individual data points, while the

horizontal bars show median concentrations at the different times.

A: Artemether B: Dihdyroartemisinin

A

B

172

The total artemisinin drug exposure in nmol/L for the two treatments was calculated

by adding both the concentrations of ARM and its metabolite DHA in subjects who

received ARTS suppositories, and ARM and its metabolite DHA in those subjects

who received i.m. ARM (Table 3.11). There was a significantly greater total

artemisinin exposure at 1 h and 2 h post-dose in the ARTS suppository group

(P<0.02), but exposure was similar at 6 and 12 h. The difference in total artemisinin

exposure was predominantly due to DHA, concentrations of which were significantly

greater in the rectal ARTS group at all time-points up to 12 h. Six patients in the

ARM group and 2 patients in the ARTS group had very low total artemisinin

exposure (< 500nmol/L) at 1 and 2 h post-dose. Clinical recovery and parasite and

fever clearance in these 8 patients were similar to those in other patients except for

one child from the i.m. ARM group. This child’s parasite density rose from 3,655/µL

to 46,225/µL during the 12 h after the first injection, fell to below the baseline level

by 24 h and was negative at 36 h.

Table 3.11. Artemisinin exposure by treatment group at selected times after drug

administration. Total artemisinin concentrations indicate sum of concentrations of

plasma artesunate + dihydroartemisinin or of artemether + dihydroartemisinin. Data

are median (range).

Artesunate suppositories


Time after dose (h) Dihydro-

artemisinin (nmol/L)

Total artemisinin

Dihydro-artemisinin

(nmol/L)

Total artemisinin

1 3509 (222-5218)*

2690 (172-9729)*

69 (21-4314) 441 (91-1192)

2 2437 (120-14613)*

3660 (169-21871)*

57 (18-4186) 438 (56-1251)

6 819 (84-2690)* 1165 (75-6318) 53 (29-509) 661 (70-2307)

12 798 (134-1192)* 278 (62-5205) 31 (22-81) 514 (67-1546)

* P<0.02

173

3.4.5. CONCLUSIONS

Data from the previous study of rectal ARTS (Section 3.2) suggested that rectal

ARTS 10-20 mg/kg was rapidly absorbed in children with uncomplicated malaria in

PNG, with plasma concentrations of drug and the active metabolite, DHA,

comparable with those achieved after parenteral administration of conventional 2-4

mg/kg ARTS doses. [182, 338] At the time this study was done, concerns had

emerged that the current standard therapy for severe malaria in PNG, i.m. ARM,

could be sub-optimal because of poor absorption. [229, 347, 355, 356] Evidence

that ARTS suppositories are effective in severe malaria in adults [235, 237, 240,

241, 343, 359] provided further justification for a trial of rectal ARTS in PNG children

with complicated P. falciparum infections. The present study demonstrated the

benefits of rectally administered ARTS for the initial treatment of severe malaria in

children, with a more rapid fall in parasite density than observed after i.m. ARM.

The lower rates of initial parasite clearance in ARM-treated patients probably reflect

the poor absorption of this drug from its i.m. depot, leading to concentrations of

ARM plus DHA that were 6 to 8-fold lower 1-2 h post-dose than the equivalent

drug/metabolite levels after rectal ARTS.

The median concentrations of ARTS, ARM and DHA in the present subjects were in

accord with those of previous studies of rectal ARTS [130, 232, 247] or i.m. ARM

[229, 347, 355, 356] when administered in equivalent mg/kg doses. The median

Cmax in 10 Vietnamese adults with severe malaria treated with an initial i.m. dose of

3.2 mg/kg ARM was 574 (range 67-1631) nmol/L and most had peak plasma DHA

concentrations <25 nmol/L. [355] A study of Thai adults with uncomplicated malaria

treated with 2 mg/kg i.m. ARM demonstrated a mean Cmax of 121 nmol/L DHA

equivalents by bioassay. [347] This was some 16 times less than that seen with

equivalent oral dosing and peak levels occurred at a median of 8 (range 4-24) h

after administration. [347] Studies in African children with severe malaria have

shown a mean Cmax of between 116-257 nmol/L as DHA equivalents. [229, 356]

Two studies in Thai adults demonstrated higher plasma levels of ARM and a much

greater rate of conversion of ARM to DHA. [181, 360] The reasons for these

apparently discrepant results are unclear but may relate to infection severity (one

Thai study was performed in healthy volunteers), ethnic factors and/or the different

174

age ranges of the subjects. These factors may have impacted on hepatic perfusion,

or on hepatic and/or extrahepatic enzymatic metabolism of the parent drug.

Alternatively, the DHA assay used in the Thai studies, one of the first such assays

developed, may have had a different specificity to other methods.

Based on in vitro parasite culture data, ARM may be intrinsically less potent than

either ARTS or DHA. [361, 362] This suggests that the total artemisinin

concentration derived by adding plasma ARM and DHA concentrations

overestimates bioactivity when compared to ARTS plus DHA. Dose-response

relationships for the artemisinin derivatives remain relatively poorly understood.

However, when ARTS is given orally, a positive in vivo relationship between mg/kg

dosage and parasite clearance rates appears to exist up to doses of 2 mg/kg,

above which a maximal effect is achieved. [351] This threshold dose equates to

peak plasma concentrations of 1400-2800 nmol/L DHA equivalents at 1-2 h post-

dose. [250] As an extension of this finding, the total plasma artemisinin

concentrations achieved in ARM-treated patients in this study at 1-2 h (medians 438

and 441 nmol/L), whilst exceeding published IC50s for ARM and DHA [97], explain

the inferior PD response in this group, particularly in individuals at the lower end of

the range of absorption. Indeed, another study has identified a subset of patients

with relatively poor outcomes in whom plasma ARM concentrations are

undetectable by either bioassay or HPLC (limits of detection approximately 35

nmol/L). [356]

It could be argued that the different PK and PD outcomes in the present study were

a function of the total doses administered. The mg/kg dose of ARTS was

approximately three times that of ARM. However, it seems unlikely that increasing

the dose of ARM would be either practical or effective. An increased dose means

an increased injection volume that would be painful and difficult to administer in

small children. Furthermore, larger injection volumes may lead to compression of

capillary beds surrounding the injection site that could further limit drug absorption.

[363] Although this could be addressed by dividing the dose into two injections at

separate sites, this increases complexity and consumables cost, adds to the

existing risks associated with i.m. injection and doubles the degree of discomfort in

most children. In any case, i.m. ARM now appears to be an inferior choice to i.m.

175

ARTS which could be more easily administered as a single injection, with superior

and more consistent absorption.

Although a number of evaluations of rectal ARTS have been performed in adult

subjects with severe malaria [235, 237, 240, 241, 343, 359], only one other very

small study has evaluated their use in strictly-defined severe malaria in children.

[341] Therefore, the current study constitutes an important demonstration of efficacy

in the group most likely to benefit from this intervention (children with severe

malaria). The study by Pengsaa et al. [341] of Thai children treated with initial doses

of either 10 or 20mg/kg showed rapid early parasite clearance in all 13 subjects

(median PCT50s 6-7 h and median PCT90s 11-12h) and no deaths. Two additional

published studies have evaluated rectal ARTS in childhood malaria of intermediate

severity. [117, 232] These also demonstrated comparable rates of parasite

clearance to those seen in the current study (also reviewed in section 3.2 of this

chapter and the Appendix) and no deaths in a total of 123 children treated. The

study by Barnes et al. [117] also used PC%12 as an outcome measure and is the

only study other than the present study to have compared rectal ARTS directly with

a conventional treatment for severe malaria (further discussed in the Appendix).

Because of differences in PK and parasite stage specificity, artemisinin drugs clear

parasites more rapidly than quinine, regardless of route of administration. [184, 185]

Nonetheless, the findings by Barnes et al. of significantly faster early parasite

clearance than treatment with parenteral quinine add to the findings of the present

study in suggesting that rectal ARTS is likely to have efficacy at least equivalent to

some existing parenteral treatments for severe malaria. The current study had

additional validity because it compared efficacy within the artemisinin group rather

than between classes of antimalarial agents.

This study was inadequately powered to detect between-treatment differences in

either death or permanent neurological sequelae. However, the mortality and rate

of neurological complications were reassuringly low, with only one such event, a

death 2 h after study entry, amongst 79 severely ill children. Indeed, the range of

complications present at study entry in this child and the rapid clinical deterioration

suggest strongly that any form of intervention would have had limited benefit.

176

Since the present study was performed, a large multi-centre South-East Asian study

of mainly non-immune adults (SEAQUAMAT) has demonstrated that i.v. ARTS is

associated with a lower mortality (15%) than i.v. quinine (22%) in severe malaria.

[116] The benefit of i.v. ARTS was not significant in subjects <15 years [116] but

this may have been a function of low numbers of patients and a lack of statistical

power. Although it is not yet clear whether the findings of SEAQUAMAT can be

extrapolated to children exposed to hyperendemic transmission (eg PNG or sub-

Saharan Africa) these data and the fact that i.v. or i.m. injection of ARTS provides

early and reliable therapeutic levels of drug and metabolite [355] without the close

observation needed after rectal administration, confirm parenteral ARTS as first-line

hospital-based treatment for severe malaria, including in children. Where facilities

do not exist to safely administer antimalarial drugs by injection, the present and

other data [117] indicate that ARTS suppositories are an appropriate alternative.

Both the drugs used in the present study were well tolerated, although a small

number of children in each group (1 in 16 overall) developed generally mild

gastrointestinal symptoms after 2-4 days of treatment. The time-course of these

symptoms relative to parasite and fever clearance, their occurrence in both groups

of patients and the fact that they have not been reported in previously-published

studies, could imply that they arose because of a reaction to artemisinin drugs in

susceptible Melanesian subjects. However, their spontaneous resolution during oral

ARTS therapy suggests that they were a delayed effect of severe malaria on

intestinal function.

Rectocaps® are manufactured to international Good Manufacturing Practice

standards and cost approximately US$0.38 and US$0.68 for 50 mg and 200 mg

suppositories, respectively. Based on the number of days of treatment required

prior to oral therapy, and the dosages administered in the present children, the

average cost of rectal treatment was US$2.44. The equivalent average cost of i.m.

ARM at US$1.50/ampoule was US$2.49 per child. Because this comparison does

not take into account consumables needed to administer injections, rectal ARTS is

at least as cost-effective as i.m. ARM for the initial treatment of severe paediatric

malaria in PNG. Nevertheless, the close observation necessary to ensure that

suppositories are retained may be problematic. It is notable that the staffing ratios

177

employed in this Clinical trial (ranging from 1:1 to 1:3) were much higher than would

be practically achieved at a busy clinic or hospital. Therefore, parenteral artemisinin

therapy may still be the therapy of choice in settings where injections can be safely

given. However, if use is intended for pre-referral emergency treatment in the

community (eg by village health worker or a child’s mother) [230], assuming that

those administering treatment are adequate trained, a high level of supervision to

detect expulsion should be possible, and suppositories represent an important

potential health intervention.

178

179

3.5. Clinical trial 5. Social and cultural acceptability of artesunate suppositories in Papua New Guinea

3.5.1. SUMMARY

Aims. To examine the pre-existing attitudes of Papua New Guinean caregivers

towards rectal drug administration in order to inform effective deployment strategies

for artesunate (ARTS) suppositories in children.

Methods Caregivers presenting with children with uncomplicated malaria

completed a preliminary questionnaire to assess their prior experience with, and

attitudes to, rectal administration. They were then offered treatment for their child

with three-days ARTS suppositories. Those who refused answered a questionnaire

to determine reasons for refusal and following treatment, those who accepted

answered a questionnaire to assess satisfaction. Willingness and ability to self-

administer, and preference for suppositories in the home or clinic, were assessed.

Social, demographic and clinical variables were examined for effects on attitudes

and preferences.

Results 131 caregivers were offered suppositories of who 71% accepted and 29%

refused. Not having spousal approval and fear of side-effects were the most

common reason cited for refusal. Caregivers who had never been to school or had

no previous experience of suppositories were more likely to refuse. Sixty-six

percent of caregivers who were asked to self-administer suppositories agreed and

did so successfully, but fathers were much less likely to agree than were mothers.

Following treatment, suppositories were perceived as effective (99%), safe (96%)

and fast-acting (91%), but problematic to administer to a struggling child (56%).

Shame, embarrassment and hygiene were not significant concerns. Sixty percent of

caregivers said they would prefer suppositories to oral treatment at home and 52%

to parenteral treatment at the health-centre.

Conclusions Overall acceptability of rectal administration is relatively high in PNG.

Deployment should be accompanied by health education that addresses the

practical aspects of administration, is appropriate for the illiterate rural poor and

directed at fathers as well as mothers.

180

3.5.2. AIMS

3.5.2.1. Primary

To examine the feasibility and acceptability of rectal administration of ARTS as a

viable health intervention at both community and health centre-level in PNG.

Specific objectives included determining the following:

1. The social and cultural acceptability of rectal administration in this

population.

2. The amenability of mothers to self-administration of suppositories to their

own children.

3. Other pre-existing perceptions and attitudes regarding rectal administration

that may require consideration for health education and deployment

strategies.

3.5.2.2. Secondary

To assess clinical response to three daily doses of rectal ARTS over a 3-day period.

3.5.3. METHODS



The primary research tool used in this study was a questionnaire designed in

collaboration with Rachael Hinton (PNG IMR, Goroka, PNG) who also performed

the preceding formative qualitative research necessary to its design. The

questionnaire itself was administered to study participants by Rachael Hinton,

Grace Pongua, Olive Oa and Alma Auwun (all of the PNG IMR, Goroka or Maprik).

Grace Pongua, Olive Oa and Alma Auwun also performed the clinical assessments

and procedures of the study. Moses Lagog (PNG IMR, Maprik) performed the

microscopy.

I conceived the study and its overall design, prepared the funding application,

obtained ethical approval and, with Rachael Hinton, performed the analysis and

181

interpretation of results. I was also a key contributor to the preparation of a

manuscript for publication in collaboration with others (publication 6).


Key health policy-makers (Professor John Vince, senior paediatrician Port Moresby

Hospital and University of PNG) were consulted at the study design stage, in

association with WHO (Steve Bjorge, WHO country malaria advisor for PNG). Both

expressed a high level of interest in the results of this study.

The social research undertaken in this study employed quantitative methods, using

a questionnaire as the primary research tool. The questionnaire itself had been

designed following a period of formative research based on qualitative methods.

This approach was chosen in order to firstly identify the areas of primary concern to

the population and then to formulate questions that would provide the greatest

insight into each particular issue. Formative research was also important in

designing questions in Tok Pisin (the lingua franca in PNG) that would be clearly

understood and for which both questions and responses would be as un-ambiguous

as possible.

The emphasis of the final questionnaire was in seeking to identify potential

impediments to deployment that may be amenable to health-education messages.

Because one possible deployment strategy is home-based management, it was

also necessary to establish to what extent caregivers would be comfortable

administering suppository treatment to their own children. The questionnaire was

also designed to examine the effect of key social indicators and the severity of the

child’s illness as covariates that may impact on attitudes and preferences.

The study group was chosen so as to be as representative as possible of the

setting in which suppositories would be most useful: Mothers of young children with

febrile illness living in remote rural areas of PNG. The design was made more

robust by actually offering treatment to mothers of children with malaria and asking

them to self-administer treatment. By doing this, it was felt that determining reasons

for treatment refusal could be particularly instructive, as would assessing

182

satisfaction with the treatment after the mother had actually experienced

suppository use in “real-life”.

3.5.3.3. Study site and participants

The present study was conducted at two geographically, linguistically and culturally

distinct sites on the north coast of PNG. The first, Kunjingini sub-health centre, is

situated in the Wosera sub-district of East Sepik. The Wosera district, with a total

population of 42,000 and a density as high as 400 persons per square kilometre, is

the most densely populated area of mainland PNG. [364] Since 1990 it has been

the site of a demographic surveillance system operated by the PNG IMR in

preparation for future malaria vaccine trials. [32, 260] Therefore, morbidity, mortality

and malaria epidemiology have been well-documented in the region. Plasmodium

falciparum transmission is hyper- to holo-endemic with a population prevalence of

parasitaemia of 60%. [32] Malaria is the second leading cause of admissions, the

third leading cause of death and the leading cause of death in children aged 0.4 to

4 years in the area. [32] The Kunjingini sub-health centre is a primary health-care

facility run in association with the local Catholic mission. It is the main source of

care for over 8,000 people, including those seeking treatment from outside the sub-

district. Private and informal drug selling is not practised to any significant extent in

the area. The health centre is accessible by road, but most patients must travel by

foot and may therefore journey several hours to reach it. The centre is staffed by

three nursing staff each day. Consultations, medicines and vaccines are provided

for a small daily or annual fee.

Mugil health centre is operated by the local Catholic mission and is situated within

the Sumkar district, Madang Province. The health centre covers a catchment area

of between 20,000 and 25,000 people. Local P. falciparum transmission is hyper-

endemic and malaria is the second leading cause of morbidity in the district and the

leading cause of death. [31] The health centre is easily accessible from the main

highway but most patients must travel long distances either by foot or by local

public transport to reach it. The centre is staffed by 11 people and sees an average

of 30 new malaria cases per day.

183

The present study was conducted during April 2004, May 2004 and February 2005

in Kunjingini and between March and April 2006 in Mugil.

3.5.3.4. Formative research and questionnaire design

Preliminary formative research employed to guide development of the subsequent

questionnaire design is detailed in Appendix B (section 7.2).

3.5.3.5. Enrolment procedures

As the study aimed to assess attitudes and treatment preferences of caregivers with

young children with malaria, it was felt that the best way to recruit a representative

sample was to identify children presenting to the health-centre with confirmed

malaria and then interview the caregiver at the time they were offered treatment

with ARTS suppositories. Eligible participants were those presenting with a child

aged 6 months to 7 years with confirmed uncomplicated P. falciparum malaria.

Children with fever (axillary temp >37.5 ºC) or a history of fever were screened

using a Paracheck® (Tulip diagnostics, Mumbai, India) rapid test for P. falciparum

malaria. Positive tests were confirmed by microscopy. Children with signs of severe

malaria [24] and children with signs of another infection or significant co-morbidity

were excluded from the study. For eligible children with a confirmed diagnosis of P.

falciparum, caregivers were informed of the study, offered suppository treatment for

their child and requested to answer a standard set of questions regardless of

whether they declined or consented to suppository treatment. This included

multiple-choice questions regarding preconceptions of effectiveness, harmful effects

and shame or embarrassment.

3.5.3.6. Treatment and clinical monitoring

If the caregiver declined rectal therapy the child received oral CQ-SP in accordance

with PNG standard treatment guidelines. [204] In this situation, additional multiple-

choice questions were asked to determine the primary reason for the caregiver’s

choice. If caregivers consented, children were treated with 10-15mg/kg ARTS

suppositories (Rectocaps�, Mepha Pharmaceuticals, Aesch-Basel, Switzerland)

administered as a combination of one to two 50mg or 200mg suppositories.

Suppositories were administered as a daily dose for three days, with a single dose

of SP (25/1.25mg/kg) administered on the final day of treatment. In order to ensure

adequate clinical response, each child was assessed daily through a symptom

184

questionnaire and a blood film for parasite density. The child was observed at the

health centre for one hour after suppository administration so that, if he/she passed

the suppository, the treatment could be re-administered. The first dose was

administered by a health worker and caregivers were asked to administer the

second and third dose under health-worker supervision.

3.5.3.7. Follow-up questionnaire

On completion of treatment, a second questionnaire was administered to

caregivers. This included multiple-choice questions to assess satisfaction with the

treatment, including perceptions of efficacy, adverse effects or other associated

problems. At this time, caregivers were also asked whether they would prefer to

administer suppositories or oral therapy as home-based treatment for their children.

3.5.3.8. Data management and analysis

Questionnaire responses, demographic and clinical data were all recorded in a

Microsoft Excel® spreadsheet. Some questionnaire responses were recorded as

categorical data (e.g. disagree strongly, disagree, not sure, agree and agree

strongly). These responses were sometimes grouped and recoded as binary

variables to facilitate data analysis. For instance negative responses (e.g. “very

unsafe” and “not very safe”) would be grouped together (“unsafe”). Key

questionnaire outcomes included whether the caregiver accepted or refused to

have their child be treated with suppositories at the time of presentation and initial

interview, whether the caregiver was able to successfully self-administer

suppositories on the second day of treatment and whether the caregiver said

he/she would prefer suppositories for either home or health-centre treatment when

questioned at the conclusion of the treatment course. Factors associated with these

key questionnaire responses (the dependent variables) were ascertained by

univariate analysis of a number of independent demographic (age and sex of child,

relationship of caregiver to the child), sociological (educational and marital status)

and clinical (temperature, respiratory rate, parasite density and Hb concentration at

baseline) variables.

Data analysis was by means of SPSS for Windows (v10.0, SPSS Inc, Chicago,

USA). Continuous variables were compared using Student’s t-test for comparison of

185

means from two independent samples. Non-normally distributed data were

compared using the Mann-Whitney U test. For binary variables, odds ratios were

determined with 2 x 2 tables and P values calculated using Pearson Chi-square

test.

3.5.4. RESULTS

3.5.4.1. Study participants

A total of 131 caregivers (61 at Kunjingini and 70 at Mugil) were offered suppository

treatment for their child. Table 3.12 details the characteristics of the children (A) and

their caregivers (B). The clinical features in the children were consistent with

uncomplicated malaria in PNG. Seventy percent of caregivers identified themselves

as the child’s mother and 24% as the father. The remaining 6% were a close female

relative such as an aunt or grandmother.

3.5.4.2. Prior knowledge and perceptions of rectal treatment

Caregivers’ responses to the initial questionnaire are summarised in Table 3.13.

Whilst 41% had heard of suppositories, only 29% had personal experience with

their use in their own or another child. The most common concerns related to

perceptions that suppositories would not dissolve and be absorbed (44%) and the

practical difficulties of administering to a reluctant child (47%). A smaller number

were concerned that administration would cause pain (37%) or interfere with the

child defecating normally (37%).

186

Table 3.12. Baseline characteristics in Papua New Guinean children with

uncomplicated malaria (A: n=131) and their caregivers (B: n=131) enrolled in

Clinical trial 5. Data are number (%), mean ± SD or median (interquartile range).

A. Children.

Variable Value

Sex (male) 61 (46.6)

Age (months) 44.9 ± 21.0

Duration of illness (days) 3 (2-6)

Weight (kg) 12.2 ± 3.1

Axillary temperature (°C) 37.6 ± 1.4

Respiratory rate (/min) 31.8 ± 9.2

Palpable spleen 58 (46)*

Haemoglobin concentration (g/L) 84 ± 18

Parasitaemia (asexual parasites/�L) 6660 (590-38640)

* 5 missing values

187

B. Presenting caregiver

Variable Value

Relationship to child:

• Mother 93 (71)

• Father 32 (24.4)

• Other female relative 6 (4.4)

Age (years) 32.4 ± 7.6

Educational status

• Never schooled 48 (36.6)

• One to 6 years schooling 57 (43.5)

• More than 6 years of schooling 26 (19.9)

Marital status

• Unmarried, widowed or separated 11 (8.3)

• Married 120 (91.6)

Number of alive children 3.8 ± 0.7

One or more children deceased 40 (30.5)

One or more adopted children 14 (10.7)

188

Table 3.13. Presenting caregiver initial knowledge and perception of rectal

administration of medicine. Data are number (%). N=131 1

Variable No Yes Unsure

Prior knowledge of suppositories

• Have you heard of or know about suppositories? 77 (59) 54 (41) 0

• Have suppositories been used on your child or another child that you know of before?

93 (71) 38 (29) 0

Concerns about suppository treatment

• Do you think suppositories can have side-effects or be unsafe to use? 2

40 (31) 62 (47) 27 (21)

• Do you think suppositories could have side-effects / problems such as they won't dissolve?

57 (43) 58 (44) 16 (12)

• Do you think suppositories could have side-effects / problems such as the child will not be able to defaecate?

59 (45) 49 (37) 20 (15)

• Do you think suppositories could have side-effects / problems such as the treatment will cause pain?

53 (41) 49 (37) 20 (15)

• Do you think suppositories could have side-effects / problems such as the child will resist the treatment being administered?

36 (28) 62 (47) 30 (23)

• Do you think suppositories could have side-effects / problems such as the treatment will cause other illnesses / the child to be sick?

64 (49) 32 (24) 34 (26)

• Do you think suppositories could have side-effects / problems such as you or child would be ashamed for the treatment to be used?

102 (78) 22 (17) 2 (2)

• Do you think suppositories could have side-effects / problems such as your spouse would be unhappy with the treatment?

69 (53) 32 (24) 30 (23)

• Do you think suppositories could have any other side-effects / problems?

103 (79) 28 (21) 0

Accepted treatment for the child 40 (30.5) 91 (69.5)

1 NB numbers do not always add to 100% due to a small number of “non-

responses” for each question. 2 10 (7.6%) felt they were very unsafe, 30 (23%) not very safe, 4 (5%) safe and 55

(42%) very safe. Twenty-seven (20.6%) were unsure and 2 (1.5%) did not respond

to the question.

189

3.5.4.3. Acceptance or refusal of suppository treatment

When offered treatment with suppositories for their children, 91 caregivers (69.5%)

accepted. Of the 40 (30.5%) who refused, 13 (33%) claimed they did so because it

was too difficult for them to attend the health centre for the three days required for

supervision of treatment. However, most (70%) said that they refused because they

“did not like the treatment route” due to a fear of side-effects and other

complications. The majority (70%) also claimed they refused because their

husband/spouse would be unhappy with the treatment. Other reasons given

included fear that the child would resist administration (45%), reluctance to self-

administer suppositories (40%) or some other reason (18%). Only 17% felt that

there would be a problem with them or their child experiencing shame associated

with the mode of administration.

Factors associated with caregiver refusal are summarised in Table 3.14. Caregivers

presenting to the Kunjingini field-site were more likely to refuse treatment than

those presenting at Mugil (odds ratio 2.97 (95% CI: 1.28-6.96), P=0.005). Other

factors associated with refusal included never having attended school (2.25: 0.98-

5.18, P=0.035), having had no initial prior knowledge of suppositories (2.81: 1.5-

6.99, P=0.012), no prior experience with their use (3.07; 1.08-9.15, P=0.019) and

having one or more children deceased (2.19: 0.93-5.16, P=0.049). Mothers were

more likely to refuse than were fathers but this was not statistically significant (2.38:

0.82-7.22, P=0.078). There was no significant difference in refusal rates according

to the age of the caregiver, the age or sex of the child or any clinical indicators of

severity (including temperature, Hb, respiratory rate or parasite density).

190

Table 3.14. Factors associated with declining to use suppositories and therefore

refusing to participate in trial. N=131

Variable Odds ratio (95% CI) P-value

Study site: Kunjingini vs Mugil 2.97 (1.28-6.96) 0.005

Sex of child: Female vs male 0.71 (0.31- 1.6) 0.37

Caregiver relationship: Mother vs father 2.38 (0.82-7.22) 0.078

Ever attended school: Never vs ever 2.25 (0.98-5.18) 0.035

Marital status: Single vs married 0.84 (0.17-3.78) 0.81

No prior knowledge of suppositories 2.81 (1.5-6.99) 0.012

Never used suppositories in your child or a child you know of

3.07 (1.08-9.15) 0.019

One or more dead children 2.19 (0.93-5.16) 0.049

One or more adopted children 0.35 (0.05-1.76) 0.16

Variable Refused Accepted P-value

Age of child (months) 42.2 ± 20.7 46.0 ± 21.1 0.35

Number of days sick 8.1 ± 13 4.7 ± 5.2 0.16*

Weight of child (kg) 11.2 ± 3.2 12.1 ± 3.1 0.11

Temperature at presentation (°C)

37.7 ± 1.4 37.2 ± 4.2 0.48

Haemoglobin level at presentation (g/L)

81 ± 17 84 ± 21 0.49

Respiratory rate (/min) 32.1 ± 7.0 31.7 ± 10.0 0.81

Age of caregiver (years) 32.1 ± 7.0 31.7 ± 7.9 0.74

Number of alive children 3.8 ± 2.0 3.9 ± 2.4 0.81

* Mann-Whitney U

191

3.5.4.4. Self-administration of suppositories

Seventy-six caregivers who attended clinic on the second day of treatment were

asked to self-administer the dose of suppositories and were then interviewed on the

third day of treatment. Fifty-three (70%) agreed to self-administer of whom 50 (94%

or 66% of total) did so successfully. Of the 23 who refused to self-administer, 12

were female caregivers (representing 21% of all female caregivers) and 11 were

male (representing 55% of all fathers). All 23 (100%) felt that health-workers were

better trained to perform the administration and 22 (96%) were afraid that they

would insert the suppository wrongly. Eleven (48%) were concerned that their child

would resist but only 4 (17%) said that they or their child would be ashamed.

Mothers were much more likely to agree to, and successfully administer,

suppositories than were fathers (OR 4.72: 1.55-14.7, P=0.002). Clinical variables

(fever and parasitaemia), educational or marital status and prior knowledge or

experience of suppository use were all not significantly associated with successful

administration.

3.5.4.5. Clinical response

Three children did not attend for one or more follow-up visits to the health centre.

One child developed signs of complicated malaria by the second day (impaired

conscious state), and was withdrawn from the study and given rescue treatment

with i.m. quinine. Another child completed three days of ARTS therapy but had

begun to deteriorate on the second day with worsening conscious state, vomiting,

convulsions, focal neurological signs and neck stiffness, despite having cleared

parasites from the blood film at 24 h. A diagnosis of meningo-encephalitis was

made and the child was treated with parenteral chloramphenicol. The child died 5

days later.

Overall, 88 children had a fully-supervised three-day course of ARTS suppositories

and a dose of SP on the final day. Apart from the child with early treatment failure

and the child who developed meningo-encephalitis, clinical response was adequate

over the 48 h of follow-up in all other children. Twenty-four h after commencement

of treatment, 94% were afebrile and the mean reduction in parasite density from

baseline was 99.3%. Of the evaluable patients, 28 (34%) had achieved a negative

192

blood slide by the second day (roughly 24 h after first dose) and a further 43 (52%)

by 2 days (48 h). Twelve (14%) had not cleared parasites by the third day (>2

days). In all but two of these, parasitaemia at 48 h was low (<200 parasites/�L).

Two patients had parasitaemias of 2000 and 2400/�L at the third day (48 h) but

both had had high initial baseline parasitaemias (145 000 and 145 140/�L).

3.5.4.6. Satisfaction following treatment course

Following completion of the treatment course, caregivers were re-interviewed to

determine their satisfaction with the treatment. The responses are summarised in

Table 3.15. Almost all felt that the treatment was effective (99%) and safe (96%).

The most common problems cited were the child struggling and resisting

administration (56%) and wanting to defaecate immediately after administration

(26%). Perceived benefits included more rapid action (92%) and greater efficacy

than other antimalarial treatments (91%). A substantial proportion (48%) felt the

route of administration was advantageous because it was an alternative to oral

administration which they found difficult in their own children. Only 9% regarded

treatment as unhygienic.

When caregivers were asked what would be their preferred mode of administration

for their child at home, 60% nominated suppository in favour of oral treatment

(31%). Caregivers who had ever attended school were much more likely to prefer

suppositories to oral treatment (OR 4.03: 1.32-12.52, P=0.005) but there were no

other factors predictive of a preference for suppositories. These included age and

sex of the child or caregiver, marital status, previous use or knowledge of

suppositories, indicators of clinical response (fever or parasitaemia at 24 and 48 h)

and markers of clinical severity at baseline.

193

Table 3.15.: Caregiver satisfaction/ acceptance following a 3-day treatment course

of ARTS suppositories (10-15mg/kg daily) to 88 Papua New Guinean children.

Data are number (%).

Question Yes No Not sure/ did not respond

Was the treatment effective? 87 (99)1 0 1 (1)

Was the treatment safe? 84 (96)2 0 4 (5)

Perceived problems with treatment:

• Child wanted to defaecate 23 (26) 65 (74) 0

• Child felt pain 16 (18) 71 (81) 1 (1)

• Child resisted administration 48 (56) 40 (46) 0

• Treatment was difficult to administer 9 (10) 79 (90) 0

• Unhygienic 8 (9) 77 (88) 3 (3)

• Family unhappy with rectal administration

3 (2) 76 (86) 9 (10)

• Fever persisted 7 (8) 78 (89) 3 (3)

• Other 10 (11) 77 (88) 1 (1)

Perceived benefits/advantages

• Child finds it difficult to take medicine orally

42 (48) 45 (51) 1 (1)

• Treatment worked quickly 81 (92) 5 (6) 2 (2)

• Treatment worked better than other treatments for malaria

80 (91) 6 (7) 2 (2)

• Had fewer side-effects than other treatment

28 (32) 55 (63) 5 (6)

• Other 22 (25) 65 (74) 1 (1)

Would you recommend this treatment to others?

80 (91) 7 (8) 1 (1)

If you had a choice of treatment to use at home, would you use suppositories (in preference to oral administration)?

53 (60) 35 (31) 4 (5)

If you had a choice of treatment to use at the health centre, would you prefer suppositories (in preference to oral administration or injections)?

52 (59) 25 (28) 3 11 (13)

1 56 (64) thought it was effective, 31 (35) very effective. 2 3 (3) thought it was safe, 81 (92) very safe. 3 11 (13) preferred tablets, 14 (16) preferred injections

194

3.5.4.7. Preferred mode of administration at home or health-centre

When caregivers were asked what would be their preferred mode of administration

for their child if it was administered by health-workers at a health centre, 52%

nominated suppository treatment in favour of either oral (13%) or injectable (16%)

treatment. The sole factor associated with a preference for suppositories over other

modes of administration was whether the caregiver felt their child was still sick at

day 2 (48 h) (OR 0: 0-0.7, P=0.003).

3.5.5. CONCLUSIONS

This study aimed to identify socio-cultural factors that could impact on the

successful deployment of ARTS suppositories as an intervention against childhood

malaria. Only one other published study has addressed this issue. [365]

The limitations of this study included its relatively small sample size which was

limited to only two cultural groups within PNG. Caution should therefore be

exercised in generalising its findings to other countries, or even to the many other

cultural groups within PNG. It was notable that the acceptance rate in this study

was significantly higher at the Mugil site than at Kunjingini. As well as cultural and

social differences, the community’s prior experience of and relationship to the

health-care system may also lead to important differences in the attitudes studied in

this research. For instance, there is little access to medicines through an informal

sector in PNG, making it very different to the situation in much of Asia and Africa.

Furthermore, the population at the Kunjingini site has had a longstanding

relationship with the PNG IMR that may have significantly influenced perceptions of

treatment for malaria. The degree of prior exposure of the population to rectal

administration of other medicines (such as paracetamol) may also be important.

The “face-to-face” nature of the interview process may also have biased the results

towards more positive responses than had they been conducted in an anonymous

fashion. Nonetheless, some of the results bear great similarity with other

experiences with suppository use. [365]

The present questionnaire responses indicate a relatively low level of existing

familiarity with rectal administration. However, prior knowledge or experience with

rectal administration was strongly associated with more positive attitudes to this

195

form of treatment. Caregivers with prior experience of rectal administration were at

least three times as likely to accept suppository treatment for their child. Experience

with rectal administration in PNG probably relates predominantly to paracetamol

suppositories (diazepam may also be given rectally to children with convulsions).

Paracetamol is more rapidly absorbed when given rectally rather than orally,

meaning that suppositories may effect an earlier reduction in fever. It is possible

that this may have contributed to positive perceptions of the efficacy of

suppositories more generally amongst caregivers. Similarly ARTS, given its intrinsic

potency, is likely to result in a more rapid killing of parasites and resolution of fever

and other clinical signs than would occur with conventional antimalarial drugs,

regardless of route of administration. [117] Although this study did not assess

definitive cure using the standard WHO 28-day test, the early clinical response was

excellent. In particular, rapid parasite clearance and resolution of fever occurred in

almost all patients. This appeared to be recognised by caregivers in this study, of

whom 92% claimed the treatment worked more rapidly than other malaria medicine.

Widespread deployment of suppositories in this population would seem likely to

reinforce positive perceptions of rectal administration, as familiarity and experience

with their rapid onset of action increased.

In the present setting, parental shame and embarrassment do not appear major

impediments to the implementation of suppositories as standard treatment for

malaria. In contrast, caregivers were more concerned with the more practical

aspects of treatment such as whether the suppository would be well absorbed or

cause side-effects, their ability to administer the suppository properly themselves

and difficulties caused by the child struggling or defaecating.

A moderate proportion of caregivers in this study refused treatment and the reasons

for their refusal are instructive. As well as a lack of prior knowledge or experience of

rectal administration, poor educational status was a significant association.

Caregivers who had never been to school were twice as likely to refuse as those

with some schooling. Caregivers most commonly cited lack of spousal approval as

well as concerns about safety and the practical challenges of administering to a

reluctant child as reasons for refusal. These issues should therefore be addressed

in health education activities accompanying deployment. The most important group

196

for health education messages to reach, the poorly educated and illiterate, would

also be the most difficult to reach effectively. The issue of spousal approval was

important and in this context the final decision-making appears to rest with the

father despite mothers having a primary role in accessing health-care for their child.

Health education activities in this context should clearly target fathers as well as

mothers.

A variety of strategies might be used to deploy suppositories either at primary

health-care, village or household level. Amongst these are the possibility that

caregivers themselves are provided with suppositories to either initiate treatment of

their sick child or at least to administer some of the doses themselves. Seventy

percent of caregivers who participated agreed to attempt to self-administer

suppositories of whom the majority (94%) were able to do so on their first attempt.

This is encouraging for the possibility of future deployment at a household level and

shows that, even without any medical training, suppositories are easy to use.

However, it was notable that fathers were much more reluctant to administer

suppositories themselves. Any household deployment strategy in PNG should aim

to utilise mothers to administer suppositories in this context.

Satisfaction with suppository treatment when assessed following administration of a

three-day treatment course was generally very good with the majority rating the

treatment as safe and rapidly effective and few raising concerns about hygiene. The

most common problems experienced were the practical difficulties of forcibly

administering treatment to a struggling child. However, almost half the caregivers

also indicated that they found it easier to administer suppositories than oral

medicine. Oral administration of antimalarial drugs can be difficult in very young

children and probably contributes to the poor compliance with even short-course

therapy in this group. [206] This may explain why, in other settings, suppositories

have proved more popular than oral formulations amongst both caregivers and

health workers for the treatment of malaria in young children (E Christophel, WHO,

personal communication). Whilst rectal administration may improve compliance, it

might also lead to excessive and inappropriate use of suppositories beyond their

suggested indications and therefore increase the theoretical risk of artemisinin

parasite resistance. Nevertheless, in terms of future deployment strategies for

197

suppositories in PNG, it was reassuring that the majority of caregivers said they

would rather treat with suppositories in preference to other routes of administration,

whether given at home by themselves or in the health centre by a health-worker.

This study has demonstrated that, in PNG, overall acceptability of rectal

administration is relatively high at least in the two sites studied. It has also indicated

that acceptance is closely linked with familiarity and therefore should improve as

use becomes more widespread. However, deployment will be more effective if

accompanied by appropriate health education which addresses concerns that

suppositories may not be well absorbed and provides practical advice and

reassurance regarding administration to a struggling child who may defaecate soon

after administration. It should be delivered in a manner suitable for the illiterate

rural poor and should be directed at fathers as well as mothers. Because

suppositories may be easier to administer than oral medication in young children,

excessive and inappropriate use could result. Although there should be careful

consideration of the possibility and implications of this, community-based

interventions using artemisinin-derived suppositories are likely to be feasible in

PNG if combined with appropriate health education activities.

198

199


3.6.1. Safety and tolerability of rectal artesunate

• Based on data accrued from Clinical trials 3, 4 and 5, the safety profile of

rectal ARTS appears acceptable. Late fever spikes described in Clinical trial

3 and constipation and abdominal distension in Clinical trial 4 were possible

drug side effects. However the latter also occurred in the i.m. ARM group,

suggesting a possible class effect of the artemisinin derivatives in this

population. All these adverse effects appeared to be minor, self-limiting and

to resolve without specific treatment.

• Two deaths in Clinical trials 4 and 5 were not thought to be adverse drug

events, the first (Clinical trial 4) being consistent with the prognosis of

severe malaria and the second (Clinical trial 5) suggestive of a

meningoencephalitic infection.

3.6.2. Bioavailability and pharmacokinetic considerations

• Absorption following rectal ARTS administration generally achieved high

levels of drug within 3 h of administration. However high inter-individual

variability was demonstrated in Clinical trial 3. It was suspected some

children may have passed suppositories without the knowledge of the

investigators but this cannot be verified and therefore the absence of

detectable drug in plasma following dosing in 7 patients was of concern.

This high inter-individual variability is consistent with data relating to all

rectally administered artemisinin derivatives (see Appendix C).

• Almost all of the total 176 patients treated in Clinical trials 3, 4 and 5 made

an acceptable parasitological and clinical recovery. The exceptions were

one child from Clinical trial 4 who died within 2 h of initial dosing (i.e. too

early to implicate pharmacological failure) and one child from Clinical trial 5

who developed signs of severity within the first 24 h and in whom death was

ascribed to meningoencephalitis. Therefore, any sub-optimal therapeutic

effect in a subset of “poor-absorbers” of rectal ARTS is unlikely to be clinical

importance in a significant proportion of patients. Clinical trial 4

200

demonstrated that the problem of variably low bioavailability and sub-optimal

therapeutic response may be of greater significance for i.m. ARM, a

commonly used injectable artemisinin therapy for severe malaria.

3.6.3. Role of �-thalassaemia in modulating pharmacokinetics and clinical response to rectally administered artesunate

• No association between �-thalassaemia genotype and either PK parameters

or clinical outcomes was found in Clinical trial 3 to support in vitro and in

vivo data from Thailand implicating possible impaired response to

artemisinin in �-thalassaemia [285-289, 350]. This genetic mutation is

almost ubiquitous in coastal PNG [42, 43, 366] so the good clinical response

in almost all of the 176 rectal ARTS-treated patients from the three Clinical

trials suggests little clinically significant effect of this mutation on clinical

outcomes for individual patients in this population.

3.6.4. Association between GPC∆∆∆∆ex3 and cerebral malaria

• An association between severe malaria with impaired conscious state and

carriage of the GPC∆ex3 mutation in Clinical trial 4a suggest a possible

protective effect specifically against cerebral malaria. Because of

methodological constraints, these results should be viewed as preliminary

and hypothesis-generating at this stage.

3.6.5. Therapeutic efficacy and equivalence of rectal artesunate to other treatments for severe malaria

• Median PC%24s were >99% in all three Clinical trials of rectal ARTS.

Significantly faster parasite clearance seen than i.m. ARM in Clinical trial 4b

is likely to reflect PK differences and to represent a genuine therapeutic

advantage of i.r. ARTS over i.m. ARM.

3.6.6. Operational feasibility of rectal artesunate as a public health intervention

• Acceptance of rectal administration of antimalarials in rural PNG appears to

be high, based on the results of Clinical trial 5. This study also suggested

that increasing familiarity with rectal administration may be associated with

increased acceptance.

201

4. PRESUMPTIVE TREATMENT OF MALARIA IN PREGNANCY

202

203

4.1. Overview of Clinical trials in this chapter The pharmacokinetic, safety and efficacy studies included in this chapter were part

of a single Clinical trial conducted in 60 Papua New Guinean women of child-

bearing age. This included 30 pregnant women and a matched control group of 30

non-pregnant women. The study design was chosen to examine the effects of the

physiological changes of pregnancy and to address the following research

questions:

1. How do the physiological changes of pregnancy effect the PK disposition of

SDOX, PYR, CQ and its primary metabolite, DECQ?

2. How well tolerated is the combination of CQ and SP in women of child-bearing

age? In particular, what are the effects of this drug combination on

haemodynamic stability, glycaemic control and Hb concentration, and how does

pregnancy impact on these parameters?

3. How effective is the combination of CQ and SP in clearing/preventing peripheral

parasitaemia when used as intermittent presumptive treatment in semi-immune

pregnant women in PNG?

4. How can knowledge of changes in PK disposition of the two drugs be used to

devise more efficacious dosing regimens for use in pregnancy?

204

205

4.2. Clinical trial 6: Pharmacokinetics, safety and efficacy of chloroquine and sulphadoxine-pyrimethamine for presumptive treatment of malaria in pregnancy

4.2.1. SUMMARY

Aims: To determine the effect of pregnancy on the pharmacokinetic disposition of

sulphadoxine (SDOX), pyrimethamine (PYR), chloroquine (CQ) and their

metabolites when co-administered for presumptive antimalarial treatment in

Melanesian women and to examine the efficacy, safety and tolerability of this

combination in pregnancy.

Methods: Thirty women in the second or third trimester of pregnancy and thirty

age-matched non-pregnant women were administered a single dose of 1500 mg

sulphadoxine-pyrimethamine (SP: approximately 28/1.4 mg/kg) and CQ 450 mg

base daily (approximately 8.5 mg/kg/day) for three days. Blood was taken at

baseline, 1, 2, 4, 6, 12, 18, 24, 30, 48 and 72 h and at 7, 10, 14, 28 and 42 days

after the first treatment dose in women from both groups. Plasma was subsequently

assayed for CQ, desethylchloroquine (DECQ), SDOX, N-acetylsulphadoxine

(NASDOX), and PYR by HPLC. Compartmental modeling was performed in each

subject for each primary drug to determine Vss/F, CL/F and t½e. Non-compartmental

analysis was used to determine AUC, Cmax and t½e for each drug and active

metabolite. Derived pharmacokinetic parameters were compared between pregnant

and non-pregnant women using non-parametric methods. Microscopy of Giemsa-

stained peripheral blood samples was performed on samples from baseline, 1, 2, 3,

7, 14, 28 and 42 days. Semi-quantitative plasmodial species-specific PCR was also

performed on samples from day 0, 7, 14, 28 and 42 as a more sensitive alternative

to microscopy. Treatment failure was defined according to WHO definitions of

adequate clinical and parasitological response (ACPR). Haemoglobin

concentrations, erect and supine blood pressures and pulse were measured at

baseline, 1, 2, 3, 7, 14, 28 and 42 days. Capillary blood glucose was determined at

baseline, 1, 2 and 3 days.

Results: Post-treatment concentrations of CQ, DECQ, SDOX and PYR were all

significantly lower in pregnant, than non-pregnant women. A two-compartment

206

model best-described the CQ data and showed that, compared with non-pregnant

controls, pregnant subjects had similar V/F (median 180 vs 156 L/kg in non-

pregnant group: P= 0.5) but significantly more rapid CL/F, (15.3 vs 10.8 mL/min/kg:

P=0.04) shorter t½e�2 (196 vs 236 h: P=0.03) and lower AUC0-� for both CQ (34

000 vs 56 000 �g.h/L: P<0.01) and DECQ (25 vs 47 �g.h/L: P<0.01). A one-

compartment model best-described the disposition of SDOX showing significantly

larger V/F (0.24 vs 0.21 L/kg: P<0.01), more rapid CL/F (0.022 vs 0.016 mL/min/kg:

P<0.01), shorter t½e (134 vs 161 h: P=0.03) and lower AUC0-� (22 000 vs 34 000

mg.h/L: P<0.01) in pregnant subjects. An acceptable compartmental model could

not be fitted to the PYR data. However, non-compartmental analysis showed

pregnant subjects to have significantly lower AUC0-� (75 000 vs 113 000 �g.h/L:

P<0.001). Median t½e was 226 h in pregnant and 214 h in non-pregnant subjects.

Five of 13 pregnant and 2 of 7 non-pregnant women with P. falciparum at baseline

experienced late parasitological failure within 28 days. Emergent infections due to

P. malariae, P. ovale and P. vivax were also detected by PCR during the follow-up

period in both pregnant and non-pregnant groups. Minor reductions in blood

pressure and heart rate occurred with nadirs occurring around the third day of

treatment.

Conclusions: Because lower plasma concentrations of CQ, DECQ, SDOX and

PYR could compromise both curative efficacy and post-treatment prophylactic

properties in pregnant patients, IPTp regimens should incorporate higher mg/kg

doses than that recommended for non-pregnant patients.

207

4.2.2. AIMS

4.2.2.1. Primary

i. To determine the effect of pregnancy on the PK disposition of CQ and its active

metabolite, DECQ. The primary PK parameters of interest would be V/F, CL/F,

AUC, Cmax and t½e.

ii. To determine the effect of pregnancy on the PK disposition of SDOX and PYR.

The primary PK parameters of interest would be V/F, CL/F, AUC, Cmax and t½e.

4.2.2.2. Secondary

i. To provide preliminary data on the efficacy of the combination of CQ and SP as

IPTp in Papua New Guinean women. Efficacy would be determined according to

WHO definitions of 28-day ACPR and also by the reappearance of any peripheral

parasitaemia (by microscopy or PCR) between day 7 and day 42.

ii. To examine possible clinical or sub-clinical toxicity of CQ plus SP administered to

pregnant women, with particular regard to:

a. Changes in erect and supine blood pressure and pulse.

b. Hb concentration

c. Capillary blood glucose concentrations.

4.2.3. METHODS



Irwin Law (School of Medicine and Pharmacology, UWA) provided much of the

clinical supervision of this study at Alexishafen Health Centre. Clinical examinations

and procedures were performed by Jovitha Lammey, Wesley Sikuma, Maria Goretti

and Servina Gommorai (all of PNG IMR, Madang) and Irwin Law. Nandau

Tarongka, Lena Lorry and Kaye Baea (PNG IMR, Madang) performed most of the

microscopy.

Madhu Page-Sharp (School of Medicine and Pharmacology, UWA) performed the

SDOX, NASDOX, PYR, CQ and DECQ assays under the supervision of Ken Ilett.

208

Pete Zimmerman and colleagues (Case Western Reserve University, Cleveland,

Ohio, USA) performed the quantitative species-specific PCR assays (by ligase

detection reaction – fluorescent microsphere assay: LDR FMA).

I conceived and designed the study, assisted in its clinical supervision and

performed the present data analysis (including all of the PK modeling) and

interpretation.


The study design chosen was a single-arm treatment cohort that included the

patient group of interest (pregnant women) and a matched control group of non-

pregnant women. The latter were carefully matched by age and place of residence

with pregnant cases in order to minimize the potential confounding effects of geo-

epidemiological, genetic, nutritional and other variables.

It was expected that approximately 30% of women in both groups would have

asexual peripheral parasitaemia on blood slide at the time of enrolment [367] (with

higher rates by PCR) and that the presence or absence of peripheral parasitaemia

(along with other covariates such as age, weight, parity and pregnancy status)

could be evaluated as an independent predictor of each PK parameter. Assessing

correlations between gestational age and each PK parameter would also be

feasible in the pregnant group.

It was considered that a 30% difference in the magnitude of any PK parameter

between pregnant and non-pregnant groups would be of clinical significance.

Therefore, the study was designed using power calculations that aimed to give a

high chance of detecting 30% between-group differences in the major four PK

parameters (Vd, CL, AUC and t½e) for all three drugs (CQ, SDOX and PYR).

Therefore, 12 separate sample size calculations were performed, with the aim of

choosing a final sample size that exceeded each of these estimates. These

calculations assumed that pregnancy would result in an increase in Vd and CL and

a reduction in AUC and t½e, and used estimates of centrality and variance of PK

parameters from previous studies of SP and CQ in non-pregnant adults [300, 368]

Based on data from Edstein et al. [368], for SDOX and PYR, estimates of per-group

209

sample size required to detect a 30% difference in mean PK parameters varied

between 4 and 12 (using � of 0.05 and � of 0.2) and between 7 and 21 (using � of

0.05 and � of 0.1). However, because of higher inter-individual variability in PK

parameters for CQ, [300] per-group sample size estimates were larger for this drug,

ranging between 7 (AUC), 14 (CL/F), 22 (t½e) and 25 (Vd) (using � of 0.05 and � of

0.2). Therefore, we decided on a sample size of 30 per group (assuming that up to

15% of patients would be lost to follow up) as providing a reasonable chance (by

exceeding sample size requirements based on � of 0.05 and � of 0.2) of detecting

clinically-significant differences in all PK parameters of interest for all three drugs.

The power of the study to detect differences in some parameters for CQ (t½e and

Vd) would be relatively less (approximately 80%) than for SDOX and PYR (>90%).


The study was conducted at Alexishafen Health Centre, Madang Province, over 6

months between February and July 2006. The local population is almost

exclusively of Melanesian race and P. falciparum is considered to be hyper-

endemic. [31] In this population, severe malaria is rare in adults and febrile illness

due to malaria in women of childbearing age is uncommon. Approximately 29% of

all women attending for a first antenatal clinic have peripheral blood P. falciparum

parasitaemia by microscopy and 42% have splenomegaly. [367] Rates of

parasitaemia and splenomegaly in non-pregnant women in the community are

similar to those of antenatal clinic attendees. [367]


4.2.3.4.1. Eligibility criteria and enrolment procedures

Women attending the antenatal clinic for the first time during their current

pregnancy were eligible for the trial provided that i) they had no features of severe

malaria [24] ii) there was no history of use of one of the study drugs in the previous

14 days, iii) there was no clinical evidence of any significant co-morbidity, iv) follow-

up was feasible and v) informed consent was obtained. Women were eligible

regardless of whether baseline microscopy demonstrated parasitaemia with any of

the four species of malaria (including mixed infections) or not.

210

After written consent had been obtained, an initial detailed clinical assessment was

performed, including estimation of gestational age by fundal height and supine and

erect blood pressure. A 3 mL blood sample was taken and baseline tests including

Hb (HaemoCue®, Angelholm, Sweden) and venous blood glucose (HaemoCue®,

Angelholm, Sweden) were performed, plasma frozen and retained for later drug

assays and the cell pellet saved for later parasite genotyping. The diagnosis was

subsequently confirmed by a skilled microscopist.

Non-pregnant controls were women of child-bearing age who were enrolled during

visits to nearby villages. To ensure that cases and controls were matched as closely

as possible with respect to ethnicity, villages in which previously enrolled pregnant

cases resided were chosen. Following an explanation of the study and its

procedures, volunteers were sought. Volunteers whose ages best matched

previously enrolled controls were given preference for enrolment in the study.

Eligibility, enrolment, treatment and follow-up procedures were identical for

pregnant cases and non-pregnant controls.

4.2.3.4.2. Drug administration, dosage and dosing times

All women received three tablets of CQ (Chloroquin®, Astra, Sydney, Australia, 450

mg CQ base) daily for three days (0, 24 and 48 h: total dose of CQ base 1350 mg)

and a single dose of SP (Fansidar®, Roche, Basel, Switzerland, 1500 mg SDOX

and 75 mg PYR) at enrolment, in line with PNG standard treatment

recommendations. [369] Doses were not adjusted for weight. Drugs were

administered as a combination of whole tablets and swallowed whole with water. All

doses were administered under direct supervision, with the exact time of

administration recorded. Food intake at the time of drug administration was not

controlled. CQ and SP tablets were from GMP compliant manufacturers.

4.2.3.4.3. Blood sampling

All women had a heparinised intravenous cannula inserted at the time of enrolment.

In addition to a baseline sample, following administration of the first dose of CQ

(450mg CQ base administered at enrolment and again at 24 and 48 h), 3 mL

samples were drawn through the cannula at approximately 1, 2, 4, 6, 12, 18, 24, 30,

48 and 72 h after administration of the first dose. Further 3 mL samples were taken

211

by venesection at 7, 10, 14, 28 and 42 days. The sampling protocol was identical in

pregnant women and non-pregnant controls. The exact time of each sample

collection was recorded.

4.2.3.4.4. Clinical monitoring

All patients were admitted to the health centre and observed as inpatients for at

least 48 h to enable supervised administration of all drug doses, intensive blood

sampling and clinical monitoring. All women had a daily assessment which included

a standardised symptom checklist (including dizziness), measurement of axillary

temperature, respiratory rate and plasma glucose. A Giemsa-stained thick blood

film was also taken at these times. Erect and supine blood pressure and pulse

were taken 2 min apart at baseline, 1, 2, 3, 7, 14 and 28 days. Postural hypotension

was defined as >20mm fall in systolic blood pressure or a >10mm fall in diastolic

pressure from supine to erect. [370] The mean arterial pressure (MAP) was

calculated by adding 1/3 of the pulse pressure (systolic minus diastolic pressure) to

the diastolic pressure. All women were given an insecticide-treated bed net (ITN)

with instructions as to how to use it at the time of discharge and were followed up

as outpatients on day 7, 10, 14, 28 and 42. At these times axillary temperature and

Hb were recorded and a Giemsa-stained thick blood film taken.


4.2.3.5.1. Microscopy , haemoglobin and glucose

Giemsa-stained thick blood smears were examined by skilled microscopists who

were blinded to the subject’s pregnancy status. The microscopist viewed >100 fields

at 1,000x magnification before a slide was considered negative. Parasite density

was determined by counting asexual parasites per 1000 leucocytes and assuming a

total leucocyte count of 8,000/�L. All slides were read independently by two

microscopists who were blinded to each others’ results. Any slide discrepant for

positivity/negativity or speciation was referred to a third microscopist. If there was

concordance between at least two of the three reads (with respect to slide positivity

or speciation), the final result would be based on the two concordant slides and the

parasite density determined by geometric mean of these two readings. If none of

the three reads were concordant, a fourth read would be performed. This process

would be continued with additional readings until at least two readings were

212

concordant. For positive slides, parasite density was calculated from the number of

asexual forms/1,000 leucocytes and the whole blood leucocyte count.


Sample collection and processing, storage, extraction and HPLC methodology for

CQ and DECQ were identical to those in clinical studies 1 and 2 and are described

in Appendix 1, section 7.12. Assay methodology for SDOX and PYR is described in

Appendix 1, section 7.15.

4.2.3.5.3. Parasite PCR assays

Given the likely poor sensitivity of microscopic examination of peripheral blood

smears in semi-immune pregnant women, quantitative species-specific PCR assays

were also performed to detect parasitaemia below the microscopic threshold. The

ligase detection reaction – fluorescent microsphere assay (LDR FMA) was used.

This novel technology is described in detail elsewhere. [371]

4.2.3.6. Pharmacodynamic outcomes

Treatment response was defined according to current WHO definitions of adequate

clinical and parasitological response (ACPR) as described in Section 2.3.3.6. [113]

WHO-defined ACPR was assessed at both day 28 and day 42. Genotyping to

distinguish re-infection from recrudescence was not performed. However, because

WHO microscopy-based definitions may under-diagnose recurrent infections in

pregnancy (due to peripheral parasitaemia below the microscopic threshold in semi-

immune women) additional semi-quantitative species-specific PCR was also

performed by LDR FMA on blood from each follow-up visit, including day 28 and

day 42, to detect possible sub-microscopic recurrent infections.

4.2.3.7. Pharmacokinetic outcomes

Chloroquine and SDOX concentrations were analysed by a two stage approach.

One-, 2- and 3-compartment models with a lag time and using weighting of 1/y2

were fitted to individual patient concentration-time data using Topfit 2.0 . [290]

Discrimination between models was decided on the basis of the overall model

goodness of fit criterion (B-value), visual inspection of the distribution of residuals,

and the Akaike Information Criterion value. Parameters of interest derived from the

model included half lives for absorption (t½abs), distribution (t½e�1) and elimination

213

(t½e�2), volume of distribution at steady-state relative to bioavailability (Vss/F),

clearance relative to bioavailability (CL/F), and areas under the concentration-time

curve (AUC) to 42 days (AUC0-42) and infinity (AUC0-�). For CQ, DECQ, SDOX,

NASDOX and PYR, AUC0-42 (log-linear trapezoidal rule), AUC42-� (C42/terminal

elimination rate constant), AUC0-� (AUC0-42 + AUC42-�) and t½e (log-linear

regression of last 3-4 data pairs) were estimated by non-compartmental analysis

using Topfit 2.0 . [291]

In order to allow comparison with previous PK studies of CQ, SDOX and PYR, all

drug concentrations and AUC calculations are expressed using weight/volume

units. In another study described in this thesis (Clinical trial 1-see Section 2.2) CQ

and DECQ AUCs are given in molar units (µmol*h/L). Therefore, for CQ and DECQ

both weight/volume and molar units are also given (the latter in parentheses) to

allow cross-reference with this previous study. A molecular weight of 319.87 for CQ

and 291.87 for DECQ was used for all conversions.

4.2.3.8. Data analysis

Student’s t-test or the Mann-Whitney U test were used to compare mean and

median values for admission characteristics and PD outcomes between pregnant

cases and non-pregnant controls. Only non-parametric methods (Mann-Whitney U

test) were used to compare PK parameters between pregnant and non-pregnant

groups. Categorical data were compared using either Pearson Chi squared or

Fisher’s exact test where appropriate. A two-tailed level of significance of P<0.05

was used for all tests (SPSS v 9.0, Chicago, IL, USA). Comparison of changes in

continuous variables (blood pressure, heart-rate, Hb and glucose) at each time-

point during follow-up were performed using one-way ANOVA. Univariate analyses

were conducted to determine potential covariates (from age, weight, height,

pregnancy status, presence versus absence of parasitaemia at baseline and parity)

for each PK parameter (t½e, t½e�2, Vss/F, CL/F and AUC0-42) with each drug (CQ,

SDOX and PYR). This was then used to determine the three most appropriate

variables for use in multiple linear regression analysis. Backwards-stepwise entry of

variables was then used to determine independent predictors of the major PK

parameters.

214

4.2.4. RESULTS


The baseline demographic, anthropometric and clinical features of the 30 pregnant

subjects and the 30 non-pregnant controls are described and compared in Table

4.1. The two groups were well-matched for age, weight, height and the presence or

absence of detectable parasitaemia by microscopy (see Table 4.1A). However, P.

ovale and mixed infections were detected more commonly in pregnant subjects by

PCR. Although close to half the subjects in each group had detectable parasitaemia

(usually with P. falciparum). No subject had a documented fever (temperature

>37.5°C) at the time of enrolment. Gametocytaemia was not detected in any subject

at baseline.

Twenty-three pregnant subjects were enrolled during the second trimester (range:

14-27 weeks). The remaining 7 patients were enrolled during the third trimester

(range: 28-31 weeks). The majority of non-pregnant controls were nulliparous (see

Table 4.1B).

Pregnant subjects had significantly higher pulse and lower Hb and glucose

concentrations consistent with known physiologic changes in pregnancy (see Table

4.1C). [372]

4.2.4.2. Efficacy outcomes

Treatment efficacy outcomes according to conventional WHO clinical and

microscopy-based criteria (28-day in vivo adequate clinical and parasitological

response) are presented in Table 4.2A for the 20 subjects with P. falciparum

detected by microscopy at baseline. Of the 4 subjects with P. vivax at baseline (2

pregnant, 2 non-pregnant) none re-developed P. vivax parasitaemia during 28 days

of follow-up and were therefore defined as ACPR. The two subjects with P.

malariae at baseline (both pregnant) both remained aparasitaemic over 28 days by

microscopy.

Results of microscopy and species-specific PCR at day 28 and 42 are shown in

Table 4.2B. Semi-quantitative species-specific PCR results over the 42-day period

of post-treatment monitoring are shown in more detail in Figure 4.1, specifically, two

215

separate primers for P. falciparum (A: Pf 72, B: Pf 73), and primer’s for P. vivax (C),

P. malariae (D) and P. ovale (E) in both pregnant and non-pregnant groups.

The emergence of new and alternative species infections was commonly detected

by both microscopy and PCR in both the pregnant and non-pregnant groups, mostly

at day 42 (see Table 4.2B and Figure 4.1).

216

Table 4.1 Baseline variables of 30 pregnant and 30 non-pregnant Papua New

Guinean women enrolled in Clinical trial 6. All data are mean ± SD, median

(interquartile range) or number (%).

A. Anthropometric variables and malaria status.

Pregnant (n=30) Non-pregnant (n=30)

P-value for difference

Age (years) 26.0 ± 5.9 25.5 ± 8.9 0.8

Weight (kg) 54.0 ± 6.4 51.8 ± 5.5 0.16

Height (cm) 151 ± 5 149 ± 5 0.17

Temperature (°C) 36.3 ± 0.5 36.2 ± 0.7 0.6

P. falciparum parasitaemia

Microscopy

PCR

13 (43)

12 (40)

7 (23)

8 (27)

0.1

0.3

P. vivax parasitaemia

Microscopy

PCR

2 (7)

9 (30)

2 (7)

7 (26)

1

0.6

P. malariae parasitaemia

Microscopy

PCR

2 (6)

4 (13)

0

1 (3)

0.5

0.2*

P. ovale parasitaemia

Microscopy

PCR

0

5 (17)

0

0 (0)

1

0.03*

Mixed parasitaemia

Microscopy

PCR

1

7 (26)

0

1 (3)

0.3

0.03*

Any asexual parasitaemia

Microscopy

PCR

16 (53)

18 (60)

9 (27)

15 (50)

0.09

0.5

* Fisher’s exact test (1-tailed)

217

B. Pregnancy status.

Pregnant (n=30)

Non-pregnant (n=30)


Estimated gestational age at enrolment (weeks)

22 (20-28) - -

Fundal height (cm) 24 (21-26) - -

Gravidity 3 (1-5) 0 (0-2) <0.001

Parity 2 (0-4) 0 (0-2) 0.03

C. Baseline haemodynamic variables, haemoglobin and glucose.

Pregnant (n=30)

Non-pregnant (n=30)


Respiratory rate (/min) 23.3 ± 5.3 22.9 ± 4.1 0.7

Heart rate supine (/min) 87.7 ± 11.7 74.9 ± 8.8 <0.001

Heart rate erect (/min) 99.4 ± 12.6 84.1 ± 9.3 <0.001

Systolic blood pressure supine (mm Hg)

98 ± 8 100 ± 13 0.5

Diastolic blood pressure supine (mm Hg)

60 ± 8 62 ± 9 0.5

Systolic blood pressure erect (mm Hg)

100 ± 11 104 ± 11 0.4

Diastolic blood pressure erect (mm Hg)

62 ± 12 67 ± 11 0.1

Haemoglobin (g/L) 80 ± 13 107 ± 19 <0.001

Glucose (mmol) 4.8 ± 2.0 5.8 ± 1.5 0.02

218

Table 4.2 Efficacy outcomes. Data are number (%).

A. Conventional WHO-defined criteria (based on clinical and microscopy follow-up

to 28 days) for 20 Papua New Guinean women with microscopy-proven P.

falciparum at baseline.

Pregnant (n=13) Non-pregnant (n=7) All (n=20)

Early treatment failure 0 0 0

Late treatment failure-clinical

0 0 0

Late treatment failure-parasitological

5 (38) 2 (28) 7 (35)

Adequate clinical and

parasitological response

8 (62)

5 (71)

13 (65)

B. Microscopy and parasite PCR-positivity at day 28 and 42 for all w. Data are

number (%).

Pregnant Non-pregnant All

Micro PCR Micro PCR Micro PCR

Day 28

P. falciparum

P. vivax

P. ovale

P. malariae

Any

(n=23)

4 (17)

0

0

0

4 (17)

(n= 22)

1* (5)

0

1 (5)

2 (9)

4 (18)

(n=25)

3 (12)

1 (4)

0

0

0

(n=24)

1 (4)

0

1 (4)

2 (8)

4 (17)

(n=48)

7 (14)

1 (2)

0

0

4 (8)

(n=46)

2 (4)

0

2 (4)

4 (9)

8 (17)

Day 42

P. falciparum

P. vivax

P. ovale

P. malariae

Any

(n=25)

2 (8)

1(4)

0

0

3 (12)

(n=22)

3 (14)

3 (14)

1 (3)

3 (14)

8 (37)

(n=21)

5 (20)

0

0

0

5 (20)

(n= 21)

1 (5)

1 (5)

1 (5)

3 (14)

5 (23)

(n=46)

7 (15)

1 (2)

0

0

8 (17)

(n=43)

4 (9)

4 (9)

2 (5)

6 (14)

13 (30)

* This patient had P. falciparum gametocytes (but not asexual parasites) detected

by microscopy at this follow-up visit.

219

Figure. 4.1 Semi-quantitative LDR-FMA results. The X axes show time in days from

commencement of treatment, Y axes the PCR signal intensity (log scale). Pf 72 pregnant women

Time (days)0 7 14 21 28 35 42

Sig

nal i

nten

sity

100

1000

10000

Pf 72 non-pregnant women

Time (days)0 7 14 21 28 35 42

Sig

nal i

nten

sity

100

1000

10000

Pf 73 pregnant women

Time (days)0 7 14 21 28 35 42

Sig

nal i

nten

sity

100

1000

10000

Pf 73 non-pregnant women

Time (days)0 7 14 21 28 35 42

Sig

nal i

nten

sity

100

1000

10000

Pv pregnant women

Time (days)0 7 14 21 28 35 42

Sig

nal i

nten

sity

100

1000

10000

Pv non-pregnant women

Time (days)0 7 14 21 28 35 42

Sig

nal i

nten

sity

100

1000

10000

Pm pregnant women

Time (days)0 7 14 21 28 35 42

Sig

nal i

nten

sity

100

1000

10000

Pm non-pregnant women

Time (days)0 7 14 21 28 35 42

Sig

nal i

nten

sity

100

1000

10000

Po pregnant women

Time (days)0 7 14 21 28 35 42

Sig

nal i

nten

sity

100

1000

10000

Po non-pregnant women

Time (days)0 7 14 21 28 35 42

Sig

nal i

nten

sity

100

1000

10000

A

B

C

D

E

220


4.2.4.3.1. Blood pressure and heart-rate

Mean erect and supine blood pressures recorded at each follow-up over the 14

days following initiation of treatment are shown in Figure 4.2 (systolic and diastolic)

and Figure 4.3 (mean arterial pressure: MAP). Mean erect and supine heart rate are

shown in Figure 4.4. Falls in MAP were observed in both pregnant subjects and

non-pregnant controls with a nadir generally occurring on the third day (i.e. at the

time of the final treatment dose of CQ: see Figure 4.2). Similarly, mean erect resting

heart rate also fell (although the nadir was later at 72 h). However, only some of the

changes in these parameters reached statistical significance by ANOVA over the 14

days of follow-up. These included erect heart rate in pregnant women (P=0.015 by

ANOVA) and erect systolic, diastolic and mean arterial pressure in non-pregnant

women (P �0.002 by ANOVA). Changes in the remaining parameters (supine heart

rate and supine and erect blood pressures in pregnant women, supine and erect

heart rate and supine blood pressures in non-pregnant women) did not reach

statistical significance. Further details are given below for each group.

a. Pregnant subjects

The observed mean fall in MAP was greatest in the supine position (mean decrease

in MAP of 5 +/- 10mm Hg at 48 h from baseline: See Figure 4.2) and appeared to

be predominantly due to a fall in diastolic blood pressure (mean fall 5mm Hg).

However, changes in these parameters over time were not statistically significant

over the entire sampling period (P=0.3 by ANOVA). At baseline, the number of

subjects with a recorded MAP (either erect or supine) <70mm Hg was 8 (27%). At

48 h, this had risen to 11 (37%). Two subjects (7%) had a recorded MAP (either

erect or supine) <60mm Hg at baseline. At 48 h, this rose to 6 (20%). Three

subjects met the criteria for postural hypotension (postural fall >20mmHg systolic or

>10mmHg diastolic) recorded at days 2, 7 and 14, respectively for each subject.

Dizziness was not recorded on the side-effect questionnaire more commonly during

follow-up visit than at the baseline assessment.

The nadir in erect mean heart rate occurred at day 3 and was 9/min less than at

baseline (P=0.16). The observed change in erect heart rate was greatest between

221

day 3 and day 7 (P=0.049 by ANOVA– see Figure 4.4). Changes in supine heart-

rate were not significant (P=0.6 by ANOVA).

b. Non-pregnant subjects

In contrast to pregnant subjects, non-pregnant women had a statistically significant

fall in all erect blood pressures (9 mmHg fall in MAP between day 0 and day 2:

P=0.007 by ANOVA). This was predominantly due to a much greater fall in systolic

blood pressure (12 mm Hg: P=0.001 by ANOVA). At baseline, the number of

subjects with a recorded MAP (either erect or supine) <70mm Hg was 6 (20%). By

48 h, this had risen to 16 (53%). No subjects had a recorded MAP <60mm Hg at

baseline but 2 (7%) were below this cut-point at 48 h. Blood pressure recordings

meeting the criteria for postural hypotension were made on 22 occasions in 18

subjects over the 14 day follow up period (two recordings at baseline, 5 at day 1, 5

at day 2, 4 at day 3, 2 at day 7 and 4 at day 14). However, there was no significant

increase in questionnaire-reported symptoms of dizziness over the follow-up period.

Changes in heart rate over time were not significant but there was a trend towards a

fall in erect heart rate (P=0.09 by ANOVA).

4.2.4.3.2. Haemoglobin

Haemoglobin concentrations over the 42 days following initiation of treatment are

shown in Figure 4.5. In both pregnant and non-pregnant groups, a small (mean 5-7

g/L, respectively) fall in Hb concentration occurred with a nadir occurring at 2-3

days followed by recovery by day 7. In both pregnant and non-pregnant subjects

haemoglobin concentrations continued to rise such that the mean values at day 42

were slightly higher (by a mean of 4 g/L in both groups) than baseline. However,

changes in Hb with time were not statistically significant by ANOVA (P>0.4).

4.2.4.3.3. Blood glucose

Blood glucose concentrations measured in the 4-days following treatment are

shown in Figure 4.6. Changes over time were not statistically significant by ANOVA

(P>0.4). One pregnant subject had a measured glucose concentration of 2.7

mmol/L at day 2 and 2.2 mmol/L at day 3 (compared with 3.1 mmol/L at baseline).

She was not symptomatic. None of the other 59 subjects had hypoglycaemia

(glucose <2.5 mmol/L) during the 72 h following initiation of treatment.

222

Figure 4.2 Mean blood pressure following treatment with chloroquine (CQ) and sulphadoxine-pyrimethamine (SP) in Papua New

Guinean women. Dosage and timing is indicated by the arrows. The mean supine blood pressure is on the left of each pair of vertical

bars and the erect on the right. The upper cross-bar indicates systolic and the lower diastolic blood pressure.

A Pregnant subjects

Time (d)

0 2 4 6 8 10 12 14

Blo

od p

ress

ure

(mm

Hg)

50

60

70

80

90

100

110

B Non-pregnant controls

Time (d)

0 2 4 6 8 10 12 14B

lood

pre

ssur

e (m

m H

g)

50

60

70

80

90

100

110

CQ 8.7 mg/kg x 3

SP 28/1.4 mg/kg x1

CQ 9.1 mg/kg x 3

SP 29/1.5 mg/kg x1

223

Figure 4.3 Mean ± SD mean-arterial pressure (MAP) following treatment with chloroquine (CQ) and sulphadoxine-pyrimethamine

(SP) in Papua New Guinean women. Dosage and timing is indicated by the arrows. The mean supine MAP is indicated by the black

line and the erect by the grey line. A significant difference for the difference in MAP compared with baseline is denoted by * (P<0.05

by ANOVA) and † (P<0.01).

A Pregnant subjects

Time (d)

0 2 4 6 8 10 12 14

Mea

n ar

teria

l blo

od p

ress

ure

(mm

Hg)

50

60

70

80

90


Time (d)

0 2 4 6 8 10 12 14

Mea

n ar

teria

l blo

od p

ress

ure

(mm

Hg)

50

60

70

80

90

CQ 8.7 mg/kg x 3

SP 28/1.4 mg/kg x1

CQ 9.1 mg/kg x 3

SP 29/1.5 mg/kg x1

* † **

224

Figure 4.4 Mean ± SD heart-rate following treatment with chloroquine (CQ) and sulphadoxine-pyrimethamine (SP) in Papua New

Guinean women. Dosage and timing is indicated by the arrows. The mean supine heart-rate is indicated by the black line and the

erect by the grey line.

A Pregnant subjects

Time (d)

0 2 4 6 8 10 12 14

Hea

rt-ra

te (/

min

)

60

80

100

120


Time (d)

0 2 4 6 8 10 12 14

Hea

rt-ra

te (/

min

)

60

80

100

120

CQ 8.7 mg/kg x 3

SP 28/1.4 mg/kg x1

CQ 9.1 mg/kg x 3

SP 29/1.5 mg/kg x1

225

Figure 4.5 Mean ± SD haemoglobin concentrations in pregnant and non-pregnant

Papua New Guinean women administered chloroquine (CQ) and sulphadoxine-

pyrimethamine (SP) over the 42 days following initiation of treatment.

Time (d)

0 10 20 30 40

Hae

mog

lobi

n co

ncen

tratio

n (g

/L)

40

60

80

100

120

140Pregnant women Non-pregnant women

CQ 9 mg/kg x 3

SP 28/1.4 mg/kg x1

226

Figure 4.6 Mean ± SD glucose concentrations in pregnant and non-pregnant Papua

New Guinean women administered chloroquine (CQ) and sulphadoxine-

pyrimethamine (SP) over the 42 days following initiation of treatment.

Time (d)

0.0 0.5 1.0 1.5 2.0 2.5 3.0

Glu

cose

con

cent

ratio

n (m

mol

/L)

3

4

5

6

7

8

9

Pregnant women Non-pregnant women

CQ 9 mg/kg x 3

SP 28/1.4 mg/kg x1

227

4.2.4.4. Pharmacokinetics of chloroquine and desethylchloroquine

4.2.4.4.1. Drug concentrations

Plasma concentrations of CQ are shown in Figure 4.7 for each subject in the

pregnant (A) and non-pregnant (B) groups. Median CQ concentrations for both

groups are compared directly in Figure 4.8. These were significantly higher in non-

pregnant, relative to pregnant, subjects throughout the post-treatment phase. At the

54-h time-point (approximating the time of peak concentrations following the final

treatment dose) median CQ levels were 1.4x higher in the non-pregnant group (298

�g/L) than the pregnant group (214 µg/L: P=0.024 by Mann-Whitney U). However,

these differences became more pronounced throughout the post-treatment phase

with a relative difference of 1.6 at day 7 (91 vs 57 µg/L: P<0.001), 1.9 at day 14 (33

vs 18 µg/L: P<0.001), 2.1 at day 28 (12 vs 5 µg/L: P<0.001) and 2.7 at day 42 (7 vs

3 µg/L: P<0.001). Overall, these differences contributed to a 35% lower median

AUC0-42 in the pregnant group (see Table 4.4).

In order to demonstrate the possible effect of pregnancy status on time-above-MIC

for CQ, a hypothetical MIC of 30 µg/L is also depicted in Figure 4.8. This figure is

based on a range of 3-32 µg quoted by Gustaffson et al., [300] with the higher end

of this range chosen to assume some degree of CQ resistance. Figure 4.8

demonstrates that, for this hypothetical MIC, the time above MIC would be

approximately 10 days in pregnant women versus approximately 17 days in non-

pregnant women. It can be deduced from the nature of the curves in Figure 4.8 that

in areas of high level CQ resistance (i.e. an in vivo MIC >30µg/L) the relative

difference in time-above-MIC between pregnant and non-pregnant subjects would

be less, but where there are more sensitive parasites (in vivo MIC <30µg/L) this

difference would be greater.

Plasma concentrations of DECQ are shown in Figure 4.9 for each subject in the

pregnant (A) and non-pregnant (B) groups and median values for both groups

compared directly in Figure 4.10. Differences between the two groups were similar

to those observed with CQ, with a 50% lower median AUC0-42 in the pregnant group

(see Table 4.4).

228

Time (h)

0 200 400 600 800 1000 1200

Con

cent

ratio

n ( µµ µµ

g/L)

1

10

100

1000

Time (h)

0 200 400 600 800 1000

Con

cent

ratio

n ( µµ µµ

g/L)

1

10

100

1000

Figure 4.7. Plot of measured (dots and grey lines) plasma chloroquine

concentrations against time in Papua New Guinean women administered 455mg

chloroquine base daily for 3 days. The solid line represents the sample median

concentrations.

A. Pregnant subjects; B. Non-pregnant subjects

A

B

229

In the pregnant group, there were 11 undetectable values at the final (day 42)

sampling point and one undetectable value at day 28. These values are not shown.

Time (h)

0 200 400 600 800 1000 1200

Con

cent

ratio

n ( µµ µµ

g/L)

1

10

100

1000

Pregnant womenNon-pregnant women

Figure 4.8. Comparison of median [interquartile range] plasma chloroquine

concentrations in pregnant (black line: n=30) and non-pregnant (grey line: n=30)

groups). Asterixes indicate P-value <0.001 for difference at that time-point (Mann-

Whitney U). The horizontal dashed line represents a hypothetical parasite in vivo

MIC of 30 µg/L for chloroquine. The arrows show the putative time-above MIC for

pregnant and non-pregnant women, assuming this hypothetical MIC.

*

*

*

*

*

230

Time (h)

0 200 400 600 800 1000

Con

cent

ratio

n ( µµ µµ

g/L)

1

10

100

1000

Time (h)

0 200 400 600 800 1000

Con

cent

ratio

n ( µµ µµ

g/L)

1

10

100

1000

Figure 4.9. Plot of measured (dots and grey lines) and plasma desethylchloroquine

concentrations against time Papua New Guinean women administered 455mg

chloroquine base daily for 3 days. The solid line represents the sample median

concentrations.


Four undetectable values at the final (42-day) sampling point (3 in pregnant, 1 in

non-pregnant subjects) are not shown.

A

B

231

Time (h)

0 200 400 600 800 1000 1200

DE

CQ

con

cent

ratio

n ( µµ µµ

g/L)

1

10

100

1000

Pregnant Non-pregnant

Figure 4.10. Comparison of median plasma desethylchloroquine concentrations in

pregnant (black: n=30) and non-pregnant (grey: n=30) groups). Asterisks indicate

P-value <0.001 for difference at that time-point (Mann-Whitney U).

4.2.4.4.2. Pharmacokinetic analysis

Of the 60 subjects originally enrolled in the study, 16 (10 pregnant and 6 non-

pregnant subjects) were excluded from the PK analysis due to incomplete data

(defined as having available drug concentration data from less than 3 of the final 4

time-points). A further 6 (0 pregnant, 6 non-pregnant) were excluded due to the

presence of CQ in the baseline plasma sample (indicating prior treatment) and 3 (2

pregnant, 1 non-pregnant) due to having inadvertently received additional doses of

CQ during the follow-up period. This left 18 pregnant and 17 non-pregnant subjects

* *

*

*

*

*

232

who were included in the PK analysis. When the excluded subjects were compared

with those included, no significant differences were found for baseline variables

including age, weight, height, presence or absence of parasitaemia, gestational

age, gravidity, parity, heart-rate, respiratory rate, blood pressure, Hb or glucose

concentration.

Using the criteria outlined previously in the methods, a 2-compartment model using

weighting of 1/y2 provided the best fit to the CQ data. Following application of the

model to each individual dataset, median PK values were derived for pregnant and

non-pregnant groups, respectively. Simulated model-derived concentration-time

curves, using these values, are shown in relationship to the raw CQ concentration-

time data in Figure 4.11 for both pregnant (A) and non-pregnant (B) subjects.

Median values for PK parameters derived from compartmental analysis of CQ are

compared for pregnant and non-pregnant subjects in Table 4.3. Median values for

non-compartmental analysis of both CQ and DECQ in pregnant and non-pregnant

subjects are compared in Table 4.4. To enable comparison with previous studies

[373, 374] AUC for the first 3 days/ 72 h (AUC0-3) was calculated in order to

determine the ratio of DECQ to CQ during the early treatment phase. This showed

median DECQ:CQ ratios of 0.48 [range: 0.27-0.88 ] in pregnant and 0.56 [0.28-

0.73] in non-pregnant women (P=0.2 for difference between pregnant and non-

pregnant by Mann-Whitney U). At 42 days (AUC0-42) the ratio was 0.68 in pregnant

versus 0.87 in non-pregnant women (P=0.4).

The model-derived parameters from the pregnant group can be used to simulate

concentration-time profiles with other hypothetical dosage regimens in an effort to

produce a profile similar to that of non-pregnant women (especially with regard to

the likely time-above MIC). Figure 4.12 shows a simulation based on a 50%

increase in each of three daily doses. Although this simulation results in a longer

likely time-above-MIC, peak concentrations are approximately 30% higher than

those seen with the non-pregnant model at conventional doses.

233

Time (h)

0 200 400 600 800 1000 1200

Con

cent

ratio

n (C

Q o

nly:

µµ µµg/

L)

1

10

100

1000

Time (h)

0 200 400 600 800 1000 1200

Con

cent

ratio

n (C

Q o

nly:

µµ µµg/

L)

1

10

100

1000

Figure 4.11. Simulated concentration-time curves based on median model-derived

pharmacokinetic parameters for plasma chloroquine (solid black line). Dots and

grey lines show actual concentration-time data for each individual patient.

A. Pregnant subjects; B. Non-pregnant subjects.

In the pregnant group, there were 11 undetectable values at the final (day 42)

sampling point and one undetectable value at day 28. These values are not shown.

A

B

234

Table 4.3 Compartmentally derived pharmacokinetic parameters for chloroquine in

pregnant versus non-pregnant subjects. Data are median [interquartile range].

Median values using alternative units are also given in parentheses.

Parameter Pregnant Non-pregnant P-value for difference (Mann Whitney-U)

t½e�1 (h) 15 [10-22] 12 [9-22] 0.77

t½e�2 (h) 196 [167-235]

(8.2 d)

236 [214-299]

(9.9 d)

0.03

t½abs (h) 0.26 [0.05-0.67] 0.54 [0.05-4.0] 0.41

V/F (L) 180 [136-261] 153 [134-199] 0.49

CL/F (mL/min/kg)

15.3 [11.2-21.7]

(0.92 L/h/kg)

10.8 [9.7-14.8]

(0.65L/h/kg)

0.04

235

Table 4.4 Non-compartmentally derived pharmacokinetic parameters for

chloroquine and desethylchloroquine in pregnant versus non-pregnant subjects.

Data are median (interquartile range). Median values using alternative units are

also given in parentheses.

Parameter Pregnant Non-pregnant P-value for difference (Mann Whitney-U)

Cmax (µg/L)

CQ

DECQ

303 [240-453]

124 [92-230]

404 [276-486]

234 [150-346]

0.2

0.06

t½e (h)

CQ

DECQ

203 [170-278]

(8.5 d)

283 [270-434]

(11.8 d)

267 [243-349]

(11.1 d)

306 [254-478]

(12.8 d)

0.06

0.28

AUC0-3 (µg*h/L)

CQ

DECQ

11 800 [9 000-17 300]

7 100 [3 300-9 100]

18 300 [13 600-19 600]

9 900 [6 100-11 000]

0.03

0.06

AUC0-42 (µg*h/L)

CQ

DECQ

31 800 [19 900-42 000]

(99.4 µmol*h/L)

21 500 [12 600-30 800]

(73.6 µmol*h/L)

49 000 [40 800 65 400]

(153 µmol*h/L)

42 800 [32 000-56 500]

(147 µmol*h/L)

0.002

0.001

AUC0-� (�g*h/L)

CQ

DECQ

31 800 [19 900-42 700]

(99.4 µmol*h/L)

28 000 [15 900-34 100]

(95.9 µmol*h/L)

50 400 [42 900-69 700]

(157 µmol*h/L)

46 600 [35 300-63 000]

(160 µmol*h/L)

0.001

0.001

236

Time (h)

0 200 400 600 800 1000

Con

cent

ratio

n ( µµ µµ

g/L)

1

10

100

1000

Non-pregnant standard dose 9mg/kg daily for 3 daysPregnant standard dose 8.5mg/kg daily for 3 daysSimulation of 12.8mg/kg daily for 3 days in pregnant women

Figure 4.12. Simulated plasma concentration-time curve for chloroquine using

model-derived parameters and based on 50% dosage increase in pregnant women

(dashed line). Modeled curves in pregnant (black line) and non-pregnant (grey line)

using conventional doses are shown for comparison.

237

Following univariate analysis to determine the three most useful putative covariates

(from age, pregnancy status, presence or absence of parasitaemia at baseline,

height, weight and parity), linear regression analysis of the total sample (including

pregnant and non-pregnant women) identified pregnancy as predicting lower AUC0-

42 (P=0.003, t=-3.9), and higher parity as predicting higher CL/F (P=0.014, t=2.6). In

the pregnant group, no correlations were found between gestational age and any of

the major PK endpoints (V/F, CL/F, t½e, AUC0-42).

4.2.4.5. Pharmacokinetics of sulphadoxine and N-acetylsulphadoxine


Plasma concentrations of SDOX are shown in Figure 4.13 for each subject in the

pregnant (A) and non-pregnant (B) groups. Molar concentrations of the primary

metabolite, NASDOX, averaged between 1-7% those of the parent drug at each

time-point (median NASDOX concentrations are also represented in Figure 4.13).

Median SDOX concentrations for both groups are compared directly in Figure 4.14.

These were significantly higher in non-pregnant, compared with pregnant subjects

throughout the post-treatment phase. At the 6-h time-point (approximating the time

of peak concentrations following the final treatment dose) median SDOX

concentrations were 1.2x higher in the non-pregnant group (142 mg/L) than the

pregnant group (122 mg/L: P=0.001 by Mann-Whitney U). However, these

differences became more pronounced throughout the post-treatment phase with a

relative difference of 1.5 at day 7 (68 vs 45 mg/L: P<0.001), 2.1 at day 14 (34 vs 16

mg/L: P<0.001), 2.8 at day 28 (9 vs 3 mg/L: P=0.003) and 2.8 at day 42 (3 vs 1

mg/L: P=0.004). Overall, these differences contributed to median AUC0-42 that was

35% lower for SDOX (and 47% lower for NASDOX) in the pregnant group (see

Table 4.6).


for SDOX, a hypothetical MIC of 30 mg/L is also depicted in Figure 4.14. Defining

the real in vivo MIC is complex, particularly as in vivo activity is heavily dependent

on synergy with PYR. However, a figure of 30mg/L has been chosen based on a

breakpoint used by Barnes et al. [159] Figure 4.14 demonstrates that the time

above MIC would be approximately 10 days in pregnant women versus

approximately 17 days in non-pregnant women. It can be deduced from the nature

238

of the curves in Figure 4.14 that this relationship would remain similar throughout a

wide range of parasite MICs.

Time (h)

0 200 400 600 800 1000 1200

Con

cent

ratio

n (m

g/L)

0.1

1

10

100

1000

Time (h)

0 200 400 600 800 1000 1200

Con

cntr

atio

n (m

g/L)

0.1

1

10

100

1000

Figure 4.13. Plot of measured plasma sulphadoxine (dots and solid grey lines)

concentrations against time (multiple zero values at day 28 and 42 not shown) in

Papua New Guinean women administered a single dose of 1500mg sulphadoxine.

The solid line represents the sample median sulphadoxine concentrations the

dashed line the sample median N-acetylsulphadoxine concentrations.


A

B

239

Time (h)

0 200 400 600 800 1000 1200

Con

cent

ratio

ns (m

g/L)

0.1

1

10

100

1000

PregnantNon-pregnant

Figure 4.14. Comparison of median [interquartile range] plasma sulphadoxine


groups). Symbols † and *indicate P-values of <0.01 or <0.001 (Mann-Whitney U),

respectively, for difference between the two groups at that time-point. The

horizontal dashed line represents a hypothetical parasite in vivo MIC of 30mg/L for

sulphadoxine. The arrows show the putative time-above MIC for pregnant and non-

pregnant women, assuming this hypothetical MIC.


Of the 60 subjects originally enrolled in the study, 8 (4 pregnant and 4 non-pregnant

subjects) were excluded from the pharmacokinetic analysis due to incomplete data


time-points). One additional pregnant subject was excluded due to finding of SDOX

in the baseline plasma sample (indicating prior treatment). This left 25 pregnant and

*

**

†

†

*

240

26 non-pregnant subjects who were included in the PK analysis. When the

excluded subjects were compared with those included, no significant differences

were found for baseline variables including age, weight, height, presence or

absence of parasitaemia, gestational age, gravidity, parity, heart-rate, respiratory

rate, blood pressure, Hb or glucose concentration.

Using the criteria outlined previously in the methods, consistent with previous

published studies [159, 375], a 1-compartment model using weighting of 1/y2

provided the best fit to the SDOX data. Following application of the model to each

individual dataset, median PK values were derived for pregnant and non-pregnant

groups, respectively. Simulated model-derived concentration-time curves, using

these values, are shown in relationship to the raw SDOX concentration-time data in

Figure 4.15 for both pregnant (A) and non-pregnant (B) subjects.

Median values for PK parameters derived from compartmental analysis of SDOX

are compared for pregnant and non-pregnant subjects in Table 4.5 and those by

non-compartmental analysis in Table 4.6.

The model-derived parameters from the pregnant group were used to simulate

concentration-time profiles with other hypothetical dosage regimens in an effort to

improve the profile to one comparable to that in non-pregnant women (especially

with regard to the likely time-above MIC). Figure 4.16 shows simulations based on a

33% increase in dose administered as a single dose (A) or a 60% increase in total

dose administered in three split doses over three days (B).

241

Time (h)

0 200 400 600 800 1000 1200

Con

cent

ratio

n (m

g/L)

0.1

1

10

100

1000

Time (h)

0 200 400 600 800 1000 1200

Con

cent

ratio

n (m

g/L

0.1

1

10

100

1000

Figure 4.15. Simulated plasma concentration-time curves based on median model-

derived pharmacokinetic parameters for sulphadoxine (solid black line). Dots and

grey lines show actual concentration-time data for each individual patient.


A

242

Table 4.5 Compartmentally derived pharmacokinetic parameters for sulphadoxine

in pregnant versus non-pregnant subjects. Data are mean ± SD or median

[interquartile range]. Alternative units are also given in parentheses.

Parameter Pregnant

(n=25)

Non-pregnant

(n=26)

P-value for difference (Mann Whitney-U)

Dose (mg/kg) 28.1 ±3.3 29.3 ± 3.1 0.18

t½e (h) 134 [83-235]

(5.6 d)

161 [64-207]

(6.7 d)

0.03

t½abs (h) 0.8 [0.05-15] 0.7 [0.05-4.6] 0.36

V/F (L/kg) 0.24 [0.17-0.41] 0.21 [0.15-0.29] 0.002

CL/F (mL/min/kg)

0.022

[0.014-0.044]

0.016

[0.010-0.031]

<0.001


sulphadoxine in pregnant versus non-pregnant subjects. Data are median

[interquartile range].

Parameter Pregnant (n=25)

Non-pregnant (n=26)

P-value for difference (Mann

Whitney-U)

Cmax (mg/L) 136 [111-146] 158 [143-188] <0.001

t½e (h) 146 [131-172] (6.1 d)

167 [144-196] (6.7 d)

0.2

AUC0-42 (mg*h/L) SDOX NASDOX

22 100 [19 400-

26 500] 800 [500-

10000]

34 100 [26 000-

39 400] 1 500 [1 400-1

800]

0.001

0.001

AUC0-� (mg*h/L) SDOX NASDOX

22 100 [19 400-

26 500]

900 [500-100]

35 200 [26 200-

40 000]

1700 [1500-2000]

<0.001

0.001

243

Time (hours)

0 200 400 600 800 1000 1200

Con

cent

ratio

n (m

g/L)

1

10

100

1000

Non-pregnant subjects 1500mg (29.4mg/kg)Pregnant subjects 1500mg (27.53mg/kg)Simulation of 2000mg in pregnancy (36.7mg/kg)

Time (h)0 200 400 600 800 1000 1200

Con

cent

ratio

n (m

g/L)

1

10

100

1000

Non-pregnant subjects 1500mg (29.4mg/kg stat)Pregnant subjects 1500mg (27.53mg/kg stat)Simulation of 2400mg in pregnancy (15mg/kg daily x 3)

Figure 4.16. Simulated plasma concentration-time curves for sulphadoxine using

model-derived parameters. The dashed lines shows simulated curves based on a

33% dosage increase administered as a single dose in pregnant women (A) and

60% total dose increase administered as 3 split doses over 48 h (B). Modeled

curves in pregnant (black line) and non-pregnant (grey line) using conventional

doses are shown for comparison.

A

B

244



height, weight and parity), linear regression analysis of the entire sample (including

pregnant and non-pregnant women) identified pregnancy as predicting higher Vss/F

(P=0.001, t=3.6) and higher CL/F (P<0.001, t=4.1). More advanced age was also

predictive of higher Vss/F (P=0.028, t=2.3). In the pregnant group, no correlations

were found between gestational age and any of the major PK endpoints (Vss/F,

CL/F, �1, AUC0-42).

4.2.4.6. Pharmacokinetics of pyrimethamine


Plasma concentrations of PYR are shown in Figure 4.17 for each subject in the

pregnant (A) and non-pregnant (B) groups. Median PYR concentrations for both

groups are compared directly in Figure 4.18. These were higher in non-pregnant,

compared with pregnant, subjects throughout the post-treatment phase. At the 4-h

time-point (approximating the time of peak concentrations following the treatment

dose) median PYR levels were 1.4x higher in the non-pregnant group (655 µg/L)

than the pregnant group (463 µg/L: P<0.001 by Mann-Whitney U). However, these

differences became more pronounced throughout the post-treatment phase with a

relative difference of 1.9 at day 10 (145 vs 75 µg/L: P<0.001), 2.2 at day 14 (88 vs

41 µg/L: P<0.001), 1.9 at day 28 (29 vs 15 µg/L: P=0.003) and 1.4 at day 42 (10 vs

7 µg/L: P=0.1). Overall, these differences contributed to a 37% lower median AUC0-

42 in the pregnant group (see Table 4.7).


for PYR, a hypothetical MIC of 40 µg/L for PYR is also depicted in Figure 4.18.

Defining the real in vivo MIC is complex, particularly as in vivo activity is heavily

dependent on synergy with SDOX. However, 40 µg/L has been chosen based on a

breakpoint used by Barnes et al. [159] Figure 4.18 demonstrates that, the time

above this MIC would be approximately 14 days in pregnant women versus

approximately 24 days in non-pregnant women. It can be deduced from the nature

of the curves in Figure 4.18. that this difference would be less when parasite

susceptibility to PYR is less.

245

Time (hours)

0 200 400 600 800 1000 1200

Con

cent

ratio

n ( µµ µµ

g/L)

1

10

100

1000

Time (h)

0 200 400 600 800 1000 1200

Con

cent

ratio

n ( µµ µµ

g/L)

1

10

100

1000

Figure 4.17. Plot of measured plasma pyrimethamine (dots and solid grey lines)

concentrations against time (one zero value in pregnant group at day 42 not shown)

in Papua New Guinean women treated with 225mg of pyrimethamine. The solid line

represents the sample median pyrimethamine concentrations at each time point.


A

B

246

Time (h)

0 200 400 600 800 1000 1200

Con

cent

ratio

n (µ

g/L)

1

10

100

1000

Figure 4.18. Comparison of median [interquartile range] plasma pyrimethamine


groups). Symbols † and * indicate P-values of <0.01 or <0.001 (Mann-Whitney U),

respectively, for difference between the two groups at that time-point. The

horizontal dashed line represents a hypothetical parasite in vivo MIC of 40µg/L for

pyrimethamine. The arrows show the putative time-above MIC for pregnant and

non-pregnant women, assuming this hypothetical MIC.


Of the 60 subjects originally enrolled in the study, 14 (8 pregnant and 6 non-

pregnant subjects) were excluded from the PK analysis due to incomplete data


time-points). Two additional pregnant subjects were excluded due to finding of PYR

in the baseline plasma sample (indicating prior treatment) and one non-pregnant

subject due to having inadvertently received a dose of SP during the follow-up

**

*

†

*

247

period. This left 20 pregnant and 23 non-pregnant subjects who were included in

the PK analysis. When the excluded subjects were compared with those included,

no significant differences were found for baseline variables including age, weight,

height, presence or absence of parasitaemia, gestational age, gravidity, parity,

heart-rate, respiratory rate, blood pressure, Hb or glucose concentration.

Using the criteria outlined previously in the methods, contrary to previous published

studies [159, 375], a 1-compartment model using weighting of 1/y2 provided a poor

fit to the data. A better fit was provided by a 2-compartment model but this also

failed to adequately describe the final data points (day 28 and day 42). The dataset

was considered insufficient to enable a 3-compartment model to be applied in a

statistically robust manner. Therefore, analysis was limited to non-compartmental

methods.

Median values for pharmacokinetic parameters of PYR derived from non-

compartmental analysis are compared for pregnant and non-pregnant subjects in

Table 4.7.


pyrimethamine in pregnant versus non-pregnant subjects. Data are median

[interquartile range]. Median values using alternative units are also given in

parentheses.

Parameter Pregnant

(n=20)

Non-pregnant

(n=23)

P-value for difference (Mann

Whitney U)

Cmax (µg/L) 508 [383-641] 765 [686-883] <0.001

t½e (h) 226 [209-312]

(9.4 days)

214 [185-218]

(8.9 days)

0.03

AUC0-42 (µg*h/L)

69 100 [52 900-81 500]

109 400 [91 200-133 400]

<0.001

AUC0-� (�g*h/L)

74 700 [55 800-86 200]

112 500 [93 300-135 900]

<0.001

248

CL (mL/min/kg)

0.3 [0.27-0.42] 0.21 [0.18-0.25] <0.001

Vd (L/kg) 7.0 [5.4-10.6] 3.8 [3.0-4.8] <0.001



height, weight and parity), linear regression analysis of the entire sample (including

pregnant and non-pregnant women) identified pregnancy as predicting higher Vss/F

(P=0.001, t=-3.6), higher CL/F (P<0.001, t=4.6) and lower AUC0-42 (P<0.001, t=-

5.6). No covariate other than pregnancy was a significant predictor of any of the

major PK endpoints (Vss/F, CL/F, �1, AUC0-42). In the pregnant group, no

correlations were found between gestational age and any of the major PK endpoints

(Vss/F, CL/F, t½e, AUC0-42).

4.2.5. CONCLUSIONS

Prior to this study, a total of less than 150 pregnant women have been enrolled in

published studies of antimalarial pharmacokinetics. [102, 374, 375] Most of these

had not employed study designs enabling comparison with a well-matched non-

pregnant control group. Therefore, the current study of 30 pregnant women and 30

non-pregnant controls represents a substantial contribution to the current body of

knowledge regarding the effects of pregnancy on the disposition of antimalarial

drugs. Because of the combination treatment used, this study had the additional

advantage of being able to simultaneously evaluate three different drugs (CQ,

SDOX and PYR) and their metabolites (DECQ and NASDOX).

This study clearly demonstrated significantly lower concentrations of all three drugs

in pregnant subjects. In general, these were approximately half those

demonstrated in non-pregnant subjects following completion of the treatment

course – the time period likely to be crucial in determining the drug’s post-treatment

suppressive effect. The study therefore demonstrates significant scope for

improving dosing regimens in order to optimise the efficacy of antimalarials when

used as IPTp.

249

4.2.5.1. Pharmacokinetics of chloroquine and desethylchloroquine in pregnancy

The PK parameters derived for both groups were broadly consistent with those of

the 20 Papua New Guinean children enrolled in Clinical trial 1 (see 2.2.4.3) and with

most previous studies of children and adults. All have shown large volumes of

distributions (ranging from 59 to 882 L/kg) [123, 124, 293, 294] consistent with the

values of 180 and 153L/kg in pregnant and non-pregnant groups, respectively. The

multi-exponential elimination profile previously observed in Clinical trial 1 and in

other published studies was also evident in this study and implies that t½e�2

estimation is dependent on the duration of the sampling schedule and on assay

sensitivity. However, the present median t½e�2s of 8.2 and 9.3 days are consistent

with values of 5.8-20.0 days generated from studies using similar assay

methodology and sampling durations. [294, 297-303] Values for CL/F for CQ

calculated in the current study (median 15.3 mL/min/kg and 10.8 mL/min/kg in

pregnant and non-pregnant groups, respectively) were within the range previously

reported (range 2.2-18.3 mL/min/kg) [294, 297-303] and similar to those seen in

children from Clinical trial 1 (13.3 mL/min/kg).

Although CQ has been used extensively in pregnancy for many decades only three

small studies have examined its PK disposition in pregnancy. [373, 374, 376] Two

of these [373, 376] employed a very limited sampling schedule that precluded

estimation of most PK parameters. A study by Massele et al. analysed whole-blood

concentrations of chloroquine in 5 blood samples taken in the 26th and then again

in the 36th weeks of gestation in women receiving weekly CQ prophylaxis. [376]

Non-compartmental methods were used to calculate AUC and CL/F (dose divided

by AUC) at each of these two gestational time-points (26 and 36 weeks). The study

demonstrated a modest increase in CL (18%) between the 26 and 36 week

evaluations. The magnitude of the difference in median CL between pregnant and

non-pregnant women in the present study (42% higher in pregnancy) was therefore

much greater than the difference between second and third trimesters observed in

the study by Massele et al. This is consistent with observations indicating that the

major pregnancy-related physiological changes likely to impact on drug disposition

become established early in gestation. [377] For instance, 70% of the increase in

cardiac output that occurs throughout pregnancy takes place within the first

250

trimester. [377] This may underlie increases in hepatic clearance becoming

prominent very early in pregnancy. Similarly, in terms of renal clearance, glomerular

filtration rate is thought to increase by 45% over pre-conception levels by as early

as the 9th week of gestation. [377]

A study by Chukwuani et al compared CQ and DECQ concentrations in 5 pregnant

and 5 non-pregnant women over 48 h following a single dose of 600mg of CQ. [373]

The very short sampling schedule meant that calculations of CL/F, V/F and

elimination half-life were not possible and therefore few comparisons can be made

with the present study. However, it was notable for observing markedly higher

concentrations of DECQ in the pregnant group, with ratios of DECQ to CQ AUC0-2

ranging from 0.21 to 0.59 in non-pregnant compared with 2.8-11.1 in pregnant

women, representing a mean 15.1-fold difference in pregnancy and raising

questions regarding toxicity of CQ when used in pregnancy. However, this

difference has not been observed in a subsequent study [374] or the present study.

A study by Lee et al. showed ratios of DECQ:CQ of 0.2 in pregnant versus 0.22 in

non-pregnant women in the first 3 days (AUC0-3). [374] Similarly, the present study

showed no difference in ratios between pregnant and non-pregnant women at 3

days (DECQ:CQ AUC0-3 ratios of 0.48 and 0.56, respectively) or at 42 days

(DECQ:CQ AUC0-42 0.68 in pregnant versus 0.87 in non-pregnant). Given the

magnitude of the difference observed in the study by Chukwuani et al., and that

their findings have not been reproduced in two subsequent studies, the possibility of

a systematic error in data analysis or drug assay should be considered as a

possible explanation.

A study by Lee et al. measured whole blood CQ concentrations to evaluate the PK

of CQ and DECQ in 12 pregnant and 15 non-pregnant Karen women with

symptomatic P. vivax infection over 42 days following 3 days treatment with

25mg/kg CQ base. [374] Although analysis was restricted to non-compartmental

methods, and drug was assayed in whole blood, rather than plasma, its very similar

design enables close comparison with the present study. The results of this study

contrasted markedly with those of the present study, with no significant differences

between pregnant and non-pregnant groups with respect to any PK parameter

evaluated (including t½e, CL/F, AUC0-3 and AUC0-�). Pregnant women enrolled in

251

the study by Lee et al. were comparable to the present study with regard to median

age (25 versus 24.5 years in the present study), total dose administered (25 versus

26 mg/kg), gestational age (27 versus 22 weeks) and parity (median of 2 in both

studies). However, the pregnant women in the study by Lee et al. were lighter

(mean 49.5 versus 54 kg in the present study) with a lower mean BMI (21.9 versus

23.6 kg/m2). In addition, the women in the non-pregnant group were more likely to

be older (median age 29 versus 21 years), lighter (mean weight 46 kg versus 51.8

kg, mean BMI 20.0 versus 23.1 kg/m2) and less likely to be nulliparous (median

parity 1 versus 0). Also, in contrast to the present study, all women had P. vivax

parasitaemia with febrile symptoms and had much higher indices of red cell mass

(mean haematocrit of 31.2% in pregnant and 37.1% in non-pregnant women,

equivalent to Hb concentrations of 104 and 124 g/L, respectively). The prevalence

of asymptomatic parasitaemia in the present group (53% in pregnant, 27% in non-

pregnant by microscopy) and much lower Hb concentrations (80 and 107 g/L)

reflect the very much higher malaria transmission intensity affecting our PNG

women. However, pregnant and non-pregnant groups seemed to be similarly well-

matched in each study (with the possible exception of age in the study by Lee et al.

[374]

It is difficult, therefore, to see how differences in population characteristics of the

two studies could have lead to such conflicting results with regards to the observed

differences between pregnant and non-pregnant groups. However, differences in

the two studies may well have related to Lee et al. assaying whole blood rather than

plasma. Because CQ concentrates in erythrocytes, concentrations measured in

whole blood are usually 2-3x higher in whole blood than in plasma. For this reason

median values for AUC0-� (122 000 and 134 000 µg*h/L in pregnant and non-

pregnant groups) and Cmax (961 and 700 µg/L) calculated by Lee et al. were

approximately 2-3x higher than those from the present study. This may have also

contributed to large absolute differences in the primary PK parameters calculated in

the two studies including non-compartmentally-derived values of clearance (0.21

L/h/kg in pregnant and 0.19 L/h/kg in non-pregnant) by Lee et al. that were 3-4x

lower than those from the present study.

The lack of difference between pregnant and non-pregnant groups in the study by

252

Lee et al., also possibly reflected a type 2 error related to its smaller sample size

and high degree of inter-individual variability in the calculated PK parameters. It is

notable that, despite Lee et al. finding no differences between pregnant and non-

pregnant groups, more advanced gestational age at enrollment was associated with

a lower AUC0-� on univariate analysis.

It is possible that differences between the two studies reflected different rates of

conversion of CQ to DECQ. Lee et al. calculated ratios of DECQ:CQ AUC0-14 of 0.26

and 0.28 for pregnant and non-pregnant women, respectively. This compares with

values of DECQ:CQ AUC0-3 and DECQ:CQ AUC0-42 ranging from 0.48-0.87 in the

present study. Given the two studies sampled different biological fluids (whole blood

versus plasma) it is possible that this related to differences in partitioning of the two

compounds in plasma and erythrocytes. It is also possible that the discrepancy

reflects genetic differences in drug metabolism between the two ethnic groups. It is

notable that compared with the findings of Lee et al., the DECQ:CQ AUC ratios

from the present study are more congruent with those calculated in previous

studies, including those from Thai adults with P. vivax malaria (DECQ:CQ AUC0-28

0.60) [301], Filipino adults with malaria (DECQ:CQ AUC0-28 0.8) [298] and a

previous study of PNG children with malaria (DECQ:CQ AUC0-42 0.71: see Section

2.2, Table 2.3). Also, the values of CL/F for CQ calculated by Lee et al. (3.5 and 3.2

mL/min/kg) are markedly lower than those derived from most studies from a wide

variety of ethno-geographic population groups (with calculated mean/median values

of CL/F ranging from 8.5-15.8 mL/min/kg – see also 2.2.4). [293, 297, 299, 300,

302, 303] Therefore, if genetic factors are to be implicated for the differences

between the findings of Lee et al. and the present study, the results from the Thai-

Burmese patients may represent the less typical scenario.

It is notable that CQ is a substrate for CYP2D6, CYP2C8 and CYP3A4. [378, 379]

Therefore, given the overall genetic heterogeneity of the cytochrome P450 system,

there is scope for significant variation in phenotypic expression both within and

between populations. [379] The high degree of inter and intra-population genotypic

heterogeneity of the Cytochrome P450 enzyme system has recently been

demonstrated in the Alexishafen population itself (from the population in Clinical trial

2: von Ahsen et al., unpublished data). This study also found two novel putative

253

non-functional alleles of the CYP2D6 gene in the Alexishafen population that have

not been described elsewhere and that were absent in another population from

coastal PNG (Kunjingini, East Sepik). However, overall, the majority (approximately

95%) had genotypes associated with either extensive or ultra-metaboliser

phenotype. This may partially explain the high CQ CL/F and high rate of conversion

of CQ to DECQ observed in the present study, when compared to the study by Lee

et al. [374] However, because DECQ has antimalarial activity that may be

equivalent or even greater than the parent compound [142], it remains uncertain

whether or not these observed differences would have clinical implications.

Although the current study demonstrated statistically significant differences

pregnant and non-pregnant groups with respect to t½e�2, CL/F and AUC of CQ, no

statistically significant difference in V/F was demonstrated between the two groups.

It is possible, however, that this represents a type 2 error, given the high degree of

variability of this parameter for CQ and the sample-size constraints of this study

(see 4.3.2.2).

4.2.5.2. Pharmacokinetics of sulphadoxine and N-acetylsulphadoxine in pregnancy

The PK parameters derived for SDOX from the non-pregnant control group in this

study were within the range of values previously reported from studies of adults

including healthy volunteers [368, 380-386], asymptomatic parasitaemic women

[375] and patients with symptomatic malaria [159, 298] that have shown similar

mean/median values for t½abs (ranging from 0.5-0.8 h), t½e (5.6-10.9 days), Vd/F

(0.08-0.37 L/kg) and CL/F (0.006-0.025 mL/min/kg). Studies of children, however,

have shown generally much higher Vd/F (0.19-1.6 L/kg), CL/F (0.023-0.49

mL/min/kg) and shorter t½e (4-8.6 days). [159, 160, 387-390] These differences in

PK disposition of SDOX between adults and children have now been clearly

demonstrated in a large comparative study and have important implications for

dosing in children. [159]

Only one other study has evaluated SDOX in pregnancy. [375] Although its design

was somewhat different, its findings were very similar to the present study with

respect to pregnancy-related differences in SDOX disposition. The study by Green

254

et al. measured concentrations of SDOX and PYR in whole blood and evaluated the

PK of SP in a group of 30 pregnant Kenyan women many of whom, like the present

subjects, had asymptomatic malaria parasitaemia. However, in contrast to the

present study, many subjects were also HIV positive and the non-pregnant

comparator group comprised 10 subjects initially enrolled in the pregnant group who

were subsequently evaluated again in the post-partum period. Nonetheless the

relative differences in PK between pregnant and post-partum subjects were very

similar to those seen between the present pregnant and non-pregnant groups, with

mean CL/F in the post-partum group (0.010 mL/min/kg) being 60% lower than in

pregnant subjects [(0.017 mL/min/kg) and mean AUC0-� being 76% higher in post-

partum subjects (40 1000 vs 22 800 mg*h/L]. This compares with relative

differences of 38% and 59% for CL/F and AUC0-�, respectively, in the present study.

The magnitude of these differences between pregnant and non-pregnant subjects is

moderate, especially when compared with those demonstrated between adults and

children. [159] However, even relatively modest differences may lead to significantly

compromised clinical efficacy, particularly with regards to the duration of post-

treatment suppression when administered as IPT (see Fig 4.14). It would appear,

therefore, that there is significant scope for improving the efficacy of IPTp using SP.

However, due to the effects of higher clearance in pregnancy, increased single-

dose therapy may have only modest effects on the duration of post-treatment

suppression (see Fig 4.16) and multiple dosing schedules are often precluded in

much of the developing world by issues of practicality and compliance.

This study demonstrated a low rate of conversion of SDOX to its primary inactive

metabolite NASDOX, with NASDOX AUC0-�s <5% those of SDOX. There was no

apparent difference in NA SDOX: SDOX ratios between pregnant and non-pregnant

subjects and the kinetics of NASDOX appear to be formation-rate limited. This is

consistent with knowledge that the concentration-time profile for NASDOX is largely

controlled by renal excretion which, in humans, occurs at a rate approximately 10-

fold greater than that for SDOX. [391] Few other studies have documented

NASDOX concentrations. However, one study of Caucasian volunteers also

demonstrated a similarly low rate of conversion (mean NASDOX AUC0-� <2% those

of SDOX). [368] Such a low rate of acetylation of the primary drug, and the

unimodal distribution of the percentage of NASDOX in this population (data not

255

shown), suggests that SDOX may be a substrate for N-acetyl transferase 1, rather

than N-acetyl transferase 2 which would be expected to produce a bimodal

distribution of the percentage of drug metabolized, according to slow and rapid

acetylator phenotypes within the population. [392, 393] The substrate specificity of

SDOX for NAT1 and NAT2 has not been studied in vitro. However, if this were to

confirm SDOX as an NAT1 substrate, it would be unlikely that metabolic clearance

of the drug would vary considerably within or between populations. This is

consistent with the relatively narrow range of values for CL/F described both within

and between studies of SDOX PK. [159, 368, 375, 380, 383-386, 390]

4.2.5.3. Pharmacokinetics of pyrimethamine in pregnant and non-pregnant women

Although we were unable to characterize a satisfactory compartmental PK model

for this dataset, plasma concentrations of PYR were significantly lower in pregnant

women throughout most of the 42-day follow-up period. There were corresponding

significant differences in non-compartmentally derived PK parameters, with the

median calculated AUC0-� being 34% lower in pregnant women. However, the PK

parameters derived in both groups in this study were markedly different from those

described in the literature, whether in adults [159, 298, 368, 375, 380, 381, 383-

386] or children [159, 160, 387, 388, 390] from a wide variety of clinical and ethno-

geographic settings. In particular, the median t½e values of 9.4 and 8.9 days for

pregnant and non-pregnant women, respectively, are higher than any reported in

the literature. They compare with values ranging from 2.9-5.1 days for adults [159,

298, 368, 375, 380, 383-386] and 2.8-4.5 days in children. [159, 160, 387, 390] In

the present study PYR concentrations did not exhibit a log-linear decline throughout

the duration of the 42-day follow-up in the manner seen for SDOX. Instead a slower

elimination phase was demonstrated between 10 and 42 days. Many previous

studies have employed sampling durations of �14 days. [160, 368, 375, 384-386]

These would have been unable to detect a late, longer elimination phase of the type

demonstrated in the present study.

Only two other studies have attempted to measure PYR concentrations out to 42

days. [159, 380] However, both may have had limitations related to assay

sensitivity. In the study by Mansor et al. a relatively small dose (approximately 0.9

256

mg/kg) was administered to healthy volunteers and, using a gas chromatography –

electron capture detection (GC-ECD) system with a LOQ of 10 µg/L that was unable

to detect drug could in plasma beyond 14 days. [380] Similarly, a large population

PK study of 139 children and 168 adults by Barnes et al excluded approximately

38% of data from their analysis because drug concentrations from these samples

were either undetectable or below the 10 µg/L LOQ of their LCMS assay. [159]

Therefore, it seems likely that the much longer t½e in the present study reflects a

combination of the long sampling duration and use of a highly sensitive assay

methodology (LOQ 2.5 µg/L). This situation appears analogous to the manner in

which the t ½e for CQ (and PQ) has become longer as more sensitive assay

methodologies have been employed over time (see 2.2.5). [135, 394] These factors

also probably contribute to the lower non-compartmentally derived median CL/F in

non-pregnant women from the present study (0.21 mL/min/kg) compared with

mean/median values ranging from 0.22-0.59 mL/min/kg in other studies of non-

pregnant adults[159, 368, 375, 380, 385, 386]) and much higher median AUC0-�

(113 000 �g*h/L) than in other studies of adults administered similar doses (34 700-

42 000 �g*h/L). [159, 375]

Only one other study, by Green et al., has evaluated PYR PK in pregnancy. [375] In

contrast to the present study, there were no significant differences in any of the PK

parameters evaluated following application of a 1-compartment model. This may

have related partly to the short sampling duration of this study (10 days) that may

have failed to adequately describe the late terminal elimination phase demonstrated

in the present study. It may also have been related to cross-over design (non-

pregnant controls were in this case a group of women initially evaluated as

pregnant cases who were evaluated again 2-3 months post-partum) and to the

relatively small size of the control group (n=10) that would have increased the

chances of a type 2 statistical error. It is important to note that whilst highly

statistically significant, the pregnancy-related differences in PK parameters

demonstrated in the present study were of a modest magnitude (34% reduction in

PYR AUC0-�). Nonetheless, given the synergistic mechanism of action of the SP

combination, together with pregnancy-related changes in SDOX disposition, these

relatively modest differences in PYR concentrations may combine to significantly

compromise the post-treatment suppressive properties of SP IPTp (see Figure

257

4.19). If higher dose regimens of SDOX are to be employed for SP IPTp, it would

seem sensible for the same ratio of SDOX:PYR be employed so that dose

escalation of the PYR component was of the same magnitude.

A substantial proportion of PYR clearance probably occurs via the renal route (with

urinary excretion accounting for 16-32% of a single oral dose). [395] Therefore, the

large increases in glomerular filtration rate that occur with pregnancy [395] may

underlie the higher CL observed in pregnant women in this study.

Because it is difficult to define an in vivo MIC for PYR, It is not clear whether the

much longer half-life of PYR demonstrated in this study has clinical implications.

However, this study suggests that it is possible that its post-treatment suppressive

properties and therefore its value as a component of IPT, may have been

underestimated up until now. [109]

4.2.5.4. Clinical implications

This study was not designed to assess the efficacy of CQ-SP as IPTp. Doing so

would require use of an appropriately matched comparator group (pregnant women

treated with placebo or an alternative treatment), evaluation of the most robust and

relevant clinical endpoints (birth-weight and neonatal mortality), and therefore, a

huge sample size. However, it seems plausible that peripheral parasitaemia

provides an acceptable surrogate marker of treatment efficacy in this instance.

Given that ITPp’s mechanism of action is thought to relate to prevention of parasite

sequestration within the placenta for a period of time following treatment, the re-

emergence of peripheral parasitaemia (regardless of whether this was

recrudescence or re-infection with a new parasite strain) seems an undesirable

outcome. Indeed, whilst microscopy has relatively poor sensitivity in detecting

placental malaria in highly endemic settings (between 27 and 42% depending on

whether placental microscopy or PCR is used as the gold standard), the presence

of peripheral parasitaemia does appear to be highly specific for microscopically or

PCR-proven placental parasitaemia (specificity 97-100%). [396] Therefore, the high

rate of re-emergent parasitaemia in pregnant women (5 of 13 pregnant women with

P. falciparum at baseline experienced WHO-defined parasitological treatment

failure by 28 days) would appear to indicate a sub-optimal response to the CQ-SP

258

combination.

The use of parasite PCR on peripheral blood in this study probably improved

diagnostic specificity of placental malaria (which is 84% if using placental PCR as

gold standard) without significantly compromising specificity (97%). [396] Therefore,

the use of this additional, more sensitive surrogate marker adds to concerns about

sub-optimal efficacy in the present study. The study was not powered to compare

efficacy in pregnant versus non-pregnant groups. However, re-emergent P.

falciparum parasitaemia also occurred within 28 days in two non-pregnant women,

consistent with observations that the efficacy of the CQ-SP combination is

compromised by parasite resistance in PNG. [59, 280] Nonetheless, the PK data

from this study suggests that there is scope for improving the concentration-time

profile of CQ, SDOX and PYR in order to optimise the duration of post-treatment

suppression when administered as IPTp (see Figures 4.12 and 4.16).

The potential benefits of dose modification are, however, subject to important

limitations. Firstly the realities of outpatient-based drug administration in the

developing world mean that ensuring compliance with lengthy treatment courses is

problematic. Therefore, IPT dosage regimens have been limited to 3 days (or

ideally, even as short as a single treatment dose). This means that for practical

purposes, improving the concentration-time profile of a drug such as CQ is limited

to increasing the amount of each individual dose, rather than the number of doses

or the duration of therapy. Increasing each individual treatment dose is likely to

counteract pregnancy-related increases in Vd. However, it will be less effective in

counteracting pregnancy-related changes in CL/F and t½e. The limiting factors with

using increased individual dosages in pregnancy are likely to be issues of toxicity

and tolerability associated with higher peak drug concentrations. This would appear

to be a particular risk with CQ, which is well known to lead to haemodynamic

instability (hypotension) at high concentrations (e.g. following IV administration).

[295] Given that pregnancy in itself leads to reductions in peripheral vascular

resistance, the risk of haemodynamic compromise appears increased in this group.

The present data suggests subtle perturbations in blood pressure and heart rate

coinciding with the times that CQ concentrations would be expected to be at their

259

highest. However, these only reached statistical significance when examining blood

pressure in non-pregnant women and were not associated with symptoms. Dose

modification might result in higher maximal CQ concentrations than those seen in

non-pregnant subjects following conventional dosing. In this instance, careful

monitoring would be mandatory. Whether or not dose-modification studies of CQ in

pregnancy would be warranted is debatable. The efficacy of this drug is severely

compromised by P. falciparum resistance throughout most of the world. Although it

is currently one of very few drugs considered safe in pregnancy, it is hoped that

other antimalarials will soon be approved for this indication. For these reasons, the

limited potential benefit that might come from dose modification may not justify

exposing pregnant women to the potential risks associated with higher drug doses.

In contrast with CQ, SDOX does not appear to be subject to dose-dependent

toxicity at conventional doses (since its major toxicity issues appear to be

idiosyncratic and extremely rare). Previous studies of children have shown it to be

well-tolerated when administered as an i.m. injections in a manner that produces

peak concentrations approximately 50% higher than those when given orally. [160]

Also, unlike CQ, pregnancy-related differences in PK disposition include changes in

Vd/F which are more readily amenable to absolute dosage increase. Nonetheless,

pregnancy-related changes in CL/F and t½e are more difficult to overcome by dose

modification of short-course IPT. However, administration as 3 daily doses may

represent an acceptable option (See Figure 4.16b), particularly if, for example,

treatment was to be administered in conjunction with another antimalarial agent

(such as CQ, AQ, PQ or azithromycin) that might conventionally be administered as

3 daily doses. In this instance, splitting the SP dose into 3 daily doses would not

increase the total duration of the treatment course nor increase the regimen’s

overall complexity.

The use of quantitative parasite-specific PCR has not been validated as an

alternative to microscopy for the evaluation of therapeutic efficacy of antimalarial

drugs in clinical trials. It is, however, likely to be more sensitive than microscopy

and therefore has theoretical advantages in settings such as coastal PNG where

adult patients have a high level of pre-existing immunity that limits peripheral

parasitaemia to levels close to, or below, the limits of detection by light microscopy.

260

In addition, reliably differentiating species of plasmodia by microscopy (especially P.

vivax and P. ovale) alone can be challenging even in expert hands and particularly

when mixed infections are common. Discordance between different microscopists

is high. PCR is likely to have advantages of providing a more objective result with

regard to speciation, greater sensitivity in detecting individual species of plasmodia

(especially as mixed infections) and overall greater sensitivity in detecting any

parasitaemia. It may, however, be compromised by issues of specificity related to

the effect of persistent .

The results of the PCR diagnostics from this study generally showed a good

concordance with those of microscopy, particularly with regard to the demonstration

of any parasitaemia (regardless of species). However, there was a high degree of

discordance between PCR and microscopy with regard to speciation. In particular,

PCR demonstrated much higher rates of P. ovale, P. malariae and mixed infections

detected (both at baseline and during the follow-up period). The results during the

follow-up suggested that microscopy may have incorrectly speciated emergent P.

ovale and P. malariae parasitaemias as being P. falciparum. Therefore, the rates of

P. falciparum treatment failure may have been overestimated. Most information

regarding the deleterious effects of malaria during pregnancy relate to P.

falciparum. [38] Although the effects of P. vivax have been studied in pregnancy

(and an association with lower birth-weight demonstrated) [397] there are no data

on the effects of P. ovale or P. malariae in pregnancy. [38] It was notable that PCR

but not microscopy suggested that pregnant women may be at higher risk of P.

ovale infections than non-pregnant women (P= 0.03). Therefore, PCR-diagnostics

might have unmasked a phenomenon that has not been previously described,

possibly because of previously microscopic mis-diagnosis in areas such as PNG

where multi-species transmission is prevalent. They suggest that further information

is needed on the effects of these species on pregnancy outcomes and on their

response to conventional antimalarial treatments in pregnancy.

4.2.5.5. Overall conclusions

This study has defined the PK of the two antimalarial drugs most extensively used

in pregnancy. By employing a sampling schedule out to 42 days, it provides insights

related to the disposition of CQ and SP during the time period when post-treatment

261

prophylactic properties are likely to be most important for the effectiveness of IPTp.

All 3 drugs (CQ, SDOX and PYR) and CQ’s active metabolite (DECQ) were present

in measurable concentrations for the full 42 days in most subjects (though levels of

each were significantly lower in the pregnant group). This suggests that each of the

4 compounds may have a role in post-treatment prophylaxis. In particular DECQ

may be just as, if not more important than CQ itself in terms of its post-treatment

prophylactic effect. Also the t½e of PYR was much longer than those previously

reported. This probably reflects shorter sampling durations and relatively low assay

sensitivity in previous studies and helps to explain why a drug that has long been

considered “short-acting”, seems to be very effective when given as IPT. [109]

Whilst the study demonstrated that drug concentrations of all three agents were

consistently lower in the pregnant group, the PK models derived from the data

suggest that this may occur through a combination of different mechanisms

including changes in Vd, CL and t½e. The relative contribution of each mechanism

and the absolute magnitude of pregnancy-related changes in disposition are not

readily predictable. Therefore, studies of this type will be needed for other new

antimalarial agents so that optimal dosage regimens can be rationally designed

specifically for use in pregnant women.

This study explored the potential value of PCR-based diagnostics as an alternative

to microscopy in the challenging setting of high-level multi-species transmission

where mixed infections are common and low-level parasitaemias are the norm in

adult populations. It has demonstrated a means of enabling more detailed

evaluations of the role of P. ovale and P. malariae in pregnancy in PNG.

The overall magnitude of pregnancy-related differences in PK parameters

demonstrated in this study was moderate. However, if the efficacy of IPT relates

predominantly to the time that drug concentrations are maintained above the

parasite MIC, even relatively modest changes in PK disposition could lead to large

differences in time above MIC and therefore the drug’s effectiveness in suppressing

or preventing infections in the post-treatment period (see Figures 4.8, 4.14 and

4.18). If more than one drug is being used simultaneously for additive or synergistic

effects, then these differences are likely to compound. This design employed for

262

this study provides a blueprint for further evaluations of antimalarial drugs in

pregnancy, especially if these are to be considered for IPTp.

263


4.3.1. Pharmacokinetics of chloroquine and desethylchloroquine in pregnancy

• Concentrations of both CQ and its active metabolite DECQ were

significantly lower in pregnant women (median CQ AUC0-� 37% lower than

in non-pregnant women) and Cl/F and t½e�2 were also significantly lower

pregnant women. This suggests scope for improving dosing regimens for

pregnancy but contrasted with data from the only other study comparing CQ

PK in pregnant and non-pregnant women. [374] Also, in contrast to another

previous study [373] no differences between pregnant and non-pregnant

women were observed with respect to the apparent rate of metabolic

conversion of CQ to DECQ.

4.3.2. Pharmacokinetics of sulphadoxine and pyrimethamine in pregnancy

• Significantly lower concentrations of both SDOX and PYR were

demonstrated in pregnant women, with significant differences in CL/F, Vd/F,

t½ e and reductions in median AUC0-� of 35% and 34% for SDOX and PYR,

respectively in pregnant subjects. The study therefore suggested significant

scope for improving dosing regimens in pregnancy. The results were highly

concordant with the one previous study of SP in pregnancy with regard to

SDOX, but not for PYR. [375]

4.3.3. Longer than expected elimination half- life of pyrimethamine

• Median t½e s of 9.4 and 8.9 days for pregnant and non-pregnant women

calculated were higher than any reported in the literature (2.8-5.1 days).

This unexpected finding probably reflected a combination of the long

sampling duration and use of a highly sensitive assay methodology.

4.3.4. Use of species-specific PCR diagnostics in pregnancy – the role of non-falciparum species in pregnant women in PNG

• Microscopy showed an uncorrected treatment failure rate of 35% in the 20

women with P. falciparum at baseline appearing to suggest poor therapeutic

efficacy of the CQ-SP IPTp regimen. However, whilst showing generally

264

good concordance in detection of any parasitaemia, PCR diagnostics

showed significant discrepancy with regard to speciation. A much higher

proportion of parasitaemias both at baseline and during follow-up were

identified as being P. ovale or P. malariae and the proportion identified as P.

falciparum was much less. This suggests that in areas with poly-species

high transmission such as PNG, conventional microscopy-based outcome

measures may overestimate P. falciparum treatment failures. Conversely,

the rate of re-emergent infections due to P. malariae and P. ovale may be

underestimated by microscopy.

265

5. CONCLUSIONS: IMPLICATIONS FOR PUBLIC

HEALTH POLICY AND FURTHER RESEARCH

266

267

The studies described in this thesis have focused on the three broad clinical groups

that carry the greatest morbidity and mortality burden from malaria in PNG: Children

with uncomplicated malaria, children with severe malaria and pregnant women with

placental malaria. At least three of the studies (Clinical trials 2, 4 and 5) have

already directly lead to health policy change in PNG. This chapter discusses their

implications for health policy both in PNG and elsewhere, as well as highlighting the

need for further research arising from their findings.

5.1. Policy implications for the management of uncomplicated malaria in Papua New Guinea

Based on the results of Clinical trial 2, current chloroquine-based first-line treatment

for malaria can no longer be considered effective for either P. falciparum or P. vivax

in PNG. Additionally, although the current second-line treatment ARTS-SP is an

ACT endorsed by the WHO [64] it was inferior to AL for P. falciparum and to DHA-

PQ for P. vivax and therefore cannot be considered for first-line treatment in PNG.

The differential species-specific drug-efficacy of the two new ACTs (AL and DHA-

PQ) demonstrated Clinical trial 2 created a conundrum for health policy-makers in

PNG. AL had superior efficacy (PCR-corrected ACPR of >95% at Day 42 on per

protocol analysis) to DHA-PQ (88%) against P. falciparum but higher rates of re-

emergent P. vivax with AL (>50% at day 28 and almost 70% at Day 42) compared

with DHA-PQ (16% at Day 28 and 31% at Day 42). In PNG microscopy services are

poor, so most treatment for malaria is administered empirically, without knowledge

of the infecting species. Therefore, standard treatments should ideally have

acceptable efficacy against all possible infecting parasite species.

The PNG Ministry of Health reviewed the results of Clinical trial 2 at a meeting

convened by WHO in December 2008. The final decision was to adopt AL as its

new first-line treatment in PNG due to its superior efficacy to DHA-PQ against P.

falciparum. However, this decision represents a compromise in that children treated

with AL who have P. vivax infections are likely to have an inferior outcome (data

from Clinical trial 2 indicating that as many as two thirds might redevelop

parasitaemia within 6 weeks of treatment) than if they were to be treated with DHA-

PQ. It reflects a pragmatic approach that acknowledges that most antimalarial

268

treatment in PNG is likely to continue to be administered on an empirical basis

(rather than based on microscopic speciation and confirmation of the diagnosis).

The choice of AL therefore assumes P. falciparum to be more important than P.

vivax, by way of being responsible for a greater proportion of acute febrile illness in

PNG (>70% of enrolled cases in Clinical trial 2) and its assumed higher intrinsic

pathogenicity and therefore higher risk of complications from inadequate treatment.

However, recent data suggest that the attributable morbidity and mortality from P.

vivax may have been underestimated. [10, 11, 13] Also, it is unclear how good

adherence to AL’s complex 3-day, 6-dose schedule (and need to co-administer with

fatty food) will be and therefore to what extent the efficacy demonstrated in Clinical

trial 2 will translate to effectiveness when deployed in the field where drug

administration cannot always be rigorously supervised. Therefore, compliance and

effectiveness evaluations are planned by the Ministry of Health to evaluate

effectiveness under un-supervised field conditions during the roll-out of AL as the

new first-line therapy. (Personal correspondence Manuel Hetzel, PNG IMR)

The unexpected results of the efficacy evaluation from Clinical trial 2 support the

WHO-recommended strategy of regular monitoring of parasite sensitivity and in vivo

response in each country. [64] It will therefore be important that in PNG, efficacy of

AL is monitored over time by in vitro tests of lumefantrine susceptibility of local

parasite strains and regular in vivo assessments of ACPR following AL’s

introduction to detect any emergence of resistance. Ideally such assessments

should be conducted in parallel with alternative antimalarial drug regimens, such

that if early signs of diminishing efficacy to AL are detected, a proactive decision

can be made to change treatment policy before treatment failure becomes

significant. Such assessments should also evaluate non-falciparum species which

are clearly a major contributor to morbidity and possibly mortality in PNG.

Policy decisions will also need to be made in the context of health-economic

factors. A health-economic analysis, utilizing data from Clinical trial 2 is planned

(Wendy Davis, UWA) to evaluate the relative cost-effectiveness of standard

treatment with either DHA-PQ or AL.

269

5.2. The problem of Plasmodium vivax in Papua New Guinea

Given the differential efficacy of DHA-PQ and AL for P. vivax, the extremely high

rates of emergent asymptomatic P. vivax parasitaemia in Clinical trials 1 and 2

suggest a need for greater understanding of the morbidity and mortality significance

of chronic relapsing P. vivax in such high transmission settings. Recent studies,

including from PNG and West Papua have lead to this being seriously re-

considered. [10, 11, 13] In particular, it is likely that P. vivax contributes to a high

burden of anaemia which is in turn a co-factor for susceptibility to, and outcome

from, other infectious diseases. This would mean that an antimalarial drug’s

capacity to cure P. vivax infections and prevent further infections following

treatment has important implications for the overall attributable morbidity and

mortality burden of malarial disease in a population. Further clinical epidemiological

studies are therefore needed to explore the relationship between P. vivax infection,

anaemia and all-cause morbidity and mortality in PNG. This information could be

used in mathematical models evaluating the impact of improved curative or

suppressive treatments for P. vivax (eg DHA-PQ) and their likely impact in terms of

overall population morbidity and mortality. A currently ongoing clinical surveillance

project, evaluating aetiology of severe childhood illness at Modilon General Hospital

(co-ordinated by Laurens Manning, PhD student, School of Medicine and

Pharmacology, UWA) may help provide additional data on the incidence of severe

manifestations of P. vivax and its contribution to severe anaemia in PNG.

5.3. Need for improved understanding of pharmacokinetics and determinants of therapeutic efficacy of piperaquine in order to guide improved dosing regimens

Given the volume of accrued efficacy data, [131-134, 141, 281, 311, 316-320, 398,

399] acceptable safety profile, [130, 311] convenient dosing schedule and

affordable cost, piperaquine-containing ACTs are probably now the leading

alternative choice to AL for the WHO-endorsed policy of adopting an ACT for first-

line treatment of uncomplicated malaria. However, the unexpectedly poor efficacy of

DHA-PQ for P. falciparum demonstrated in Clinical trial 2 demonstrates a need to

270

clarify the reasons behind this and to identify determinants of efficacy of this

combination. Possible reasons explored in Chapter 2 (see 2.2.5 and 2.3.5) included

PK factors, cross-resistance of local parasite strains to CQ and very high baseline

parasitaemias in PNG children. Confirmatory data is required to support these

findings in order to identify sub-groups at higher risk of treatment failure and to

predict situations where DHA-PQ’s efficacy may be compromised. Pharmacokinetic

analyses or at least measurement of drug concentrations at appropriate times (eg

day 7) [134, 314] should be incorporated into future Phase III/IV studies of

treatment efficacy in order to explore relationships between drug-levels and efficacy

in other settings. Concomitant in vitro assays of parasite drug sensitivity will also

help to explore the role of cross-resistance with other chemically-related drugs in

other regions.

The PK data from Clinical trial 1 suggest that the ability to cure high parasitaemia

infections might be further compromised by PQ’s rapid distribution phase and

consequent rapid, early post-treatment fall in plasma concentrations that may have

resulted in a proportion of DHA-PQ treated children with plasma PQ concentrations

insufficient to clear the parasite residuum (see Figure 1.1). From Clinical trial 2 it

appears that it is a subset of patients with the highest parasitaemias for whom these

issues and the need to optimise therapy become paramount. Clinical outcome was

related to PQ AUC in Clinical trial 1 and to both AUC and day 7 PQ concentrations

in previous studies. [314] Others have suggested dosage modification in children,

or routine co-administration of food with each dose in order to improve

bioavailability and therefore clinical efficacy. [314, 315] However, the PK analysis in

Clinical trial 1 suggests that this may be problematic. Application of the 2-

compartment model developed to describe this data suggests that given the rapid

and extensive distribution of PQ, higher doses may lead to only modest increases in

plasma concentrations during the elimination phase, but to significantly higher peak

concentrations. Therefore, considering the practical constraints of short-course

therapy, increased PQ doses may risk toxicity associated with greater peak

concentrations without significantly improving curative or post-treatment

suppressive effects. Larger-scale PK-PD studies are needed to explore the

association between plasma PQ levels and clinical outcome and dose escalation

271

studies should employ careful safety monitoring given the risks associated with

higher peak drug concentrations.

Because of its long t½e PQ has been advocated as a possible candidate for use in

IPT as the effectiveness of this strategy probably relies on a long post-treatment

prophylactic or suppressive effect. [109] The results of Clinical trial 1 raise concerns

about its use in this context. Also because IPT has its major application in infants

and pregnant women and because drug disposition can vary significantly with age

and in pregnancy, it will be important that the PK of PQ be specifically evaluated in

pregnant women and infants prior to use as part of IPT strategies.

5.4. Place of rectal artesunate in public health policy in Papua New Guinea

The studies evaluating the potential of ARTS suppositories for treating severe

malaria have added to the existing knowledge regarding this pharmaceutical

preparation and mode of administration (publications 1, 2, 5 and 6). The overall

favourable absorption characteristics, clinical efficacy (including equivalence with

existing standard treatments for severe malaria) safety profile and socio-cultural

acceptability suggest it is likely to be a viable and effective strategy for the

treatment of severe paediatric malaria in PNG. Following a national meeting on

antimalarial drug policy convened by the PNG Ministry of Health and WHO in

December 2008, based on evidence accrued in these studies, ARTS suppositories

are to be included in the national pharmacopoeia. New standard treatment

guidelines for PNG will recommend the use of rectal ARTS as emergency first-line

management of severely ill children with suspected or possible severe malaria

presenting to primary health-care facilities. However, no recommendation was

made for pre-emptive treatment to be administered by non-medically trained

persons in the community. Pre-referral treatment in the community has been

shown to be feasibly and effectively administered by non-medially trained persons

in Africa and Asia. [230] However it may be premature to recommend this in PNG at

this time. A significant commitment to operational and health-systems research is

still required to determine feasible and effective strategies by which rectal

artemisinins can be integrated into primary and community health-care systems in

PNG. Refining optimal deployment strategies will require research to determine who

272

to train as drug prescriber/administrators (e.g. mothers, designated “village malaria

workers”, traditional healers etc.), how to train them, and how to maintain an

effectively integrated relationship with the formal health system.

5.5. Further research needs for rectal artemisinins

5.5.1. Safety

Although there is little convincing evidence of neurotoxicity manifesting in human

subjects, [194, 400, 401] animal toxicity data [187-189] still merits careful clinical

surveillance. It is not known whether neurotoxicity could possibly manifest as an

uncommon event in certain vulnerable groups such as children (in whom the

developing CNS may be at greater risk) or specific ethnic groups that have not been

assessed in previous human studies. [194, 400, 401] The absorption-dependent

kinetics of rectal ARTS together with its highly variable bioavailability may also

mean that a subset of patients might experience higher and more sustained drug

concentrations than would ordinarily be seen with conventional routes of

administration, particularly if repeated doses are administered in a child apparently

slow to respond. This would lead to a drug-concentration-time profile more similar to

those associated with neurotoxicity in animal models. [189]

Rare side-effects are difficult to detect in most clinical trials due to sample size

constraints. It is important therefore that when large, Phase IV trials are conducted,

that rigorous prospective safety monitoring is employed, including documentation of

any new neurological sign or symptom. Ideally, children who succumb to an

unexpected death with neurological features (such as the patient in Clinical trial 5)

would have an autopsy and histology of brainstem tissue performed to determine

whether neurodegenerative changes of the kind seen in animal models had

occurred.

5.5.2. Bioavailability and pharmacokinetic considerations

Although overall absorption following rectal ARTS administration generally achieved

high levels of drug within 3 h of administration, the degree of inter-individual

variability demonstrated in Clinical trial 3 and other studies of rectal artemisinins

(see Appendix C) has important implications. A possible relationship between

273

dosage and parasite clearance rates for ARTS suppositories (see Appendix C)

indicate that a drug-concentration-clinical response relationship may exist within the

therapeutic dosing range and, therefore, that there may be further scope for

optimizing therapy. In order to improve the consistency of therapeutic response,

dosing on the higher end of the therapeutic range (eg 20mg/kg rather than the 5-

10mg/kg recommended by WHO [231]) should be employed in future studies.

Together with careful safety monitoring, such studies should measure drug

concentrations to determine the effects of dosage regimens on interindividual

variability in bioavailability and especially whether the lower limits of this range are

improved. Such considerations should also become vital in the evaluation of other

new artemisinin products that become available for rectal administration.

5.5.3. Future comparative trials

Understanding of PK-PD relationships for artemisinin drugs in general is poor.

Therefore, it is difficult to define the optimal drug concentrations that should be

aimed for in the early phase of treatment. Further elucidating such PK-PD

relationships will require analysis of large datasets of artemisinin-treated patients

that include appropriate drug concentration-time data and standardized measures

of clinical efficacy (eg parasite clearance times). However, in the absence of such

information, parenteral ARTS is likely to become the “gold-standard” by which other

severe malaria treatments, including rectal artemisinin preparations, should be

judged. [116] Therefore, the ideal rectal formulation should be considered one that

consistently and rapidly achieves concentrations of drug or active metabolite

comparable to those of i.v. or i.m. ARTS.

There are inadequate PK and efficacy data for alternative rectal artemisinin

preparations including DHA and ARM (see Appendix C). These alternative agents

need to be assessed in comparative head-to-head trials of different formulations.

Ideally, such efficacy assessments would also use the new “gold standard”, i.v.

ARTS as a comparator arm and, for outcome measurements, use the most

important clinical endpoint, mortality. However, using mortality as an endpoint in

such a trial is unlikely to be feasible due to the large sample size that would be

required for adequate statistical power. Markers of early parasite clearance (eg

274

PC%12 and PCT50) and defining the extent and variability of drug absorption

probably represent the most appropriate surrogate markers to use in this instance.

5.5.4. Drug development: Need for a co-formulated ACT suppository

A significant concern with community-based rectal artemisinin use is that

unregulated use by non-medically trained prescribers, may lead to incomplete

treatment. [231] In particular, failure to ensure a definitive combination treatment

including an appropriately effective long-acting second antimalarial could have

significant implications. These might include treatment failure in the individual

patient or the development of artemisinin resistance in parasite populations. A

possible solution might be the development of a co-formulated suppository

containing both ARTS and a second long-half-life drug from another antimalarial

drug class. If the technical challenges of ensuring adequate bioavailability of both

components of the co-formulation could be met, this would mean that combination

therapy was guaranteed whenever a suppository is used. Chances of definitive cure

in individual patients would be increased and the population risk of developing

artemisinin resistance would, in theory, be less. The challenges involved in the

development of such a drug would include:

1. The choice of the second antimalarial. As described in Chapter 2, all available

alternatives are imperfect whether due to pre-existing drug resistance, cost,

tolerability, or PK shortcomings.

2. The technical challenges of ensuring high bioavailability of the partner drug.

Most candidate partner drugs (such as MQ, LUM and PQ) are lipophillic and

have poor water solubility, so that achieving a pharmaceutical preparation that is

consistently well-absorbed across the rectal mucosa might be difficult.

Clinical evaluations of such a preparation would first need to focus on bioavailability

in healthy volunteers. Because partner drugs may have narrower therapeutic

margins than the artemisinins, the degree of variability in bioavailability of the

partner drug in such preparations may be even more important than that seen with

ARTS.

275

5.6. Importance of erythrocyte polymorphisms in Papua New Guinea and implications for further research

Studies in this thesis indicated that �-thalassaemia is unlikely to be a primary

determinant of PK disposition or clinical response to artemisinin derivatives.

Therefore, further enquiry into its role is likely to be of only secondary interest.

However, inclusion of �-thalassaemia status as a covariate in larger studies of

artemisinin derivatives may nonetheless be worthwhile. If multivariate analysis were

to demonstrate a statistically significant association with clinical efficacy outcomes

(even if this was of modest effect size) it might still provide insights into the mode of

action of the artemisinins at a cellular level.

The findings from Clinical trial 4a suggesting an association between features of

severe malaria and carriage of the GPC∆ex3 mutation, are of preliminary interest,

suggesting a possible protective effect specifically against cerebral malaria.

However, the design of this study, limited as it was only to subjects with severe

malaria, was imperfect and subject to selection bias and confounding. Therefore,

these results should be viewed as hypothesis-generating at this stage. An

adequately designed case-control study comparing GPC∆ex3 mutation prevalence

in children with cerebral malaria versus carefully matched healthy community

controls would provide the best possible level of evidence to further test the

hypothesis generated in Clinical trial 4a, that GPC∆ex3 mutation is protective

against cerebral malaria. However, at least two other erythrocyte polymorphisms (�-

thalassaemia and SLC4A1�27) are highly prevalent in this population and have

been associated with protection from severe or cerebral malaria. Numerous

additional factors such as the time taken to access medical care and institution of

treatment prior to presentation would also be important determinants. These factors

could vary geographically as could the population prevalence of GPC∆ex3 mutation

[46, 272] and could therefore be confounding variables in the analysis. A robust

study design would require careful inclusion of putative protective/risk factors in a

multi-variate analysis. The sample size requirements for such a study are likely to

be very large and therefore to require a high level of resources. However, firm

evidence of a protective role of GPC∆ex3 would be of great interest in gaining

276

understandings of the pathogenesis of severe malaria at a cellular level. A currently

ongoing clinical surveillance project, evaluating aetiology of severe childhood illness

at Modilon General Hospital (co-ordinated by Laurens Manning, School of Medicine

and Pharmacology, UWA) will utilise a case control design including children with

severe malaria, uncomplicated malaria and matched healthy community controls.

This study design and its much larger sample size, will have greater power to test

the hypothesis generated in Clinical trial 4a, that GPC∆ex3 mutation is protective

against cerebral malaria.

5.7. Improved dosing regimens for intermittent preventive treatment of malaria in pregnancy

The studies in Chapter 4 demonstrated significant scope for modifying dosage

regimens to improve the efficacy of CQ, SDOX and PYR when used as IPTp.

However, whether or not dose-modification studies of CQ in pregnancy would be

warranted is debatable, given its efficacy is severely compromised by P. falciparum

resistance in PNG and throughout most of the world. There may also be a risks of

toxicity related to higher maximal drug concentrations if higher drug doses are

trialed in pregnant women, particularly given the known effects of CQ on

haemodynamic stability that could be potentiated by pregnancy-related changes in

cardiovascular physiology. Therefore, the limited potential benefit from improving

the efficacy of a drug already highly compromised by resistance may not justify

exposing pregnant women to the potential risks associated with higher drug doses,

even in the highly regulated setting of a clinical trial. Nonetheless CQ remains one

of very few antimalarials that are both affordable and considered safe in pregnancy.

It is also still recommended as treatment for malaria in pregnancy due to non-

falciparum species and therefore likely to be in continued use for some time in

many parts of the world.

The efficacy of IPT was first demonstrated with SP and the majority of evidence

supporting the IPT intervention pertains to use of this drug. [402] Because of this

and issues of affordability and safety of alternative antimalarials, SP currently

remains the most important drug for use in IPTp. Therefore, in contrast to CQ,

optimization of SP dosage in pregnancy remains a high priority. Because the

efficacy of SP is dependent on synergy between its two components, [109] and

277

because the magnitude of pregnancy-related reductions in drug concentrations

demonstrated in Clinical trial 6 was of a similar magnitude for both SDOX and PYR,

dose modifications of SP should probably maintain the same ratio of SDOX to PYR

(20:1 based on mg doses). Re-development of co-formulations other than the

current standard products (i.e. those containing 500mg SDOX and 25mg PYR) is

probably not necessary at this stage.

There are clearly significant limitations to re-designing dosage regimens for IPTp.

Firstly, increasing individual dosages will lead to higher maximal drug

concentrations that may have consequences for toxicity and tolerability. Secondly,

in contrast to the effects of higher Vd, the effects of higher CL in pregnancy will be

more difficult to overcome with dose escalation alone. Thirdly, whilst split dosing

schedules or more frequent administration of IPT may compensate for higher CL,

this may be precluded in much of the developing world by issues of practicality and

compliance.

The PK models generated from Clinical trial 6 were used to simulate concentration-

time profiles for various hypothetical dosage modifications for use in pregnancy.

These certainly resulted in profiles that would be expected to improve their efficacy

when used as IPTp. However, proposed dosage modifications will need to be

evaluated in clinical trials with careful monitoring in order to detect toxicity

associated with higher drug concentrations, particularly in the case of CQ.

The PK simulations also showed that pregnancy-related changes in disposition may

occur by a combination of effects on Vd, CL and t½e. The relative contribution of

each mechanism and their absolute magnitude may not necessarily be predictable.

Therefore, studies employing a similar design to that of Clinical trial 6 will be

needed for other new antimalarial agents so that optimal dosage regimens can be

rationally designed specifically for use in pregnant women. Clinical trial 6 has

provided a study design template for further evaluations of drugs being considered

for use as IPTp in PNG, including azithromycin (co-ordinated by Sam Salman, PhD

student, School of Medicine and Pharmacology, UWA) and PQ (co-ordinated by

Brioni Moore, PhD student, School of Medicine and Pharmacology, UWA).

278

The results of Clinical trial 6 have only recently been finalized (publications

submitted) and have not yet had an impact on health policy in PNG. Given that CQ

is so highly compromised by parasite resistance in PNG, it is likely that the CQ-SP

combination will be replaced. Currently, investigation of alternative agents

(azithromycin and PQ) for use in pregnancy in PNG, are underway. These will

require additional reproductive toxicity data (PQ), efficacy evaluations (both PQ and

azithromycin) and understanding of pregnancy related changes in PK disposition

necessary to optimise dosing (both drugs). However, even if one of these is

considered acceptable for national treatment policy, it is likely that the prevailing

approach of combination therapy will continue and that co-administration of a

second drug will be considered desirable in order to improve efficacy and limit the

development of resistance. Given that the list of available partner drugs is limited by

issues of cost, lack of available safety data in pregnancy and that the majority of

evidence supporting IPT pertains to SP, it is likely that SP will still be considered as

a partner drug in this instance. Because PQ and azithromycin are likely to be

administered in split doses over 2-3 days, SP dose modification would probably be

most appropriately be administered as a 50% total dose increase administered as 3

split doses (shown in Fig. 4.16 B).

5.8. Use of species-specific PCR diagnostics in pregnancy – the role of non-falciparum species in pregnant women in PNG

Species-specific PCR diagnostics used in Clinical trial 6 showed generally good

concordance with microscopy but were significantly discrepant with regard to

speciation, suggesting that microscopy overestimated P. falciparum treatment

failures and under-diagnosed emergent infections with P. ovale and P. malariae.

The importance of these two species in pregnancy is essentially unknown and

merits further investigation. The PCR diagnostics in Clinical trial 6 suggested that

pregnant women may be at higher risk of P. ovale infections, a phenomenon that

may previously have been missed due to microscopic mis-diagnosis. These findings

require further confirmation to investigate the previously neglected role of P. ovale

in pregnancy. Species-specific PCR based technologies of the type used in Clinical

trial 6 may have an important role when used in larger efficacy studies evaluating

IPT interventions, especially in areas like PNG where there is high transmission of

279

multiple species. These might serve to both more accurately define the morbidity

impact of these species, and to evaluate the efficacy of IPTp drugs in relation to

these species. The high rate of PCR detection of P. ovale and P. malariae in

Clinical trial 6 suggests that efficacy of CQ-SP is poor for these species in PNG.

More efficacy data is required for these species for both conventional existing

antimalarials and for newer drugs.

280

281

6. REFERENCES

282

283

1. Lopez, A.D., et al., Global and regional burden of disease and risk factors, 2001: systematic analysis of population health data. Lancet, 2006. 367(9524): p. 1747-57.

2. Trape, J., The public health impact of chloroquine resistance in Africa. Am J Trop Med Hyg, 2001. 64(90010): p. 12-17.

3. Murphy, S. and J. Breman, Gaps in the childhood malaria burden in Africa: cerebral malaria, neurological sequelae, anemia, respiratory distress, hypoglycemia, and complications of pregnancy. Am J Trop Med Hyg, 2001. 64(90010): p. 57-67.

4. Breman, J., The ears of the hippopotamus: manifestations, determinants, and estimates of the malaria burden. Am J Trop Med Hyg, 2001. 64(90010): p. 1-c-11.

5. Carter, R. and K.N. Mendis, Evolutionary and historical aspects of the burden of malaria. Clin Microbiol Rev, 2002. 15(4): p. 564-94.

6. Singh, B., et al., A large focus of naturally acquired Plasmodium knowlesi infections in human beings. Lancet, 2004. 363(9414): p. 1017-24.

7. Snow, R.W., E.L. Korenromp, and E. Gouws, Pediatric mortality in Africa: plasmodium falciparum malaria as a cause or risk? Am J Trop Med Hyg, 2004. 71(2 Suppl): p. 16-24.

8. Snow, R.W., et al., The global distribution of clinical episodes of Plasmodium falciparum malaria. Nature, 2005. 434(7030): p. 214-7.

9. White, N.J., Malaria, in Manson's Tropical Diseases, G. Cook and A. Zumla, Editors. 2003, Elsevier. p. 1205-1296.

10. Genton, B., et al., Plasmodium vivax and mixed infections are associated with severe malaria in children: a prospective cohort study from Papua New Guinea. PLoS Med, 2008. 5(6): p. e127.

11. Kochar, D.K., et al., Severe Plasmodium vivax malaria: a report on serial cases from Bikaner in northwestern India. Am J Trop Med Hyg, 2009. 80(2): p. 194-8.

12. Price, R.N., et al., Vivax malaria: neglected and not benign. Am J Trop Med Hyg, 2007. 77(6 Suppl): p. 79-87.

13. Tjitra, E., et al., Multidrug-resistant Plasmodium vivax associated with severe and fatal malaria: a prospective study in Papua, Indonesia. PLoS Med, 2008. 5(6): p. e128.

14. Kitua, A.Y., et al., The role of low level Plasmodium falciparum parasitaemia in anaemia among infants living in an area of intense and perennial transmission. Trop Med Int Health, 1997. 2(4): p. 325-33.

15. Sen, R., et al., Clinico-haematological profile in acute and chronic Plasmodium falciparum malaria in children. J Commun Dis, 1994. 26(1): p. 31-8.

16. Guerra, C.A., R.W. Snow, and S.I. Hay, Mapping the global extent of malaria in 2005. Trends Parasitol, 2006. 22(8): p. 353-8.

17. WHO, The Abuja Declaration. The African Summit of Roll Back Malaria, April 2000. 2000. 18. WHO, The use of antimalarial drugs. Report of a WHO Informal Consultation. 2001:

Geneva. 19. White, N.J., et al., Averting a malaria disaster. Lancet, 1999. 353(9168): p. 1965-7. 20. Rosenberg, R., et al., An estimation of the number of malaria sporozoites ejected by a

feeding mosquito. Trans R Soc Trop Med Hyg, 1990. 84(2): p. 209-12. 21. Cowman, A.F. and B.S. Crabb, Invasion of red blood cells by malaria parasites. Cell, 2006.

124(4): p. 755-66. 22. Simpson, J.A., et al., Population dynamics of untreated Plasmodium falciparum malaria

within the adult human host during the expansion phase of the infection. Parasitology, 2002. 124(Pt 3): p. 247-63.

23. Rogerson, S.J., G.E. Grau, and N.H. Hunt, The microcirculation in severe malaria. Microcirculation, 2004. 11(7): p. 559-76.

24. WHO, Severe falciparum malaria Trans R Soc Trop Med Hyg, 2000. 94 Suppl 1: p. S1-90. 25. Allen, S.J., et al., Severe malaria in children in Papua New Guinea. Qjm, 1996. 89(10): p.

779-88. 26. English, M., et al., Acidosis in severe childhood malaria. Qjm, 1997. 90(4): p. 263-70. 27. Marsh, K., et al., Indicators of life-threatening malaria in African children. N Engl J Med,

1995. 332(21): p. 1399-404.

284

28. Gillies, M.T., Anopheline mosquitoes and biomomics, in Malaria: Principles and Practice of Malariology, W.H. Wernsdorfer and I. McGregor, Editors. 1988, Churchill Livingstone: Edinburgh. p. 453-485.

29. Muller, I., et al., The epidemiology of malaria in Papua New Guinea. Trends Parasitol, 2003. 19(6): p. 253-9.

30. Molineaux, L., et al., The epidemiology of malaria and its measurement, in Malaria: Principles and Practice of Malariology, W.H. Wernsdorfer and I. McGregor, Editors. 1988, Churchill Livingstone. p. 999-1090.

31. Cattani, J.A., et al., The epidemiology of malaria in a population surrounding Madang, Papua New Guinea. Am J Trop Med Hyg, 1986. 35(1): p. 3-15.

32. Genton, B., et al., The epidemiology of malaria in the Wosera area, East Sepik Province, Papua New Guinea, in preparation for vaccine trials. I. Malariometric indices and immunity. Ann Trop Med Parasitol, 1995. 89(4): p. 359-76.

33. Imbert, P., et al., Severe malaria among children in a low seasonal transmission area, Dakar, Senegal: influence of age on clinical presentation. Trans R Soc Trop Med Hyg, 1997. 91(1): p. 22-4.

34. Snow, R.W., et al., Relation between severe malaria morbidity in children and level of Plasmodium falciparum transmission in Africa. Lancet, 1997. 349(9066): p. 1650-4.

35. Marsh, K. and R.W. Snow, Host-parasite interaction and morbidity in malaria endemic areas. Philos Trans R Soc Lond B Biol Sci, 1997. 352(1359): p. 1385-94.

36. Pasvol, G., D.J. Weatherall, and R.J. Wilson, Effects of foetal haemoglobin on susceptibility of red cells to Plasmodium falciparum. Nature, 1977. 270(5633): p. 171-3.

37. Nosten, F., R. McGready, and T. Mutabingwa, Case management of malaria in pregnancy. Lancet Infect Dis, 2007. 7(2): p. 118-25.

38. Desai, M., et al., Epidemiology and burden of malaria in pregnancy. Lancet Infect Dis, 2007. 7(2): p. 93-104.

39. Ticconi, C., et al., Effect of maternal HIV and malaria infection on pregnancy and perinatal outcome in Zimbabwe. J Acquir Immune Defic Syndr, 2003. 34(3): p. 289-94.

40. Ladner, J., et al., HIV infection, malaria, and pregnancy: a prospective cohort study in Kigali, Rwanda. Am J Trop Med Hyg, 2002. 66(1): p. 56-60.

41. Ayisi, J.G., et al., The effect of dual infection with HIV and malaria on pregnancy outcome in western Kenya. Aids, 2003. 17(4): p. 585-94.

42. Flint, J., et al., High frequencies of alpha-thalassaemia are the result of natural selection by malaria. Nature, 1986. 321(6072): p. 744-50.

43. Zimmerman, P.A., et al., Erythrocyte polymorphisms and malaria parasite invasion in Papua New Guinea. Trends Parasitol, 2003. 19(6): p. 250-2.

44. Cortes, A., et al., Adhesion of Plasmodium falciparum-infected red blood cells to CD36 under flow is enhanced by the cerebral malaria-protective trait South-East Asian ovalocytosis. Mol Biochem Parasitol, 2005. 142(2): p. 252-7.

45. Allen, S.J., et al., a+- Thalassemia protects chlidren against disease caused by other infections as well as malaria. Proc. Natl. Acad. Sci. USA, 1997. 94: p. 14736 - 14741.

46. Patel, S., C. King, and C. Mgone, Gerbich antigen blood group and band 3 polymorphisms in two malaria holoendemic regions of Papua New Guinea. Am J Hematol, 2004. 75: p. 1-5.

47. Maier, A.G., et al., Plasmodium falciparum erythrocyte invasion through glycophorin C and selection for Gerbich negativity in human populations. Nat Med, 2003. 9(1): p. 87-92.

48. Roper, C., et al., Intercontinental spread of pyrimethamine-resistant malaria. Science, 2004. 305(5687): p. 1124.

49. White, N.J., Antimalarial drug resistance. J Clin Invest, 2004. 113(8): p. 1084-92. 50. Hastings, I.M., W.M. Watkins, and N.J. White, The evolution of drug-resistant malaria: the

role of drug elimination half-life. Philos Trans R Soc Lond B Biol Sci, 2002. 357(1420): p. 505-19.

51. Ekland, E.H. and D.A. Fidock, Advances in understanding the genetic basis of antimalarial drug resistance. Curr Opin Microbiol, 2007. 10(4): p. 363-70.

52. Korsinczky, M., et al., Mutations in Plasmodium falciparum cytochrome b that are associated with atovaquone resistance are located at a putative drug-binding site. Antimicrob Agents Chemother, 2000. 44(8): p. 2100-8.

285

53. Triglia, T., et al., Amplification of the multidrug resistance gene pfmdr1 in Plasmodium falciparum has arisen as multiple independent events. Mol Cell Biol, 1991. 11(10): p. 5244-50.

54. Plowe, C.V., et al., Mutations in Plasmodium falciparum dihydrofolate reductase and dihydropteroate synthase and epidemiologic patterns of pyrimethamine-sulfadoxine use and resistance. J Infect Dis, 1997. 176(6): p. 1590-6.

55. Reeder, J.C., et al., Point mutations in the dihydrofolate reductase and dihydropteroate synthetase genes and in vitro susceptibility to pyrimethamine and cycloguanil of Plasmodium falciparum isolates from Papua New Guinea. Am J Trop Med Hyg, 1996. 55(2): p. 209-13.

56. Nagesha, H.S., et al., Mutations in the pfmdr1, dhfr and dhps genes of Plasmodium falciparum are associated with in-vivo drug resistance in West Papua, Indonesia. Trans R Soc Trop Med Hyg, 2001. 95(1): p. 43-9.

57. Jiang, H., et al., Current understanding of the molecular basis of chloroquine-resistance in Plasmodium falciparum. J Postgrad Med, 2006. 52(4): p. 271-6.

58. Bray, P.G., et al., Defining the role of PfCRT in Plasmodium falciparum chloroquine resistance. Mol Microbiol, 2005. 56(2): p. 323-33.

59. Casey, G.J., et al., Molecular analysis of Plasmodium falciparum from drug treatment failure patients in Papua New Guinea. Am J Trop Med Hyg, 2004. 70(3): p. 251-5.

60. Nguyen, M.H., et al., Treatment of uncomplicated falciparum malaria in southern Vietnam: can chloroquine or sulfadoxine-pyrimethamine be reintroduced in combination with artesunate? Clin Infect Dis, 2003. 37(11): p. 1461-6.

61. Cox-Singh, J., et al., Application of a multi-faceted approach for the assessment of treatment response in falciparum malaria: a study from Malaysian Borneo. Int J Parasitol, 2003. 33(13): p. 1545-52.

62. Price, R.N., et al., Mefloquine resistance in Plasmodium falciparum and increased pfmdr1 gene copy number. Lancet, 2004. 364(9432): p. 438-47.

63. Price, R.N., et al., Molecular and pharmacological determinants of the therapeutic response to artemether-lumefantrine in multidrug-resistant Plasmodium falciparum malaria. Clin Infect Dis, 2006. 42(11): p. 1570-7.

64. WHO, Guidelines for the treatment of malaria. 2006. 65. Redd, S.C., et al., Usefulness of clinical case-definitions in guiding therapy for African

children with malaria or pneumonia. Lancet, 1992. 340(8828): p. 1140-3. 66. Moody, A., Rapid diagnostic tests for malaria parasites. Clin Microbiol Rev, 2002. 15(1): p.

66-78. 67. Schellenberg, D., et al., African children with malaria in an area of intense Plasmodium

falciparum transmission: features on admission to the hospital and risk factors for death. Am J Trop Med Hyg, 1999. 61(3): p. 431-8.

68. Waller, D., et al., Clinical features and outcome of severe malaria in Gambian children. Clin Infect Dis, 1995. 21(3): p. 577-87.

69. Dzeing-Ella, A., et al., Severe falciparum malaria in Gabonese children: clinical and laboratory features. Malar J, 2005. 4: p. 1.

70. Mockenhaupt, F.P., et al., Manifestation and outcome of severe malaria in children in northern Ghana. Am J Trop Med Hyg, 2004. 71(2): p. 167-72.

71. Molyneux, M.E., et al., Clinical features and prognostic indicators in paediatric cerebral malaria: a study of 131 comatose Malawian children. Q J Med, 1989. 71(265): p. 441-59.

72. Taylor, T.E., A. Borgstein, and M.E. Molyneux, Acid-base status in paediatric Plasmodium falciparum malaria. Q J Med, 1993. 86(2): p. 99-109.

73. Milner, D.A., Jr., et al., Sampling of supraorbital brain tissue after death: improving on the clinical diagnosis of cerebral malaria. J Infect Dis, 2005. 191(5): p. 805-8.

74. Taylor, T.E., et al., Differentiating the pathologies of cerebral malaria by postmortem parasite counts. Nat Med, 2004. 10(2): p. 143-5.

75. Beare, N.A., et al., Perfusion abnormalities in children with cerebral malaria and malarial retinopathy. J Infect Dis, 2009. 199(2): p. 263-71.

76. Lewallen, S., et al., Using malarial retinopathy to improve the classification of children with cerebral malaria. Trans R Soc Trop Med Hyg, 2008. 102(11): p. 1089-94.

286

77. English, M., et al., Deep breathing in children with severe malaria: indicator of metabolic acidosis and poor outcome. Am J Trop Med Hyg, 1996. 55(5): p. 521-4.

78. English, M., et al., Lactic acidosis and oxygen debt in African children with severe anaemia. Qjm, 1997. 90(9): p. 563-9.

79. Krishna, S., et al., Lactic acidosis and hypoglycaemia in children with severe malaria: pathophysiological and prognostic significance. Trans R Soc Trop Med Hyg, 1994. 88(1): p. 67-73.

80. Newton, C.R., et al., The prognostic value of measures of acid/base balance in pediatric falciparum malaria, compared with other clinical and laboratory parameters. Clin Infect Dis, 2005. 41(7): p. 948-57.

81. Planche, T., et al., A prospective comparison of malaria with other severe diseases in African children: prognosis and optimization of management. Clin Infect Dis, 2003. 37(7): p. 890-7.

82. Rogerson, S.J., et al., Malaria in pregnancy: pathogenesis and immunity. Lancet Infect Dis, 2007. 7(2): p. 105-17.

83. Gensini, G.F., A.A. Conti, and D. Lippi, The contributions of Paul Ehrlich to infectious disease. J Infect, 2007. 54(3): p. 221-4.

84. Schlitzer, M., Malaria Chemotherapeutics Part I: History of Antimalarial Drug Development, Currently Used Therapeutics, and Drugs in Clinical Development. ChemMedChem, 2007. 2(7): p. 944-986.

85. Davis, T.M., et al., Piperaquine: a resurgent antimalarial drug. Drugs, 2005. 65(1): p. 75-87.

86. Boudreau, E.F., et al., Type II mefloquine resistance in Thailand. Lancet, 1982. 2(8311): p. 1335.

87. Nosten, F., et al., Cardiac effects of antimalarial treatment with halofantrine. Lancet, 1993. 341(8852): p. 1054-6.

88. Winstanley, P., Chlorproguanil-dapsone (LAPDAP) for uncomplicated falciparum malaria. Trop Med Int Health, 2001. 6(11): p. 952-4.

89. Alloueche, A., et al., Comparison of chlorproguanil-dapsone with sulfadoxine-pyrimethamine for the treatment of uncomplicated falciparum malaria in young African children: double-blind randomised controlled trial. Lancet, 2004. 363(9424): p. 1843-8.

90. Fanello, C.I., et al., High risk of severe anaemia after chlorproguanil-dapsone+artesunate antimalarial treatment in patients with G6PD (A-) deficiency. PLoS One, 2008. 3(12): p. e4031.

91. Hsu, E., Reflections on the 'discovery' of the antimalarial qinghao. Br J Clin Pharmacol, 2006. 61(6): p. 666-70.

92. Clinical studies on the treatment of malaria with qinghaosu and its derivatives. China Cooperative Research Group on qinghaosu and its derivatives as antimalarials. J Tradit Chin Med, 1982. 2(1): p. 45-50.

93. Bruce-Chwatt, L.J., Qinghaosu: a new antimalarial. Br Med J (Clin Res Ed), 1982. 284(6318): p. 767-8.

94. Jiang, J.B., et al., Antimalarial activity of mefloquine and qinghaosu. Lancet, 1982. 2(8293): p. 285-8.

95. Bunnag, D., et al., Clinical trial of artesunate and artemether on multidrug resistant falciparum malaria in Thailand. A preliminary report. Southeast Asian J Trop Med Public Health, 1991. 22(3): p. 380-5.

96. Nosten, F., et al., Treatment of multidrug-resistant Plasmodium falciparum malaria with 3-day artesunate-mefloquine combination. J Infect Dis, 1994. 170(4): p. 971-7.

97. Nosten, F., et al., Effects of artesunate-mefloquine combination on incidence of Plasmodium falciparum malaria and mefloquine resistance in western Thailand: a prospective study. Lancet, 2000. 356(9226): p. 297-302.

98. Adjuik, M., et al., Artesunate combinations for treatment of malaria: meta-analysis. Lancet, 2004. 363(9402): p. 9-17.

99. Davis, A., Clinical trials in parasitic diseases. Transactions of the Royal Society of Tropical Medicine and Hygiene, 2004. 98(3): p. 139-141.

287

100. Benet, L.Z., D.L. Kroetz, and L.B. Sheiner, Pharmacokinetics, in Goodman and Gliman's The Pharmacological Basis of Therapeutics, J.G. Hardiman, et al., Editors. 1996, McGraw-Hill: New York. p. 3-27.

101. White, N.J., et al., Quinine pharmacokinetics and toxicity in cerebral and uncomplicated Falciparum malaria. Am J Med, 1982. 73(4): p. 564-72.

102. Ward, S.A., et al., Antimalarial drugs and pregnancy: safety, pharmacokinetics, and pharmacovigilance. Lancet Infect Dis, 2007. 7(2): p. 136-44.

103. Simpson, J.A., L. Aarons, and N.J. White, How can we do pharmacokinetic studies in the tropics? Trans R Soc Trop Med Hyg, 2001. 95(4): p. 347-51.

104. Ducharme, J. and R. Farinotti, Clinical pharmacokinetics and metabolism of chloroquine. Focus on recent advancements. Clin Pharmacokinet, 1996. 31(4): p. 257-74.

105. Winstanley, P., et al., Towards optimal regimens of parenteral quinine for young African children with cerebral malaria: the importance of unbound quinine concentration. Trans R Soc Trop Med Hyg, 1993. 87(2): p. 201-6.

106. Winstanley, P.A., et al., Towards optimal regimens of parenteral quinine for young African children with cerebral malaria: unbound quinine concentrations following a simple loading dose regimen. Trans R Soc Trop Med Hyg, 1994. 88(5): p. 577-80.

107. Gabrielsson, J. and D. Weiner, Pharmacokinetic and Pharmacodynamic Data Analysis: Concepts and Applications. 4 ed. 2000, Stockholm: Swedish Pharmaceutical Press.

108. White, N.J., Assessment of the pharmacodynamic properties of antimalarial drugs in vivo. Antimicrob Agents Chemother, 1997. 41(7): p. 1413-22.

109. White, N.J., Intermittent presumptive treatment for malaria. PLoS Med, 2005. 2(1): p. e3. 110. Newton, P., et al., Pharmacokinetics of quinine and 3-hydroxyquinine in severe falciparum

malaria with acute renal failure. Trans R Soc Trop Med Hyg, 1999. 93(1): p. 69-72. 111. Sukontason, K., et al., Plasma quinine concentrations in falciparum malaria with acute

renal failure. Trop Med Int Health, 1996. 1(2): p. 236-42. 112. Janse, C.J., et al., Comparison of in vivo and in vitro antimalarial activity of artemisinin,

dihydroartemisinin and sodium artesunate in the Plasmodium berghei-rodent model. Int J Parasitol, 1994. 24(4): p. 589-94.

113. WHO, Assessment and monitoring of antimalarial drugs for the treatment of acute uncomplicated falciparum malaria, in WHO/HTM/RBM/2003.50. 2003.

114. Noedl, H., et al., Evidence of artemisinin-resistant malaria in western Cambodia. N Engl J Med, 2008. 359(24): p. 2619-20.

115. A meta-analysis using individual patient data of trials comparing artemether with quinine in the treatment of severe falciparum malaria. Trans R Soc Trop Med Hyg, 2001. 95(6): p. 637-50.

116. Dondorp, A., et al., Artesunate versus quinine for treatment of severe falciparum malaria: a randomised trial. Lancet, 2005. 366(9487): p. 717-25.

117. Barnes, K.I., et al., Efficacy of rectal artesunate compared with parenteral quinine in initial treatment of moderately severe malaria in African children and adults: a randomised study. Lancet, 2004. 363(9421): p. 1598-605.

118. White, N.J., Cardiotoxicity of antimalarial drugs. Lancet Infect Dis, 2007. 7(8): p. 549-58. 119. Nies, A.S. and S.P. Spielberg, Principles of Therapeutics, in Goodman and Gillman's

Pharmacological Basis of Therapeutics J.G. Hardman, et al., Editors. 1996, McGraw-Hill: New York. p. 43-62.

120. Amin, A.A., et al., The difference between effectiveness and efficacy of antimalarial drugs in Kenya. Trop Med Int Health, 2004. 9(9): p. 967-74.

121. Yeung, S. and N.J. White, How do patients use antimalarial drugs? A review of the evidence. Trop Med Int Health, 2005. 10(2): p. 121-38.

122. Schlitzer, M., Antimalarial Drugs - What is in Use and What is in the Pipeline. Arch Pharm (Weinheim), 2008. 341(3): p. 149-63.

123. White, N.J., Clinical pharmacokinetics of antimalarial drugs. Clin Pharmacokinet, 1985. 10(3): p. 187-215.

124. Krishna, S. and N.J. White, Pharmacokinetics of quinine, chloroquine and amodiaquine. Clinical implications. Clin Pharmacokinet, 1996. 30(4): p. 263-99.

288

125. Sowunmi, A., O. Walker, and L.A. Salako, Pruritus and antimalarial drugs in Africans. Lancet, 1989. 2(8656): p. 213.

126. Flaurenson, I., et al., Chloroquine overdose in Papua New Guinea. Bmj, 1993. 307(6903): p. 564-5.

127. Looareesuwan, S., et al., Cardiovascular toxicity and distribution kinetics of intravenous chloroquine. Br J Clin Pharmacol, 1986. 22(1): p. 31-6.

128. Schuurkamp, G.J., et al., Chloroquine-resistant Plasmodium vivax in Papua New Guinea. Trans R Soc Trop Med Hyg, 1992. 86(2): p. 121-2.

129. Schwartz, I.K., E.M. Lackritz, and L.C. Patchen, Chloroquine-resistant Plasmodium vivax from Indonesia. N Engl J Med, 1991. 324(13): p. 927.

130. Karunajeewa, H., et al., Safety evaluation of fixed combination piperaquine plus dihydroartemisinin (Artekin) in Cambodian children and adults with malaria. Br J Clin Pharmacol, 2004. 57(1): p. 93-9.

131. Denis, M.B., et al., Efficacy and safety of dihydroartemisinin-piperaquine (Artekin) in Cambodian children and adults with uncomplicated falciparum malaria. Clin Infect Dis, 2002. 35(12): p. 1469-76.

132. Tran, T.H., et al., Dihydroartemisinin-piperaquine against multidrug-resistant Plasmodium falciparum malaria in Vietnam: randomised clinical trial. Lancet, 2004. 363(9402): p. 18-22.

133. Mayxay, M., et al., An open, randomized comparison of artesunate plus mefloquine vs. dihydroartemisinin-piperaquine for the treatment of uncomplicated Plasmodium falciparum malaria in the Lao People's Democratic Republic (Laos). Trop Med Int Health, 2006. 11(8): p. 1157-65.

134. Smithuis, F., et al., Efficacy and effectiveness of dihydroartemisinin-piperaquine versus artesunate-mefloquine in falciparum malaria: an open-label randomised comparison. Lancet, 2006. 367(9528): p. 2075-85.

135. Tarning, J., et al., Pitfalls in estimating piperaquine elimination. Antimicrob Agents Chemother, 2005. 49(12): p. 5127-8.

136. Roshammar, D., et al., Pharmacokinetics of piperaquine after repeated oral administration of the antimalarial combination CV8 in 12 healthy male subjects. Eur J Clin Pharmacol, 2006. 62(5): p. 335-41.

137. Sim, I.K., T.M. Davis, and K.F. Ilett, Effects of a high-fat meal on the relative oral bioavailability of piperaquine. Antimicrob Agents Chemother, 2005. 49(6): p. 2407-11.

138. Hung, T.Y., et al., Population pharmacokinetics of piperaquine in adults and children with uncomplicated falciparum or vivax malaria. Br J Clin Pharmacol, 2004. 57(3): p. 253-62.

139. Sheng, N., W. Jiang, and H.L. Tang, Pre-clinical toxicological study of new anitmalarial agents: II. Piperaquine phosphate and its compound 'Preventative No.3'. Ti Erh Chun i Ta Hsueh Hsueh Pao (Acad J Second Military College), 1981. 1: p. 40-6.

140. Karema, C., et al., Safety and efficacy of dihydroartemisinin/piperaquine (Artekin((R))) for the treatment of uncomplicated Plasmodium falciparum malaria in Rwandan children. Trans R Soc Trop Med Hyg, 2006.

141. Ratcliff, A., et al., Two fixed-dose artemisinin combinations for drug-resistant falciparum and vivax malaria in Papua, Indonesia: an open-label randomised comparison. Lancet, 2007. 369(9563): p. 757-65.

142. Basco, L.K. and P. Ringwald, In vitro activities of piperaquine and other 4-aminoquinolines against clinical isolates of Plasmodium falciparum in Cameroon. Antimicrob Agents Chemother, 2003. 47(4): p. 1391-4.

143. Russell, B., et al., Determinants of in vitro drug susceptibility testing of Plasmodium vivax. Antimicrob Agents Chemother, 2008. 52(3): p. 1040-5.

144. Davis, T.M., H.A. Karunajeewa, and K.F. Ilett, Artemisinin-based combination therapies for uncomplicated malaria. Med J Aust, 2005. 182(4): p. 181-5.

145. Ezzet, F., R. Mull, and J. Karbwang, Population pharmacokinetics and therapeutic response of CGP 56697 (artemether + benflumetol) in malaria patients. Br J Clin Pharmacol, 1998. 46(6): p. 553-61.

146. Ezzet, F., et al., Pharmacokinetics and pharmacodynamics of lumefantrine (benflumetol) in acute falciparum malaria. Antimicrob Agents Chemother, 2000. 44(3): p. 697-704.

289

147. Ashley, E.A., et al., How much fat is necessary to optimize lumefantrine oral bioavailability? Trop Med Int Health, 2007. 12(2): p. 195-200.

148. Ashley, E.A., et al., Pharmacokinetic study of artemether-lumefantrine given once daily for the treatment of uncomplicated multidrug-resistant falciparum malaria. Trop Med Int Health, 2007. 12(2): p. 201-8.

149. White, N.J., M. van Vugt, and F. Ezzet, Clinical pharmacokinetics and pharmacodynamics and pharmacodynamics of artemether-lumefantrine. Clin Pharmacokinet, 1999. 37(2): p. 105-25.

150. van Agtmael, M.A., et al., Multiple dose pharmacokinetics of artemether in Chinese patients with uncomplicated falciparum malaria. Int J Antimicrob Agents, 1999. 12(2): p. 151-8.

151. McGready, R., et al., The pharmacokinetics of artemether and lumefantrine in pregnant women with uncomplicated falciparum malaria. Eur J Clin Pharmacol, 2006. 62(12): p. 1021-31.

152. van Vugt, M., et al., No evidence of cardiotoxicity during antimalarial treatment with artemether-lumefantrine. Am J Trop Med Hyg, 1999. 61(6): p. 964-7.

153. Omari, A.A.A., C. Preston, and P.A. Garner, Artemether-lumefantrine (six-dose regimen) for treating uncomplictaed falciparum malaria (Cochrane Review). in Cochrane Database Syst Rev. , U. Software, Editor. 2005: Oxford.

154. Denis, M.B., et al., Efficacy of artemether-lumefantrine for the treatment of uncomplicated falciparum malaria in northwest Cambodia. Trop Med Int Health, 2006. 11(12): p. 1800-7.

155. Sisowath, C., et al., In vivo selection of Plasmodium falciparum pfmdr1 86N coding alleles by artemether-lumefantrine (Coartem). J Infect Dis, 2005. 191(6): p. 1014-7.

156. Hastings, I.M. and S.A. Ward, Coartem (artemether-lumefantrine) in Africa: the beginning of the end? J Infect Dis, 2005. 192(7): p. 1303-4; author reply 1304-5.

157. Fogg, C., et al., Adherence to a six-dose regimen of artemether-lumefantrine for treatment of uncomplicated Plasmodium falciparum malaria in Uganda. Am J Trop Med Hyg, 2004. 71(5): p. 525-30.

158. Piola, P., et al., Supervised versus unsupervised intake of six-dose artemether-lumefantrine for treatment of acute, uncomplicated Plasmodium falciparum malaria in Mbarara, Uganda: a randomised trial. Lancet, 2005. 365(9469): p. 1467-73.

159. Barnes, K.I., et al., Sulfadoxine-pyrimethamine pharmacokinetics in malaria: pediatric dosing implications. Clin Pharmacol Ther, 2006. 80(6): p. 582-96.

160. Winstanley, P.A., et al., The disposition of oral and intramuscular pyrimethamine/sulphadoxine in Kenyan children with high parasitaemia but clinically non-severe falciparum malaria. Br J Clin Pharmacol, 1992. 33(2): p. 143-8.

161. Bjorkman, A. and P.A. Phillips-Howard, Adverse reactions to sulfa drugs: implications for malaria chemotherapy. Bull World Health Organ, 1991. 69(3): p. 297-304.

162. Bwijo, B., et al., High prevalence of quintuple mutant dhps/dhfr genes in Plasmodium falciparum infections seven years after introduction of sulfadoxine and pyrimethamine as first line treatment in Malawi. Acta Trop, 2003. 85(3): p. 363-73.

163. Takechi, M., et al., Therapeutic efficacy of sulphadoxine/pyrimethamine and susceptibility in vitro of P. falciparum isolates to sulphadoxine-pyremethamine and other antimalarial drugs in Malawian children. Trop Med Int Health, 2001. 6(6): p. 429-34.

164. Faye, B., et al., Efficacy and tolerability of four antimalarial combinations in the treatment of uncomplicated Plasmodium falciparum malaria in Senegal. Malar J, 2007. 6: p. 80.

165. Mohamed, A.O., et al., The efficacies of artesunate-sulfadoxine-pyrimethamine and artemether-lumefantrine in the treatment of uncomplicated, Plasmodium falciparum malaria, in an area of low transmission in central Sudan. Ann Trop Med Parasitol, 2006. 100(1): p. 5-10.

166. Mukhtar, E.A., et al., A comparative study on the efficacy of artesunate plus sulphadoxine/pyrimethamine versus artemether-lumefantrine in eastern Sudan. Malar J, 2007. 6: p. 92.

167. Barnadas, C., et al., Plasmodium vivax dhfr and dhps mutations in isolates from Madagascar and therapeutic response to sulphadoxine-pyrimethamine. Malar J, 2008. 7: p. 35.

290

168. Desai, M.R., et al., Randomized, controlled trial of daily iron supplementation and intermittent sulfadoxine-pyrimethamine for the treatment of mild childhood anemia in western Kenya. J Infect Dis, 2003. 187(4): p. 658-66.

169. Schellenberg, D., et al., Intermittent treatment for malaria and anaemia control at time of routine vaccinations in Tanzanian infants: a randomised, placebo-controlled trial. Lancet, 2001. 357(9267): p. 1471-7.

170. Shulman, C.E., et al., Intermittent sulphadoxine-pyrimethamine to prevent severe anaemia secondary to malaria in pregnancy: a randomised placebo-controlled trial. Lancet, 1999. 353(9153): p. 632-6.

171. Bakyaita, N., et al., Sulfadoxine-pyrimethamine plus chloroquine or amodiaquine for uncomplicated falciparum malaria: a randomized, multisite trial to guide national policy in Uganda. Am J Trop Med Hyg, 2005. 72(5): p. 573-80.

172. Hwang, J., et al., Chloroquine or amodiaquine combined with sulfadoxine-pyrimethamine for uncomplicated malaria: a systematic review. Trop Med Int Health, 2006. 11(6): p. 789-99.

173. Staedke, S.G., et al., Combination treatments for uncomplicated falciparum malaria in Kampala, Uganda: randomised clinical trial. Lancet, 2004. 364(9449): p. 1950-7.

174. Zongo, I., et al., Amodiaquine, sulfadoxine-pyrimethamine, and combination therapy for uncomplicated falciparum malaria: a randomized controlled trial from Burkina Faso. Am J Trop Med Hyg, 2005. 73(5): p. 826-32.

175. Zongo, I., et al., Artemether-lumefantrine versus amodiaquine plus sulfadoxine-pyrimethamine for uncomplicated falciparum malaria in Burkina Faso: a randomised non-inferiority trial. Lancet, 2007. 369(9560): p. 491-8.

176. Chemical Abstracts Services SciFinder Scholar database, American Chemical Society, Washington DC, USA.

177. Ilett, K. and K. Batty, Artemisinin and its dervatives., in Antimicrobial Therapy and Vaccines, V.L. Yu, et al., Editors. 2005, ESun Technologies LLC, : Pitsburg PA. p. 981-1002.

178. Binh, T.Q., et al., Oral bioavailability of dihydroartemisinin in Vietnamese volunteers and in patients with falciparum malaria. Br J Clin Pharmacol, 2001. 51(6): p. 541-6.

179. Titulaer, H.A., et al., The pharmacokinetics of artemisinin after oral, intramuscular and rectal administration to volunteers. J Pharm Pharmacol, 1990. 42(11): p. 810-3.

180. Ilett, K.F., et al., The pharmacokinetic properties of intramuscular artesunate and rectal dihydroartemisinin in uncomplicated falciparum malaria. Br J Clin Pharmacol, 2002. 53(1): p. 23-30.

181. Karbwang, J., et al., Pharmacokinetics and bioavailability of oral and intramuscular artemether. Eur J Clin Pharmacol, 1997. 52(4): p. 307-10.

182. Davis, T.M., et al., Pharmacokinetics and pharmacodynamics of intravenous artesunate in severe falciparum malaria. Antimicrob Agents Chemother, 2001. 45(1): p. 181-6.

183. Hien, T.T. and N.J. White, Qinghaosu. Lancet, 1993. 341(8845): p. 603-8. 184. McIntosh, H.M. and P. Olliaro, Artemisinin derivatives for treating uncomplicated malaria.

Cochrane Database Syst Rev, 2000(2): p. CD000256. 185. McIntosh, H.M. and P. Olliaro, Artemisinin derivatives for treating severe malaria.

Cochrane Database Syst Rev, 2000(2): p. CD000527. 186. Price, R., et al., Adverse effects in patients with acute falciparum malaria treated with

artemisinin derivatives. Am J Trop Med Hyg, 1999. 60(4): p. 547-55. 187. Brewer, T.G., et al., Neurotoxicity in animals due to arteether and artemether,. Transactions

of the Royal Society of Tropical Medicine and Hygiene, 1994. 88(Supplement 1): p. 33-36. 188. Brewer, T.G., et al., Fatal neurotoxicity of arteether and artemether. Am J Trop Med Hyg,

1994. 51(3): p. 251-9. 189. Li, Q.G., et al., Neurotoxicity and efficacy of arteether related to its exposure times and

exposure levels in rodents. Am J Trop Med Hyg, 2002. 66(5): p. 516-25. 190. Gordi, T. and E.I. Lepist, Artemisinin derivatives: toxic for laboratory animals, safe for

humans? Toxicol Lett, 2004. 147(2): p. 99-107.

291

191. Genovese, R.F., D.B. Newman, and T.G. Brewer, Behavioral and neural toxicity of the artemisinin antimalarial, arteether, but not artesunate and artelinate, in rats. Pharmacol Biochem Behav, 2000. 67(1): p. 37-44.

192. Miller, L.G. and C.B. Panosian, Ataxia and slurred speech after artesunate treatment for falciparum malaria. N Engl J Med, 1997. 336(18): p. 1328.

193. Toovey, S. and A. Jamieson, Audiometric changes associated with the treatment of uncomplicated falciparum malaria with co-artemether. Transactions of the Royal Society of Tropical Medicine and Hygiene, 2004. 98(5): p. 261-267.

194. Kissinger, E., et al., Clinical and neurophysiological study of the effects of multiple doses of artemisinin on brain-stem function in Vietnamese patients. Am J Trop Med Hyg, 2000. 63(1-2): p. 48-55.

195. Davis, T.M., G.O. Edwards, and J.S. McCarthy, Artesunate and cerebellar dysfunction in falciparum malaria. N Engl J Med, 1997. 337(11): p. 792; author reply 793.

196. Davis, T.M., et al., Penetration of dihydroartemisinin into cerebrospinal fluid after administration of intravenous artesunate in severe falciparum malaria. Antimicrob Agents Chemother, 2003. 47(1): p. 368-70.

197. Johann-Liang, R. and R. Albrecht, Safety evaluations of drugs containing artemisinin derivatives for the treatment of malaria. Clin Infect Dis, 2003. 36(12): p. 1626-7; author reply 1627-8.

198. Clark, R.L., Embryotoxicity of the artemisinin antimalarials and potential consequences for use in women in the first trimester. Reprod Toxicol, 2009.

199. McGready, R., et al., Artemisinin antimalarials in pregnancy: a prospective treatment study of 539 episodes of multidrug-resistant Plasmodium falciparum. Clin Infect Dis, 2001. 33(12): p. 2009-16.

200. Price, R.N., et al., Effects of artemisinin derivatives on malaria transmissibility. Lancet, 1996. 347(9016): p. 1654-8.

201. Meshnick, S.R., Artemisinin: mechanisms of action, resistance and toxicity. Int J Parasitol, 2002. 32(13): p. 1655-60.

202. Eckstein-Ludwig, U., et al., Artemisinins target the SERCA of Plasmodium falciparum. Nature, 2003. 424(6951): p. 957-61.

203. Afonso, A., et al., Malaria parasites can develop stable resistance to artemisinin but lack mutations in candidate genes atp6 (encoding the sarcoplasmic and endoplasmic reticulum Ca2+ ATPase), tctp, mdr1, and cg10. Antimicrob Agents Chemother, 2006. 50(2): p. 480-9.

204. Standard Treatment of Common Illnesses of Children in Papua New Guinea. 7 ed. 2000: Papua New Guinea Department of Health.

205. Weber, M.W., et al., Evaluation of an algorithm for the integrated management of childhood illness in an area with seasonal malaria in the Gambia. Bull World Health Organ, 1997. 75 Suppl 1: p. 25-32.

206. Amin, A.A., et al., The difference between effectiveness and efficacy of antimalarial drugs in Kenya. Tropical Medicine and International Health, 2004. 9(9): p. 967-974.

207. White, N., Antimalarial drug resistance and combination chemotherapy. Philos Trans R Soc Lond B Biol Sci, 1999. 354(1384): p. 739-49.

208. Kamya, M.R., et al., Artemether-Lumefantrine versus Dihydroartemisinin-Piperaquine for Treatment of Malaria: A Randomized Trial. PLoS Clin Trials, 2007. 2(5): p. e20.

209. Molyneux, M.E., et al., Reduced hepatic blood flow and intestinal malabsorption in severe falciparum malaria. Am J Trop Med Hyg, 1989. 40(5): p. 470-6.

210. Roche, R.J., et al., Quinine induces reversible high-tone hearing loss. Br J Clin Pharmacol, 1990. 29(6): p. 780-2.

211. Touze, J.E., et al., The effects of antimalarial drugs on ventricular repolarization. Am J Trop Med Hyg, 2002. 67(1): p. 54-60.

212. White, N.J., et al., Quinine loading dose in cerebral malaria. Am J Trop Med Hyg, 1983. 32(1): p. 1-5.

213. Newton, P.N., et al., Early treatment failure in severe malaria resulting from abnormally low plasma quinine concentrations. Trans R Soc Trop Med Hyg, 2006. 100(2): p. 184-6.

214. Arya, T.V., et al., Spontaneous and quinine induced hypoglycaemia in severe falciparum malaria. Trop Geogr Med, 1989. 41(1): p. 73-5.

292

215. White, N.J., Clinical pharmacokinetics and pharmacodynamics of artemisinin and derivatives. Trans R Soc Trop Med Hyg, 1994. 88 Suppl 1: p. S41-3.

216. Bouchaud, O., et al., Severe cardiac toxicity due to halofantrine: importance of underlying heart disease. J Travel Med, 2002. 9(4): p. 214-5.

217. Krishna, S., et al., Pharmacokinetics, efficacy and toxicity of parenteral halofantrine in uncomplicated malaria. Br J Clin Pharmacol, 1993. 36(6): p. 585-91.

218. Taylor, W.R. and N.J. White, Antimalarial drug toxicity: a review. Drug Saf, 2004. 27(1): p. 25-61.

219. Faiz, M.A., et al., A randomized controlled trial comparing artemether and quinine in the treatment of cerebral malaria in Bangladesh. Indian J Malariol, 2001. 38(1-2): p. 9-18.

220. Karbwang, J., et al., Comparison of artemether and quinine in the treatment of severe falciparum malaria in south-east Thailand. Trans R Soc Trop Med Hyg, 1995. 89(6): p. 668-71.

221. Murphy, S., et al., An open randomized trial of artemether versus quinine in the treatment of cerebral malaria in African children. Trans R Soc Trop Med Hyg, 1996. 90(3): p. 298-301.

222. Olumese, P.E., et al., Comparative efficacy of intramuscular artemether and intravenous quinine in Nigerian children with cerebral malaria. Acta Trop, 1999. 73(3): p. 231-6.

223. Pe Than, M. and S. Tin, A controlled clinical trial of artemether (qinghaosu derivative) versus quinine in complicated and severe falciparum malaria. Trans R Soc Trop Med Hyg, 1987. 81(4): p. 559-61.

224. Seaton, R.A., et al., Randomized comparison of intramuscular artemether and intravenous quinine in adult, Melanesian patients with severe or complicated, Plasmodium falciparum malaria in Papua New Guinea. Ann Trop Med Parasitol, 1998. 92(2): p. 133-9.

225. Taylor, T.E., et al., Intramuscular artemether vs intravenous quinine: an open, randomized trial in Malawian children with cerebral malaria. Trop Med Int Health, 1998. 3(1): p. 3-8.

226. Tran, T.H., et al., A controlled trial of artemether or quinine in Vietnamese adults with severe falciparum malaria. N Engl J Med, 1996. 335(2): p. 76-83.

227. van Hensbroek, M.B., et al., A trial of artemether or quinine in children with cerebral malaria. N Engl J Med, 1996. 335(2): p. 69-75.

228. Walker, O., et al., An open randomized comparative study of intramuscular artemether and intravenous quinine in cerebral malaria in children. Trans R Soc Trop Med Hyg, 1993. 87(5): p. 564-6.

229. Mithwani, S., et al., Population pharmacokinetics of artemether and dihydroartemisinin following single intramuscular dosing of artemether in African children with severe falciparum malaria. Br J Clin Pharmacol, 2004. 57(2): p. 146-52.

230. Gomes, M.F., et al., Pre-referral rectal artesunate to prevent death and disability in severe malaria: a placebo-controlled trial. Lancet, 2009. 373(9663): p. 557-66.

231. WHO, Use of rectal artesmisinin based suppositories in the management of severe malaria. Report of a WHO Informal Consultation, 27-28 March 2006. 2007.

232. Krishna, S., et al., Bioavailability and preliminary clinical efficacy of intrarectal artesunate in Ghanaian children with moderate malaria. Antimicrob Agents Chemother, 2001. 45(2): p. 509-16.

233. Benakis, A., et al., Pharmacokinetic study of a new pharmaceutical form of artesunate (Plasmotrim-200 Rectocaps) administered in healthy volunteers by the rectal route. Jap J Trop Med Hyg, 1996. 24 (Suppl 1): p. 39-45.

234. Gomez-Landires, E.A., M.H. Jurado, and N. Cambon, Randomised efficacy and safety study of two 3-day artesunate rectal capsule/mefloquine regimens versus artesunate alone for uncomplicated malaria in Ecuadorian children. Acta Tropica, 2003. 89: p. 47-53.

235. Thwe, Y., et al., Artesunate suppository-mefloquine tablets (Plasmotrim, Rectocaps, Mefloquine, Lactab) in the treatment fo severe falciparum malaria. Jap J Trop Med Hyg, 1996. 24: p. 25-32.

236. Looareesuwan, S., P. Wilairatana, and M. Andrial, Artesunate suppository for the treatment of severe falciparum malaria in Thailand. Jap J Trop Med Hyg, 1996. 24: p. 13-15.

237. Awad, M.I., et al., Descriptive study on the efficacy and safety of artesunate suppository in combination with other antimalarials in the treatment of severe malaria in Sudan. Am J Trop Med Hyg, 2003. 68(2): p. 153-8.

293

238. Halpaap, B., et al., Plasma levels of artesunate and dihydroartemisinin in children with Plasmodium falciparum malaria in Gabon after administration of 50-milligram artesunate suppositories. Am J Trop Med Hyg, 1998. 58(3): p. 365-8.

239. Gomez, E.A., M.H. Jurado, and N. Cambon, Randomised efficacy and safety study of two 3-day artesunate rectal capsule/mefloquine regimens versus artesunate alone for uncomplicated malaria in Ecuadorian children. Acta Trop, 2003. 89(1): p. 47-53.

240. Looareesuwan, S., et al., A comparative clinical trial of sequential treatments of severe malaria with artesunate suppository followed by mefloquine in Thailand. Am J Trop Med Hyg, 1997. 57(3): p. 348-53.

241. Looareesuwan, S., et al., Efficacy and tolerability of a sequential, artesunate suppository plus mefloquine, treatment of severe falciparum malaria. Ann Trop Med Parasitol, 1995. 89(5): p. 469-75.

242. Sabchareon, A., et al., Comparative clinical trial of artesunate suppositories and oral artesunate in combination with mefloquine in the treatment of children with acute falciparum malaria. Am J Trop Med Hyg, 1998. 58(1): p. 11-6.

243. Wilairatana, P., et al., Artesunate suppositories: an effective treatment for severe falciparum malaria in rural areas. Ann Trop Med Parasitol, 1997. 91(7): p. 891-6.

244. WHO, FDA Briefing Document for the Anti-Infective Drug Products Advisory Committee: Artesunate Rectal Capsules. 2002.

245. TDR/GEN/01.5, Special Programme for Research and Training in Tropical Diseases. 15th Programme Report. Progress 1999-2000.

246. Kidane, G. and R.H. Morrow, Teaching mothers to provide home treatment of malaria in Tigray, Ethiopia: a randomised trial. Lancet, 2000. 356(9229): p. 550-5.

247. Simpson, J.A., et al., Population pharmacokinetics of artesunate and dihydroartemisinin following intra-rectal dosing of artesunate in malaria patients. PLoS Med, 2006. 3(11): p. e444.

248. Ashton, M., et al., Artemisinin kinetics and dynamics during oral and rectal treatment of uncomplicated malaria. Clin Pharmacol Ther, 1998. 63(4): p. 482-93.

249. Khanh, N.X., et al., Declining concentrations of dihydroartemisinin in plasma during 5-day oral treatment with artesunate for Falciparum malaria. Antimicrob Agents Chemother, 1999. 43(3): p. 690-2.

250. Newton, P., et al., Antimalarial bioavailability and disposition of artesunate in acute falciparum malaria. Antimicrob Agents Chemother, 2000. 44(4): p. 972-7.

251. McGready, R., et al., The pharmacokinetics of atovaquone and proguanil in pregnant women with acute falciparum malaria. Eur J Clin Pharmacol, 2003. 59(7): p. 545-52.

252. McGready, R., et al., Pharmacokinetics of dihydroartemisinin following oral artesunate treatment of pregnant women with acute uncomplicated falciparum malaria. Eur J Clin Pharmacol, 2006. 62(5): p. 367-71.

253. Na Bangchang, K., et al., Mefloquine pharmacokinetics in pregnant women with acute falciparum malaria. Trans R Soc Trop Med Hyg, 1994. 88(3): p. 321-3.

254. Parise, M.E., et al., Efficacy of sulfadoxine-pyrimethamine for prevention of placental malaria in an area of Kenya with a high prevalence of malaria and human immunodeficiency virus infection. Am J Trop Med Hyg, 1998. 59(5): p. 813-22.

255. van Eijk, A.M., et al., Effectiveness of intermittent preventive treatment with sulphadoxine-pyrimethamine for control of malaria in pregnancy in western Kenya: a hospital-based study. Trop Med Int Health, 2004. 9(3): p. 351-60.

256. Attenborough, R. and M.P. Alpers, Human Biology in Papua New Guinea: The Small Cosmos. 1993, USA: Oxford University Press.

257. Mueller, I., et al., Malaria control in Papua New Guinea results in complex epidemiological changes. P N G Med J, 2005. 48(3-4): p. 151-7.

258. Maitland, K., T.N. Williams, and C.I. Newbold, Plasmodium vevax and P. falciparum: Biological interactions and the possibility of cross-species immunity. Parasitol Today, 1997. 13(6): p. 227-31.

259. National Health Plan 2001-2010. 2000, Papua New Gunea Ministry of Health.

294

260. Genton, B., et al., The epidemiology of malaria in the Wosera area, East Sepik Province, Papua New Guinea, in preparation for vaccine trials. II. Mortality and morbidity. Ann Trop Med Parasitol, 1995. 89(4): p. 377-90.

261. Reeder, J.C., John C Reeder--Director of the Papua New Guinea Institute of Medical Research. Interview by Pam Das. Lancet Infect Dis, 2004. 4(6): p. 376-80.

262. Cattani, J.A., et al., Small-area variations in the epidemiology of malaria in Madang Province. P N G Med J, 1986. 29(1): p. 11-7.

263. Moir, J.S., et al., Mortality in a rural area of Madang Province, Papua New Guinea. Ann Trop Med Parasitol, 1989. 83(3): p. 305-19.

264. Allen, S.J., et al., Causes of preterm delivery and intrauterine growth retardation in a malaria endemic region of Papua New Guinea. Arch Dis Child Fetal Neonatal Ed, 1998. 79(2): p. F135-40.

265. Brabin, B. and C. Piper, Anaemia- and malaria-attributable low birthweight in two populations in Papua New Guinea. Ann Hum Biol, 1997. 24(6): p. 547-55.

266. Imbert, P., et al., Severe falciparum malaria in children: a comparative study of 1990 and 2000 WHO criteria for clinical presentation, prognosis and intensive care in Dakar, Senegal. Trans R Soc Trop Med Hyg, 2002. 96(3): p. 278-81.

267. Varandas, L., et al., Independent indicators of outcome in severe paediatric malaria: maternal education, acidotic breathing and convulsions on admission. Ann Trop Paediatr, 2000. 20(4): p. 265-71.

268. Jarolim, P., et al., Deletion in erythrocyte band 3 gene in malaria-resistant Southeast Asian ovalocytosis. Proc Natl Acad Sci U S A, 1991. 88(24): p. 11022-6.

269. Genton, B., et al., Ovalocytosis and cerebral malaria. Nature, 1995. 378(6557): p. 564-5. 270. Mgone, C.S., et al., Occurrence of the erythrocyte band 3 (AE1) gene deletion in relation to

malaria endemicity in Papua New Guinea. Trans R Soc Trop Med Hyg, 1996. 90(3): p. 228-31.

271. Chong, S.S., et al., Single-tube multiplex-PCR screen for common deletional determinants of alpha-thalassemia. Blood, 2000. 95(1): p. 360-2.

272. Patel, S.S., et al., The association of the glycophorin C exon 3 deletion with ovalocytosis and malaria susceptibility in the Wosera, Papua New Guinea. Blood, 2001. 98(12): p. 3489-91.

273. Ewers, W.H., Robert Koch, his work in New Guinea and hiscontribution to malariology. Papua New Guinea Medical Journal, 1972. 15: p. 117-24.

274. al-Yaman, F., et al., Resistance of Plasmodium falciparum malaria to amodiaquine, chloroquine and quinine in the Madang Province of Papua New Guinea, 1990-1993. P N G Med J, 1996. 39(1): p. 16-22.

275. Yung, A.P. and N.M. Bennett, Chloroquine-resistant falciparum malaria acquired in Papua New Guinea. Med J Aust, 1976. 2(22): p. 845.

276. Marfurt, J., et al. In vivo and molecular monitoring of antimalarial drug resistance in Papua New Guinea. in Medicine and health in the tropics. 2005. Marseilles, France.

277. Mehlotra, R.K., et al., Evolution of a unique Plasmodium falciparum chloroquine-resistance phenotype in association with pfcrt polymorphism in Papua New Guinea and South America. Proc Natl Acad Sci U S A, 2001. 98(22): p. 12689-94.

278. Rieckmann, K.H., D.R. Davis, and D.C. Hutton, Plasmodium vivax resistance to chloroquine? Lancet, 1989. 2(8673): p. 1183-4.

279. Whitby, M., et al., Chloroquine-resistant Plasmodium vivax. Lancet, 1989. 2(8676): p. 1395. 280. Marfurt, J., et al., Low Efficacy of Amodiaquine or Chloroquine Plus Sulfadoxine-

Pyrimethamine against Plasmodium falciparum and P. vivax Malaria in Papua New Guinea. Am J Trop Med Hyg, 2007. 77(5): p. 947-54.

281. Ratcliff, A., et al., Therapeutic response of multidrug-resistant Plasmodium falciparum and P. vivax to chloroquine and sulfadoxine-pyrimethamine in southern Papua, Indonesia. Trans R Soc Trop Med Hyg, 2007. 101(4): p. 351-9.

282. Sumawinata, I.W., et al., Very high risk of therapeutic failure with chloroquine for uncomplicated Plasmodium falciparum and P. vivax malaria in Indonesian Papua. Am J Trop Med Hyg, 2003. 68(4): p. 416-20.

283. Stace, J., et al., Serum levels of quinine following intramuscular administration to children. P N G Med J, 1983. 26(1): p. 21-4.

295

284. Darlow, B., et al., Fansidar-resistant falciparum malaria in Papua New Guinea. Lancet, 1980. 2(8206): p. 1243.

285. Charoenteeraboon, J., et al., Inactivation of artemisinin by thalassemic erythrocytes. Biochem Pharmacol, 2000. 59(11): p. 1337-44.

286. Ittarat, W., et al., Effects of alpha-thalassemia on pharmacokinetics of the antimalarial agent artesunate. Antimicrob Agents Chemother, 1998. 42(9): p. 2332-5.

287. Kamchonwongpaisan, S., et al., Resistance to artemisinin of malaria parasites (Plasmodium falciparum) infecting alpha-thalassemic erythrocytes in vitro. Competition in drug accumulation with uninfected erythrocytes. J Clin Invest, 1994. 93(2): p. 467-73.

288. Senok, A.C., et al., Thalassaemia trait, red blood cell age and oxidant stress: effects on Plasmodium falciparum growth and sensitivity to artemisinin. Trans R Soc Trop Med Hyg, 1997. 91(5): p. 585-9.

289. Vattanaviboon, P., P. Wilairat, and Y. Yuthavong, Binding of dihydroartemisinin to hemoglobin H: role in drug accumulation and host-induced antimalarial ineffectiveness of alpha- thalassemic erythrocytes. Mol Pharmacol, 1998. 53(3): p. 492-6.

290. Heinzel, G. and P. Tanswell, Compartmental analysis methods manual, in Topfit 2.0 Pharmacokinetic and phamacodynamic data analysis system for the PC, G. Heinzel, R. Woloszcak, and P. Thomann, Editors. 1993: Stuttgart. p. 5-140.

291. Thomann, P., Non-compartmental analysis methods manual, in TopFit 2.0 Pharmacokinetic and pharmacodynamic data analysis system for the PC, G. Heinzel, R. Woloszcak, and P. Thomann, Editors. 1993, Gustav Fisher: Stuttgart. p. 5-66.

292. Hung, T.Y., Pharmacokinetics of piperaquine in humans, in School of Medicine and Pharmacology. 2003, University of Western Australia: Perth.

293. Frisk-Holmberg, M., et al., The single dose kinetics of chloroquine and its major metabolite desethylchloroquine in healthy subjects. Eur J Clin Pharmacol, 1984. 26(4): p. 521-30.

294. Adjepon-Yamoah, K.K., et al., Whole-blood single-dose kinetics of chloroquine and desethylchloroquine in Africans. Ther Drug Monit, 1986. 8(2): p. 195-9.

295. Looareesuwan, S., et al., Cardiovascular toxicity and distribution kinetics of intravenous chloroquine. Br J Clin Pharmacol, 1986. 22(1): p. 31-6.

296. White, N.J., et al., Chloroquine treatment of severe malaria in children. Pharmacokinetics, toxicity, and new dosage recommendations. N Engl J Med, 1988. 319(23): p. 1493-500.

297. Aderounmu, A.F., et al., Comparison of the pharmacokinetics of chloroquine after single intravenous and intramuscular administration in healthy Africans. Br J Clin Pharmacol, 1986. 22(5): p. 559-64.

298. Bustos, D.G., et al., Pharmacokinetics of sequential and simultaneous treatment with the combination chloroquine and sulfadoxine-pyrimethamine in acute uncomplicated Plasmodium falciparum malaria in the Philippines. Trop Med Int Health, 2002. 7(7): p. 584-91.

299. Edwards, G., et al., Pharmacokinetics of chloroquine in Thais: plasma and red-cell concentrations following an intravenous infusion to healthy subjects and patients with Plasmodium vivax malaria. Br J Clin Pharmacol, 1988. 25(4): p. 477-85.

300. Gustafsson, L.L., et al., Disposition of chloroquine in man after single intravenous and oral doses. Br J Clin Pharmacol, 1983. 15(4): p. 471-9.

301. Na-Bangchang, K., et al., The pharmacokinetics of chloroquine in healthy Thai subjects and patients with Plasmodium vivax malaria. Br J Clin Pharmacol, 1994. 38(3): p. 278-81.

302. Walker, O., et al., The disposition of chloroquine in healthy Nigerians after single intravenous and oral doses. Br J Clin Pharmacol, 1987. 23(3): p. 295-301.

303. Wetsteyn, J.C., et al., The pharmacokinetics of three multiple dose regimens of chloroquine: implications for malaria chemoprophylaxis. Br J Clin Pharmacol, 1995. 39(6): p. 696-9.

304. Looareesuwan, S., et al., High rate of Plasmodium vivax relapse following treatment of falciparum malaria in Thailand. Lancet, 1987. 2(8567): p. 1052-5.

305. Hutagalung, R., et al., A randomized trial of artemether-lumefantrine versus mefloquine-artesunate for the treatment of uncomplicated multi-drug resistant Plasmodium falciparum on the western border of Thailand. Malar J, 2005. 4: p. 46.

306. Dupont, W.D. and W.D. Plummer, Jr, Power and sample size calculations. A review and computer program. . Control Clin Trials, 1990. 11: p. 116-28.

296

307. WHO, The WHO child growth standards. 2005, Geneva, Switzerland: World Health Organisation.

308. Cattamanchi, A., et al., Distinguishing recrudescence from reinfection in a longitudinal antimalarial drug efficacy study: comparison of results based on genotyping of msp-1, msp-2, and glurp. Am J Trop Med Hyg, 2003. 68(2): p. 133-9.

309. Felger, I. and H.P. Beck, Genotyping of Plasmodium falciparum. PCR-RFLP analysis. Methods Mol Med, 2002. 72: p. 117-29.

310. Davis, T.M., Antimalarial drugs and glucose metabolism. Br J Clin Pharmacol, 1997. 44(1): p. 1-7.

311. Myint, H.Y., et al., Efficacy and safety of dihydroartemisinin-piperaquine. Trans R Soc Trop Med Hyg, 2007. 101(9): p. 858-66.

312. Davis, T.M., et al., In Vitro Interactions between Piperaquine, Dihydroartemisinin, and Other Conventional and Novel Antimalarial Drugs. Antimicrob Agents Chemother, 2006. 50(8): p. 2883-5.

313. Fivelman, Q.L., I.S. Adagu, and D.C. Warhurst, Effects of piperaquine, chloroquine, and amodiaquine on drug uptake and of these in combination with dihydroartemisinin against drug-sensitive and -resistant Plasmodium falciparum strains. Antimicrob Agents Chemother, 2007. 51(6): p. 2265-7.

314. Price, R.N., et al., Clinical and pharmacological determinants of the therapeutic response to dihydroartemisinin-piperaquine for drug-resistant malaria. Antimicrob Agents Chemother, 2007. 51(11): p. 4090-7.

315. Price, R.N., G. Dorsey, and F. Nosten, Antimalarial therapies in children from Papua New Guinea. N Engl J Med, 2009. 360(12): p. 1254; author reply 1255.

316. Zongo, I., et al., Randomized comparison of amodiaquine plus sulfadoxine-pyrimethamine, artemether-lumefantrine, and dihydroartemisinin-piperaquine for the treatment of uncomplicated Plasmodium falciparum malaria in Burkina Faso. Clin Infect Dis, 2007. 45(11): p. 1453-61.

317. Ashley, E.A., et al., Randomized, controlled dose-optimization studies of dihydroartemisinin-piperaquine for the treatment of uncomplicated multidrug-resistant falciparum malaria in Thailand. J Infect Dis, 2004. 190(10): p. 1773-82.

318. Ashley, E.A., et al., A randomized, controlled study of a simple, once-daily regimen of dihydroartemisinin-piperaquine for the treatment of uncomplicated, multidrug-resistant falciparum malaria. Clin Infect Dis, 2005. 41(4): p. 425-32.

319. Grande, T., et al., A randomised controlled trial to assess the efficacy of dihydroartemisinin-piperaquine for the treatment of uncomplicated falciparum malaria in peru. PLoS ONE, 2007. 2(10): p. e1101.

320. Janssens, B., et al., A randomized open study to assess the efficacy and tolerability of dihydroartemisinin-piperaquine for the treatment of uncomplicated falciparum malaria in Cambodia. Trop Med Int Health, 2007. 12(2): p. 251-9.

321. Tangpukdee, N., et al., An open randomized clinical trial of Artekin vs artesunate-mefloquine in the treatment of acute uncomplicated falciparum malaria. Southeast Asian J Trop Med Public Health, 2005. 36(5): p. 1085-91.

322. Checchi, F., et al., Supervised versus unsupervised antimalarial treatment with six-dose artemether-lumefantrine: pharmacokinetic and dosage-related findings from a clinical trial in Uganda. Malar J, 2006. 5: p. 59.

323. Dorsey, G., et al., The impact of age, temperature, and parasite density on treatment outcomes from antimalarial clinical trials in Kampala, Uganda. Am J Trop Med Hyg, 2004. 71(5): p. 531-6.

324. Kochar, D.K., et al., Plasmodium vivax malaria. Emerg Infect Dis, 2005. 11(1): p. 132-4. 325. Naqvi, R., et al., Outcome in severe acute renal failure associated with malaria. Nephrol

Dial Transplant, 2003. 18(9): p. 1820-3. 326. Ozen, M., et al., Cerebral malaria owing to Plasmodium vivax: case report. Ann Trop

Paediatr, 2006. 26(2): p. 141-4. 327. Verret, W.J., et al., The effect of varying analytical methods on estimates of anti-malarial

clinical efficacy. Malar J, 2009. 8: p. 77.

297

328. Greenhouse, B., et al., Impact of Transmission Intensity on the Accuracy of Genotyping To Distinguish Recrudescence from New Infection in Antimalarial Clinical Trials. Antimicrob. Agents Chemother., 2007. 51(9): p. 3096-3103.

329. Juliano, J.J., S.M. Taylor, and S.R. Meshnick, Polymerase Chain Reaction Adjustment in Antimalarial Trials: Molecular Malarkey? J Infect Dis, 2009. 200(1): p. 5-7.

330. Snounou, G. and H.P. Beck, The use of PCR genotyping in the assessment of recrudescence or reinfection after antimalarial drug treatment. Parasitol Today, 1998. 14(11): p. 462-7.

331. Babalola, C.P., et al., Pharmacokinetics of quinine in African patients with acute falciparum malaria. Pharm World Sci, 1998. 20(3): p. 118-22.

332. Nies, A.S. and S.P. Spielberg, Principles of Therapeutics, in Goodman and Gilman's The Pharmacological Basis of Therapeutics, J.G. Hardman, et al., Editors. 1996, McGraw-Hill: New York. p. 43-62.

333. Dean, A., et al., Epi Info. 1997, Centers for Disease Control and Prevention: Atlanta, Georgia, USA.

334. Beal, S.L. and L.B. Sheiner, NONMEM User Guide. 1992, San Fransisco, USA: NONMEM Project Group, Univesity of California.

335. Ilett, K.F., et al., Glucuronidation of dihydroartemisinin in vivo and by human liver microsomes and expressed UDP-glucuronosyltransferases. Drug Metab Dispos, 2002. 30(9): p. 1005-12.

336. Lee, I.S. and C.D. Hufford, Metabolism of antimalarial sesquiterpene lactones. Pharmacol Ther, 1990. 48(3): p. 345-55.

337. Batty, K.T., et al., A pharmacokinetic and pharmacodynamic study of artesunate for vivax malaria. Am J Trop Med Hyg, 1998. 59(5): p. 823-7.

338. Batty, K.T., et al., A pharmacokinetic and pharmacodynamic study of intravenous vs oral artesunate in uncomplicated falciparum malaria. Br J Clin Pharmacol, 1998. 45(2): p. 123-9.

339. White, N.J. and S. Krishna, Treatment of malaria: some considerations and limitations of the current methods of assessment. Trans R Soc Trop Med Hyg, 1989. 83(6): p. 767-77.

340. McIntosh, H.M., Chloroquine or amodiaquine combined with sulfadoxine-pyrimethamine for treating uncomplicated malaria. Cochrane Database Syst Rev, 2001(4): p. CD000386.

341. Pengsaa, K., et al., Life-saving rectal artesunate for complicated malaria in children. Southeast Asian J Trop Med Public Health, 2005. 36(3): p. 597-601.

342. Benakis, A., et al., Pharmacokinetics/Pharmacodynamics findings after repeated administration of ARTESUNATE thermostable suppositories (RECTOCAPS) in Vietnamese patients with uncomplicated malaria. Eur J Drug Metab Pharmacokinet, 2006. 31(1): p. 41-5.

343. Bhatt, K.M., et al., Efficacy and tolerability of a sequential artesunate suppository-mefloquine treatment of severe falciparum malaria. Jpn J. Trop. Med. Hyg., 1996. 24(Suppl. 1): p. 59-63.

344. Barradell, L.B. and A. Fitton, Artesunate. A review of its pharmacology and therapeutic efficacy in the treatment of malaria. Drugs, 1995. 50(4): p. 714-41.

345. White, N.J., Artemisinin: current status. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1994. 88(Supplement 1): p. 3-4.

346. Hassan Alin, M., et al., Multiple dose pharmacokinetics of oral artemisinin and comparison of its efficacy with that of oral artesunate in falciparum malaria patients. Trans R Soc Trop Med Hyg, 1996. 90(1): p. 61-5.

347. Silamut, K., et al., Artemether bioavailability after oral or intramuscular administration in uncomplicated falciparum malaria. Antimicrob Agents Chemother, 2003. 47(12): p. 3795-8.

348. Zhang, S.Q., et al., Multiple dose study of interactions between artesunate and artemisinin in healthy volunteers. Br J Clin Pharmacol, 2001. 52(4): p. 377-85.

349. Svensson, U.S., et al., Characterisation of the human liver in vitro metabolic pattern of artemisinin and auto-induction in the rat by use of nonlinear mixed effects modelling. Biopharm Drug Dispos, 2003. 24(2): p. 71-85.

350. Vattanaviboon, P., et al., Membrane heme as a host factor in reducing effectiveness of dihydroartemisinin. Biochem Pharmacol, 2002. 64(1): p. 91-8.

298

351. Angus, B.J., et al., Oral artesunate dose-response relationship in acute falciparum malaria. Antimicrob Agents Chemother, 2002. 46(3): p. 778-82.

352. Organisation, R.B.M.W.H., Position of WHO's Roll Back Malaria Department on malaria treatment policy. Statement. 2004.

353. Allen, S.J., et al., Prevention of cerebral malaria in children in Papua New Guinea by southeast Asian ovalocytosis band 3. Am J Trop Med Hyg, 1999. 60(6): p. 1056-60.

354. Tanner, M.J., Molecular and cellular biology of the erythrocyte anion exchanger (AE1). Semin Hematol, 1993. 30(1): p. 34-57.

355. Hien, T.T., et al., Comparative pharmacokinetics of intramuscular artesunate and artemether in patients with severe falciparum malaria. Antimicrob Agents Chemother, 2004. 48(11): p. 4234-9.

356. Murphy, S.A., et al., The disposition of intramuscular artemether in children with cerebral malaria; a preliminary study. Trans R Soc Trop Med Hyg, 1997. 91(3): p. 331-4.

357. Berkley, J., et al., Bacteraemia complicating severe malaria in children. Trans R Soc Trop Med Hyg, 1999. 93(3): p. 283-6.

358. Karunajeewa, H.A., et al., Disposition of artesunate and dihydroartemisinin after administration of artesunate suppositories in children from Papua New Guinea with uncomplicated malaria. Antimicrob Agents Chemother, 2004. 48(8): p. 2966-72.

359. Gomez Landires, E.A., Efficacy of artesunate suppository followed by oral mefloquine in the treatment of severe falciparum malaria in endemic areas whre resistance to chloroquine exists in Ecuador. Jpn J. Trop. Med. Hyg., 1996. 24(Suppl. 1): p. 17-24.

360. Karbwang, J., et al., Pharmacokinetics of intramuscular artemether in patients with severe falciparum malaria with or without acute renal failure. Br J Clin Pharmacol, 1998. 45(6): p. 597-600.

361. Ringwald, P., J. Bickii, and L.K. Basco, In vitro activity of antimalarials against clinical isolates of Plasmodium falciparum in Yaounde, Cameroon. Am J Trop Med Hyg, 1996. 55(3): p. 254-8.

362. Ringwald, P., J. Bickii, and L.K. Basco, In vitro activity of dihydroartemisinin against clinical isolates of Plasmodium falciparum in Yaounde, Cameroon. Am J Trop Med Hyg, 1999. 61(2): p. 187-92.

363. Kokwaro, G., Bioavailability of artemether. Br J Clin Pharmacol, 1997. 44: p. 305. 364. Winkvist, A., Water spirits, medicine-men and witches: avenues to successful reproduction

among the Abelam, Papua new Guinea. , in The anthropology of pregnancy loss., R. Cecil, Editor. 1996, Berg: Oxford. p. 59-73.

365. Kaona, F. and M. Tuba, A qualitative study to identify community structures for management of severe malaria: a basis for introducing rectal artesunate in the under five years children in Nakonde District of Zambia. BMC Public Health, 2005. 5(1): p. 28.

366. Allen, S.J., et al., alpha+-Thalassemia protects children against disease caused by other infections as well as malaria. Proc Natl Acad Sci U S A, 1997. 94(26): p. 14736-41.

367. Brabin, B.J., et al., A longitudinal study of splenomegaly in pregnancy in a malaria endemic area in Papua New Guinea. Trans R Soc Trop Med Hyg, 1988. 82(5): p. 677-81.

368. Edstein, M.D., Pharmacokinetics of sulfadoxine and pyrimethamine after Fansidar administration in man. Chemotherapy, 1987. 33(4): p. 229-33.

369. Gerhardy, C.L. and M. Garrett, Obstetrics and Gynaecology for Nurses and Midwives. 5 ed, ed. G.D.L. Mola and M. Voigt. 2002, Madang, Papua New Guinea: Lutheran School of Nursing.

370. Kochar, M.S., Management of postural hypotension. Curr Hypertens Rep, 2000. 2(5): p. 457-62.

371. McNamara, D.T., et al., Diagnosing infection levels of four human malaria parasite species by a polymerase chain reaction/ligase detection reaction fluorescent microsphere-based assay. Am J Trop Med Hyg, 2006. 74(3): p. 413-21.

372. Chesnutt, A.N., Physiology of normal pregnancy. Crit Care Clin, 2004. 20(4): p. 609-15. 373. Chukwuani, M.C., et al., Evidence for increased metabolism of chloroquine during the early

third trimester of human pregnancy. Trop Med Int Health, 2004. 9(5): p. 601-5. 374. Lee, S.J., et al., Chloroquine pharmacokinetics in pregnant and nonpregnant women with

vivax malaria. Eur J Clin Pharmacol, 2008. 74(10): p. 987-982.

299

375. Green, M.D., et al., Pharmacokinetics of sulfadoxine-pyrimethamine in HIV-infected and uninfected pregnant women in Western Kenya. J Infect Dis, 2007. 196(9): p. 1403-8.

376. Massele, A.Y., et al., Chloroquine blood concentrations and malaria prophylaxis in Tanzanian women during the second and third trimesters of pregnancy. Eur J Clin Pharmacol, 1997. 52(4): p. 299-305.

377. Hodge, L.S. and T.S. Tracy, Alterations in drug disposition during pregnancy: implications for drug therapy. Expert Opin Drug Metab Toxicol, 2007. 3(4): p. 557-71.

378. Kim, K.A., et al., Cytochrome P450 2C8 and CYP3A4/5 are involved in chloroquine metabolism in human liver microsomes. Arch Pharm Res, 2003. 26(8): p. 631-7.

379. Projean, D., et al., In vitro metabolism of chloroquine: identification of CYP2C8, CYP3A4, and CYP2D6 as the main isoforms catalyzing N-desethylchloroquine formation. Drug Metab Dispos, 2003. 31(6): p. 748-54.

380. Mansor, S.M., et al., Single dose kinetic study of the triple combination mefloquine/sulphadoxine/pyrimethamine (Fansimef) in healthy male volunteers. Br J Clin Pharmacol, 1989. 27(3): p. 381-6.

381. Obua, C., et al., Pharmacokinetic interactions between chloroquine, sulfadoxine and pyrimethamine and their bioequivalence in a generic fixed-dose combination in healthy volunteers in Uganda. Afr Health Sci, 2006. 6(2): p. 86-92.

382. Sarikabhuti, B., et al., Plasma concentrations of sulfadoxine in healthy and malaria infected Thai subjects. Acta Trop, 1988. 45(3): p. 217-24.

383. Schwartz, D.E., et al., Multiple-dose pharmacokinetics of the antimalarial drug Fansimef (pyrimethamine + sulfadoxine + mefloquine) in healthy subjects. Chemotherapy, 1987. 33(1): p. 1-8.

384. Wang, N.S., et al., Pharmacokinetics of the combination pyrimethamine with sulfadoxine and mefloquine (FANSIMEF) in Chinese volunteers and the relative bioavailability of a lacquered tablet. Chemotherapy, 1990. 36(3): p. 177-84.

385. Weidekamm, E., et al., Plasma concentrations in pyrimethamine and sulfadoxine and evaluation of pharmacokinetic data by computerized curve fitting. Bull World Health Organ, 1982. 60(1): p. 115-22.

386. Weidekamm, E., et al., Single-dose investigation of possible interactions between the components of the antimalarial combination Fansimef. Chemotherapy, 1987. 33(4): p. 259-65.

387. Corvaisier, S., et al., Population pharmacokinetics of pyrimethamine and sulfadoxine in children treated for congenital toxoplasmosis. Antimicrob Agents Chemother, 2004. 48(10): p. 3794-800.

388. Dzinjalamala, F.K., et al., Association between the pharmacokinetics and in vivo therapeutic efficacy of sulfadoxine-pyrimethamine in Malawian children. Antimicrob Agents Chemother, 2005. 49(9): p. 3601-6.

389. Obua, C., et al., Population pharmacokinetics of chloroquine and sulfadoxine and treatment response in children with malaria: suggestions for an improved dose regimen. Br J Clin Pharmacol, 2008. 65(4): p. 493-501.

390. Trenque, T., et al., Population pharmacokinetics of pyrimethamine and sulfadoxine in children with congenital toxoplasmosis. Br J Clin Pharmacol, 2004. 57(6): p. 735-41.

391. Hekster, C.A. and T.B. Vree, Clinical pharmacokinetics of sulphonamides and their N4-acetyl derivatives. Antibiot Chemother, 1982. 31: p. 22-118.

392. Cook, I.F., J.P. Cochrane, and M.D. Edstein, Race-linked differences in serum concentrations of dapsone, monoacetyldapsone and pyrimethamine during malaria prophylaxis. Trans R Soc Trop Med Hyg, 1986. 80(6): p. 897-901.

393. Williams, R.L., et al., Acetylator phenotype and response of individuals infected with a chloroquine-resistant strain of Plasmodium falciparum to sulfalene and pyrimethamine. Am J Trop Med Hyg, 1975. 24(5): p. 734-9.

394. Tett, S.E. and D.J. Cutler, Apparent dose-dependence of chloroquine pharmacokinetics due to limited assay sensitivity and short sampling times. Eur J Clin Pharmacol, 1987. 31(6): p. 729-31.

395. Smith, C.C. and J. Ihrig, Persistent excretion of pyrimethamine following oral administration. Am J Trop Med Hyg, 1959. 8(1): p. 60-2.

300

396. Mockenhaupt, F.P., et al., Diagnosis of placental malaria. J Clin Microbiol, 2002. 40(1): p. 306-8.

397. Nosten, F., et al., Effects of Plasmodium vivax malaria in pregnancy. Lancet, 1999. 354(9178): p. 546-9.

398. Yeka, A., et al., Artemether-lumefantrine versus dihydroartemisinin-piperaquine for treating uncomplicated malaria: a randomized trial to guide policy in Uganda. PLoS ONE, 2008. 3(6): p. e2390.

399. Hasugian, A.R., et al., Dihydroartemisinin-piperaquine versus artesunate-amodiaquine: superior efficacy and posttreatment prophylaxis against multidrug-resistant Plasmodium falciparum and Plasmodium vivax malaria. Clin Infect Dis, 2007. 44(8): p. 1067-74.

400. Hutagalung, R., et al., A case-control auditory evaluation of patients treated with artemether-lumefantrine. Am J Trop Med Hyg, 2006. 74(2): p. 211-4.

401. Van Vugt, M., et al., A case-control auditory evaluation of patients treated with artemisinin derivatives for multidrug-resistant Plasmodium falciparum malaria. Am J Trop Med Hyg, 2000. 62(1): p. 65-9.

402. Menendez, C., U. D'Alessandro, and F.O. ter Kuile, Reducing the burden of malaria in pregnancy by preventive strategies. Lancet Infect Dis, 2007. 7(2): p. 126-35.

403. Karunajeewa, H.A., et al., Pharmacokinetics and efficacy of piperaquine and chloroquine in melanesian children with uncomplicated malaria. Antimicrob Agents Chemother, 2008. 52(1): p. 237-43.

404. Karunajeewa, H.A., et al., The pharmacokinetic properties of sulfadoxine-pyrimethamine in pregnancy. Antimicrob Agents Chemother, 2009.

405. van Tulder, M., et al., Updated method guidelines for systematic reviews in the cochrane collaboration back review group. Spine, 2003. 28(12): p. 1290-9.

406. Makundi, E.A., et al., Role of traditional healers in the management of severe malaria among children below five years of age: the case of Kilosa and Handeni Districts, Tanzania. Malar J, 2006. 5: p. 58.

407. Awad, M.I., et al., Pharmacokinetics of artesunate following oral and rectal administration in healthy Sudanese volunteers. Trop Doct, 2004. 34(3): p. 132-5.

408. Karunajeewa, H.A., et al., Safety and therapeutic efficacy of artesunate suppositories for treatment of malaria in children in Papua New Guinea. Pediatr Infect Dis J, 2003. 22(3): p. 251-6.

409. Karunajeewa, H.A., et al., Artesunate suppositories versus intramuscular artemether for treatment of severe malaria in children in Papua New Guinea. Antimicrob Agents Chemother, 2006. 50(3): p. 968-74.

410. Navaratnam, V., et al., Comparative pharmacokinetic study of oral and rectal formulations of artesunic acid in healthy volunteers. Eur J Clin Pharmacol, 1998. 54(5): p. 411-4.

411. Andoh, J., et al., Clinical trial of a new form of pharmaceutical artesunate: Plasmotrim and its active metabolite dihydroartemisinine. Unpublished.

412. Arnold, K., et al., A randomized comparative study of artemisinine (qinghaosu) suppositories and oral quinine in acute falciparum malaria. Trans R Soc Trop Med Hyg, 1990. 84(4): p. 499-502.

413. Birku, Y., E. Makonnen, and A. Bjorkman, Comparison of rectal artemisinin with intravenous quinine in the treatment of severe malaria in Ethiopia. East Afr Med J, 1999. 76(3): p. 154-9.

414. Cao, X.T., et al., Comparison of artemisinin suppositories, intramuscular artesunate and intravenous quinine for the treatment of severe childhood malaria. Trans R Soc Trop Med Hyg, 1997. 91(3): p. 335-42.

415. Guo, X.B. and L.C. Fu, [Comparative study of artemisinin suppositories and piperaquine phosphate in the treatment of falciparum malaria]. Zhong Xi Yi Jie He Za Zhi, 1989. 9(8): p. 475-7, 453.

416. Ha, V., et al., Severe and complicated malaria treated with artemisinin, artesunate or artemether in Viet Nam. Trans R Soc Trop Med Hyg, 1997. 91(4): p. 465-7.

417. Hien, T.T., et al., Comparison of artemisinin suppositories with intravenous artesunate and intravenous quinine in the treatment of cerebral malaria. Trans R Soc Trop Med Hyg, 1992. 86(6): p. 582-3.

301

418. Koopmans, R., et al., The pharmacokinetics of artemisinin suppositories in Vietnamese patients with malaria. Trans R Soc Trop Med Hyg, 1998. 92(4): p. 434-6.

419. Koopmans, R., et al., The pharmacokinetics of artemisinin after administration of two different suppositories to healthy Vietnamese subjects. Am J Trop Med Hyg, 1999. 60(2): p. 244-7.

420. Li, G.Q., et al., Observation on the efficacy of qinghaosu suppository in 100 cases of falciparum malaria. J Tradit Chin Med, 1985. 5(3): p. 159-61.

421. Li, G.Q. and G. Xingbo. Clinical studies on artemisinin suppositories, artesunate and artemether. in World Conference on Clinical Pharmacology and Therapeutics. 1989. Mannheim-Heidelberg.

422. Li, Y.Q., [Effect of qinghaosu by rectal administration in the treatment of vivax malaria]. Ji Sheng Chong Xue Yu Ji Sheng Chong Bing Za Zhi, 1984. 2(4): p. 279.

423. Tran, T.H., et al., Single dose artemisinin-mefloquine treatment for acute uncomplicated falciparum malaria. Trans R Soc Trop Med Hyg, 1994. 88(6): p. 688-91.

424. Esamai, F., et al., Rectal dihydroartemisinin versus intravenous quinine in the treatment of severe malaria: a randomised clinical trial. East Afr Med J, 2000. 77(5): p. 273-8.

425. Wilairatna, P., et al., Clinical trial of sequential treatments of moderately severe and severe malaria with dihydroartemisinin suppository followed by mefloquine in Thailand. Am J Trop Med Hyg, 2000. 63(5-6): p. 290-4.

426. Zhao, K.C. and Z.Y. Song, [Pharmacokinetics of dihydroqinghaosu in human volunteers and comparison with qinghaosu]. Yao Xue Xue Bao, 1993. 28(5): p. 342-6.

427. Aceng, J.R., J.S. Byarugaba, and J.K. Tumwine, Rectal artemether versus intravenous quinine for the treatment of cerebral malaria in children in Uganda: randomised clinical trial. Bmj, 2005. 330(7487): p. 334.

428. Teja-Isavadharm, P., et al., Comparative bioavailability of oral, rectal, and intramuscular artemether in healthy subjects: use of simultaneous measurement by high performance liquid chromatography and bioassay. Br J Clin Pharmacol, 1996. 42(5): p. 599-604.

429. Hien, T.T., et al., Comparative effectiveness of artemisinin suppositories and oral quinine in children with acute falciparum malaria. Trans R Soc Trop Med Hyg, 1991. 85(2): p. 210-1.

430. Hien, T.T., An overview of the clinical use of artemisinin and its derivatives in the treatment of falciparum malaria in Viet Nam. Trans R Soc Trop Med Hyg, 1994. 88 Suppl 1: p. S7-8.

431. Barennes, H., et al., Safety and efficacy of rectal compared with intramuscular quinine for the early treatment of moderately severe malaria in children: randomised clinical trial. Bmj, 2006. 332(7549): p. 1055-9.

302

1

7. APPENDICES

2

3

7.1. Appendix A: Drug assay methodology 7.1.1. PIPERAQUINE

Materials, methods, assay validation and quality control for the PQ assays used in

studies in this thesis are described in detail in Publication 8. [403] Briefly, with CQ

used as internal standard, 1 ml samples were alkalinized with 0.1 mL of 1 M NaOH

and extracted with hexane:isoamyl alcohol (90:10). After centrifugation supernatant

was back extracted into 0.1 mL 0.05 M HCl by manual shaking and HCl layers

aspirated and re-centrifuged. The HPLC system comprised of Hewlett Packard

model 1100 with a gradient pump, auto-sampler and a variable wavelength UV

detector (Agilent Technology, Waldbronn, Germany).Aliquots (25 µL) were injected

onto on a Chromolith® Performance column (100 mm x 4.6 mm i.d.; E Merck GmbH,

Darmstadt, Germany) at 30oC with a mobile phase of 6% v/v acetonitrile in 50 mM

K2HPO4 buffer (pH 2.5) pumped at 2 mL/min. Analytes were detected at 340 nm

and quantified using Chemstation Software (Version 9, Agilent Technology,

Waldbronn, Germany). The usual linear assay range was 3-1867 nmol/L.

The limits of quantification (LOQ) and detection (LOD) were 3 nmol/L and 1.3

nmol/L, respectively. All samples were assayed within the frozen storage stability

limits previously established in our laboratory of 12 months. [292]

7.1.2. CHLOROQUINE

Materials, methods, assay validation and quality control for the CQ assays used in

studies in this thesis are also described in detail in Publication 8. [403] Briefly, using

PQ as internal standard, 1 ml samples were alkalinized with 0.1 mL of 1 M NaOH

and extracted with 8 mL of t-butyl methyl ether. Subsequent back-extraction, re-

centrifugation, injection onto a Chromolith® Performance column, detection at 340

nm and quantification using a Hewlett Packard model 1100 HPLC system and

Chemstation Software were identical to those used for PQ (above). The LOQ and

LOD for CQ were 6 and 3 nmol/L, and for DECQ were 4 and 2 nmol/L. All samples

were assayed within the frozen storage stability limits previously established in our

laboratory of 6 months for CQ (unpublished data).

4

7.1.3. ARTESUNATE AND DIHYDROARTEMISININ

Blood samples were immediately centrifuged at 5000 RPM for 5 min and separated

plasma stored frozen (–80oC) in fluoride oxalate tubes prior to transportation on dry

ice to Perth where assays were performed by liquid chromatography-mass

spectrometry (LC-MS). Materials, methods, assay validation and quality control for

these assays are described in detail in publication 2. [358] Briefly, thawed plasma

was first centrifuged at 14,000 g for 1 min to precipitate protein and the supernatant

subjected to solid phase extraction as previously described. [338] Assays used an

Agilent 1100 Series LC system (Aglient Technologies, Forest Hill, NSW, Australia).

Analytes were resolved on a Zorbax SB-C8 column (3.5 nm, 2.1 mm [inner

diameter] x 100mm, Aglient Technologies) with a Eclipse XDB-C8 guard column (5

mm; 2.1 mm [inner diameter] by 12.5 mm; Aglient Technologies). A mobile phase of

55% 0.01M LiOH.H2O and 45% acetonitrile was pumped at 200 �L/min. Eluting

peaks were detected using UV absorbance (210nm) and an MSD Trap (Agilent

1100 Series).

The mass spectrometer was operated using electro-spray ionization in positive ion

mode. ARTS and DHA both produced ion fragmentation patterns with one major

ion at 245 m/z. Nitrogen was used as the nebulizer gas at a pressure of 40 psi and

a vaporizer temperature of 325oC. The limits of quantitation and limits of detection

were 80 nmol/L and 50 nmol/L, and 50 nmol/L and 10 nmol/L, respectively for

ARTS and �-DHA. All samples were assayed within the frozen storage stability

limits previously established for ARTS and DHA (up to 12 months).

7.1.4. ARTEMETHER

Artemether and DHA after i.m. ARM were quantified by gas chromatography-mass

spectrometry. [355] Method performance assessed as intra- and inter-day relative

standard deviations over the relevant concentration ranges, was similar to that of

previous studies. [355]

7.1.5. SULPHADOXINE AND PYRIMETHAMINE

Materials, methods, assay validation and quality control for assays of SDOX and

PYR used in this thesis are described in detail in publication 14. [404] Briefly, for

5

SDOX, plasma standards and samples (50 µL) were aliquoted and spiked with

internal standard sulfamethazine (4 mg/L) and 1 mL of acetonitrile then mixed by

manual shaking for 5 min. After centrifugation at 1500 g for 10 min, 0.5 mL of

organic phase were transferred into round-bottomed borosilicate glass tubes and

evaporated to dryness at 45°C in a rotary evaporator. Residues were reconstituted

in 200 µL of the HPLC mobile phase and 50 µL aliquots were injected onto the

HPLC column.

Pyrimethamine was extracted as described by Trenque et al., [390] with the

following modifications. Briefly, plasma standards and samples (200 µL) were

spiked with internal standard (midazolam; 1000 µg/L) and mixed with 200 µL of 1M

NaOH and 1 mL of t-butylmethyl ether and manually shaken for 10 min. After

centrifugation at 1500 g for 10 min, 0.08 mL of the organic phase was transferred

into round-bottomed borosilicate glass tubes and evaporated to dryness at 45°C in

a rotary evaporator. Residues were reconstituted in 100 µL of the mobile phase and

50 µL aliquots were injected onto the HPLC column.

The HPLC system comprised of Hewlett Packard model 1100 with a gradient pump,

auto-sampler and a variable wavelength UV detector (Agilent Technology,

Waldbronn, Germany). Separation was performed on a LichrospherTM RP Select B

column (5 µm, 250 mm x 4 mm id; E Merck Gmbh, Damstadt, Germany) at 30°C.

For SDOX and NASDOX, the mobile phase of 25% v/v acetonitrile in 0.01% v/v

phosphoric acid and 0.01% w/v NaCl and was pumped at 1.5 mL/min. For PYR, the

mobile phase contained 30% v/v acetonitrile in 0.01% v/v phosphoric acid and

0.01% w/v NaCl and was pumped at 1 mL/min. All analytes were detected by their

UV absorbance at 270 nm and analysis of chromatograms undertaken using

Chemstation Software (Version 9, Agilent Technology, Waldbronn, Germany).

The LOQs for SDOX, NASDOX and PYR were 0.1 mg/L, 0.02 mg/L and 2.5 µg/L,

respectively. The LODs for SDOX, NASDOX and PYR were 0.05 mg/L, 0 .01 mg/L

and 1 µg/L, respectively.

6

7

7.2. Appendix B: Design of Questionnaire Tool for Use in Clinical Trial 5

Preliminary formative research was conducted by Rachael Hinton, PNG IMR in 6

villages in the Wosera district during November 2003 in order to inform the design

of a questionnaire for caregivers. Forty women with children under the age of 5

years were randomly selected from the PNG IMR demographic surveillance system

and asked to participate in in-depth interviews. An initial general enquiry into

mothers’ perceptions and practices relating to childhood fevers was used to provide

a context in which to investigate issues that may be important if ARTS suppositories

were to be introduced in the area and to define the terminology to be used in the

subsequent questionnaire. Questioning was open-ended and care was taken to

ensure that questions about suppositories were placed in the context of other

malaria treatment modalities, often by way of ranking exercises. For example,

mothers were shown 4 different treatment modes for malaria (suppository, tablet,

syrup and syringe), asked to rank them in order of preference, strength and

efficacy, potential for side-effects and preference for self-administration and to give

reasons for their preferences. Issues raised by mothers during the ranking

exercises included lack of knowledge of suppositories, fear of side-effects,

uncertainties regarding self-administration, concern about older children physically

resisting suppository administration and concerns about the effectiveness of

conventional treatments for malaria. These were incorporated into multiple-choice

options in the questionnaire for caregivers along with responses relating to shame,

embarrassment and perceptions of efficacy and harmful effects of suppositories.

8

9

7.3. Appendix C: Systematic review of pharmacokinetics, safety and efficacy of rectally administered artemisinin derivatives

7.3.1. SUMMARY

Aims: To review the pharmacokinetics (PK), efficacy, and safety of rectally

administered artesunate (ARTS), artemisinin (QHS), dihydroartemisinin (DHA), and

artemether (ARM).

Methods: Searches were performed of the MEDLINE, EMBASE, Cochrane

Database of Clinical Reviews, Global Health, Web of Science, and CINAHL

computerized databases up to December 2006, along with review of unpublished

data from conference proceedings, pharmaceutical companies, and regulatory

applications. Studies in languages other than English were translated. Studies were

included involving rectal administration of an artemisinin derivative to healthy

volunteers or patients with measurement of plasma drug concentrations or rates of

initial parasite clearance. Both single-arm and comparative trials were included.

Primary efficacy outcome measures included PC%12 and PC%24. Pharmacokinetic

variables included Cmax, tmax and AUC. Weighted means were calculated from

available data.

Results: Forty-five studies were identified, of which 39 eligible studies were

included in the review. Thirty-two studies provided valid clinical efficacy data: 19 of

ARTS, 10 of QHS, 2 of DHA, and 1 of ARM. All demonstrated prompt parasite

clearance, with evidence of a dose-dependent effect for ARTS. Mortality rates in

severe malaria (weighted means, 0%-13%) were consistent with those expected.

Eight studies compared rectal QHS with conventional parenteral treatment (quinine,

ARM, or ARTS) for severe malaria. Despite similar clinical outcomes, rectal

artemisinin derivatives initiated parasite clearance more rapidly than parenteral

treatment (percentage of baseline at 12 h, �27% vs �56%, respectively). Eighteen

PK studies were identified, including 13 of ARTS. There was marked inter-individual

variability in most PK variables, but ARTS achieved an earlier tmax and higher Cmax

and AUC than other artemisinin derivatives.

Conclusions Available rectal preparations of artemisinin derivatives differ in their

10

PK disposition. Most available evidence pertains to ARTS and QHS. Despite

marked inter-individual variability in bioavailability, rectal preparations appear to

have acceptable therapeutic efficacy, including in severe illness.

7.3.2. AIMS

To systematically review the evidence (published and unpublished) relating to the

PK safety and efficacy of all artemisinin derivatives (ARTS, QHS, ARM and DHA)

formulated for rectal administration.

7.3.3. METHODS

7.3.3.1. Acknowledgements


This study was conceived in collaboration with Tim Davis (TD). Laurens Manning

(LM: School of Medicine and Pharmacology, University of Western Australia)

performed simultaneous duplicate data searches and Tim Davis duplicated the

methodological quality assessment scoring of clinical trials. Tim Davis, Ken Ilett,

and Ivo Mueller (PNG IMR Goroka) assisted in editing the final published

manuscript (see publication 7).

I formulated and designed the study selection criteria, data extraction and data

synthesis strategies, performed the data collection, analysis and interpretation and

prepared the initial manuscript draft for publication (publication 7).


This study was a systematic review. A number of different pharmaceutical

preparations of different artemisinin derivatives are now available and have been

used in a variety of dosing regimens. However, it is unclear how the different

preparations compare with one another with respect to PK properties and efficacy.

There is a need to review and synthesise available data in order to formulate

recommendations as to the choice of drug and the regimen to use. The study was

designed to identify all possible relevant studies in order to maximize the quantity of

data available for analysis and to include non-English language publications and

un-published studies in order to minmise selection and publication biases. The

review included studies already included in this thesis (clinical studies 3 and 4b).

The PK and clinical outcomes chosen were those that were well-defined and

11

standardised between studies, and those thought to be of greatest clinical

importance in the setting in which rectal administration is likely to be most valuable.

Calculation of weighted means for each outcome enabled the data to be

summarised in a manner that took into account the sample sizes of each individual

study.

7.3.3.3. Data sources

Searches were performed of MEDLINE, EMBASE, Cochrane database of clinical

reviews, Global Health, WEB of SCIENCE and CINAHL computerized databases to

December 2006 with a strategy that used both key words and MeSH term searches

comprising (artesunate or artesunic acid, or dihydroartemisinin or dihydroqinghaosu

or artemisinin or qinghaosu or artemether) AND (rectal or suppository or

suppositories). Pharmaceutical companies were contacted for unpublished data,

conference proceedings reviewed and reference lists from retrieved articles used to

identify other relevant studies. Foreign language articles were translated into

English.

7.3.3.4. Study selection

Inclusion criteria were i) human clinical studies including healthy volunteers or

patients of any age, regardless of Plasmodium species or clinical status, in which

subject number, age and clinical status were specified, ii) rectal administration of an

artemisinin derivative, and iii) measurement of either plasma drug concentrations or

parasite clearance assessed from serial blood smears taken at least every 6 h. The

search strategy was conducted simultaneously and independently by two

investigators (HK and LM) and then cross-checked. Both single-arm and

comparative trials were included. For comparative clinical trials, methodological

quality was assessed independently by two reviewers (HK. and TD) using the

Maastricht-Amsterdam score list. [405] The 11 items relating to internal validity were

used and studies fulfilling >6 of these criteria were considered of high quality.

7.3.3.5. Outcome assessment

Primary efficacy variables of interest included PC%12 and PC%24, and/or PCT50. If

not reported, PC%12 and PC%24 were derived from raw data by dividing 12-hour

and 24-hour parasite densities by the baseline densities if available, or by

interpolation of graphical representations. A PC%12 of <60% has been used as a

breakpoint in comparing early parasite clearance between different antimalarial

12

regimens. [117] PCT, mortality and reported adverse events were also included.

Because raw data were not always available, measures of central tendency and

variance used in the original articles were reproduced as stated rather than

converted to a standardised methodology. Where calculations were performed on

raw data, mean ± SD and, for non-normally distributed data, median (range) were

used.

The PK variables of interest were maximal concentration (Cmax), time to Cmax (tmax),

and area under the plasma concentration-time curve (AUC). All drug concentrations

were converted to nanomoles using molecular weights of 284.4, 282.4 and 298.4

for DHA, QHS and ARM, respectively. Where studies did not provide mg/kg doses,

these were approximated by dividing the total administered dose by either the

reported mean subject body weight or, for adults, an assumed adult weight of 65 kg

for Caucasian or African subjects, or 50 kg for South-East Asian subjects.

7.3.3.6. Data synthesis

In order to summarize the efficacy and PK data for each drug, data from each study

were used to generate a weighted mean for each parameter according to the

formula (�1.n1 + �2.n2 + �3.n3…+ �t.nt) / (n1+ n2+ n3…+ nt) where � = the individual

study mean, n = the number of subjects in the study and t = the total number of

studies included. Where mean values were unavailable or could not be calculated

from published data, median values were substituted for �. Mortality data were

obtained from studies of moderately severe or severe malaria.

7.3.4. RESULTS

7.3.4.1. Description of included studies

Twenty-five studies of ARTS suppositories were identified. Of these, two were

excluded due to absence of PK or efficacy endpoints, [365, 406] leaving 23 studies.

Of these, 18 evaluated Plasmotrim Rectocaps® (Mepha Pharmaceuticals, Aesch-

Basel, Switzerland) administered as combinations of 50 and 200 mg suppositories.

[117, 232, 233, 235-242, 341-343, 358, 359, 407-410] Five were unpublished

(Joseph Andoh, MD, et al., unpublished data, February 10 March-March 19, 1998),

[244, 411] four of which were sourced from a submission made by WHO to the FDA

in 2002, [244] which included two studies detailing bioequivalence and clinical data

13

relating to another formulation of ARTS suppositories available as 100 and 400 mg

doses (WHO/TDR and Scanpharm A/S, Denmark).

Fourteen studies of rectally administered QHS were identified. [248, 412-423] Three

were excluded, one because it involved patients with P. vivax [422] and the others,

one of which was unpublished, [421] because they provided inadequate subject

details. [415, 421] Of four studies of DHA [180, 424-426] and two of ARM, [427,

428] all were included except for one DHA study with inadequate subject details.

[426]

14

Table 7.1: Pharmaceutical preparations of artemisinin derivatives for rectal

administration.

1 A further 2 identified studies of artesunate, 3 of artemisinin and 1 of

dihydroartemisinin were excluded from this review. 2 Details of these studies are described in a WHO FDA application from 2002 [244];

one additional unpublished study by Andoh [411] used the Mepha product 3 The manufacturer of the suppositories is not specified in many of these studies 4 This formulation has not been marketed for rectal administration

Generic Drug

Current Manufacturer (s)

Trade name, dose strength and formulation

Published studies

Unpublished studies

Mepha, Aesch-Basel, Switzerland

Plasmotrim Rectocaps ® 50mg and 200mg suppositories

18 3 2

Artesunate1

Scanpharm, Denmark

100mg, 200mg. 400mg suppositories

0 2 2

Artemisinin 1

Mediplantex, Vietnam

Vidipha, Vietnam

Polyethylene glycol (PEGS) 500mg suppositories

Fatty (FS) 400mg suppositories

113 0

Dafra Pharma, Belgium

Rectal artemether suppogels ®

1 0 Artemether

Kunming Pharmaceuticals, China

Artemether in arachis oil (parenteral preparation) 4

1 0

Beijing Holley-Cotec China

Cotecxin ®

3 0 Dihydroartemisinin 1

GVS Laboratories, India

Alaxin ® 0 0

15

7.3.4.2. Clinical efficacy

7.3.4.2.1. Artesunate

A total of 356 ARTS-treated patients were enrolled in studies of severe malaria and

287 in studies of moderately-severe infection (Table 7.2 A). For studies of severe

disease, the definitions of severity varied and were sometimes poorly specified,

complicating between-study comparisons. Because there was significant

heterogeneity in dosing regimens for rectal ARTS, these data were dichotomized

into initial doses of i) <5 mg/kg and ii) >5 mg/kg. Several studies employed a low-

dose regimen with 1.8-4.4 mg/kg administered repeatedly over the first 12 h (Table

7.2 A i). [235, 237, 238, 240, 342, 343] These showed generally slower parasite

clearance during the first 12 h and a weighted mean PCT that was twice as long as

that seen with the higher-dose regimen (Table 7.2 A ii). [130, 232, 239, 244, 341,

359, 409-411] There were 6 deaths in ARTS-treated patients enrolled in studies of

severe malaria (1.7%) and 2 in studies of moderately-severe infection (0.7%).

7.3.4.2.2. Artemisinin

The QHS studies identified involved 350 patients and, with the exception of one

African study of severely ill adults, [413] were performed in Vietnam or China in

patients with infections ranging from uncomplicated to severe (Table 7.2 B). [248,

412-414, 416-418, 420, 423] Few calculated, or provided data necessary to

calculate, PC%12 or PC%24. However, in the 6 studies that reported PCT50 the

mean/median value ranged from 7-11.3 h, comparable with that for rectal ARTS.

The weighted mean mortality in artemisinin-treated patients was relatively high at

12.9%, but included a study of children with cerebral malaria which reported a

mortality of 5%. [414]

7.3.4.2.3. Dihydroartemisinin

There have been only two studies of DHA suppositories involving a total of 180

patients (Table 7.2 C). [424, 425] One of these [425] showed much slower initial

parasite clearance than seen in studies of other preparations, with the mean

parasite density rising to 30% above baseline at 12 h before falling to 70% at 24 h.

The dose used in this study was relatively low (3 mg/kg) and variable, with some

patients receiving as little as 2.3 mg/kg. A similar dose was used in the second

study of dihydroartemisinin [424] but no direct comparison could be made because

16

markers of early parasite clearance were not determined. There were no deaths

from severe malaria in either of these studies.

7.3.4.2.4. Artemether

Only one trial of ARM suppositories is available [427] and is summarised in Table

7.2 D.

17

Table 7.2: Clinical efficacy of rectally administered artemisinin derivatives. Data are either mean ± SD or median (range). A. Artesunate. i) Low-dose artesunate regimens (<5 mg/kg initial dose).

Study author Country n Age Severity Dose1 PCT (h) PCT50 (h) PC%12 PC%24 Mortality

Halpaap, 1998 [238]

Gabon 12 C (7-12) U Approx.1.8 mg/kg at 0 and 4 h

NC 13.8% (0-167)2

0.1% (0-4.5)2

-

Benakis, 2006 [342]

Vietnam 12 A U Approx. 3 mg/kg at 0,3, 6 and 9 h

48 (24-84) NC 40.1% (0.5-84.6)2

10.8% (0-100)2

-

Looareesuwan, 1995 [241]9

Thailand 30 A S Approx. 4 mg/kg at 0, 4, 8, 12, 24, 36, 48, 60 h

50.4 (29-75)

11 NC NC 0

Bhatt, 1996 [343] Kenya 23 A S Approx. 3.3 mg/kg at 0, 4, 8,12, 24 h

32.5 ± 6 NC 70.0% 45.0% 3 (13.0%)

13 A S Approx. 4.2 mg/kg bd for 3 days

35.01 ± 21.1

NC NC NC 0 Thwe, 1996 [235]

Myanmar

18 A S Approx. 4.4 mg/kg) bd for 2 days

42.5 ± 53.9

NC NC NC 0

32 A S Approx. 3.9 mg/kg at 0, 4, 8, 12, 24, 36, 48, 60 h

47.3 ± 12.4

11.1

(±4.7)

60%

(approx)2

5%

(approx)2

0 Looareesuwan, 1997 [240]

Thailand

31 A S Approx. 3.6 mg/kg at 0,12, 24, 36, 48, 72 h

55.3 ± 17.4

12.8

(±6.2)

80% (approx)2

15% (approx)2

0

Awad, 2003 [237] Sudan 100 A S Approx. 3.1 mg/kg tds for 3 days

31.5 ± 10.1

14

(approx)2

63.8% 11.2% 1

Weighted means

3.4 mg/kg initial dose 40.3 12.9 62.1% 13.9% 1.6% (S)

18

ii) High-dose artesunate regimens (>5 mg/kg initial dose). Study author Country n Age Severity Dose1 PCT PCT50 PC%12 PC%24 Mortality

Sabchareon, 1998 [242]

Thailand 26 C (5-12) U 15 (10-19) mg/kg daily for 3 days

42.2 ± 20.6

16.8 ± 11.7

NC NC -

Gomez Landires, 2003 [239]

Ecuador 150 C (2-12) U 3.9-13.6 mg/kg daily, bd or tds over 3-6

days

8.9 NC 3% (approx.) NC -

Karunajeewa, 2003 [408] (Clinical trial 3)

PNG 47 C (5-10) U 13 mg/kg at 0 and 12 h

20

(2->24)

7

(2-20)

8.5%

(0-812)

0%

(0-1.2)

-

12 C MS 9.3 mg/kg single dose 20

(16-42)

8

(3.75-9.25)

15%

(0-52)

NC 0 Krishna, 2001[232]

Ghana

24 C MS 18.9 mg/kg single dose

22

(16-32)

7

(3.75-14.75)

22%

(0-96)

NC 0

WHO 2002 (unpublished) [244]

Thailand 44 C MS 10 mg/kg NC NC 25.6% 0% 1 (2.3%)


Thailand 24 A MS 10-20 mg/kg NC NC 52.1% NC 0


Thailand 69 A MS 10 mg/kg NC NC 25%-34.5% 0% 0

Malawi 87 C MS 10 mg/kg single dose 36 NC 26.7% 0.1% 0 Barnes, 2004 [117]

South Africa 27 A MS 10 mg/kg single dose 49 NC 12.8% 1.0% 1 (3.7%)

19

ii) High-dose artesunate regimens (>5 mg/kg initial dose: continued). Study author Country n Age Severity Dose1 PCT PCT50 PC%12 PC%24 Mortality

Andoh (unpublished) [411]

Cote d’Ivoire 25 C U (n=10),

S (n=15)

8 mg/kg at 0 and 12 h NC NC 2% 0% 0

10 A U 8 NC 0.0% 0.0% - Gomez Landires, 1996 [359]

Ecuador

30 A S

Approx. 9 mg/kg on day 1, then 3 mg/kg bd

on days 2 and 3 23.4 NC 9.6% (0.5-69.2)2

0% (0-3.5)2

1 (2.5%)

5 C S 20 mg/kg then 10 mg/kg at 12 h then

daily

48 (12-60) 7 (3.4-18.2)

NC NC 0 Pengsaa, 2005 [341]

Thailand

8 C S 10 mg/kg at 0, 4, 12

and 24 h then daily 24 (12-48) 6 (3.5-

14.1) NC NC 0

Karunajeewa, 2006 [409] (Clinical trial 4b)

PNG 41 C (1-10) S 13 mg/kg at 0 and 12 h

30.3 ± 14.2

9.1 ± 4.9 17% (0-190) 0.5% (0-40)

1 (2.4%)

Weighted means

10.8 mg/kg (initial dose)

23.7 9.1 16.0% 0.15% 0.7 % (MS), 2.0% (S)

20

B. Artemisinin. Study author Country n Age Severity Dose1 PCT PCT50 PC%12 PC%24 Mortality

Arnold, 1990 [412]

Vietnam 32 A (mostly males)

U Approx. 12 mg/kg at 0 and 4 h

41.8 ± 10.1 11.3 ± 6.4 40% (approx)2

15% (approx)2

-

Hien, 1991 [429] Vietnam 30 C U 10-20 mg/kg at 0 and 4 h

25.8 ± 78.9 6.2 ± 3.2 NC NC -

Hien, 1994 [430] Vietnam 27 C U 15mg/kg 35.3 ± 24.5 8.0 ± 5.3 NC NC -

Ashton 1998 [248]

Vietnam 15 A (m) U Approx. 10 mg/kg at 0 and 4 h (PEGS)

29 ± 11 9.4 ± 5.2 NC NC -

Koopmans 1998 [418]

Vietnam 8 A (m) U Approx. 12.2 mg/kg (FS)

24 (24-72) NC NC NC -

Li 1985 [420] China 100 A & C U, MS & S Approx. 12 mg/kg at 0 and 4 h

53.2 ± 15.6 NC NC NC -

Hien 1992 [417] Vietnam 18 A S (cerebral malaria)

Approx. 12 mg/kg at 0 and 4 h

37.9 ± 17.4 9.7 ± 7.2 Approx. 30%

Approx. 5%

5 (28%)

Cao 1997 [414] Vietnam 37 C S 40 mg/kg then 20 mg/kg at 4 h

48 (8-84) 7.0 (2.3-28.4)

NC NC 2 (5%)

Ha 1997 [416] Vietnam 51 A S 24 mg/kg 30 NC NC NC 9 (17.6%)

Birku 1999 [413] Ethiopia 32 A S Approx. 13.1 mg/kg at 0 then 8.7mg/kg at 12

h

27.5 ± 4.7 12 Approx. 45%

Approx. 7%

2 (6.1%)

Weighted mean 17.2 mg/kg (initial dose)

39.7 9.0 39.8% 9.7% 12.9% (S)

21

C. Dihydroartemisinin.

Study author Country n Age Severity Dose1 PCT PCT50 PC%12 PC%24 Mortality

Esamai 2000 [424]

Kenya 30 C & A S Approx 2.5 mg/kg 38.4 ± 6.5 NC NC NC 0

Wilairatana 2000 [425]

Thailand 150 C & A MS & S Approx. 3 mg/kg (2.3-6 mg/kg)

46.1 ± 15.7 Approx. 182 131.5 ± 802%

71.5 ± 128.5%

0

D. Artemether. Study author Country n Age Severity Dose1 PCT PCT50 PC%12 PC%24 Mortality

Aceng 2005 [427]

Uganda 51 C S (cerebral) Approx. 7.3mg/kg 54.2 ± 33.6 NC NC NC 6 (11.7%)

Abbreviations: A = Adult, C = Child, (m) = male, U = Uncomplicated malaria, MS = Moderately severe malaria, S = Severe malaria,

NC = not calculated, PNG = Papua New Guinea, FS = fatty-base suppository, PEGS = Polyethylene glycol-base suppository.

1 Dose administered in the first 12 h of treatment (as a single dose unless otherwise stated). 2 Values approximated by interpolation from graphical representations of parasite clearance in the published text.

22

7.3.4.3. Comparative studies

A number of studies have directly compared rectal artemisinin drugs with

conventional therapies, including two studies using ARTS, [117, 409] four using

QHS, [413, 414, 416, 417] and one each involving DHA [424] and ARM. [427] In

each of these 8 studies, the similarity of the groups at baseline, compliance with

therapy, withdrawal/drop-outs and the timing of outcome assessments appeared

acceptable, and the acute measures of efficacy (primarily parasite clearance)

implied an intention-to-treat analysis, meaning a score of 5/5 for each. For all

except the study by Hien et al. [417] (in which the artemisinin suppositories became

unavailable during the trial), the baseline characteristics of the groups were similar,

suggesting that seven studies were of high quality (score >6). The study by Aceng

et al. [427] was the only one that employed patient blinding (with a rectal placebo)

and it had the highest score of 7/11. None of the studies reported allocation

concealment or blinding of the care providers or outcome assessors, and co-

interventions were either not stated or were not comparable between the groups.

Barnes et al. [117] compared rectally administered ARTS with parenteral quinine in

children and adults with moderately severe malaria in Malawi and South Africa. By

12 h, 92% of children and 96% of adults treated with rectal ARTS had achieved

reductions in parasitaemia to <60% of baseline, compared with 14% and 38% of

quinine-treated children and adults, respectively, a difference that was maintained

at 24 h. Children treated with rectal ARTS also had more rapid clearance of fever

than quinine-treated children, but there were no statistically significant differences

seen with other markers of clinical response, including time to return to per os

status (ability to drink or take tablets orally). A study by Karunajeewa et al. [409]

(Clinical trial 4b) compared rectal ARTS with i.m. ARM, a commonly-used

alternative to parenteral quinine for severe malaria, in 79 children with severe

malaria in PNG. This showed statistically significant differences in PC%12, PCT50

and PCT90 in favor of rectal ARTS. The authors concluded, from plasma drug

concentration data, that this was likely due to relatively low and erratic absorption of

i.m. ARM. There were no between-treatment differences in clinical outcomes,

including measures of fever clearance and return to per os status.

23

Four studies have compared QHS suppositories with either i.v. quinine or parenteral

ARTS. [413, 414, 416, 417] The three using quinine as a comparator all showed

significantly faster parasite clearance (PCT50, PCT90 and PCT) in the QHS

group.[413, 414, 416, 417] However, other clinical markers including fever

clearance, coma duration and mortality were similar. Regardless of whether i.v. or

i.m. ARTS was used as the comparator, no significant differences were seen with

rectal artemisinin for parasitologic or other outcomes.

A study by Esamai et al. [424] that compared DHA suppositories to i.v. quinine

demonstrated a more rapid PCT with the artemisinin derivative, but markers of early

parasite clearance were not calculated and there was no statistically significant

difference in fever clearance. Aceng et al. [427] compared rectally administered

ARM with i.v. quinine in 103 Ugandan children with cerebral malaria. There were no

significant differences between treatment groups for clinical (FCT, time to regaining

consciousness and time to per os status) and parasitologic outcomes.


Nineteen PK studies, 6 in healthy volunteers and 13 in adults or children with

uncomplicated, moderately severe or severe malaria, were identified and are

summarised in Table 7.3. There was marked inter-individual variability in most PK

parameters, but conventional doses of ARTS achieved an earlier tmax and higher

Cmax than other artemisinin derivatives. The 14 PK studies of rectal ARTS (Table

7.3 A) included healthy adult volunteers, [233, 407] adults with uncomplicated

malaria, [342] and children with uncomplicated, [238, 242, 358] moderately severe

[232] or severe [341] malaria (Table 7.3 A i) and ii)). No study was able to

demonstrate a relationship between any PK parameter and clinical outcome. Few

studies were of sufficient size to enable examination of the effect of covariates on

PK.

The results of three PK studies of rectal QHS performed in mostly male Caucasian

or Vietnamese subjects are summarised in Table 7.3 B. [248, 418, 419] Patients

were either healthy volunteers or had uncomplicated malaria. No study involved

children or adults with moderate or severe disease. The one study of DHA

suppositories (Table 7.3 C) was performed in adults with uncomplicated malaria.

[180] The study of rectal ARM (Table 7.3 D) [428] used a parenteral formulation

24

(Kunming Pharmaceuticals, China) dissolved in arachis oil and rectally self-injected

by healthy Caucasian volunteers. There is a suppository formulation of ARM [427]

but its PK properties are unknown.

7.3.4.5. Safety and tolerability of rectal artemisinin derivatives

More than 1600 patients, including >600 children, have received a rectal artemisinin

drug in published or unpublished trials (Tables 7.2 and 7.3). Adverse events have

not been reported in a standardised manner. Minor adverse events have included

dizziness, transient fever occurring approximately 24 h after parasite clearance, and

gastrointestinal symptoms (nausea, vomiting, abdominal pain, constipation or

diarrhoea). [241, 408, 409, 413, 420, 421] It is unclear whether these symptoms

relate to the artemisinin drugs, a co-administered partner drug or malaria itself.

No symptoms suggestive of neurotoxicity have been described in any study

evaluating a rectally-administered artemisinin derivative, including those using

ARTS at doses of up to 20 mg/kg [232, 341] and QHS at doses up to 40 mg/kg.

[414] Although tenesmus has been described in up to 26% of those administered

QHS suppositories, [413, 420, 421] no study has described rectal bleeding or anal

irritation that would indicate a local inflammatory reaction.

25

Table 7.3. Pharmacokinetics of rectally administered artemisinin derivatives. Data are either mean (CV%) or median (range). A. Artesunate (because ARTS is rapidly metabolized to DHA in vivo, the pharmacokinetic parameters reported below refer to plasma DHA concentrations). i) Low-dose artesunate regimens (<5 mg kg initial dose).

Study Country n Age group

Disease status

Dose1 tmax (h) Cmax (nM) AUC2 (µµµµmol.h/l)

Benakis 1996 [233] Switzerland 6 A HV 2.7 mg/kg 1.5 840 2.4 (91%)

Halpaap 1998 [238] Gabon 12 C (7-12) U Approx .1.8 mg/kg at 0 and 4 h

1.1 (51%) 630 (56%) NC

Navaratnam 1998 [410]

Malaysia 12 A HV Approx 3.1 mg/kg 1.8 (36%) 1370 (46%) 3.4 (63%)

Awad 2004 [407] Sudan 12 A HV Approx. 3.1 mg/kg) 2.0 (32%) 770 (53%) 4.2 (71%)

Benakis 2006 [342] Vietnam 12 A U Approx. 3 mg/kg) at 0, 3, 6 and 9 h

NC 970 NC

Weighted mean 2.7mg/kg (initial dose) 1.61 930 3.5

26

ii) High-dose artesunate regimens (>5 mg/kg initial dose).


Disease status

Dose1 tmax (h) Cmax (nM) AUC2 (µmol.h/l)

Sabchareon 1998 [242]

Thailand 19 C (5-12) U 15 (10-19) mg/kg NC 2380 (88%)

2.6

(0.2-8.7)

10 C (2-7) MS 9.3 mg/kg 1.7 (0.9-3.2) 2400

(800-5800)

9.8

(1.4-28.2)

Krishna 2001 [232]

Ghana

16 C (2-7) MS 18.9mg/kg 1.8 0.6-3.3) 3100

(700-6800)

13.2

(2.9-26.2))


NS 40 A (m) HV 6.8mg/kg NC 1430 (56%) 5.2 (70.1%)

10mg/kg 2.8 (61%) 3860 (68%) 15.1 (106%) WHO 2002 (unpublished) [244]

Thailand 48 A MS

20mg/kg 4.2 (50%) 7100 (80%) 40.1 (104%)


South Africa 27 A MS 10mg/kg 4.4 (52%) 3010 (79%) 14.3 (76%)

Karunajeewa 2004 [358] (Clinical trial 3)

PNG 47 C (5-10) U 13mg/kg 2.3 2530 NC

Pengsaa 2005 [341] Thailand 2 C S 20mg/kg 1.2 3890

NC

Simpson 2006 [247], 3c

Thailand, Ghana,

Malawi, South Africa

179 A (n=38), C (n=141)

MS 10mg/kg 3 (approx) NC NC

Weighted mean 12.4mg/kg 3.0 3010 14.2

27

B. Artemisinin (based on plasma QHS concentrations).


Disease status

Dosea tmax

(h)

Cmax (nM) AUC (µmol.h/l)

Koopmans 1998 [418]

Vietnam 8 A (m) U Approx. 12.2mg/kg (FS) 6.5 (60%) 380 (56%) 3.8 (63%)

8 A HV Approx. 7.2mg/kg (FS) 7.1 (30%) 350 (102%) 3.6 (126%) Koopmans 1999 [419]

Vietnam

6 A HV Approx 10mg/kg (PEGS) 6.7 (30%) 270 (56%) 2.6 (62%)

Ashton 1998 [248] Vietnam 15 A (m) U Approx. 10mg/kg (PEGS)

4.0 (2.0-10) 650 (50%) 3.1 (60%)

Weighted mean 9.9mg/kg 5.6 470 3.3

28

C. Dihydroartemisinin (based on plasma DHA concentrations). Study Country n Age

group Disease status

Dose1 tmax (h) Cmax (nM) AUC (µmol.h/l)

Ilett 2002 [180] Vietnam 12 A U 3 mg/kg 4.0 (2.6-6.0) 750 (550-1110) 3.4 (38%)

D. Artemether (based on plasma artemether concentrations). Study Country n Age

group Disease status

Dose1 tmax (h) Cmax (nM) AUC0-6 (µmol.h/l)

Teja-Isavadharm 1996 [428]

Malaysia 8 A HV 5mg/kg (lipid solution) 3.1 (68%) 250 (62%) 0.9 (78%)

Abbreviations: A = Adult, C = Child, (m) = male, HV = healthy volunteer, U = Uncomplicated malaria, MS = Moderately severe

malaria, S = Severe malaria, NS not specified, NC = not calculated, PNG = Papua New Guinea, FS = fatty-base suppository, PEGS

= Polyethylene glycol-base suppository.

1 Dosage indicates dose administered in the first 12 h of treatment (as a single dose unless otherwise stated). 2 All AUC calculations are AUC0-� except for study by Sabchaeron et al. [242] (AUC0-12) and Teja-Isavadharm et al. [428] (AUC0-

6). 3 The study by Simpson et al. [247] was a population PK study that included study subjects from studies by Krishna et al. [232]

and WHO, [244] and has not been included in the weighted mean calculations.

29

7.3.5. CONCLUSIONS

Artemisinin derivatives formulated for rectal administration have the potential to be

widely used in tropical countries and a myriad of preparations now exist (see Table

7.1). There is, therefore, a need for increased awareness of the pharmacologic

properties and efficacy of rectally-administered artemisinin compounds. Although it

is likely that ARTS suppositories will remain the most widely prescribed drug in this

respect, [245] issues of availability, cost and sociopolitical factors may result in an

increase in use of these alternative derivatives.

Similar to experience with oral and parenteral artemisinin formulations, the

introduction of rectal derivatives into clinical use has not followed classical

pathways of rational drug development. Rectal absorption of these compounds

must be prompt and reliable if they are to have a rapid parasiticidal effect, but may

not necessarily be equivalent for different preparations. Supportive PK and PD data

have lagged behind the widespread availability of a variety of formulations that are

administered using empirically-derived dosing regimens. The reasons for this

include the practical challenges of performing PK studies in representative patient

samples in the rural tropics, and the lack of stringent control of drug licensing in

many developing countries. Because pharmaceutical factors can influence PK

disposition, preparations of the same derivative from different manufacturers may

not be bioequivalent.

The data reviewed showed a mean/median tmax of 1-4 h for ARTS compared with

values of 4-7 h for QHS and DHA, and 3 h for ARM. Similarly, ARTS administered

at >10mg/kg doses resulted in mean/median Cmax and AUC values 3-10 times

higher than those achieved with QHS (7-12 mg/kg), DHA (3 mg/kg) or ARM (5

mg/kg). These values suggest a therapeutic advantage for ARTS, but more data

are needed for DHA and ARM. Although it is unclear whether earlier and higher

plasma concentrations confer therapeutic advantage, it is notable that initial

parasite clearance was slower for low-dose ARTS regimens (weighted mean PC%12

62.3% vs 16.1% for high-dose; Table 7.2 A), with the slowest in one study of low-

dose rectal DHA. [425]

30

As previously noted with ARTS, the most striking aspect of the PK studies reviewed

was the degree of inter-individual variability in PK parameters. In addition to the

wide variation of bioavailability of ARTS reported by Krishna et al. [232], coefficients

of variation for Cmax and AUC are also wide for the other rectally-administered

artemisinin derivatives. This adds to concerns that a small proportion of patients are

at risk of sub-therapeutic drug concentrations.

The present review suggests that there are no additional safety concerns

associated with rectal administration of other rectal artemisinin derivatives. Although

tenesmus and a burning sensation were reported in some studies of rectal QHS,

[420] there has been no evidence suggesting proctitis of the type that occurs with

rectal quinine. [431] It is reassuring that there were no descriptions of neurotoxic

symptoms in the approximately 1600 subjects, more than one-third of whom were

children, treated with rectal artemisinin compounds in studies reviewed here. This

includes doses of up to 20 mg/kg ARTS and 40 mg/kg QHS. [232, 341, 414]

A number of comparative trials provide some evidence that rectal preparations may

initiate parasite clearance more rapidly than conventional treatment including

parenteral quinine and ARM. [117, 409, 413, 414, 416, 417, 424, 427] None of the

comparative trials of rectal artemisinin derivatives were powered to assess mortality

as an endpoint, although it is noteworthy that other clinical markers (including fever

clearance and coma duration) were similar to those with the comparator drugs.

Because of the large sample sizes required for a non-inferiority study, it is unlikely

there will ever be a trial comparing a rectal artemisinin with conventional treatment

for severe malaria using mortality as an endpoint. Therefore, markers of early

parasite clearance and PK parameters (including their variability) will remain the

most useful endpoints in future trials.

This review had limitations. First, the weighted mean estimates of efficacy and PK

parameters were often from small numbers of studies with incomplete data. This

made formal assessment of statistical heterogeneity difficult. The derived results

must be interpreted with caution, but the summary data from individual studies cited

in the Tables support the overall conclusions. Second, the definition of severe

(including cerebral) malaria included a variable number of components and different

31

severity cut-points. Although the PK properties of the artemisinin drugs do not

appear to be influenced by disease severity, [182] a greater number of studies and

more uniform definition of severity would have facilitated an analysis of parasite

clearance parameters by clinical status.

Available data support ARTS at doses of >10 mg/kg and QHS 12 mg/kg as having

good clinical efficacy regardless of factors such as age, gender and ethnicity.

Artesunate appears the better option given its more rapid absorption and higher

plasma concentrations at these doses. However, the present review highlights the

need for more pharmacologic and efficacy data for all rectal artemisinin drugs,

including studies involving i) head-to-head comparisons of the initial clinical efficacy

of different formulations and dose regimens (especially involving DHA and ARM), ii)

PK analysis including bioavailability and its predictors and effect on parasite

clearance, and iii) well-characterized severely ill patients with i.v. ARTS as

comparator. Given the concerns regarding lower limits of bioavailability and pending

further research particularly in the context of severe malaria, prescribing should be

consistent with current WHO guidelines that restrict use to pre-transfer treatment

when oral therapy is not possible and where injectable antimalarial drugs are

unavailable, followed by prompt definitive treatment (i.v. ARTS/quinine or a full

course of an approved oral ACT) as dictated by the clinical state of the patient.

Documents

Clinical Pharmacology of the Treatment of Malaria in Papua