Transactions of the Royal Society of Tropical Medicine and Hygiene (2007) 101, 777—785

avai lab le at www.sc iencedi rec t .com

journa l homepage: www.e lsev ierhea l th .com/ journa ls / t rs t

Randomized, controlled, double-blind trial withivermectin on Loa loa microfilaraemia: efficacy ofa low dose (∼25 �g/kg) versus current standarddose (150 �g/kg)

Joseph Kamgnoa, Sebastien D.S. Pionb, Mathurin C. Tejiokemc,Nana A.Y. Twum-Dansod, Bjorn Thyleforsd, Michel Boussinesqe,∗

a National Onchocerciasis Task Force, Ministry of Public Health, BP 4794, Yaounde, Cameroonb Department of Infectious Disease Epidemiology, Imperial College Faculty of Medicine, St Mary’s Campus, Norfolk Place,London W2 1PG, UKc Laboratoire d’Epidemiologie et de Sante Publique, Centre Pasteur du Cameroun, BP 1274, Yaounde, Cameround Mectizan Donation Program, 750 Commerce Drive, Suite 400, Decatur, GA 30030, USAe Institut de Recherche pour le Developpement (IRD), Departement Societes et Sante, UR-024, 213 rue La Fayette,75480 Paris Cedex 10, France

Received 14 July 2006; received in revised form 28 March 2007; accepted 28 March 2007Available online 24 May 2007

KEYWORDSOnchocerciasis;Loiasis;Loa loa;Ivermectin;Adverse effects;Clinical controlledtrial

Summary Neurological serious adverse events (SAEs) following ivermectin treatment mayoccur in individuals harbouring high Loa loa microfilarial densities and are of major concern inthe context of mass ivermectin distributions organized in Africa for onchocerciasis and lymphaticfilariasis control. As those SAEs are induced by the rapid and massive microfilaricidal effect of astandard dose of ivermectin (150 �g/kg), we performed a randomized, controlled, double-blindtrial to determine whether ivermectin given as: (a) a single low dose of 1.5 mg (i.e. 25 �g/kg fora 60 kg person); or (b) two doses of 1.5 mg given at a 2 week interval leads to a more progressivedecrease in Loa microfilarial loads compared with the standard dosage. A low dose of ivermectinbrought about a significantly smaller decrease in Loa microfilaraemia than the standard dose.However, this decrease was not sufficiently different from that obtained after the standard doseto be acceptable to public health programmes, which require a wide safety margin. A secondlow dose of ivermectin given 15 days after the first dose did not lead to a further decreasein Loa microfilaraemia. Lastly, the variability in the response observed in the group treatedwith 25 �g/kg suggests that even lower doses would have no effect on a significant number of

patients. Ivermectin given at a low dose (≤25 �g/kg) is probably not adequate to prevent theoccurrence of post-treatment neurological SAEs.© 2007 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rightsreserved.

∗ Corresponding author. Tel.: +33 1 42 49 38 15; fax: +33 1 42 49 38 15.E-mail address: boussinesq@ird.fr (M. Boussinesq).

0035-9203/$ — see front matter © 2007 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.trstmh.2007.03.018

7

1

I(ip(rAtAs2Stamcwytp

tttiaslomcottrt(awc2ei

2

2

TH1Ova2hAdv

r(iptay1ptmtfprtahce

2

Ttisar6i3

waibt

mmtpmfwGmriww1Gt

78

. Introduction

vermectin treatment administered at the standard dose150 �g/kg) may induce serious adverse events (SAEs),ncluding an encephalopathy, which may be fatal, in thoseatients harbouring a very high Loa loa microfilaraemiaBoussinesq et al., 2003; Gardon et al., 1997). Althoughare, these neurological SAEs impede the progress of thefrican Programme for Onchocerciasis Control (APOC) andhe Programme to Eliminate Lymphatic Filariasis (PELF) infrica in areas where loiasis is co-endemic with onchocercia-is (Haselow et al., 2003; Malecela-Lazaro and Twum-Danso,004). The pathogenic mechanisms associated with theseAEs are not fully understood. However, the presence inhe retina of haemorrhages and other lesions suggestive ofn obstructive process suggests that the encephalopathiesight be due to a similar phenomenon in the cerebral micro-

irculation (Fobi et al., 2000). It is postulated that treatmentith ivermectin induces a fairly rapid and massive paral-sis of the blood-dwelling Loa microfilariae (mf), and thathis would lead to diffuse embolisms, and other pathologicalrocesses, in the cerebral capillaries.

Several small-scale clinical trials have been conductedo identify a treatment that would progressively decreasehe high Loa microfilaraemia below the risk threshold, sohat the neurological SAEs would be prevented. By organiz-ng mass treatments with such a regimen in L. loa-endemicreas, it would be possible subsequently to administertandard doses of ivermectin against onchocerciasis and/orymphatic filariasis (LF) with very little risk of such SAEsccurring. Assuming a dose—effect relationship of iver-ectin on L. loa, we conducted, in 1999, a first trial

omparing the effects of a standard dose (150 �g/kg) andf a low dose of 3 mg (i.e. 50 �g/kg for a 60 kg person) onhe microfilaraemia. The follow-up of the patients showedhat Loa microfilaraemias tended to decrease slightly lessapidly in those subjects treated with the low dose, buthe difference between the two groups was not significantKamgno et al., 2000). Here, we present the results ofn additional randomized, controlled, double-blind trial inhich the microfilaricidal efficacy of a standard dose wasompared to even lower doses: a single dose of 1.5 mg [i.e.5 �g/kg for a 60 kg person, half the dose tested by Kamgnot al. (2000)], and two doses of 1.5 mg given at a 2 weeknterval.

. Patients and methods

.1. Selection of patients

he trial was conducted in seven villages located in the Awaeealth District, in the Central Province of Cameroon, some00 km east of Yaounde (namely Awae I, Elat, Minkomila,doudouma, Olanguina, Sanguela and Zili). Previous sur-eys had shown that the area is highly endemic for loiasisnd Mansonella perstans filariasis (prevalences in adults:

4—30% and 21—49%, respectively) (Kamgno et al., 2001) butypoendemic for onchocerciasis (Boussinesq et al., 1998a).survey using an immunochromatographic card test (ICT)

emonstrated that infection with Wuchereria bancrofti isery rare in the area (Kale, unpublished data), and the

2

Dwe

J. Kamgno et al.

esults of a study on the diethylcarbamazine patch-testBoussinesq et al., 1998a) suggest that the level of endemic-ty for Mansonella streptocerca filariasis is very low. After areliminary visit, aimed at informing the populations abouthe objective of the trial, a survey was organized to evalu-te the Loa microfilarial load of all volunteers aged 15 to 70ears old. Capillary blood was taken by fingerprick between0:00 and 16:00 h, and a calibrated blood smear (50 �l) wasrepared from each individual. After Giemsa staining, allhe Loa mf present on each slide were counted under aicroscope, and the microfilarial loads were expressed as

he number of mf per ml of blood (mf/ml). To be eligibleor the trial, the subjects had to be 15 to 70 years old, toresent no condition for which ivermectin treatment is notecommended as part of mass treatment (pregnancy, lac-ation of a child less than 7 days old and serious illness),nd to harbour 100—15 000 Loa mf/ml. In addition, afteraving received individually further information on the pro-edures of the trial, the eligible subjects had to providexplicit consent to participate in the latter.

.2. Treatment regimens

he eligible patients who volunteered to participate in therial were randomly allocated into three groups after strat-fication on the basis of their Loa microfilaraemia (fivetrata: 100—1000; 1001—4000; 4001—8000; 8001—12 000;nd 12 001—15 000 Loa mf/ml). The patients in Group 1eceived one single dose of 1.5 mg ivermectin (25 �g/kg for a0 kg person); those in Group 2 received two doses of 1.5 mgvermectin given with a 15 day interval; and those in Groupreceived one single standard dose of 150 �g/kg.Just before the first treatment, each individual was

eighed and asked whether they had previously taken anntifilarial therapy. For ethical reasons, all the individualsnvolved in the trial were followed-up in a similar manner,ut the data from patients who had received an antifilarialreatment were excluded from analyses.

The treatments were administered in a double-blindanner, using capsules preliminarily prepared by a phar-acist who was not involved in the rest of the trial. Four

ypes of capsules were prepared, which contained a placeboowder, 1.5 mg ivermectin, 3 mg ivermectin or 6 mg iver-ectin, respectively. These capsules were indistinguishable

rom each other. Two rounds of treatment were organizedith a 15 day interval. During the first round, patients inroups 1 and 2 received one capsule containing 1.5 mg iver-ectin and one capsule of placebo, and patients in Group 3

eceived, according to their weight, two capsules with 6 mgvermectin (for those who weighed ≥65 kg), or one capsuleith 6 mg and another with 3 mg ivermectin (for those whoeighed 45—64 kg). Fifteen days later, patients in Groupsand 3 received two capsules of placebo, and patients inroup 2 one capsule with 1.5 mg ivermectin and one con-aining placebo.

.3. Monitoring of adverse reactions

uring the week following each treatment, all the villagesere visited daily to record and treat any possible adverseffects to treatment. The patients were informed that the

ctRI

Dinw

3

3r

Onlb4fm3y

tragt51

tGsTlola

3fi

ToipGit

sw

Post-ivermectin Loa-related adverse effects

team would be in the village at a given hour, and that allreactions would be treated free of charge.

2.4. Biological follow-up

Calibrated blood smears were prepared from each patient onday 0 (D0) just before administering the first treatment, andthen on days 3, 7, 15 (just before the second treatment), 22,30 and 60. All the blood collections were performed between10:00 and 16:00 h, and efforts were made for each village tobe visited at the same hour during the different examinationrounds. On D3, 5 ml venous blood was collected for mea-suring the C-reactive protein (CRP), an indicator of acuteinflammation that has been shown to increase after iver-mectin treatment in patients harbouring a significant Loamicrofilaraemia (Boussinesq et al., 2003). The CRP levelswere not measured on D0.

2.5. Statistical analyses

Adverse reactions to ivermectin typically occur within 3—4days following treatment, and one might assume that theyare related to the effects of the drug on the Loa mf duringthis period. Focusing on this short-term effect of ivermectin,we performed a first series of analyses based on the Loamicrofilarial counts measured on D0, D3, D7 and D15. As thepatients in Groups 1 and 2 were similar, in terms of treat-ment received, until D15 (they had received only one lowdose of ivermectin), they were combined in this analysis.

A second series of analyses, focusing on the medium-termeffect of ivermectin, was then performed. It included thedata collected on D0, D15, D30 and D60. Within this period,the three groups of patients differed in terms of treatmentreceived, and they were thus considered separately in theanalyses. Then, we performed analyses aimed at evaluatingwhether the second low dose received by the patients inGroup 2 brought about a further decrease in the Loa micro-filaraemia, when compared to those of patients in Group 1.To this end, we compared the changes in microfilarial loadsbetween D15 and D22, and between D15 and D30, in thegroups who had received one (Group 1) or two (Group 2)low doses of ivermectin.

The values of microfilaraemia given in the text, tablesand figures, as well as those considered in the analyses, areexpressed as per ml of blood. For both analyses of short-and medium-term effects of treatment, a first step consistedin comparing, separately for each day of examination, themicrofilarial loads between the treatment groups. This wasdone using Mann-Whitney U tests for those analyses of short-term effect of ivermectin on Loa microfilaraemia (includingonly two groups), and Kruskal-Wallis tests for the analy-ses of the medium-term effect of the drug (performed onthree groups). Then, to compare the overall changes withtime in the Loa counts between the treatment groups, weperformed an analysis of variance on repeated measures.The model was defined with the repeated measurements

of Loa microfilaraemia as within-subjects factor, and thegroup of treatment as between-subjects factor. As this pro-cedure requires normally or almost normally distributeddata, we applied a log (x+1) transformation to the micro-filarial counts x (the value 1 was added to x to include those

ctshw

779

ases for which the microfilarial counts were equal to 0 afterreatment). These analyses were performed using the GLMepeated Measures procedure provided by SPSS 10.0 (SPSSnc., Chicago, IL, USA).

In all analyses, for a given follow-up period (D0—D3—D7,0—D3—D7—D15, D0—D30, D0—D60 and D15—D22—D30), we

ncluded only those persons who attended all the exami-ations. As indicated above, individuals previously treatedith an antifilarial therapy were excluded from the analyses.

. Results

.1. General information and incidence of adverseeactions

verall, 792 persons were examined during the prelimi-ary survey, and 200 (25.3%) of them presented with a L.oa microfilaraemia. From a total of 117 individuals eligi-le for the trial, 95 agreed to participate (47 males and8 females). Group 1 included 35 people (16 males and 19emales; median age = 49.0 years), Group 2, 29 persons (15ales and 14 females; median age = 55.0 years), and Group

, 31 persons (16 males and 15 females; median age = 39.0ears).

The weights of the patients in Groups 1 and 2, i.e.hose who received one or two doses of 1.5 mg ivermectin,anged between 41 and 74 kg (median: 60 kg; quartiles: 52nd 65 kg). Consequently, the exact doses received in theseroups ranged between 20.3 and 36.6 �g/kg. In Group 3,he patients weighed 44 to 74 kg (median: 58 kg; quartiles:2 and 66 kg), and the doses administered ranged between40.6 and 204.6 �g/kg.

The proportions of patients who complained of post-herapeutic adverse events were 25.7, 36.6 and 42.4% inroups 1, 2 and 3, respectively. These proportions were notignificantly different among the groups (�2 = 2.17; P = 0.34).he most frequent adverse events were itching (16 cases),

umbar pain (15 cases) and headache (9 cases), but nonef these reactions was serious. All of them resolved fol-owing symptomatic treatment with paracetamol and/or anntihistamine, and in a few cases indomethacin.

.2. Efficacy of low and standard doses within therst week following treatment

his analysis was restricted to those individuals examinedn D0, D3 and D7. In Groups 1 and 2 taken together (64ndividuals initially), nine persons missed at least one of thearasitological examinations: two in Group 1 and seven inroup 2. In Group 3, only two individuals did so. Thus, 55

ndividuals were considered in the low-dose group and 29 inhe standard-dose group.

The Mann-Whitney U test showed that the two groups hadimilar Loa microfilarial loads before treatment (P = 0.113),hereas on D3 and D7, the microfilaraemias were signifi-

antly lower in the group treated with the standard dosehan in the one treated with the low dose (Table 1). Analy-is of variance confirmed that the ‘treatment group’ factorad a significant effect on Loa microfilaraemias during theeek following treatment (F = 5.26, d.f. = 1, P = 0.024), the

780 J. Kamgno et al.

Table 1 Comparison of Loa microfilarial loads in two groups within the first week following treatment with different doses ofivermectin

Day Low dosea (n = 55) Standard doseb (n = 29) P c

AM SD GM WM Reduction ratefrom D0 (%)

AM SD GM WM Reduction ratefrom D0 (%)

0 6998 5140 4875 4876 — 5208 4242 3344 3345 — 0.1133 3377 2895 2052 2054 51.7 1871 1934 973 975 64.1 0.0077 2804 2543 1639 1641 59.9 1636 1672 780 783 68.6 0.020

AM: arithmetic mean; GM: geometric mean; WM: Williams mean.

ared

dg

wbtvotwivt1istbdPnn

ait

Ffita

itehfifgtrtrMrros

mpTi

a 1.5 mg ivermectin, i.e. 25 �g/kg for a 60 kg person.b 150 �g/kg ivermectin.c For each day of examination, the microfilarial loads were comp

ecrease in microfilarial loads being less in the low-doseroup than in the standard-dose group (Table 1).

To complete the analyses of the parasitological data,e assessed whether there was a significant relationshipetween the pre-treatment microfilarial load and the post-reatment reduction rate of the Loa microfilaraemia. Aisual inspection of the data did not suggest the existencef such a relationship (Figure 1). In Groups 1+2, the reduc-ion rate between D0 and D3 was 53.0% in the 27 patientsith the highest microfilarial loads (>6000 mf/ml) and 47.2%

n the 28 harbouring the lowest initial microfilaraemias (i.e.ery comparable values). Similarly, in the group treated athe standard dose, the reduction rates were 63.4% in the5 patients with the highest loads, and 67.3% in the 14ndividuals with the lowest microfilaraemias (i.e. also veryimilar values). However, the correlation between the ini-ial microfilaraemia and the reduction rate of the countsetween D0 and D3 was significantly negative in the low-ose group (Spearman’s correlation coefficient rs = −0.273;= 0.030; n = 63). Conversely, the relationship was not sig-ificant in the standard-dose group (rs = −0.148; P = 0.435;

= 30).

Visual inspection of Figure 1 also suggests that there washigher inter-individual variability in the reduction rates

n the group treated with a low dose than in the grouphat received the standard dose. It shows that, although

igure 1 Relationship between the pre-treatment Loa micro-laraemia [number of microfilariae (mf) per ml of blood] andhe reduction rate in the microfilaraemia between D0 and D3fter treatment. Each square represents one single individual.

cDaiud1P

3fi

Aitsb1

latab

using the Mann-Whitney U test.

n the former group some individuals ‘responded’ well toreatment, with reduction rates exceeding 80%, some oth-rs did not show any decrease in their microfilaraemia. Thisigher variability of response in the low-dose group was con-rmed by the calculations of the variance to mean ratiosor the reduction rates, which were 11.4 in the low-doseroup and 3.7 in the standard-dose group. As noted above,he exact dose received by those patients in Groups 1 and 2anged between 20.3 and 36.6 �g/kg. To investigate whetherhose patients who responded less to treatment had actuallyeceived a lower dose of ivermectin, we compared, using theann-Whitney U test, the distributions of the exact doses

eceived: (1) by those 11 patients for whom the reductionates between D0 and D3 were less than 30%; and (2) by thether individuals in Groups 1 and 2. The difference was notignificant (P = 0.063).

The mean CRP concentrations (normal value: <6 mg/l)easured on D3 was 7.31 mg/l (SD = 11.00) in Groups 1+2 (60atients) and 8.94 mg/l (SD = 9.65) in Group 3 (29 patients).he difference between the two groups was not signif-

cant (Mann-Whitney U = 754, P = 0.31). Spearman’s rankorrelation tests showed that the CRP concentrations at3 were significantly correlated with the microfilaraemiat D0 in the Groups 1+2 (rs = 0.273; P = 0.035) but notn Group 3 (rs = 0.338; P = 0.073). Similarly, the CRP val-es at D3 were significantly correlated with the absoluteecrease in microfilaraemia between D0 and D3 in Groups+2 (rs = 0.347; P = 0.0076), but not in Group 3 (rs = 0.308;= 0.105) (Figure 2).

.3. Efficacy of low and standard doses within therst 2 weeks after treatment

mong the 55 people in Groups 1 and 2 who were exam-ned on D0, D3 and D7, four (two from each group) missedhe examination at D15. In Group 3, only one did so. Con-equently, the analysis of changes in microfilarial loadsetween D0 and D15 was done on 51 patients from Groups+2 combined, and 28 patients from Group 3.

The Mann-Whitney U tests showed that the microfilarial

oads on D0 did not differ significantly between Groups 1+2nd Group 3, whereas they differed significantly at each ofhe three post-treatment examination rounds (Table 2). Itlso appeared that the Loa microfilaraemia remained sta-le between D7 and D15 in the group treated with the low

Post-ivermectin Loa-related adverse effects 781

Table 2 Comparison of Loa microfilarial loads in two groups within the first two weeks following treatment with differentdoses of ivermectin

Day Low dosea (n = 51) Standard doseb (n = 28) P c

AM SD GM WM Reduction ratefrom D0 (%)

AM SD GM WM Reduction ratefrom D0 (%)

0 7131 5292 4868 4869 — 5115 4290 3244 3245 — 0.0893 3382 2963 2005 2007 52.6 1810 1941 929 931 64.6 0.0067 2839 2606 1634 1636 60.2 1458 1395 723 725 71.5 0.01115 3328 2899 —d 1738 53.3 1190 1184 594 596 76.7 0.0005

AM: arithmetic mean; GM: geometric mean; WM: Williams mean.a 1.5 mg ivermectin, i.e. 25 �g/kg for a 60 kg person.b 150 �g/kg ivermectin.c For each day of examination, the microfilarial loads were comparedd GM not calculable because one patient had a microfilaraemia equal t

Figure 2 Relationship between the absolute decrease in theLoa microfilaraemia between D0 and D3 after treatment [num-

t((lDat

3D

OuamywDsb1DiaPm

4

SieRagai

ber of microfilariae (mf) per ml of blood], and the concentrationof C-reactive protein (CRP) measured on D3. Each square rep-resents one single individual.

dose of ivermectin, whereas it continued to decrease dur-ing this period in the group treated with the standard dose.The analysis of variance confirmed that the treatment grouphad an influence on the values of the Loa microfilaraemia(F = 7.62; d.f. = 1; P = 0.007).

3.4. Efficacy of the three treatment regimens onLoa microfilaraemia after D15

We first considered the patients who were examined bothpre-treatment and on D30 (30 individuals in Group 1, 28 inGroup 2 and 29 in Group 3) (Table 3). The Kruskal-Wallistests showed that the microfilarial loads did not differ signifi-cantly among the three groups before treatment (P = 0.251),but they did so on D30 (P = 0.013). The ANOVA on repeatedmeasures confirmed that the effect of the three regimenson the Loa microfilaraemia differed significantly (F = 3.14;d.f. = 2; P = 0.048). The post-hoc tests, using the Bonferroni

corrections, showed that the microfilarial loads did not dif-fer significantly between Group 1 and Group 2 (P = 0.391),or between Group 2 and Group 3 (P = 0.99). By contrast,the microfilarial loads recorded in Groups 1 and 3 differedslightly (P = 0.045).

cate2

using the Mann-Whitney U test.o 0.

Another analysis was made on those individuals who par-icipated in the examinations performed on D0 and on D6030 individuals in Group 1, 27 in Group 2 and 29 in Group 3)Table 4). The Kruskal-Wallis tests showed that the microfi-arial loads in the three groups did not differ significantly on0 (P = 0.352), but did so on D60 (P = 0.016). However, thenalysis of variance did not show any difference among thehree groups (F = 2.319; d.f. = 2; P = 0.105).

.5. Effect of a second dose of ivermectin given at15

ne of the main objectives of the present study was to eval-ate whether a second low dose of ivermectin given 15 daysfter a first dose would lead to a further decrease in Loaicrofilaraemia. The statistical tests in this part of the anal-

sis were restricted to those individuals in Groups 1 and 2ho participated in the examination rounds on D15, D22 and30 (n = 24 in Group 1; n = 17 in Group 2). These analyseshowed that, after D15, the microfilarial loads increased inoth groups. The increase rates between D15 and D22 were4.0% in Group 1, and 17.3% in Group 2. Between D15 and30, the increase rates were 14.1% in Group 1, and 8.0%

n Group 2. The ANOVA on repeated measures (on D15, D22nd D30) did not show any group effect (F = 3.13; d.f. = 1;= 0.083). This result suggests that the second dose of iver-ectin had no significant effect on Loa microfilaraemia.

. Discussion

ince the early 1990s, several tens of cases of ivermectin-nduced encephalopathy have been reported from areasndemic for loiasis, especially in Cameroon, Democraticepublic of Congo and southern Sudan. Even if these eventsre very rare [the maximum incidence rate reported from aiven area is 1.9 per 10 000 patients treated (Twum-Dansond Meredith, 2003)], their possible fatal outcome and theirmpact on participation in community-directed treatments

onstitute an impediment to reaching APOC’s objectives. Inddition, if ivermectin mass treatments can possibly lowerhe intensity of transmission of W. bancrofti to such anxtent that transmission can be interrupted (Stolk et al.,003), the drug has no or little effect on the morbidity

782J.

Kamgno

etal.

Table 3 Comparison of Loa microfilarial loads measured on days 0 and 30 in three groups following treatment with different doses of ivermectin

Day Group 1a (n = 30) Group 2b (n = 28) Group 3c (n = 29) P d

AM SD WM Reduction ratefrom D0 (%)

AM SD WM Reduction ratefrom D0 (%)

AM SD WM Reduction ratefrom D0 (%)

0 7490 5498 5150 — 5807 4765 3795 — 5219 4229 3414 — 0.25130 3811 4010 1978 49.1 2443 2736 929 57.9 1403 1639 565 73.1 0.013

AM: arithmetic mean; WM: Williams mean.a Group 1: one low dose of ivermectin (1.5 mg) received on D0.b Group 2: two low doses of ivermectin (1.5 mg) received on D0 and D15.c Group 3: one standard dose of ivermectin (150 �g/kg) received on D0.d For each day of examination, the microfilarial loads were compared using the Kruskal-Wallis test.

Table 4 Comparison of Loa microfilarial loads measured on days 0 and 60 in three groups following treatment with different doses of ivermectin

Day Group 1a (n = 30) Group 2b (n = 27) Group 3c (n = 29) P d

AM SD WM Reduction ratefrom D0 (%)

AM SD WM Reduction ratefrom D0 (%)

AM SD WM Reduction ratefrom D0 (%)

0 7199 5551 4851 — 5361 4217 3583 — 5203 4246 3328 — 0.35260 3777 3527 1964 47.5 2470 3007 850 53.9 2005 2498 705 61.5 0.016

AM: arithmetic mean; WM: Williams mean.a Group 1: one low dose of ivermectin (1.5 mg) received on D0.b Group 2: two low doses of ivermectin (1.5 mg) received on D0 and D15.c Group 3: one standard dose of ivermectin (150 �g/kg) received on D0.d For each day of examination, the microfilarial loads were compared using the Kruskal-Wallis test.

anatryartidsHapwtwaopptilpft2lwiWurttiTtowilti

Talttdiswj

Post-ivermectin Loa-related adverse effects

related to LF. Due to this absence of benefit at an individuallevel, it is presently inconceivable to implement large-scaleivermectin treatments against LF in areas where loiasis is co-endemic. A number of studies have been developed in thelast 10 years to find solutions to the problem, and especiallyto prevent the occurrence of the Loa-related SAEs.

Assuming that the SAEs are caused by a rapid effect ofivermectin on great numbers of L. loa mf, several clinicaltrials have been performed to identify a treatment thatwould lower the Loa microfilarial loads progressively. Dif-ferent regimens of albendazole (Kamgno and Boussinesq,2002; Klion et al., 1993; Tabi et al., 2004; Tsague-Dongmoet al., 2002) or ivermectin (Kamgno et al., 2000) have beentested, and the best results were obtained with a daily treat-ment of albendazole administered for 3 weeks (Klion et al.,1993). Nonetheless, such a long regimen is not applicablefor large-scale treatments and, to prevent the neurologicalSAEs hampering the progress of mass treatment programmesfor onchocerciasis control and LF elimination in Africa, iden-tifying an appropriate operational treatment is still needed.

The present study demonstrated that a dose of 1.5 mgivermectin (25 �g/kg for a 60 kg person) brings about, withinthe days following treatment, a smaller decrease in Loamicrofilaraemia than the standard dose of 150 �g/kg. How-ever, the difference in the reduction rates recorded betweenD0 and D3 in these two groups (51.7 and 64.1%, respec-tively), though significant, is probably not sufficient for thelow-dose regimen to be regarded as an appropriate meansof preventing the neurological SAEs induced by ivermectinin individuals with high L. loa microfilaraemias.

One may wonder what a satisfactory reduction rate wouldbe, given our objective. Knowing that the neurological SAEsoccur only in those patients harbouring very high Loa micro-filarial loads, and that the relative reduction rate in themicrofilaraemia does not vary very much with the initialload (see Figure 1), one may assume that the risk is relatedto the absolute, rather than to the relative, number of mfthat disappear from the circulation. Based on this assump-tion, the crucial issue is thus to evaluate the minimumdecrease, between D0 and D3, that can provoke a neuro-logical SAE. To our knowledge, there are only three casesof Loa encephalopathy for whom the microfilaraemias havebeen measured before treatment and within the first daysafter treatment: in the first one, the values at D0, D1, D2and D3 were 162 920, 90 520, 7290 and 1880 mf/ml, respec-tively; in the second case, the microfilaraemias on D0 andD4 were 152 940 and 17 700 mf/ml; and in the third patient,the values on D0 and D4 were 50 520 and 1420 mf/ml, respec-tively (Boussinesq et al., 1998b). From the latter case, andassuming that the decrease continues between D3 and D4,one could postulate that a Loa encephalopathy may occurwhen 30 000 mf/ml disappear from the peripheral circula-tion between D0 and D3; and knowing that some individualsharbour some 200 000 mf/ml, one may estimate that thetreatment regimen we are seeking should bring about areduction of only 15% (= 30 000/200 000) of the initial Loamicrofilaraemia within the first 3 days. Very clearly, the

results of the present trial show that the effect of 1.5 mgivermectin (a dose of 25 �g/kg for a 60 kg person) farexceeds this value.

Further, several other results obtained during this studyare of particular importance in the context of identifying

etasb

783

n appropriate treatment to prevent the post-ivermectineurological SAEs. Firstly, we observed that the percent-ge reduction in the microfilarial load 3 days after eitherhe low dose or the standard treatment were fairly similaregardless of the initial level of microfilaraemia. A reanal-sis of the data collected by Kamgno et al. (2000) led tosimilar result: in that trial, in which the patients’ load

anged between 80 and 3560 mf/ml, the reduction rate inhe loads between D0 and D3 was not influenced by thenitial microfilaraemia, and this was true both after a lowose of 50 �g/kg (rs = 0.413; P = 0.235; n = 10) and after atandard dose of ivermectin (rs = 0.390; P = 0.188; n = 13).owever, it seems from the data collected by Ducorps etl. (1995) that this proportional relationship between there- and post-treatment microfilaraemias no longer existshen the microfilarial loads exceed a given level. In that

rial, a total of 108 patients were selected, including 32ith a microfilarial load above 30 000 mf/ml (Ducorps etl., 1995). All of them were treated with a single dosef ivermectin at 200 �g/kg. When one considers all theseatients together, the correlation coefficient between there-treatment microfilaraemia and the reduction rate ofhe loads between D0 and D3 was rs = 0.314 (P = 10−3),.e. the reduction rate depended on the initial microfi-arial load. However, when one considers separately theatients harbouring high and low loads, no correlation wasound between pre-treatment microfilaraemia and post-reatment reduction rate in those who harboured less than0 000 mf/ml (rs = 0.176; P = 0.182; n = 59) — a result simi-ar to that recorded in the present study — whereas thereas a significant correlation (rs = 0.443; P = 0.030; n = 24)

n those patients who had more than 35 000 Loa mf/ml.hen one considers the results of this trial at an individ-

al level, it seems that, despite a wide variation, the highereduction rates were recorded in patients harbouring morehan 30 000 mf/ml. For instance, still in the same trial, allhose seven patients who showed a reduction rate exceed-ng 95% presented pre-treatment loads above 30 000 mf/ml.hese findings are important because they demonstratehat, when performing trials aiming at studying the effectsf ivermectin on Loa loa, the results obtained from patientsith relatively low loads cannot be extrapolated to those

ndividuals who harbour more than 30 000 mf/ml. This hasogistical consequences, because the selection process ofrial patients is obviously more complicated when one has todentify individuals harbouring very high microfilarial loads.

One may also try to explain the phenomenon observed.he fact that the reduction rates in the microfilaraemiasre more or less constant in the subpopulation harbouringess than 30 000 mf/ml may perhaps be due to the fact thathere is, in the global parasite reservoir, a given propor-ion of Loa mf that are sensitive to ivermectin, and that areestroyed when they are exposed to the drug. Besides this,n those individuals harbouring very high loads, one may con-ider the possibility that a certain proportion of mf, whichould normally not be affected by the treatment, are sub-

ected to phenomena that lead to their disappearance. For

xample, one may imagine that they are ‘trapped’ withinhe mass of ivermectin-susceptible, and thus paralyzed, mf,nd ‘dragged down’ by the latter in the bloodstream and theubsequent destruction process. Another explanation mighte that when a large number of mf are damaged by the

7

dil

tmismpILdidi(cilwltGnon‘

Gataftdosswndlhmte

ctt(gttotrhMba

2ai

5

AoasmrLtdplconfsitdwttr

codaoactst2

AtJdfi

AEWtcp

F

84

rug, and destroyed, this induces such a strong immunolog-cal response that it would lead to the destruction of someess ‘sensitive’ mf.

A second interesting finding from the present study is thathe inter-individual variability in the mf reduction rates wasore substantial in the groups treated with a low dose of

vermectin (Groups 1+2) than in the group treated with atandard dose (Group 3). This may mean that, among thosef that are sensitive to a standard dose of ivermectin, aroportion would be less, or not, affected by lower doses.ndeed, the present study, and also previous ones (Richard-enoble et al., 1988), have demonstrated that there is aose-effect of ivermectin on the L. loa microfilaraemia, andt has been suggested that the interval between the minimalose that exerts its full effect on the Loa mf, and the max-mal dose that has no effect at all, might be fairly narrowKamgno et al., 2000). The results of the present trial indi-ate that when the low dose has been used (±25 �g/kg),n some patients, it has no effect at all on the microfi-araemia; and one may suppose that this lack of responseould be more frequent among patients treated with even

ower doses of ivermectin. This is supported by the facthat, even if the exact doses received by those patients inroups 1 and 2 who responded the least to treatment didot differ from the doses received by the other individualsf these groups, the result of the comparison (P = 0.063) wasear the limits of significance, and that only one of the 11low-responders’ had received a dose exceeding 26 �g/kg.

Another interesting, and unexpected, result is that, inroup 2, the second dose of ivermectin did not bring aboutny further decrease in the Loa microfilaraemia. In spite ofhe fact that the number of such patients included in thenalysis was fairly low, it appears that the decrease, if any,ollowing the second dose (after D15) was far less markedhan that recorded between D0 and D7 (i.e. after the firstose). This reinforces the hypothesis proposed above, thatnly a proportion of Loa mf are sensitive to ivermectin. Ifuch be the case, the first dose would eliminate most of theensitive mf in patients in Group 2, and the second doseould thus have no effect on the remaining mf. An alter-ative hypothesis would be that the effect of ivermectinepends upon the host’s immunological status, and that theatter has been modified after the first dose. Ali et al. (2002)ave shown that the response of Onchocerca volvulus to iver-ectin may vary significantly between patients, according

o their immunocompetence. A similar phenomenon mightxist with L. loa.

Lastly, we observed that the CRP values at D3 werelearly increased compared with the standard concentra-ion. Although not significantly different, the CRP levelsended to be higher in Group 3 than in the low-dose groupGroup 1+2). As shown, at an individual level, in the latterroup, the increase in the CRP levels is probably linked tohe absolute number of mf killed after treatment. A similarrend has also been described after ivermectin treatmentf patients infected with O. volvulus (Njoo et al., 1994). Inhe latter study, the increase in the CRP levels was also cor-

elated with adverse reactions and microfilarial densities. Itas been demonstrated that, in onchocerciasis patients, thisazzotti-type reaction was linked to the presence of Wol-achia in the parasite (Keiser et al., 2002). As these bacteriare absent from L. loa (Buttner et al., 2003; McGarry et al.,

dM

C

J. Kamgno et al.

003), our results suggest that the inflammatory processesre also, partly, related to a reaction against the parasitetself.

. Conclusions

single dose of ivermectin at 25 �g/kg (for a 60 kg person),r two such doses given with a 15 day interval, brings aboutlower reduction rate in the Loa microfilaraemia than a

tandard dose of 150 �g/kg. However, given the relativelyarked effect of these low doses, one may assume that such

egimens cannot be regarded as appropriate to prevent theoa-related neurological SAEs that may follow ivermectinreatment. In addition, some of our results suggest that theose of 25 �g/kg might be near the upper limit of the thera-eutic index of ivermectin against L. loa. Consequently, it isikely that a treatment with even lower doses would be asso-iated with a great inter-individual variability in the effectf the drug on the Loa microfilaraemia, a result that wouldot meet what is expected from the regimen we are seekingor use in public health programmes, which require a wideafety margin. Lastly, even if the prevention of the post-vermectin neurological SAEs is indeed a priority, it is clearhat the application of large-scale treatments with loweroses of ivermectin should be considered with caution, inell-defined areas: such an intervention might encourage

he selection of O. volvulus parasites that are less sensi-ive to ivermectin, possibly ending with the emergence of aesistance.

The results of the present study make it more urgent toonduct further trials to evaluate the effects of other drugsn Loa microfilaraemia. Studies are ongoing using singleoses of albendazole repeated at 2 month intervals, or somentimalarials that have shown to have a significant effect onther filarial species (Guderian et al., 1991; VandeWaa etl., 1989). While awaiting practical solutions to lower effi-iently, safely and on a large-scale the reservoir of L. loa inhose areas targeted for large-scale control of onchocercia-is and/or LF, efforts should be pursued to delineate, usinghe rapid assessment (RAPLOA) method (Takougang et al.,002), the areas at risk of post-ivermectin Loa-related SAEs.

uthors’ contributions: JK, NAYT-D, BT and MB designedhe study protocol; JK and MCT conducted the field trial;K, SDSP and MB carried out the statistical analyses andrafted the manuscript. All authors read and approved thenal manuscript; JK is guarantor of the paper.

cknowledgements: We thank the inhabitants of Awae I,lat, Minkomila, Odoudouma, Olanguina, Sanguela and Zili.e are also grateful to the Awae Health District team for

heir assistance. We thank Dr Eric Sunjio, who prepared theapsules used during the trial, and Aurelia Vessiere, whorepared the individual dosages for each patient.

unding: Mectizan Donation Program. SDSP thanks the Fon-

ation de France and the Fondation pour la Rechercheedicale for financial support.

onflicts of interest: None declared.

K

K

K

M

M

N

R

S

T

T

T

T

Post-ivermectin Loa-related adverse effects

Ethical approval: National Ethics Committee of Cameroon.

References

Ali, M.M.M., Mukhtar, M.M., Baraka, O.Z., Homeida, M.M.A.,Kheir, M.M., Mackenzie, C.D., 2002. Immunocompetence maybe important in the effectiveness of Mectizan® (ivermectin)in the treatment of human onchocerciasis. Acta Trop. 84,49—53.

Boussinesq, M., Kamgno, J., Eloundou Onomo, P., Befidi, R., Gardon,J., 1998a. Evaluation du patch a la DEC chez les sujets infectespar Loa loa. Document du Laboratoire mixte CPC/ORSTOMd’Epidemiologie et de Sante publique no 98-16, Centre Pasteurdu Cameroun, Yaounde.

Boussinesq, M., Gardon, J., Gardon-Wendel, N., Kamgno, J.,Ngoumou, P., Chippaux, J.P., 1998b. Three probable cases ofLoa loa encephalopathy following ivermectin treatment foronchocerciasis. Am. J. Trop. Med. Hyg. 58, 461—469.

Boussinesq, M., Gardon, J., Gardon-Wendel, N., Chippaux, J.P.,2003. Clinical picture, epidemiology, and outcome of Loa-associated adverse events related to mass ivermectin treatmentof onchocerciasis in Cameroon. Filaria J. 2 (Suppl. I), S4.

Buttner, D.W., Wanji, S., Bazzocchi, C., Bain, O., Fischer, P., 2003.Obligatory symbiotic Wolbachia endobacteria are absent fromLoa loa. Filaria J. 2, 10.

Ducorps, M., Gardon-Wendel, N., Ranque, S., Ndong, W., Boussi-nesq, M., Gardon, J., Schneider, D., Chippaux, J.P., 1995. Effetssecondaires du traitement de la loase hypermicrofilaremique parl’ivermectine. Bull. Soc. Pathol. Exot. 88, 105—112.

Fobi, G., Gardon, J., Santiago, M., Demanga-Ngangue, Gardon-Wendel, N., Boussinesq, M., 2000. Ocular findings aftertreatment of patients with high Loa loa microfilaremia. Oph-thalmic Epidemiol. 7, 27—39.

Gardon, J., Gardon-Wendel, N., Demanga-Ngangue, Kamgno, J.,Chippaux, J.P., Boussinesq, M., 1997. Serious reactions aftermass treatment of onchocerciasis with ivermectin in an areaendemic for Loa loa infection. Lancet 350, 18—22.

Guderian, R.H., Anselmi, M., Beck, B.J., Mackenzie, C.D., Williams,J.F., Proano S., J.R., Cooper, P.J., 1991. The effect of antimalar-ial chloroquine therapy and prophylaxis on concurrent infectionwith Onchocerca volvulus in Ecuador. Trans. R. Soc. Trop. Med.Hyg. 85, 634—638.

Haselow, N.J., Akame, J., Evini, C., Akongo, S., 2003. Program-matic and communication issues in relation to serious adverseevents following ivermectin treatment in areas co-endemic foronchocerciasis and loiasis. Filaria J. 2 (Suppl. I), S10.

Kamgno, J., Boussinesq, M., 2002. Effect of a single dose (600 mg)

of albendazole on Loa loa microfilaraemia. Parasite 9, 59—63.

Kamgno, J., Gardon, J., Boussinesq, M., 2000. Essai deprevention des encephalopathies a Loa loa post-ivermectinepar l’administration d’une faible dose initiale. Med. Trop. 60,275—277.

V

785

amgno, J., Tsague-Dongmo, L., Gardon, J., Boussinesq, M., 2001.Donnees complementaires sur l’endemie a Loa loa dans laProvince du Centre du Cameroun. Bull. Liais. Doc. OCEAC 34,23—27.

eiser, P.B., Reynolds, S.M., Awadzi, K., Ottesen, E.A., Taylor, M.J.,Nutman, T.B., 2002. Bacterial endosymbionts of Onchocercavolvulus in the pathogenesis of posttreatment reactions. J.Infect. Dis. 185, 805—811.

lion, A.D., Massougbodji, A., Horton, J., Ekoue, S., Lanmasso, T.,Ahouissou, N.L., Nutman, T.B., 1993. Albendazole in human loia-sis: results of a double-blind, placebo-controlled trial. J. Infect.Dis. 168, 202—206.

alecela-Lazaro, M., Twum-Danso, N., 2004. Towards a strategicplan for research to support the global program to eliminate lym-phatic filariasis. Summary of immediate needs and opportunitiesfor research on lymphatic filariasis. 2. 4 Program implementa-tion. Am. J. Trop. Med. Hyg. 71 (Suppl. 5), 16—19.

cGarry, H.F., Pfarr, K., Egerton, G., Hoerauf, A., Akue, J.P., Eny-ong, P., Wanji, S., Klager, S.L., Bianco, A.E., Beeching, N.J.,Taylor, M.J., 2003. Evidence against Wolbachia symbiosis in Loaloa. Filaria J. 2, 9.

joo, F.L., Hack, C.E., Oosting, J., Luyendijk, L., Stilma, J.S., Kijl-stra, A., 1994. C-reactive protein and interleukin-6 are elevatedin onchocerciasis patients after ivermectin treatment. J. Infect.Dis. 170, 663—668.

ichard-Lenoble, D., Kombila, M., Rupp, E.A., Pappayliou, E.S.,Gaxotte, P., Nguiri, C., Aziz, M.A., 1988. Ivermectin in loiasisand concomitant O. volvulus and M. perstans infections. Am. J.Trop. Med. Hyg. 39, 480—483.

tolk, W.A., Swaminathan, S., van Oortmarssen, G., Das, P.K.,Habbema, J.D.F., 2003. Prospects for elimination of Bancrof-tian filariasis by mass drug treatment in Pondicherry, India: asimulation study. J. Infect. Dis. 188, 1371—1381.

abi, T.E., Befifi-Mengue, R., Nutman, T.B., Horton, J., Folefack,A., Pensia, E., Fualem, R., Fogako, J., Gwanmesia, P., Quakyi, I.,Leke, R., 2004. Human loiasis in a Cameroonian village: a double-blind, placebo-controlled, crossover clinical trial of a three-dayalbendazole regimen. Am. J. Trop. Med. Hyg. 71, 211—215.

akougang, I., Meremikwu, M., Wandji, S., Yenshu, E.V., Aripko, B.,Lamlenn, S.B., Eka, B.L., Enyong, P., Meli, J., Kale, O., Remme,J.H., 2002. Rapid assessment method for prevalence and inten-sity of Loa loa infection. Bull. World Health Organ. 80, 852—858.

sague-Dongmo, L., Kamgno, J., Pion, S.D.S., Moyou-Somo, R.,Boussinesq, M., 2002. Effects of a three-day regimen of albenda-zole (800 mg daily) on Loa loa microfilaraemia. Ann. Trop. Med.Parasitol. 96, 707—715.

wum-Danso, N.A.Y., Meredith, S.E.O., 2003. Variation in incidenceof serious adverse events after onchocerciasis treatment with

ivermectin in areas of Cameroon co-endemic for loiasis. Trop.Med. Int. Health 8, 820—831.

andeWaa, E.A., Bennett, J.L., Williams, J.F., Satti, M.Z., Geary,T.G., 1989. Anti-filarial effects of nine quinoline-containingdrugs on adult filariae in vitro. J. Parasitol. 75, 367—372.