NEUROLOGICAL PROGRESS The Continuing Problem of Human

The Continuing Problem of Human AfricanTrypanosomiasis (Sleeping Sickness)

Peter G. E. Kennedy, MD, PhD, DSc

Human African trypanosomiasis, also known as sleeping sickness, is a neglected disease, and it continues to pose a major threatto 60 million people in 36 countries in sub-Saharan Africa. Transmitted by the bite of the tsetse fly, the disease is caused byprotozoan parasites of the genus Trypanosoma and comes in two types: East African human African trypanosomiasis caused byTrypanosoma brucei rhodesiense and the West African form caused by Trypanosoma brucei gambiense. There is an early or hemo-lymphatic stage and a late or encephalitic stage, when the parasites cross the blood–brain barrier to invade the central nervoussystem. Two critical current issues are disease staging and drug therapy, especially for late-stage disease. Lumbar puncture toanalyze cerebrospinal fluid will remain the only method of disease staging until reliable noninvasive methods are developed, butthere is no widespread consensus as to what exactly defines biologically central nervous system disease or what specific cerebro-spinal fluid findings should justify drug therapy for late-stage involvement. All four main drugs used for human African trypano-somiasis are toxic, and melarsoprol, the only drug that is effective for both types of central nervous system disease, is so toxic thatit kills 5% of patients who receive it. Eflornithine, alone or combined with nifurtimox, is being used increasingly as first-linetherapy for gambiense disease. There is a pressing need for an effective, safe oral drug for both stages of the disease, but this willrequire a significant increase in investment for new drug discovery from Western governments and the pharmaceutical industry.

Ann Neurol 2008;64:116–127

Human African trypanosomiasis (HAT), which is alsoknown as sleeping sickness, is one of the “neglecteddiseases,” a group that includes visceral leishmaniasis,schistosomiasis, and Chagas’ disease.1 Although thesediseases kill or disable hundreds of thousands of peoplein underdeveloped tropical regions, current treatmentfor them is often antiquated, highly toxic, and fre-quently ineffective. The pharmaceutical industry andWestern governments have until recently shown littleinterest in developing new drugs for these diseases be-cause this is associated with little or no prospect ofgenerating significant short- or long-term financialgain. Although there has been an entirely understand-able emphasis in recent times on combating suchglobal killers as malaria, acquired immune deficiencysyndrome, and tuberculosis, it should be appreciatedthat HAT is a major threat to the health of 60 millionpeople in 36 countries in sub-Saharan Africa.2 More-over, HAT is the world’s third most important para-sitic disease affecting human health after malaria andschistosomiasis, as defined by the global burden of par-asitic disease, calculated as the disability adjusted lifeyears lost.1

HAT is caused by protozoan parasites of the genusTrypanosoma, single-celled organisms that remain in ex-

tracellular form in the host. There are two forms of thehuman disease, the East African variant caused byTrypanosoma brucei rhodesiense and the West Africanform caused by Trypanosoma brucei gambiense3 (Fig 1).If untreated, the disease is always fatal. Transmission ofthe disease in both humans and cattle is by the bite ofthe blood-sucking tsetse fly of the Glossina species.4 In-festation by the tsetse fly covers 10 million square ki-lometers, one third of Africa’s landmass, which is anarea slightly larger than the United States.1 African an-imal trypanosomiasis, important in domestic livestocksuch as cattle that suffer from a wasting disease callednagana, as well as wild animals, was first shown to becaused by Trypanosoma brucei by David Bruce in 1899while investigating a major outbreak of nagana in Zu-luland.5,6 Subsequent work by Aldo Castellani enabledthe identification of trypanosomes in the blood and ce-rebrospinal fluid (CSF) in human patients with HATin 1903,6,7 and parasites causing the two human dis-ease variants were identified during the period 1902 to1910.6 Both animals and humans can act as reservoirsof parasites capable of causing the human disease, butthe detailed mechanisms by which this occurs are notfully understood. Animal trypanosomiasis has a majorhuman and economic impact because it adversely af-

From the Department of Neurology, Division of Clinical Neuro-sciences, Faculty of Medicine, University of Glasgow Institute ofNeurological Sciences, Southern General Hospital, Glasgow, GS1,4TF, Scotland, UK.

Received Mar 11, 2008, and in revised form Apr 23. Accepted forpublication May 1, 2008.

Published online in Wiley InterScience (www.interscience.wiley.com).DOI: 10.1002/ana.21429

Address correspondence to Dr Kennedy, Southern General Hospi-tal, Glasgow G51 4TF, United Kingdom.E-mail: [email protected]

NEUROLOGICAL PROGRESS

116 Published 2008 by Wiley-Liss, Inc., through Wiley Subscription Services

fects livestock production and farming. During the firsthalf of the twentieth century, HAT caused by T.b.gambiense decimated entire communities in central Af-rica,8 but then the disease was almost brought undercontrol during the 1960s primarily as a result of highlyeffective surveillance programs. But HAT then quicklyreemerged with a progressive increase in the numbersof new cases and deaths. The World Health Organiza-tion (WHO) provided estimates during the period1986 to 2004 about the disease that have been widelyquoted, with an annual prevalence of 300,000 to500,000 cases.2,9,10 Factors causing this increase wereprimarily war and famine, which resulted in severe dis-ruption of disease surveillance and treatment, especiallyin Uganda, Angola, Sudan, and the Congos where thedisease occurred in epidemics.11 Although it is still dif-ficult to provide accurate estimates of disease incidenceand prevalence, more recent WHO estimates have sug-gested that as a result of more efficient surveillance, asignificant improvement has occurred with currently aslow as 70,000 existing cases, mainly infected with T.b-.gambiense.10 However, HAT has already demonstratedits ability to recur even after it had been virtuallybrought under control. About 50 cases of HAT occurannually outside of Africa,12 usually as a result ofWestern travelers returning to North America or Eu-rope from the East African game reserves; therefore, allphysicians need to be aware of the key features of thedisease and its most appropriate drug therapy.

Outline of Parasite-Host BiologyDetails of parasite-host biology have been given else-where.3,4 Although animals are the main reservoir forT.b. rhodesiense parasites, humans are the main reser-voir for T.b. gambiense parasites.1,4 In brief, the tsetsefly vector feeds on an infected animal or human, afterwhich the ingested trypanosomes undergo a number ofbiochemical and structural alterations in the fly’s mid-gut. Infective forms of the trypanosome then reach thefly’s salivary glands from which they are transmitted tothe human host through biting. A fly remains infectivefor life, and the whole infective cycle is probably com-pleted successfully in only 1 in 10 flies.4 Approximately5 to 15 days after infection, a painful skin lesion calleda trypanosomal chancre may develop at the site of thebite.4 The parasites spread in the host bloodstream 1 to3 weeks after the initial bite, and invade the lymphnodes and systemic organs including the liver, spleen,heart, endocrine system, and eyes in what is termed theearly, stage 1, or hemolymphatic stage.3,4 If untreated,within a few weeks in the case of rhodesiense infection,or many months in the case of gambiense infection, theparasites will cross the blood–brain barrier (BBB) andenter the central nervous system (CNS), which marksthe late, stage 2, or encephalitic stage of the disease.4,13

The entire tempo of the disease is faster in the moreaggressive rhodesiense infection compared with thechronic gambiense infection, probably as a result of thegreater adaptation of the latter parasite to the host.1,3

Much is known about the molecular biology of thetrypanosome, and the entire T. brucei genome was se-quenced in 2005.14 It has about 9,000 genes, about10% of which are variable surface glycoprotein (VSG)genes encoding the VSG that are distributed on theentire surface of the trypanosome.3 During infection,the trypanosome is able to rapidly switch the expres-sion of the VSG genes in and out of the expressionsite, the result of which is antigenic variation in whichthe surface VSG genes change so fast that the parasiteis able to constantly evade the host’s immune re-sponse.15 For this reason, it has not been possible, sofar, to develop a vaccine for HAT.

Clinical Features of the DiseaseThere is seldom a clear clinical distinction between theearly and late stages of HAT that may appear to runinto each other. Patients in the early, or hemolym-phatic, stage may report nonspecific symptoms such asmalaise, headache, weight loss, arthralgia, and fatigue,and also have episodes of fever accompanied by rigorsand vomiting, which may be misdiagnosed as malar-ia.13,16 There may also be generalized lymphadenopa-thy, and enlargement of posterior cervical lymph nodesis typical of gambiense disease (“Winterbottom’s sign”).Other symptoms and signs may correspond to partic-ular organ involvement. Thus, there may be several dif-

Fig 1. Distribution of East and West African sleeping sicknessin sub-Saharan Africa. Green areas represent Trypanosomabrucei gambiense infection; brown areas represent Trypano-soma brucei rhodesiense infection. (Modified from Atouguiaand Kennedy.)

Kennedy: Human African Trypanosomiasis 117

ferent kinds of skin rash, as well as pruritus, especiallyin European patients in whom a macular, irregular, ev-anescent rash has been described as occurring in theshoulders, trunk, and upper legs.17 Derangement ofliver function, hepatomegaly, a mainly hemolytic ane-mia, splenomegaly and cardiac dysfunction such astachycardia, myocarditis, pericarditis, and congestivecardiac failure have been described.16,17 Many types ofendocrine dysfunction may occur such as loss of libidoand impotence, problems with menstrual function andfertility (abortion, premature births or stillbirths, steril-ity), hair loss, gynecomastia, orchitis, and testicular at-rophy.4,16 Parasite invasion of the eye may result iniritis, conjunctivitis, iridocyclitis, keratitis, and choroi-dal atrophy.4,16 Patients are also prone to facial edema.

The neurological features of late- or encephalitic-stage HAT are also protean and are summarized in Ta-ble 1. The various psychiatric and mental symptomssuch as anxiety, lassitude and indifference, agitation, ir-ritability, mania, sexual hyperactivity, suicidal tenden-cies, and hallucinations could all be misdiagnosed ei-ther as only early-stage disease or as manifestations of aprimary psychiatric illness, but they are unlikely to oc-cur in isolation. The neurological symptoms and signs,which may develop insidiously, have been describedpreviously4,13,16,17 and are summarized here. A largenumber of motor features may occur in late-stage HATwith virtually every motor system at risk. Tremors inthe hands and tongue are common, choreiform move-ments of the head, limbs, and trunk may occur, as maypyramidal weakness of the limbs.4 Lower limb paralysismay also occur as a result of spinal cord involvement(myelopathy or myelitis) or peripheral motor neuropa-thy. Patients may also show cerebellar ataxia with walk-ing difficulties and slurred speech. Muscle fasciculationmay also be a feature.4 The most characteristic sensoryfeature of late-stage HAT is a painful limb hyperesthe-sia that, when it has a deep quality, is called Kerandel’ssign and is common in European patients, a quarter ofwhom experience it, but is unusual in African suffer-ers.17 Abnormal reflexes indicating frontal lobe involve-ment such as pout reflex and palmomental reflexes mayalso be present. The visual system may also be affectedby late-stage disease producing diplopia, optic neuritis,papilledema, and subsequent optic atrophy.13,16

The characteristic sleep disturbances that occur inthe encephalitic stage give the disease its commonname. There is a disruption of the normal sleep/wakecycle so the patient sleeps during the day but has noc-turnal insomnia.18 Uncontrollable urges to sleep occurwithout warning, and this becomes continuous in thefinal stages. Recently, it has been shown using poly-somnography that there is an alteration of sleep struc-ture in these patients with the frequent onset of sleeponset of rapid eye movements.18 Normally, REM sleepoccurs at the end of stage 4 sleep; these abnormalities

resolve with effective treatment. If untreated, or unsuc-cessfully treated, the natural course of HAT is for thepatient to progressively deteriorate with increasingsleep disturbances, cerebral edema, incontinence, men-tal deterioration, seizures, and finally, death. The entire

Table 1. Neurological Features of Human AfricanTrypanosomiasis

Psychiatric and mental featuresLassitude and mental disturbancesAnxiety and irritabilityBehavior disturbances (eg, violence, suicidaltendencies)Uncontrolled sexual impulsesHallucinations, delirium

Sleep disturbancesReversal of normal sleep/wake cycleDaytime somnolenceNocturnal insomniaUncontrollable urges to sleepAlteration of sleep structure with sleep onset ofREM sleep

Motor disturbancesPyramidal weaknessExtrapyramidal features (tremors and abnormalmovements)Myelopathy and myelitisMuscle fasciculationSlurred speechCerebellar ataxiaPeripheral motor neuropathyPout reflexPalmarmental reflexes

Sensory disturbancesPruritusDeep hyperesthesiaAlso anesthesia, paraesthesia

Visual involvementDiplopiaOptic neuritisPapilledemaOptic atrophy

Drug inducedPeripheral neuropathyPosttreatment reactive encephalopathyMultifocal inflammatory syndromeSeizures

Modified from Atouguia JLM, Kennedy PGE. Neurologicalaspects of human African trypanosomiasis. In: Davis LE,Kennedy PGE, eds. Infectious diseases of the nervous system.Oxford: Butterworth-Heinemann, 2000:321-372, bypermission.

118 Annals of Neurology Vol 64 No 2 August 2008

course of the disease to death takes several weeks inrhodesiense and several months or even years in gambi-ense disease. Because of the relative paucity of detailedsystematic studies of the neurological features of HAT,it is difficult to say precisely how frequently the varioussymptoms and signs occur individually or in combina-tion. However, an early insight into this issue was pro-vided by the classic study of Duggan and Hutching-ton17 of 109 cases of HAT in Europeans. They foundthat in patients with both early- and late-stage disease,the percentage of patients showing somnolence was37.8%, headache was 24.5%, hyperesthesia was 26.6%,tremor and abnormal movements was 25.7%, psychi-atric symptoms was 20.1%, ataxia was 16.6%, andslurred speech was 10.6%. Blum and colleagues19 re-cently provided considerable clarification of neurologi-cal feature frequency in HAT. In what is the most ex-tensive analysis performed to date, these authorsdefined the frequency of specific neurological involve-ment in a total of 2,541 patients with late-stage HATover a period of 3 years. They reported that the per-centage of patients showing a sleep disorder was, aswould be expected, high at 74.4%, motor weaknesswas 35.4%, gait disturbance was 22%, tremor was21.2%, headache was 78.7%, behavior disturbance was25%, speech impairment was 14.2%, and abnormalmovements was 10.7%.19 As the authors pointed out,the reasons for the high variability of neurologicalsymptoms and signs in different geographic areas andin the previous publications of HAT have yet to beclarified.

Diagnostic ConsiderationsThe correct diagnosis of HAT can be suspected in theappropriate context that occurs when an individual de-velops a fever and suggestive symptoms in a HAT-endemic or epidemic region. The most importantdifferential diagnosis is malaria, especially when inap-propriate antimalarial treatment is given to a patientwho actually has HAT and who then shows a tempo-rary reduction of an intermittent fever.4,13 This dan-gerous situation may be complicated by the fact thatHAT and malaria may occur together in the same pa-tient. Other infectious diseases that may enter into thedifferential include human immunodeficiency virus in-fection, tuberculosis, toxoplasmosis, viral encephalitis,brucellosis, lymphoma, typhoid fever, and hookworm.4

The physician confronted with a patient with HATwould expect to find a number of nonspecific periph-eral blood abnormalities such as a mainly hemolyticanemia, increased erythrocyte sedimentation rate,thrombocytopenia, abnormal liver function tests, in-creased IgM antibodies, and possibly a range of auto-antibodies resulting from autoimmune responses.4 Thedefinitive method of establishing a diagnosis of HAT isby demonstrating the presence of trypanosomes in the

peripheral blood or lymph-node aspirates (Fig 2). Thisis easier in rhodesiense infection because of the persis-tently high parasitemia that generally occurs than ingambiense disease, where the parasitemia tends to below. In the latter case, parasitological confirmation us-ing concentration techniques is usually preceded by se-rological suspicion established using the card agglutina-tion trypanosomiasis test, which is simple, quick, andeasy to perform.20 Problems in management may arise,however, when the card agglutination trypanosomiasistest is equivocally positive. Recently, new molecular di-agnostic techniques for diagnosing HAT have beentested. Thus, DNA amplification techniques such aspolymerase chain reaction21 and loop-mediated ampli-fication22 have been used to detect trypanosomes inpatients’ peripheral blood, CSF, or both. Polymerasechain reaction may have a high sensitivity rate, but it isalso associated with problems with test reproducibility,an issue that appears to be less of a problem with loop-mediated amplification.22

One of the most important issues in HAT is thecorrect staging of the disease so that the early and latestages can be distinguished reliably. This is absolutelycritical because the current treatment of late-stage dis-ease, when the parasites have invaded the CNS, is sotoxic23 (see later). Because there are no reliable markersof early-stage disease, all patients who are suspected ofhaving CNS involvement, including all those with apositive card agglutination trypanosomiasis test, mustundergo a lumbar puncture to examine the CSF. Thephysician dealing with such a case might expect a typ-ical late-stage HAT CSF to show a pleocytosis, mainlylymphocytes, with a white blood cell (WBC) count be-tween 0 and 300/�l, and at times even as high as1,000/�l or more, a moderate increase in protein con-centration between 40 and 200mg/100ml, and an in-crease in the immunoglobulin concentration, especiallyIgM caused by the strong IgM intrathecal synthesis.4,24

To give some idea of the most frequently encounteredCSF findings, in a major study of 181 patients with

Fig 2. Giemsa-stained light photomicrograph showing the pres-ence of Trypanosoma brucei parasites (original magnification�1,000), which were found in a thin film blood smear.(Courtesy Centers for Disease Control and Prevention/Dr MaeMelvin, Centers for Disease Control Public Health ImageLibrary)


late second stage gambiense disease, the median CSFWBC count was 93/�l, with an interquartile range of22 to 266/�l and a maximum of 1,430/�l.25 In thesame study, the median CSF protein was 78.7mg/100ml, with an interquartile range of 45.4 to106.5mg/100ml, and the greatest value was 203.8mg/100ml.25 But there is not a universal agreement as towhat criteria define CNS involvement (apart from thedetection of trypanosomes in the CSF, which is seldomeasy). Moreover, there is also some disagreement as towhich CSF criteria actually determine whether the pa-tient should be treated with late-stage–specificdrugs.3,24 The WHO criteria are the most commonlyused and define late-stage HAT as the presence of try-panosomes in the CSF and/or more than fiveWBCs/�l in the CSF.9 However, some clinicians inWest Africa use a higher cutoff WBC count of morethan 20,25 and others have suggested, reasonably in myview, a midway figure of 10 WBCs/�l.26 The CSFIgM has also been shown to be a useful indicator ofCNS involvement.25 There is an urgent need for amore reliable diagnostic test for late-stage HAT that ischeap, reliable, easy to perform, field adaptable, andwith a high sensitivity and specificity.27 A particularproblem with introducing a new test is that there iscurrently no “gold standard” test with which to com-pare it.27 Currently, there is no noninvasive test thatcan reliably distinguish early- from late-stage disease.

There have been a few case reports of the use ofmagnetic resonance imaging (MRI) in CNS HAT, al-though these have been, unsurprisingly, in patients re-turning or managed in Western hospitals. MRI find-ings are nonspecific, although consistent findings haveincluded diffuse hyperintensities in the basal ganglia,internal and external capsules, asymmetric white matterabnormalities (Fig 3), and ventricular enlarge-ment.4,28,29 Based on understandably limited data,MRI appears to be more sensitive than computed to-mography in detecting HAT-associated abnormali-ties,28 although computed tomography has been re-ported as showing such changes as focal low densitiesin the internal capsules and centrum semiovale and ce-rebral edema.4,28 Recent reports on two patients in theAmerican neurological literature have indicated thatMRI may be useful in the diagnosis of HAT in pa-tients returning from visits to Africa, and may also helpin distinguishing different CNS syndromes seen inCNS disease.30–32 As expected, patients with late-stageHAT have an abnormal electroencephalogram, withthree different types of nonspecific abnormalities thatboth mirror the severity of the disease and improvemarkedly with successful treatment.4 These electroen-cephalographic types are a sustained low- voltage back-ground similar to that seen during light sleep, paroxys-mal waves, or various types of delta wave and rapid

intermittent high-voltage delta bursts between periodsof lower-voltage delta activity.4

Current Treatment of Early- and Late-StageDiseaseCurrent drug therapy for both early- and late-stageHAT is unsatisfactory with a heavy reliance on fourmain highly toxic drugs, most of which would haveprobably failed current rigorous safety standards.33

Three of the drugs, suramin, pentamidine and melar-soprol, were developed during the first half of thetwentieth century, and the last drug to be registered forHAT, namely, eflornithine (also known as DFMO),was registered in 1981.1,3,23 Another drug named ni-furtimox is registered for Chagas’ disease and is cur-rently under close evaluation in promising trials ofcombination chemotherapy. This depressing paucity ofsafe and effective drugs for HAT has recently attractedincreasing attention from nongovernmental organiza-tions and other funding bodies.

The standard treatment for early-stage rhodesiensedisease is intravenous suramin given as five injectionsover 3 weeks. Important adverse effects include ana-phylactic shock, renal failure, and skin lesions. Early-stage gambiense disease is treated with intramuscularpentamidine given as daily injections over 7 to 10 days.Important adverse effects of pentamidine include hypo-glycemia, hyperglycemia, and hypotension.3,4 Thesedrugs are usually effective if treatment is started early.The treatment of late-stage HAT is even more prob-lematic because the only drug that is effective for bothtypes of the disease is the highly toxic arsenical melar-soprol (Mel B), which was first used for such patients

Fig 3. Magnetic resonance imaging scan of a 13-year-old pa-tient with central nervous system human African trypanosomi-asis 3 years after successful completion of multiple treatmentsfor numerous relapses. The scan shows ventricular enlargement(especially of the frontal horns) and diffuse white matterchanges, which are prominent in the right frontal horn (ar-row) and periventricular regions. (Reprinted from Atouguiaand Kennedy.)


in 1949.3,4,23 Until recently, the standard regimen wastwo to four courses of intravenous injections (three in-jections per course, with 1-week intervals betweenthem), but a more recent 10-day continuous melarso-prol regimen has been introduced for treatment ofgambiense disease34 and is now favored in many cen-ters. Although melarsoprol is usually effective, it is fol-lowed by a severe posttreatment reactive encephalopa-thy (PTRE) in about 10% of patients, half of whomdie of it. This overall mortality rate of 5% of all pa-tients receiving melarsoprol is one of the greatest prob-lems in this field and highlights the extreme impor-tance of correct staging of the disease. Althoughtreating patients who do not, in fact, have CNS diseasewith melarsoprol will lead to an unnecessary 5% riskfor death, not treating a patient who does actually haveCNS disease will inevitably lead to death because ofthe 100% fatality rate of HAT if untreated.1,3,13,35 Pa-tients who acquire PTRE (usually after the first treat-ment course or near the end of the continuous treat-ment course) may experience development of deepcoma with seizures, convulsive status epilepticus, orrapid progressive coma in the absence of seizures orcerebral edema.4,13 Patients with the PTRE are treatedwith intravenous corticosteroids, anticonvulsants, andintensive general medical support, and those who sur-vive need to continue the melarsoprol course. The pos-sible role of corticosteroid pretreatment to preventand/or ameliorate the PTRE is currently unclear,3,13

although the author would probably choose this optionfor himself. Recently, it has been shown that a combi-nation regimen of melarsoprol and nifurtimox for gam-biense disease is more effective than standard melarso-prol monotherapy regimens.36 Melarsoprol may alsoproduce cardiac arrhythmias, skin lesions, agranulocy-tosis, peripheral neuropathy, and as recently described,a steroid-responsive multifocal inflammatory ill-ness.3,30,31

There are alternative treatment options for late-stage gambiense disease, although not for rhodesiensedisease for which melarsoprol remains the only effec-tive drug. Treatment failure with melarsoprol, how-ever, is well recognized.37 The drug eflornithine is be-ing increasingly used for CNS gambiense disease,

either alone or in combination with other drugs. Ef-lornithine, which is expensive, became an orphandrug even after it had been shown to be effectiveagainst gambiense disease in 1981.1,3 But after an in-novative partnership among Medecin Sans Frontieres,WHO, and the pharmaceutical industry, the drug wasagain made available for HAT use in sub-Saharan Af-rica1 and is being used increasingly as first-line ther-apy and alternative therapy. It does, however, have tobe given by daily intravenous injection over at least14 days, and its adverse effects include bone marrowtoxicity, seizures, and gastrointestinal symptoms.3,4,23

Recent clinical trials have tested various combinationsof drugs for late-stage gambiense disease, and theemerging results are increasingly favoring the use of anifurtimox-eflornithine regimen as the most promis-ing first-line therapy.38,39 The only new drug on theimmediate horizon for HAT is the diamidine deriva-tive DB 289, a promising development funded by theBill and Melinda Gates Foundation.3,13 This drug isgiven orally, can treat only early-stage disease, and hasbeen under recent evaluation in a Phase 3 clinicaltrial. However, clinical trials with this drug (manu-factured by Immtech Pharmaceuticals, New York,NY) have recently been put on hold because of pos-sible liver toxicity, and further development of thisdrug has unfortunately been discontinued. A sum-mary of current drug therapy for HAT is shown inTable 2.

After successful treatment, all patients need to befollowed up with blood and CSF analyses at 6-monthintervals for 2 years, after which the patient is con-sidered cured if these tests are normal.4,13 Treatmentfailures do occur, however, and it has recently beenshown that the presence of intrathecal IgM synthesisand increased CSF IgM IL-10 concentrations are sig-nificantly associated with the failure of treatment inearly-stage gambiense disease.40 Analysis of sleep struc-ture with polysomnography may also prove useful indetecting patients with relapses. Follow-up of patientsin the field is problematic, and as has recently beenpointed out, counting all patients who never turn upfor follow-up CSF examination can hardly be a goodmeasure of cure rates, especially when half the pa-

Table 2. Summary of Drug Therapy in Sleeping Sickness

Disease First-Line Therapy Alternative Therapy

Early-stage Trypanosoma brucei rhodesiense Suramin None

Early-stage Trypanosoma brucei gambiense Pentamidine Suramin

Late-stage T.b. rhodesiense Melarsoprol None

Late-stage T.b. gambiense Melarsoprol Eflornithine � nifurtimoxa

aIt is likely that eflornithine with or without nifurtimox may soon be the preferred first-line therapy for treating late-stage gambiensedisease.


tients never show up.8 Patients who have been suc-cessfully treated for late-stage disease may still experi-ence development of long-term neurological sequelaesuch as weakness, ataxia, cognitive impairment, epi-lepsy, and psychiatric disorders.4

NeuropathogenesisPathological data in HAT have been obtained from arelatively small number of autopsy studies. The keyfindings in CNS HAT are an extensive meningoen-cephalitis, widespread infiltration of white matter withinflammatory cells such as macrophages, lymphocytes

Fig 4. Brain pathology in central nervous system (CNS) hu-man African trypanosomiasis. (A) Late-stage disease in a pa-tient dying 3 to 5 months after first injection of melarsoprol.Many large astrocytes are located in white matter. Stained forglial fibrillary acidic protein by immunoperoxidase. Originalmagnification �400. (B) Morular (Mott) cells (arrows) ob-served in the brain of a patient with CNS trypanosomiasiswho had not received melarsoprol. Morular cells are plasmacells filled with immunoglobulin. Hematoxylin and eosin (HE)stain. Original magnification �400. (C) Posttreatment reac-tive encephalopathy (PTRE) in a patient 9 days after receivingmelarsoprol. Ischaemic cell changes (arrows) are seen in neu-rons in the hippocampus. HE stain. Original magnification�250. (D) PTRE with acute hemorrhagic leukoencephalopa-thy in a patient 9 days after receiving melarsoprol. There isfibrinoid necrosis in an arteriole (arrow) and focal hemor-rhage in the pons. Martius scarlet blue stain. Original magni-fication �250. (Reprinted from Adams and colleagues,41 bypermission.)

Fig 5. A schematic of the sleeping sickness mouse model of cen-tral nervous system (CNS) disease. PTRE � posttreatment reac-tive encephalopathy. (Reprinted from Kennedy,1 by permission.)


and plasma cells, extensive perivascular cuffing, astro-cyte and macrophage activation, and a small amount ofdemyelination.41–43 A pathognomonic finding in thewhite matter is the presence of morular or Mott cells,which are plasma cells containing IgM eosinophilic in-clusions41,42 (Fig 4).

The neuropathogenesis of HAT has been describedin considerable detail elsewhere,3,42– 45 and only fewkey aspects are mentioned here. Not surprisingly,only limited data have been obtained from patients inthe African field, primarily from analysis of immunefactors in blood and CSF where cause and effect canbe difficult to establish. Recently, it was shown intwo sites in Uganda that a greater plasma interferon-�level correlated with a greater severity of neurologicalimpairment in rhodesiense HAT patients, indicating apathological role of this cytokine.46 Patients with late-stage rhodesiense disease have been shown to have sig-nificantly increased levels of the counterinflammatorycytokine IL-10 in the plasma and CSF, declining tonormal levels after treatment.47 It is possible that it isthe balance of cytokines that is a key determinant ofthe outcome of CNS disease.

Much of our mechanistic knowledge in this area

has been derived from animal models, in particular, ahighly reproducible mouse model of HAT that closelymimics human late-stage disease and also allows theidentification of new therapeutic targets44,45,48 (Fig5). Subcurative therapy of infected mice with thedrug berenil leads to an exacerbation of CNS diseasewith neuropathological features that show strong sim-ilarities with the human PTRE.49 This approach hasallowed the identification of key players in the gener-ation of the neuroinflammatory response such as theroles of astrocyte activation, the neuropeptide sub-stance P, and the balance of proinflammatory andcounterinflammatory cytokines, including tumor ne-crosis factor-� and IL-10, respectively50 –53 (summa-rized in Fig 6). Knock-out mice lacking specific in-flammatory factors have also been used in severallaboratories.54,55 It appears likely that there is a com-plex array of CNS-cytokine interactions that ulti-mately determine the CNS damage in HAT.3,45

Moreover, several drugs and novel drug combinationshave been tested in the mouse model.56 –58 Rodentmodels have also provided insights into the mecha-nism of BBB traversal by trypanosomes, an area ofcrucial importance. For example, trypanosomes show

Fig 6. Schematic representation of possible immunopathological pathways leading to brain dysfunction in human African trypanoso-miasis, based on data and concepts from both human and animal data. Cytokines shown in red probably have important roles inneuropathogenesis. The schematic emphasizes the central importance of early astrocyte activation, cytokine responses, and macrophageactivation. There are likely to be multiple factors acting together to produce brain damage and also multiple potential sources ofdifferent cytokines. IFN � interferon; MHC � major histocompatibility complex; NK � natural killer; NO � nitric oxide;Tltf � trypanosome-derived lymphocyte triggering factor; TNF � tumor necrosis factor; VSG � variable surface glycoprotein. (Re-printed from Kennedy,3 by permission.).


early invasion in brain areas that lack a BBB, such asthe pineal gland and median eminence.59 A seminalstudy showed that in interferon-� knock-out mice,parasites accumulated in the perivascular compart-ment, confined between the endothelial and paren-chymal basement membranes, thus confirming thecritical role of interferon-� in parasite entry into theCNS.60 The same group has shown that, intrypanosome-infected rats, there is immunologicaland biochemical disruption of the suprachiasmaticnuclei of the hypothalamus with resultant disorgani-zation of normal circadian rhythms61,62 (Fig 7).These findings strongly suggest a credible neuro-pathological basis for the altered sleep/wake cyclesseen in HAT.

ConclusionsHAT continues to be a major health problemthroughout sub-Saharan Africa and is likely to be sofor the foreseeable future. The two key related issuesin disease management are disease staging and drug

therapy. Regrettably, progress in both of these areascontinues to be modest. In the absence of a noninva-sive method of staging, it is inevitable that distinctionbetween early- and late-stage HAT will continue torely entirely on CSF examination. But not only isthere disagreement as to what biologically constitutesCNS disease, there is also an absence of a consensusas to what are the correct grounds for making thera-peutic choices. Current drug therapy, especially forlate-stage HAT, is unacceptably toxic, and remark-ably, there are no new drugs at all on the horizon fortreating CNS disease. If there were available a non-toxic oral drug for late-stage HAT, then most of thestaging problems would be immediately obviated.These difficult problems are a direct result of manydecades of underinvestment in this neglected disease,a trend that is beginning to be addressed. However,they must be seen in the context of the wider effortsto control the disease at the level of the tsetse fly vec-tor1,63 because disruption of the man/tsetse fly con-tact through various technologies probably holds the

Fig 7. (A, B) Images of the hypothalamic suprachiasmatic nucleus (which plays a role of circadian pacemaker in the mammalianbrain) in sections processed for glial fibrillary acidic protein (GFAP) immunoreactivity. (A) Control noninfected rat. (B) Rat in-fected with Trypanosoma brucei brucei; note the astrocytic activation (shown by hypertrophy and increased immunoreactivity ofastrocytes) in the infected animal. (C–E) Images of the cingulate cortex of a rat infected with Trypanosoma brucei brucei (pro-cessed for double immunohistochemistry to show GFAP-immunolabeled astrocytes (red: C) and the proinflammatory cytokine tumornecrosis factor (TNF)-� (green: D); (E) merging of two images (yellow) and, therefore, the colocalization of the two markers. (C)Note the activation of astrocytes, (D) the induction of TNF-� expression in cells of the brain parenchyma, and (E) that TNF-� isexpressed in astrocytes. Scale bars � 100�m (A, B); 25�m (C–E). oc � optic chiasm; 3v � third ventricle. (Courtesy Maria Pal-omba, Gigliola Grassi-Zucconi and Marina Bentivoglio, University of Verona, Verona, Italy)


key to the potential ultimate control of sleeping sick-ness.

The authors research is supported by grants from the WellcomeTrust and the Medical Research Council.

I thank Drs J. Ndung’u and V. Lejon for their helpful comments onthe manuscript, and Dr J. Atouguia for advice. I am grateful to ProfM. Bentivoglio for generously providing the unpublished Figure 7.This article is dedicated to the memory of my recently departedfather, Philip Kennedy, a man of great compassion, outstanding in-telligence, and unquenchable humor.

References1. Kennedy PGE. The fatal sleep. Edinburgh, United Kingdom:

Luath Press Limited, 2007.2. Epidemiology and control of African trypanosomiasis. Report

of a WHO expert committee. Geneva, Switzerland: WorldHealth Organization Technical Report Series, 1986;739.

3. Kennedy PGE. Human African trypanosomiasis of the CNS:current issues and challenges. J Clin Invest 2004;113:496–504.

4. Atouguia JLM, Kennedy PGE. Neurological aspects of humanAfrican trypanosomiasis. In: Davis LE, Kennedy PGE, eds. In-fectious diseases of the nervous system. Oxford: Butterworth-Heinemann, 2000:321–372.

5. Vickerman K. Landmarks in trypanosome research. In: Hide G,Mottram JC, Coombs GH, Holmes PH, eds. Trypanosomiasisand leishmaniasis. Oxford: CAB International, 1997:1–37.

6. Williams BI. African trypanosomiasis. In Cox FEAG, ed. TheWellcome Trust illustrated history of tropical diseases. London:The Wellcome Trust, 1996:178–191.

7. Bentivoglio M, Grassi-Zucconi G, Kristensson K. From try-panosomes to the nervous system, from molecules to behaviour:a survey, on the occasion of the 90th anniversary of Castellani’sdiscovery of the parasites in sleeping sickness. Ital J Neurol Sci1994;15:77–89.

8. Pepin J. Combination therapy for sleeping sickness: a wake-upcall. J Infect Dis 2007;195:311–313.

9. Control and surveillance of African trypanosomiasis. Report ofa WHO expert committee. Geneva, Switzerland: World HealthOrganization Technical Report Series, 1998;881.

10. WHO African trypanosomiasis (sleeping sickness) Fact sheetNo 259. Revised August 2006. www.who.int/mediacentre/factsheets/fs259/en

11. Kuzoe FA. Current situation of African trypanosomiasis. ActaTrop 1993;54:153–162.

12. Lejon V, Boelaert M, Jannin J, et al. The challenge of Trypano-soma brucei gambiense sleeping sickness diagnosis outside Af-rica. Lancet Infect Dis 2003;3:804–808.

13. Kennedy PGE. Sleeping sickness—human African trypanosomi-asis. Pract Neurol 2005;5:260–267.

14. Berriman M, Ghedin E, Hertz-Fowler C, et al. The genome ofthe African trypanosome Trypanosoma brucei. Science 2005;309:416–422.

15. Donelson JE. Antigenic variation and the African trypanosomegenome. Acta Trop 2002;85:391–404.

16. Apted FIC. Clinical manifestations and diagnosis of sleepingsickness. In: Mulligan HW ed. The African trypanosomiases.London: George Allen & Unwin, 1970:551–583.

17. Duggan AJ, Hutchington MP. Sleeping sickness in Europeans:a review of 109 cases. J Trop Med Hyg 1966;69:124–131.

18. Buguet A, Bisser S, Josenando T, et al. Sleep structure: a newdiagnostic tool for stage determination in sleeping sickness.Acta Trop 2005;93:107–117.

19. Blum J, Schmid C, Burri C. Clinical aspects of 2541 patientswith second stage human African trypanosomiasis. Acta Trop2006;97:55–64.

20. Truc P, Lejon V, Magnus E, et al. Evaluation of the micro-CATT, CATT/Trypanosoma brucei gambiense, and LATEX/T.b. gambiense methods for serodiagnosis and surveillance ofhuman African trypanosomiasis in West and Central Africa.Bull World Health Organ 2002;80:882–886.

21. Jamonneau V, Solano P, Garcia A, et al. Stage determinationand therapeutic decision in human African trypanosomiasis:value of polymerase chain reaction and immunoglobulin Mquantification on the cerebrospinal fluid of sleeping sickness pa-tients in Cote d’Ivoire. Trop Med Int Health 2003;8:589–594.

22. Njiru ZK, Mikosza ASJ, Armstrong T, et al. Loop-mediatedisothermal amplification (LAMP) method for rapid detection ofTrypanosoma brucei rhodesiense. PLos Negl Trop Dis 2008;2:147–154.

23. Fairlamb AH. Chemotherapy of human African trypanosomiasis:current and future prospects. Trends Parasitol 2003;19:488–494.

24. Lejon V, Buscher P. Cerebrospinal fluid in human Africantrypanosomiasis: a key to diagnosis, therapeutic decision andpost-treatment follow-up. Trop Med Intl Health 2005;10:395–403.

25. Lejon v, Reiber H, Legros D, et al. Intrathecal immune re-sponse pattern for improved diagnosis of central nervous systeminvolvement in trypanosomiasis. J Infect Dis 2003;187:1475–1483.

26. Chappuis F, Loutan L, Simarro P, et al. Options for field di-agnosis of Human African Trypanosomiasis. Clin MicrobiolRev 2005;18:133–146.

27. Kennedy PGE. Diagnosing central nervous system try-panosomiasis: two stage or not to stage? Trans R Soc Trop MedHyg 2008;102:306–307.

28. Sabbah P, Brosset C, Imbert P, et al. Human Africantrypanosomiasis: MRI. Neuroradiology 1997;39:708–710.

29. Gill DS, Chatha DS, del Carpio-O’Donovan R. MR Imagingfindings in African trypanosomiasis. Am J Neuroradiol 2003;24:1383–1385.

30. Braakman HMH, van de Molengraft FJJM, Hubert WWA,Boerman DH. Lethal African trypanosomiasis in a traveler:MRI and neuropathology. Neurology 2006;66:1094–1096.

31. Kumar N, Orenstein R, Uslan DZ, et al. Melarsoprol-associated multifocal inflammatory CNS illness in Africantrypanosomiasis. Neurology 2006;66:1120–1121.

32. Kennedy PGE. Human African trypanosomiasis: in and out ofAfrica. Neurology 2006;66:962–963.

33. Fairlamb AH. Future prospects for the chemotherapy of humantrypanosomiasis. 1. Novel approaches to the chemotherapy oftrypanosomiasis. Trans R Soc Trop Med Hyg 1990;84:613–617.

34. Burri C, Nkunku S, Merolle A, et al. Efficacy of new, conciseschedule for melarsoprol in treatment of sleeping sicknesscaused by Trypanosoma brucei gambiense: a randomised trial.Lancet 2000;355:1419–1425.

35. Pepin J, Milord F. The treatment of human African trypano-somiasis. Adv Parasitol 1994;33:1–47.

36. Bisser S, N’Siesi FX, Lejon V, et al. Equivalence trial of Me-larsoprol and Nifurtimox monotherapy and combination ther-apy for the treatment of second-stage Trypanosoma brucei gam-biense sleeping sickness. J Infect Dis 2007;195:322–329.

37. Legros D, Evans S, Maiso F, et al. Risk factors for treatmentfailure after melarsoprol for Trypanosoma brucei gambiensetrypanosomiasis in Uganda. Trans R Soc Trop Med Hyg 1994;93:439–442.


38. Priotto G, Kasparian S, Ngouama D, et al. Nifurtimox-Eflornithine combination therapy for second-stage Trypano-soma brucei gambiense sleeping sickness: a randomized clinicaltrial in Congo. Clin Infect Dis 2007;45:1435–1442.

39. Checchi F, Piola P, Ayikoru H, et al. Nifurtimox plus Eflorni-thine for late-stage sleeping sickess in Uganda: a case series.PLos Negl Trop Dis 2007;1:64–69.

40. Lejon V, Robays J, N’Siesi FX, et al. Treatment failure relatedto intrathecal immunoglobulin M (IgM) synthesis, cerebro-spinal fluid IgM, and interleukin-10 in patients withhemolymphatic-stage sleeping sickness. Clin Vaccine Immunol2007;14:732–737.

41. Adams JH, Haller L, Boa FY, et al. Human African trypano-somiasis (T./b.gambiense): a study of 16 fatal cases of sleepingsickness with some observations on acute reactive arsenical en-cephalopathy. Neuropathol Appl Neurobiol 1986;12:81–94.

42. Pentreath VW, Kennedy PGE. Pathogenesis of human Africantrypanosomiasis. In: Maudlin I, Holmes PH, Miles MA, eds.The Trypanosomiases. Oxford: CAB International, 2004:283–301.

43. Cook GC. Protozoan and helminthic infections. In: LambertHP, ed. Infections of the central nervous system. Philadelphia:BC Dekker, 1991:264–282.

44. Kennedy PGE. The pathogenesis and modulation of the post-treatment reactive encephalopathy in a mouse model of humanAfrican trypanosomiasis. J Neuroimmunol 1999;100:36–41.

45. Kennedy PGE. Diagnostic and neuropathogenesis issues in hu-man African trypanosomiasis. Intl J Parasitol 2006;36:505–512.

46. MacLean L, Odiit M, MacLeod A, et al. Spatially and geneti-cally distinct African trypanosome virulence variants defined byhost interferon-y response. J Infect Dis 2007;196:1620–1628.

47. MacLean L, Odiit M, Sternberg JM. Nitric oxide and cytokinesynthesis in human African trypanosomiasis. J Infect Dis 2001;184:1086–1090.

48. Kennedy PGE. Animal models of human African trypanosomi-asis—very useful or too far removed? Roy Soc Trop Med Hyg2007;101:1061–1062.

49. Hunter CA, Jennings FW, Adams JH, et al. Sub-curative che-motherapy and fatal post-treatment reactive encephalopathies inAfrican trypanosomiasis. Lancet 1992;339:956–958.

50. Hunter CA, Jennings FW, Kennedy PGE, Murray M. Astro-cyte activation correlates with cytokine production in centralnervous system of Trypanosoma brucei brucei-infected mice.Lab Invest 1992;67:635–642.

51. Kennedy PGE, Rodgers J, Jennings FW, et al. A Substance Pantagonist, RP-67,580 ameliorates a mouse meningoencepha-litic response to Trypanosoma brucei brucei. Proc Natl AcadSci USA 1997;94:4167–4170.

52. Hunter CA, Gow JW, Kennedy PGE, et al. Immunopathologyof experimental African sleeping sickness: detection of cytokinemRNA in the brains of Trypanosoma brucei brucei-infectedmice. Infect Immun 1991;5:4636–4640.

53. Sternberg JM, Rodgers J, Bradley B, et al. MeningoencephaliticAfrican trypanosomiasis: brain IL-10 and IL-6 are associatedwith protection from neuroinflammatory pathology. J Neuro-immunol 2005;167:81–89.

54. Kennedy PGE, Rodgers J, Bradley B, et al. Clinical and neu-roinflammatory responses to meningoencephalitis in substanceP receptor knockout mice. Brain 2003;126:1683–1690.

55. Barkhulzen M, Magez S, Atkinson RA, Brombacher F.Interleukin-12p70-dependent interferon-gamma production iscrucial for resistance in African trypanosomiasis. Infect Dis2007;196:1253–1260.

56. Hunter CA, Jennings FW, Kennedy PGE, Murray M. The useof azathioprine to ameliorate post-treatment encephalopathy as-sociated with African trypanosomiasis. Neuropathol Appl Neu-robiol 1992;18:619–662.

57. Jennings FW, Gichuki CW, Kennedy PGE, et al. The role ofthe polyamine inhibitor eflornithine in the neuropathogenesisof experimental murine African trypanosomiasis. NeuropatholAppl Neurobiol 1997;23:225–234.

58. Jennings FW, Rodgers J, Bradley B, et al. Human Africantrypanosomiasis: potential therapeutic benefits of an alternativesuramin and melarsoprol regimen. Parasitol Intl 2002;51:381–388.

59. Schultzberg M, Ambatsis M, Samuelsson EB, et al. Spread oftrypanosoma brucei to the nervous system: early attack on cir-cumventricular organs and sensory ganglia. J Neurosci Res1988;21:56–61.

60. Masocha W, Robertson B, Rottenberg ME, et al. Cerebral ves-sel laminins and IFN-� define trypanosoma brucei brucei pen-etration of the blood-brain barrier. J Clin Invest 2004;114:689–694.

61. Bentivoglio M, Grassi-Zucconi G, Olsson T, Kristensson K.Trypanosoma brucei and the nervous system. Trends Neurosci1994;17:325–329.

62. Bentivoglio M, Kristensson K. Neural-immune interactions indisorders of sleep-wakefulness organization. Trends Neurosci2007;30:645–652.

63. Barrett MP, Burchmore RJS, Stich A, et al. The trypanosomi-ases. Lancet 2003;362:1469–1480.


Documents

NEUROLOGICAL PROGRESS The Continuing Problem of Human