Sa

BAMOa

b

c

d

e

f

a

ARRAA

KLLSERS

1

dcl

f

0d

Transactions of the Royal Society of Tropical Medicine and Hygiene 105 (2011) 333–340

Contents lists available at ScienceDirect

Transactions of the Royal Society ofTropical Medicine and Hygiene

journa l homepage: ht tp : / /www.e lsev ier .com/ locate / t rs tmh

eroprevalence of Leishmania infantum in a rural area of Senegal:nalysis of risk factors involved in transmission to humans

abacar Fayea,∗, Bruno Buchetonb, Anne Laure Banulsc, Massila Wagué Senghord,bdoul Aziz Niangd, Souleymane Diedhioua, Oumar Konatéa, Michel Mainack Dionee,allorie Hidec, Sandra Mellul f, Romy Knecht f, Pascal Delaunayf, Pierre Marty f,umar Gayea

Service de Parasitologie–Mycologie, Faculté de Médecine et Pharmacie, Université Cheikh Anta Diop, BP 5005 Dakar Fann, Dakar, SenegalInteractions hôtes-vecteurs-parasites dans les trypanosomoses, UMR 177 IRD-CIRAD, 01 BP 171, Bobo-Dioulasso, Burkina FasoGénétique et Evolution des Maladies Infectieuses, IRD/CNRS/UM1 (UMR 2724), Montpellier F-34394, FranceInstitut Fondamental d’Afrique Noire (IFAN), Université Cheikh Anta Diop, Dakar, SenegalInternational Trypanotolerance Centre (ITC), P.M.B. 14, Banjul, The GambiaLaboratoire de Parasitologie–Mycologie, Hôpital de l’Archet 2, CHU de Nice, France

r t i c l e i n f o

rticle history:eceived 9 February 2010eceived in revised form 2 February 2011ccepted 2 February 2011vailable online 4 May 2011

eywords:eishmaniasiseishmania infantumerologypidemiology

a b s t r a c t

Whereas Leishmania infantum, the agent of visceral leishmaniasis (VL), is well known inNorth Africa, very limited data exist on its spread in West Africa, where mainly cutaneousleishmaniasis has been widely reported. Nevertheless, dogs infected with L. infantum wererecently found in the Mont Rolland District in Senegal. To provide a better understandingof L. infantum epidemiology in this area, clinical and serological surveys were carried out todetermine the seroprevalence of L. infantum-specific antibodies in the human population.In parallel, an analysis of environmental and individual factors associated with Leishma-nia antigen seropositivity was conducted to identify potential risk factors for exposure.Although no cases of VL were detected within this study, a large part of the population(73/315; 23%) was exposed to infection, with a strong age effect (being >40 years old

isk factorsenegal

increased the risk of being seropositive). Moreover, the presence of Nebedaye trees (Moringaoleifera) and infected dogs in the household were factors increasing the risk of exposure inhousehold members. These results may provide important information to identify the stillunknown sandfly species involved in transmission.

© 2011 Royal Society of Tropical Medicine and Hygiene. Published by Elsevier Ltd. All rights reserved.

. Introduction

Leishmaniasis is composed of a diverse group of

iseases, ranging from cutaneous leishmaniasis (CL) to vis-eral leishmaniasis (VL), which is almost always fatal ifeft untreated. These diseases are caused by protozoa from

∗ Corresponding author. Tel.: +221 338 251 998;ax: +221 338 253 668..

E-mail address: bfaye67@yahoo.fr (B. Faye).

035-9203/$ – see front matter © 2011 Royal Society of Tropical Medicine and Hoi:10.1016/j.trstmh.2011.02.009

the genus Leishmania that are transmitted to their hostby infected sandfly bites. It is estimated that >350 mil-lion people are at risk for Leishmania infection. The annualincidence is estimated to be 2 million new cases per year,of which 500 000 are VL caused by parasites from theLeishmania donovani complex (L. donovani and L. infan-tum/chagasi). Whereas L. donovani is known to be endemic

in East Africa and L. infantum in North Africa, very little dataexist on the presence of VL in West Africa where CL has beenwidely reported, mainly due to L. major. In East Africa, themain vector of parasites from the L. donovani complex is

ygiene. Published by Elsevier Ltd. All rights reserved.

334 B. Faye et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 105 (2011) 333–340

es in th

Figure 1. Location of study villag

Phlebotomus orientalis and the parasite can be found in thewild fauna and occasionally in domestic dogs,1 whereas inNorth Africa dogs are usually described as the main reser-voir of L. infantum, with P. perniciosus being the principalinsect vector. The finding of L. infantum-infected dogs in theMont Rolland District in Senegal by Faye et al.2 raises sev-eral questions. First, whereas entomological studies havecharacterised the sandfly fauna in Senegal, no known Phle-botomus sp., the vector of L. infantum, was reported3 (M.W.Senghor et al., unpublished data). Second, to what extentare humans also infected with L. infantum and what is theirepidemiological role in the life cycle of L. infantum in thispart of Africa?

Importantly, most human infections by L. infantumare subclinical or asymptomatic, usually associated withstrong cell-mediated immunity.4,5 In the Mediterraneanbasin, for example, very few cases are diagnosed eachyear in immunocompetent subjects, even though canineleishmaniasis is endemic and >20% of the populationdisplays antibodies directed against Leishmania 14 kDa-and 18 kDa-specific antigens.6–10 Together with the factthat the symptoms of VL (fever, splenomegaly, anaemia,adenopathies) can be confused with several other parasitic(e.g. malaria, schistosomiasis) or systemic (leukaemia) dis-eases, this may explain why the presence of human VL inSenegal has remained unnoticed.

To provide a better understanding of L. infantumepidemiology in this region of Senegal, clinical and sero-logical surveys were carried out in villages of the Mont

Rolland District to determine the seroprevalence of L.infantum-specific antibodies in the human population. Fur-thermore, an analysis of environmental and individualfactors associated with Leishmania antigen seropositivity

e Mont Rolland District, Senegal.

was conducted to identify potential risk factors forexposure.

2. Materials and methods

2.1. Study area and population

The study population was composed of the inhabitantsof five villages (Ndiaye Bopp, Ndiaye Bopp 2, Fouloum,Nguith Fall and Mont Rolland) from the Mont Rolland Dis-trict in Senegal (Figure 1) in which previous parasitologicaland serological surveys had shown dogs infected with L.infantum.2 The study was done during 2007–2008 and sam-pling was carried out in two stages. In the first stage, aserological survey was performed in the village of NdiayeBopp where the prevalence of canine leishmaniasis was thehighest.2 Prior to the survey, a public awareness campaignwas carried out by the local health authorities, and inter-views on the local radio were used to explain the aim ofthe study. All individuals present on the given day wereincluded in this first cohort. Fingerprick blood sampleswere collected on Whatman 3MM paper from 315 subjects(203 females and 112 males) living in 101 different house-holds scattered throughout different areas of the village.For all study subjects, individual data (gender, age) andwhether a dog was present in the household was recorded.When a dog was present in the household, its Leishma-nia serological and parasitological status was assessed.For each dog, a clinical examination was conducted by a

veterinary surgeon to check for the most frequent symp-toms observed in canine leishmaniasis. Popliteal lymphnodes were sampled when swollen, and blood (5 ml) wastaken either from the saphenous or jugular vein. After

of Tropi

cdtSNpMeat

ts(pstlsItisrFh

2

24Wetaltmi1erwlVwdf

2i

iplFfTt

B. Faye et al. / Transactions of the Royal Society

entrifugation of the blood, serum was used for serologicaliagnosis by means of the rK39 rapid diagnostic dipstickest (DiaMed-IT Leish; DiaMed AG, Cressiers sur Morat,witzerland). Lymph node fluid was inoculated into twoNN tubes with 0.75 ml of penicillin (100 000 IU/ml ofhysiological serum). The cultures were placed at 26 ◦C.icroscopic examination of the cultures was performed

very 3 days for 6 weeks. A culture was considered neg-tive when no Leishmania promastigotes were seen afterhis time.

In the second stage aimed at studying environmen-al risk factors of exposure, fingerprick blood from 131ubjects living in 12 households where it was knownfrom previous surveys) that a Leishmania-infected dog wasresent was sampled. This sampling approach was cho-en because it ensured that the parasite was present inhe household and thus seropositivity in humans was moreikely to be associated with environmental factors affectingandfly species distribution within or around the house.n this case, all members present in the household at theime of the survey were included. Three households (n = 36ndividuals) were from Ndiaye Bopp and had already beenampled during the first survey. Others were from sur-ounding villages: Ndiaye Bopp 2 (n = 23; one household);ouloum (n = 20; three households); Nguith Fall (n = 21; oneousehold); and Mont Rolland (n = 31; four households).

.2. Diagnosis and serological assays

One drop of blood (corresponding to approximately0 �l of equivalent serum) was eluted overnight at 5 ◦C in00 �l of PBS 1×. Eluted serum was then treated using theestern blot technique as described previously by Marty

t al.9 Western blots were carried out on 410 sampleso assess the serological reactivity to specific L. infantumntigens. Western blot analysis in cases of acute clinicaleishmaniasis due to L. infantum should present the simul-aneous presence of antibodies against five antigens with

olecular weights of 14, 18, 21, 23 and 31 kDa,8–10 whereasn asymptomatic immunocompetent individuals only the4 kDa and/or 18 kDa bands are usually detected by West-rn blotting.6,7 All individuals with a positive immunoblotesult (presence of at least one band, 14 kDa or 18 kDa)ere subsequently requested to come to the Mont Rol-

and District Health Center for a clinical examination ofL symptoms (fever, adenopathies, hepatosplenomegaly,eight loss, etc.). For all these individuals, an rK39 rapidiagnostic dipstick test (DiaMed-IT Leish) was also per-ormed on this occasion.

.3. Evaluation of risk factors for exposure to Leishmanianfection

For subjects from the first survey in Ndiaye Bopp,nformation was collected regarding age, gender and theresence in the household of a dog as well as the dog’s sero-

ogical and parasitological status for Leishmania infection.

or all households with an infected dog, environmentalactors that may influence sandfly density were recorded.hese factors included the location of the household withinhe village (in the centre or at the periphery), main

cal Medicine and Hygiene 105 (2011) 333–340 335

household activity (agriculture, animal breeding, trade),home construction characteristics (presence of cement,thatched grass roof), number of different domestic ani-mals (cows, donkeys, horses, sheep, goats, dogs, chickens),number of hen houses, presence of granaries and veg-etation type within and 20 m around the house [neem(Azadirachta indica), Nebedaye tree (Moringa oleifera),Acacia spp., mango tree (Mangifera indica), cherry tree(Phyllanthus acidus), ironwood (Prosopis africana), balsamspurge (Euphorbia balsamifera) and baobab species]. Mostof the environmental factors were coded as binary variables(present/absent).

2.4. Statistical analysis

Univariate and multivariate analyses were performedwith JMP 5.1.2 software (SAS Institute Inc., Cary, NC, USA)by means of logistic regression to determine which factorsaffect serological positivity determined by immunoblots.On the population of Ndiaye Bopp (n = 315), the fac-tors tested were (i) gender and (ii) age group, includingfive categories [<10 years (n = 88); 10–20 years (n = 65);21–40 years (n = 39); 41–60 years (n = 60); and >60 years(n = 63)].

On the population sample living in households with aninfected dog (n = 131), the effect of environmental factorsrecorded was tested. For stepwise multivariate analysis,variables were included in the model when the associatedP-value was ≤0.2 and were removed from the model whenthe associated P-value became ≥0.1 as the result of enter-ing a new variable. Since individuals in family clusters arenot independent and may have influenced the result of theregression analysis, the relationship between the preva-lence of seropositivity found in each household and thefactors identified by the regression analysis was also anal-ysed by t-test.

3. Results

3.1. Human exposure to Leishmania infection

Previous surveys in the Mont Rolland area have shownthat canine leishmaniasis due to L. infantum was endemic inthe area. Results of the serological survey carried out on 315study subjects from Ndiaye Bopp village according to ageare shown in Figure 2. Overall, 23% (n = 73 study subjects)of the population displayed specific antibodies directedagainst the 14 kDa and 18 kDa Leishmania-specific antigensas revealed by immunoblots. Univariate logistic regressionanalysis showed that the seroprevalence was similar (vary-ing between 12.3% and 15.9%) in subjects 40 years old andyounger and was not significantly different between agegroups (P = 0.81). Seroprevalence increased significantlyfor older age groups, being 28.3% for the 41–60-year-oldsubjects and reaching 44.4% in subjects aged >60 years(P = 0.019 and P < 0.0001, respectively, compared with theiryounger counterparts). No significant differences in sero-

prevalence were noted between females and males (0.24and 0.21, respectively).

Despite the high prevalence of seropositive subjects,none of them presented Leishmania-specific clinical signs

336 B. Faye et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 105 (2011) 333–340

ing to ablots. Se

Figure 2. Seroprevalence of Leishmania-specific antigen positivity accordBopp (n = 315; 112 males and 203 females) was evaluated using Westernwas considered positive (n = 73; 23%).

that could not be attributed to other infection or systemicdiseases. Furthermore, all the subjects were negative on therK39 rapid dipstick test, a test that is almost always pos-itive during active disease.11 Therefore, for these subjectsthe reactivity observed on immunoblots was instead dueto either past or asymptomatic infection.

3.2. Risk factors and prevalence of Leishmaniaseropositivity

Results of the univariate and multivariate analyses car-ried out on the population of Ndiaye Bopp village are shownin Table 1. As noted in Section 3.1, being >40 years old wasassociated with an increased seroprevalence in the univari-ate analysis (P < 0.0001). Presence of a dog in the householdwas also significantly associated with higher seropositiv-ity (P = 0.014) and this effect was further increased whenconsidering only the presence of dogs with canine leish-maniasis (P = 0.0001). In the multivariate analysis, only age>40 years and the presence of a dog with canine leishma-niasis remained in the model after the stepwise procedure.Being >40 years old was associated with a 4.01 odds ratio

(OR) (P < 0.0001) of being seropositive, and the presenceof an infected dog was associated with an increased riskof exposure to Leishmania infection with an OR of 5.56(P < 0.0001).

Table 1Risk factors influencing Leishmania serological positivity in Ndiaye Bopp villagea

Variable Univariate analysis

P-valueb OR

Age (>40 years) <0.0001 3.37Sex (female/male) 0.433 –Dog (present/absent) 0.014 2.09Dog with canine leishmaniasis (present/absent) 0.0001 4.29

OR: odds ratio.a Statistical analysis was performed by univariate analysis and stepwise logistic rof Ndiaye Bopp village. OR and 95% CI are indicated for variables with P-values ≤b P-values of <0.05 are considered significant.

ge and gender. Fingerprick eluted serum from the population of Ndiayerum containing antibodies reactive to the 14 kDa and/or 18 kDa antigens

3.3. Environmental factors and their impact onperidomiciliary transmission of Leishmania

In an attempt to identify potential risk factors influenc-ing exposure of humans to infected sandfly bites, the rela-tionship between environmental factors registered for thestudy subjects living with an infected dog and their serolog-ical status was analysed (Table 2). In this type of setting, wewere sure that the parasite was present within the house-hold and thus that its transmission to humans could beinfluenced by the presence of environmental factors affect-ing the sandfly fauna around the house. No association wasevidenced for the main activity of the household or thehabitat type (data not shown). Other covariates tested areshown in Table 2. In the univariate analysis, being >40 yearsold remained the most significant associated factor(P = 0.0003). Presence of certain animals such as sheep andgoats showed a borderline significant association (P = 0.050and P = 0.059, respectively). Regarding the vegetation type,presence of Nebedaye and cherry trees were significantlyassociated with the risk of having a Leishmania-positiveserological status (P = 0.003 and P = 0.022, respectively).In the stepwise multivariate analysis, only the age

group >40 years (OR = 4.30, P = 0.0008) and the presenceof Nebedaye trees (OR = 3.07, P = 0.008) remained sig-nificantly associated with an increased risk of having

Multivariate analysis

95% CI P-valueb OR 95% CI

1.97–5.87 <0.0001 4.01 2.21–7.45– – – –1.14–3.76 – – –2.06–9.00 <0.0001 5.56 2.53–12.45

egression on data collected from 315 inhabitants living in different areas0.2.

B. Faye et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 105 (2011) 333–340 337

Table 2Analysis of domiciliary risk factors influencing Leishmania serological positivity in households with a dog affected by canine leishmaniasisa

Variable Univariate analysis Multivariate analysis

P-valueb OR 95% CI P-valueb OR 95% CI

Age >40 years 0.0003 4.56 1.99–10.82 0.0008 4.30 1.82–10.48Sex (female/male) 0.131 1.76 0.84–3.78 NE – –Animals

Cows (present/absent) 0.172 0.56 0.23–1.27 NE – –Sheep (present/absent) 0.050 0.39 0.13–1.00 NE – –Goats (present/absent) 0.059 0.35 0.09–1.03 NE – –Horses (present/absent) NS – – – – –Donkeys (present/absent) NS – – – – –No. of hen houses (≤4/>4) NS – – – – –No. of animals (≤5/>5) 0.129 0.56 0.26–1.18 NE – –

Vegetation around the household (20 m)Neem (present/absent) NS – – – – –Nebedaye trees (present/absent) 0.003 3.30 1.49–7.45 0.008 3.07 1.32–7.24Acacia spp. (present/absent) NS – – – – –Mango trees (present/absent) 0.190 2.00 0.69–5.67 NE – –Cherry trees (present/absent) 0.022 2.54 1.14–5.70 NE – –Ironwood (present/absent) NS – – – – –Balsam spurge (present/absent) NS – – – – –No. of trees (≤10/>10) 0.169 1.67 0.80–3.53 NE – –

OR: odds ratio; NS: not significant; NE: not entered in the model during the stepwise logistic regression.a gistic rea s with Pb

atrppTdtcati

Fmed

Statistical analysis was performed by univariate analysis and stepwise lodog with canine leishmaniasis. OR and 95% CI are indicated for variableP-values of <0.05 are considered significant.

positive serological status. As individuals in familial clus-ers are not independent and may have influenced theesults of the logistic regression analysis, the householdrevalence of seropositive individuals according to theresence of Nebedaye trees was also analysed (Figure 3).he three households in which the presence of Nebe-aye was recorded were among the households displayinghe highest prevalence of seropositivity (>40%), signifi-

antly higher than in households where these trees werebsent (P = 0.01). Interestingly, in these three householdshe prevalence of seropositivity was also particularly highn children: 8 (47%) of 17 children under the age of 10 years

igure 3. Prevalence of Leishmania-seropositive subjects in households accordingean prevalence calculated over the 12 households with an infected dog. The li

ach diamond represents the 95% CI for each group, and overlap marks are drawniamond is proportional to the number of households in each of the tested group

gression on data collected from 131 subjects living in 12 households with-values ≤0.2.

were seropositive. It is noteworthy that these three house-holds were all located in Ndiaye Bopp, although they werescattered in different locations within the village.

Among households with an infected dog, the preva-lence of seropositivity in humans was highest in NdiayeBopp, whereas the lowest seroprevalence was observedfor Ndiaye Bopp 2 village (Table 3) despite the presence ofseveral Leishmania-infected dogs.2 This village is inhabited

by Fulani who settled some 10 years ago a few kilome-tres downhill from Ndiaye Bopp, in an area with very littlevegetation except for scattered baobabs located severalhundred metres from the houses.

to the presence of Nebedaye trees. The dashed line indicates the overallne across each diamond represents the group mean, the vertical span of

(√

2CI)/2 above and below the group mean. The horizontal span of eachs.

338 B. Faye et al. / Transactions of the Royal Society of Tropical Medicine and Hygiene 105 (2011) 333–340

Table 3Leishmania seroprevalence in households with a dog with canine leishmaniasis in different villages from the Mont Rolland District

Village Ethnic group No. of households N Seroprevalence (%)

Ndiaye Bopp Serer 3 36 52.8Ndiaye Bopp 2a Fulani 1 23 17.4Fouloum Serer 3 20 35Nguith Fall Wolof 1 21 33.3Mont Rolland Serer 4 31 19.3Total 12 131 32.8

o dema

a Ndiaye Bopp 2 village was considered a single household since there are nsuch as keeping animals and gardening.

4. Discussion

Human CL caused by L. major is known to be endemicin West Africa, with Arvicanthis niloticus as the animalreservoir, and important outbreaks of the disease wererecently described in Senegal and Burkina Faso.12–15 Incontrast, very scarce data are available on the existenceof VL and to our knowledge only a few cases have beendescribed in The Gambia16 and Ivory Coast.17 In addition,historical studies carried out in Senegal in the 1960s and1970s have evidenced the existence of canine leishma-niasis, with a particularly high prevalence in the Thièsarea in Senegal.18,19 With the tools available at that time[multilocus enzyme electrophoresis (MLEE), with six enzy-matic systems], strains isolated from dogs were shownto be similar to those isolated from dogs and VL patientsin the Mediterranean basin and different from those iso-lated locally from patients with CL or from A. niloticus.20,21

Recently, epidemiological surveys were carried out in vil-lages of the Mont Rolland District (30 km from Thiès) toassess the situation of canine leishmaniasis in this area. Aspreviously described, a high prevalence of canine leishma-niasis was found, and strains isolated from dogs were typedas L. infantum.2

Although no clinical cases of VL were detected duringthis study, a high prevalence of seropositivity was evi-denced by means of serological reactivity to the 14 kDaand 18 kDa L. infantum-specific antigens. Seroprevalencewas 23% in the village of Ndiaye Bopp among 315 subjectsrandomly selected from the different areas of the village.Immunoblotting proved to be a very sensitive techniquein detecting Leishmania-specific antigen, even at very lowlevels,22 and similar seroprevalence levels are found inthe Mediterranean basin.7 In a L. infantum-endemic area,asymptomatic infection proved to be the rule23,24 and canlast for years before clinical disease appears, generally dueto immune deficiencies caused by HIV or immunosup-pressive therapies.25–27 In the present study, relying onlyon serological analysis of fingerprick blood, it is not pos-sible to say whether the high prevalence of Leishmaniaseropositivity is due to asymptomatic infection or to pre-vious contact with Leishmania parasites. However, it doesshow that humans are highly exposed to bites from infectedsandflies in this area. Such a high seroprevalence is unlikelyto be caused by cross-reactions with antibodies directed

against L. major. Such cross-reactions have indeed beenobserved for the rK39 recombinant antigen; however, itonly concerns a limited fraction of patients with CL (usually<10%).28,29 Furthermore, this study was based on serologi-

rcations between houses and inhabitants share spaces for village activities

cal reactivity to the 14 kDa and 18 kDa Leishmania-specificantigens and no CL lesions or scars were evidenced in anyof the 97 subjects who were positive on Western blot anal-ysis. The fact that no VL disease was evidenced in humansprobably results from the fact that, as in the Mediterraneanbasin, disease is rare in immunocompetent subjects. How-ever, the results suggest that VL may occur in Senegalbut could be misdiagnosed (e.g. as relapsing episodes ofmalaria or leukaemia) since VL is not routinely diagnosedin this country. In this context, special attention should bepaid to HIV patients as HIV–Leishmania co-infection hascaused serious public health problems both in Europe andAfrica.27,30,31 Although the HIV prevalence is relatively lowin Senegal, recent observations indicate an increase in theHIV-seropositive population, mainly in young adults mov-ing from rural areas to the capital.32

Among factors influencing an individual’s serologicalstatus, being >40 years old quadrupled the risk (OR = 4.01,P < 0.0001) of being reactive to 14 kDa and/or 18 kDaLeishmania-specific antigens. This strong age effect couldbe explained by the fact that the time spent in the areaincreases the chance of being bitten by an infected sand-fly. In addition, a significant effect of the presence of adog in the household was noted. This effect was increasedand highly significant when only the presence of infecteddogs was considered and was associated with an OR of5.56 (P < 0.0001) in the multivariate analysis. Presence ofan infected dog may thus constitute an important risk fac-tor for human infection by L. infantum. It also suggeststhat transmission may take place in the peridomicil-iary or intradomiciliary space, notably when an infecteddog is present. However, these are preliminary data andfurther work such as infectiousness studies, genetic anal-ysis of strains circulating within a household, and largercase–control studies including other factors such as humanactivities or mobility are now required to characterisebetter the transmission of L. infantum to humans in thisendemic area.

This observation implied that the vector can also befound near or around homes and prompted us to studythe environmental factors that could influence the den-sity of sandflies in the peridomiciliary space. For this study,it was decided to include only households with infecteddogs, given that in this kind of setting environmental fac-tors associated with the presence of the putative vector

may be associated with Leishmania serological positiv-ity in humans. Only the presence of Nebedaye trees (M.oleifera) remained associated with an increased risk ofbeing seropositive (OR = 3.07, P = 0.008) along with age in

of Tropi

tfttamtBwitNdv2dvttsoocdwNBadveuvIacdfc

MamtdihaamtsastptvwwT

B. Faye et al. / Transactions of the Royal Society

he multivariate analysis. This observation could resultrom feeding preferences of the L. infantum sandfly vec-or on the gum of this particular tree. However, althoughhe three households harbouring Nebedaye trees weremong the households with the highest observed Leish-ania seroprevalence (Figure 3), the sample size used for

his study was small and they were all located in Ndiayeopp village. Further studies conducted at a larger scaleill therefore be required to confirm this hypothesis. It is

nteresting to note that a higher prevalence of seroposi-ivity in households with an infected dog was observed indiaye Bopp village (52.8%), demonstrating that the con-itions of transmission to humans are probably met in thisillage. In contrast, seroprevalence was low in Ndiaye Bopp(17.4%), a village located only several kilometres away,

espite the high prevalence of canine leishmaniasis in thisillage inhabited by Fulani herders. Our interpretation ishat in this village, dogs probably become infected whileending herds at night in pastoral areas, but less transmis-ion to human occurs due to the low density or absencef the vector around the village owing to the absencef vegetation. Although the sandfly fauna has been wellharacterised in the study area and comprises numerousifferent species, no known vector species of L. infantumas identified (M.W. Senghor et al., unpublished data).evertheless, high sandfly densities were found in Ndiayeopp village, with a predominance of Sergentomyia dubiand S. schwetzi species. Phlebotomus spp. were at a very lowensity since only a few individuals of P. duboscqi, the usualector of L. major, were captured in this village. Furtherntomological surveys are ongoing to identify the as yetnknown vector specie(s). Results of the serological sur-eys carried out in this study could help to achieve this goal.ndeed, capture of sandflies present in households with (i)n infected dog and (ii) where the prevalence of serologi-ally positive individuals is particularly high and does notiffer between children and adults (presumably resultingrom intradomiciliary transmission) should increase thehance of identifying L. infantum-infected sandflies.

In conclusion, this seroepidemiological study in theont Rolland District in Senegal has revealed that, in

ddition to a high prevalence of L. infantum canine leish-aniasis, a high proportion of humans are also exposed

o infection by these parasites. Presence of an infectedog increased the risk of exposure in household inhab-

tants; therefore, control of canine leishmaniasis shouldelp reduce the contact of humans with L. infantum par-sites in this area. Larger-scale studies in Senegal and event the scale of West Africa are now required to define aap of L. infantum distribution in this part of the con-

inent. In areas where canine leishmaniasis and humaneropositivity to the 14 kDa and 18 kDa Leishmania-specificntigens are present, special attention should be paid toubjects presenting clinical signs of VL and diagnosis ofhis disease should be sought, most particularly in HIVatients. Finally, the results of this serological study illus-rate how such studies may help to identify the insect

ector species, especially in leishmaniasis-endemic areashere large numbers of sandfly species are usually presentith only one or a few involved in Leishmania transmission.

he results presented here constitute the first baseline data

cal Medicine and Hygiene 105 (2011) 333–340 339

on the epidemiology of L. infantum in West Africa. Owing tothe relatively small sample size used in this study, furtherlarger case–control studies combined with entomologicalstudies are now required to confirm these findings and todescribe better the life cycle of L. infantum in these areas.

Authors’ contributions: BF and BB contributed equally tothe study. OG conceived the study; BF, BB and ALB designedthe study, performed field data collection, and preparedthe first draft of the manuscript; BB analysed the data;MWS, AAN, SD, OK and MMD participated in the field datacollection; MH validated the data; SM, RK, PD and PMparticipated in the laboratory analysis of the samples andinterpreted the data. All authors contributed to revision ofthe manuscript and read and approved the final version.ALB is guarantor of the paper.

Acknowledgements: Sincere thanks are extended to MrLouis Mbengue, Les Sœurs de l’Immaculée Conception deMont Rolland, the population of the rural community ofMont Rolland, Dr Marieme Kane and the laboratory staff atthe Hôpital Régional de Thiès, the health region of Thiès,the medical officer (Dr Y. Dial) and the health district ofTivaouane. The authors thank Linda Northrup for reviewingthe English of this manuscript.

Funding: Agence Universitaire de la Francophonie(AUF), Paris, France; Centre National de la Recherche Sci-entifique (CNRS), Paris, France; Institut de Recherche pourle Développement (IRD), Marseille, France; and AgenceNationale de la Recherche (ANR), Paris, France

Conflicts of interest: None declared.

Ethical approval: The National Ethics Committee forHealth Research of Senegal approved the study and theSenegal Ministry of Health gave administrative authoriza-tion. Agreement to carry out the study was also obtainedfrom the Mont Rolland District authorities, the village com-mittees and the adults who participated in the study. Forchildren, informed consent was obtained from their par-ents or carers.

References

1. Dereure J, El-Safi SH, Bucheton B, Boni M, Kheir MM, Davoust B,et al. Visceral leishmaniasis in eastern Sudan: parasite identifica-tion in humans and dogs; host–parasite relationships. Microbes Infect2003;5:1103–8.

2. Faye B, Banuls AL, Bucheton B, Dione MM, Bassanganam O, HideM, et al. Canine visceral leishmaniasis caused by Leishmania infan-tum in Senegal: risk of emergence in humans? Microbes Infect2010;12:1219–25.

3. Ba Y, Trouillet J, Thonnon J, Fontenille D. Phlebotomines of Senegal(Diptera: Psychodidae): population and population dynamics of theMont-Rolland Region [in French]. Parasite 1998;5:143–50.

4. Ajdary S, Alimohammadian MH, Eslami MB, Kemp K, Kharazmi A.Comparison of the immune profile of non-healing cutaneous leish-maniasis patients with those with active lesions and those who have

recovered from infection. Infect Immun 2000;68:1760–4.

5. Sacks DL. B cell dependent T lymphocyte responses in leishmaniasis.Mem Inst Oswaldo Cruz 1988;83:506–13.

6. Kubar J, Quaranta JF, Marty P, Lelièvre A, Le Fichoux Y, Aufeuvre JP.Transmission of L. infantum by blood donors. Nat Med 1997;3:368.

of Tropi

340 B. Faye et al. / Transactions of the Royal Society

7. Le Fichoux Y, Quaranta JF, Aufeuvre JP, Lelievre A, Marty P, SuffiaI, et al. Occurrence of Leishmania infantum parasitemia in asymp-tomatic blood donors living in an area of endemicity in southernFrance. J Clin Microbiol 1999;37:1953–7.

8. Marty P, Lelievre A, Quaranta JF, Rahal A, Gari-Toussaint M, Le FichouxY. Use of the leishmanin skin test and Western blot analysis for epi-demiological studies in visceral leishmaniasis areas: experience in ahighly endemic focus in Alpes-Maritimes (France). Trans R Soc TropMed Hyg 1994;6:658–9.

9. Marty P, Lelièvre A, Quaranta JF, Suffia I, Eulalio M, Gari-ToussaintM, et al. Detection by Western blot of four antigens characterizingacute clinical leishmaniasis due to Leishmania infantum. Trans R SocTrop Med Hyg 1995;6:690–1.

10. Suffia JF, Quaranta MC, Eulalio M, Ferrua B, Marty P, Le Fichoux Y,et al. Human T-cell activation by 14 and 18 kDa nuclear proteins ofLeishmania infantum. Infect Immun 1995;63:3765–71.

11. Badaro R, Benson D, Eulalio MC, Freire M, Cunha S, Netto EM, et al.rK39: a cloned antigen of Leishmania chagasi that predicts activevisceral leishmaniasis. J Infect Dis 1996;173:758–61.

12. Develoux M, Diallo S, Dieng Y, Mane I, Huerre M, Pratlong F, et al.Diffuse cutaneous leishmaniasis due to Leishmania major in Senegal.Trans R Soc Trop Med Hyg 1996;90:396–7.

13. Guiguemdé RT, Sawadogo OS, Bories C, Traore KL, Nezien D, NikiemaL, et al. Leishmania major and HIV co-infection in Burkina Faso. TransR Soc Trop Med Hyg 2003;2:168–9.

14. Ndongo S, Dieng MT, Dia D, Sy TN, Leye A, Diop MT, et al. Cutaneousleishmaniasis in hospital area: epidemiological and clinical aspects,about 16 cases [in French]. Dakar Med 2004;49:207–10.

15. Traoré KS, Sawadogo NO, Traoré A, Ouedraogo JB, Traoré KL,Guiguemdé TR. Preliminary study of cutaneous leishmaniasis in thetown of Ouagadougou from 1996 to 1998 [in French]. Bull Soc PatholExot 2001;94:52–5.

16. Walters JH. A case of indigenous kala-azar in the Gambia. Trans R SocTrop Med Hyg 1949;43:287–92.

17. Eholié SP, Tanon AK, Folquet-Amorissani M, Doukouré B, AdoubrynKD, Yattara A, et al. Three new cases of visceral leishmaniasis in Côted’Ivoire [in French]. Bull Soc Pathol Exot 2008;1:60–1.

18. Ranque P, Bussiéras J, Chevalier JL, Quilici M, Mattei X. Presentimportance of dog leishmaniasis in Senegal. Value of immunologic

diagnosis. Possible incidence in human pathology [in French]. BullAcad Natl Med 1970;154:510–2.

19. Desjeux P, Bray RS, Dedet JP, Chance M. Differentiation of canine andcutaneous leishmaniasis strains in Senegal. Trans R Soc Trop Med Hyg1982;76:132–3.

cal Medicine and Hygiene 105 (2011) 333–340

20. Chance ML. The six diseases of WHO. Leishmaniasis. Br Med J (ClinRes Ed) 1981;6301:1245–7.

21. Ranque P, Quilici M, Camerlynck P. Avicanthis niloticus (rodent, Muri-dae), a reservoir of the virus causing leishmaniasis in Senegal [inFrench]. Bull Soc Pathol Exot Filiales 1974;67:167–75.

22. Mary C, Lamouroux D, Dunan S, Quilici M. Western blot analysis ofantibodies to Leishmania infantum antigens: potential of the 14-kDand 16-kD antigens for diagnosis and epidemiologic purposes. Am JTrop Med Hyg 1992;6:764–71.

23. Riera C, Fisa R, Udina M, Gállego M, Portus M. Detection of Leishmaniainfantum cryptic infection in asymptomatic blood donors living in anendemic area (Eivissa, Balearic Islands, Spain) by different diagnosticmethods. Trans R Soc Trop Med Hyg 2004;2:102–10.

24. Riera C, Fisa R, López-Chejade P, Serra T, Girona E, Jiménez M, et al.Asymptomatic infection by Leishmania infantum in blood donorsfrom the Balearic Islands (Spain). Transfusion 2008;7:1383–9.

25. Rosenthal E, Marty P, le Fichoux Y, Cassuto JP. Clinical manifestationsof visceral leishmaniasis associated with HIV infection: a retro-spective study of 91 French cases. Ann Trop Med Parasitol 2000;1:37–42.

26. Rosenthal E, Marty P, del Giudice P, Pradier C, Ceppi C, Gastaut JA,et al. HIV and Leishmania coinfection: a review of 91 cases with focuson atypical locations of Leishmania. Clin Infect Dis 2000;4:1093–5.

27. Alvar J, Aparicio P, Aseffa A, Den Boer M, Canavate C, Dedet JP,et al. The relationship between leishmaniasis and AIDS: the second10 years. Clin Microbiol Rev 2008;2:334–59.

28. Braz RF, Nascimento ET, Martins DR, Wilson ME, Pearson RD, Reed SG,et al. The sensitivity and specificity of Leishmania chagasi recombi-nant K39 antigen in the diagnosis of American visceral leishmaniasisand in differentiating active from subclinical infection. Am J Trop MedHyg 2002;67:344–8.

29. Hartzell JD, Aronson NE, Weina PJ, Howard RS, Yadava A, WortmannGW. Positive rK39 serologic assay results in US servicemen withcutaneous leishmaniasis. Am J Trop Med Hyg 2008;79:843–6.

30. Niamba P, Traoré A, Goumbri-Lompo O, Labrèze C, Traoré-Barro F,Bonkoungou M, et al. Cutaneous Leishmania in HIV patient in Oua-gadougou: clinical and therapeutic aspects [in French]. Ann DermatolVenereol 2006;6:537–42.

31. Valcárcel Y, Bastero R, Anegón M, González S, Gil A. The epidemiology

of hospital admissions due to leishmaniasis in Spain (1999-2003) [inSpanish]. Enferm Infecc Microbiol Clin 2008;5:278–81.

32. Conseil National de Lutte contre le SIDA. Plan stratégique de lutte con-tre le Sida 2007-2011. Senegal: Ministère de Santé et de la PréventionMédicale du Sénégal; 2007.