SEARCH

SEARCH BY CITATION

Keywords:

  • Leishmania (Viannia) guyanensis ;
  • Phlebotominae;
  • environmental exposure;
  • Bayesian analysis
  • Leishmania (Viannia) guyanensis ;
  • phlébotomie;
  • exposition à l'environnement;
  • analyse bayésienne
  • Leishmania (Viannia) guyanensis ;
  • Phlebotominae;
  • Exposición Ambiental;
  • análisis Bayesiano

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Objective

Cutaneous leishmaniasis (CL) is more frequently reported in men than in women; this may be due to male-biased exposure to CL vectors, female-biased resistance against the disease or both. We sought to determine whether gender-specific exposure to vector habitats explains male-biased CL incidence in two human populations of central Amazonia.

Methods

We compared the CL incidence in one population of field researchers (N = 166), with similar exposure for males and females, and one population of rural settlers (N = 646), where exposure is overall male-biased. We used a combination of questionnaires and clinical data to quantify CL cases, and modelled disease incidence in a Bayesian framework.

Results

There was a moderately higher incidence of CL among men than among women in both populations, but male bias decreased as exposure time increased. Disease incidence was overall higher among field researchers, suggesting that they are an important but understudied CL risk group.

Conclusion

Our comparison of two contrasting populations provided epidemiological evidence that CL incidence can be male-biased even when exposure is comparable in both sexes.

Objectif

La leishmaniose cutanée (LC) est plus fréquemment rapportée chez les hommes que chez les femmes, cela peut être dû à l'exposition particulière des hommes aux vecteurs de la LC, la résistance particulière des femmes contre la maladie, ou les deux. Nous avons cherché à déterminer si l'exposition spécifique au sexe aux habitats des vecteurs explique l'incidence biaisée de la LC chez les hommes dans deux populations humaines du centre de l'Amazonie.

Méthodes

Nous avons comparé l'incidence de la LC dans une population de chercheurs sur le terrain (= 166), avec une exposition similaire pour les hommes et les femmes à celle d'une population de colons ruraux (= 646), chez laquelle l'exposition est globalement particulière pour les hommes. Nous avons utilisé une combinaison de questionnaires et des données cliniques pour quantifier les cas de LC et avons modélisé l'incidence de la maladie dans un cadre bayésien.

Résultats

Il y avait une incidence modérément plus élevée de LC chez les hommes que chez les femmes dans les deux populations, mais la particularité masculine diminuait avec l'augmentation du temps d'exposition. L'incidence de la maladie était globalement plus élevée chez les chercheurs sur le terrain, ce qui suggère qu'ils sont un groupe à risque important pour la LC, mais peu étudié.

Conclusion

Notre comparaison de deux populations contrastées fourni la preuve épidémiologique que l'incidence de la LC peut être liée au sexe masculin même lorsque l'exposition est comparable entre les deux sexes.

Objetivo

La leishmaniosis cutánea (LC) se reporta más frecuentemente entre hombres que entre mujeres; esto podría deberse a una exposición sesgada de los hombres a los vectores de la LC, a un sesgo en la resistencia femenina frente a la enfermedad, o a ambos. Hemos buscado determinar si la exposición género-específica a los hábitats vectoriales explican el sesgo masculino en la incidencia de LC en os poblaciones humanas de la Amazonía central.

Métodos

Hemos comparado la incidencia de LC en una población de investigadores de campo (N = 166), con exposición similar para hombres y mujeres, y una población de habitantes rurales (N = 646), en donde la exposición en general tiene un sesgo masculino. Hemos utilizado una combinación de cuestionarios y datos clínicos para cuantificar casos de LC, y modelado la incidencia de enfermedad en un marco Bayesiano.

Resultados

Había una incidencia moderadamente mayor de LC entre hombres que entre mujeres en ambas poblaciones, pero el sesgo masculino disminuía a medida que el tiempo de exposición aumentaba. La incidencia de enfermedad en general era más alta entre trabajadores de campo, lo que sugería que son un grupo de riesgo importante pero poco estudiado

Conclusión

Nuestra comparación de dos poblaciones contrastables aportaba evidencia epidemiológica de que la incidencia de LC puede tener un sesgo masculino aún cuando la exposición es comparable en ambos sexos.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Sex is a relevant infectious disease risk-factor, as males and females can have different probabilities of acquiring an infection and expressing clinical signs or symptoms of a disease (Roberts et al. 2001). In South America, cutaneous leishmaniasis (CL) is a zoonotic disease transmitted by sandflies (Phlebotominae), with a wide spectrum of clinical manifestations; the most typical are ulcerative skin lesions developed near or at the vector's bite site (Reithinger et al. 2007). Most epidemiological studies show that clinical CL is more frequent among men than women (e.g., Armijos et al. 1997; Davies et al. 2000; Wasserberg et al. 2002; Calvopiña & Hashiguchi 2004; Guerra et al. 2006; Guerra-Silveira & Abad-Franch 2013), and there is experimental evidence suggesting differential vulnerability of males and females to Leishmania spp. infection among laboratory animals (Travi 2002; Snider et al. 2009; Klein & Roberts 2010). In the present study, we investigated sex-biased vulnerability to clinical CL in human populations with differential patterns of gender-related exposure to vector habitats.

In central Amazonia and other humid-forest regions of Latin America in which CL is endemic, exposure to Leishmania vectors depends upon contact with forested habitats that maintain sandfly populations. Men are usually more exposed than women because they engage more often in forest-related activities such as hunting, logging or military training (Lainson 1988; Davies et al. 2000; Guerra et al. 2003). Additionally, androgens can partially suppress the immune response in males (Snider et al. 2009; Klein & Roberts 2010), and human population data show that infectious disease incidence (Guerra-Silveira & Abad-Franch 2013) as well as parasite-induced mortality (Owens 2002) is higher among men than women. Thus, there is evidence supporting both differential exposure and hormone-mediation as explanations for the higher incidence of parasitic infections in men than in women.

The relative importance of differential exposure and sex-related physiological factors in shaping CL incidence in human populations remains obscure. Here, we asked whether CL incidence is sex-biased even in the absence of differential exposure. We addressed this question by comparing clinical CL incidence in two populations, one with and one without gender-related differential exposure. Both populations occupied similar upland forest sites where Lutzomyia umbratilis is the principal CL vector (Killick-Kendrick 1990), and the vast majority of CL cases is caused by Leishmania (Viannia) guyanensis (Lainson et al. 1994). Our analyses tested the null hypothesis that, for both populations, male–female differences in incidence are solely due to differences in exposure.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Study areas and populations

The first population was composed of rural settlers from the Rio Pardo settlement, established in 1996 in an area of primary upland rainforest and extending over 280 km2, 155 km north of Manaus, Brazil (Figure 1). In a census carried out in 2008, Rio Pardo had 647 settlers in 225 households (L. Soares, F. Abad-Franch, unpublished data). The second population pertained to the research sites of the Biological Dynamics of Forest Fragments Project (BDFFP), which span about 500 km2 and 70 km north of Manaus (Figure 1). The BDFFP is a long-term deforestation experiment initiated in 1979 (Bierregaard et al. 2001). Since then, at least 300 field researchers, including field assistants, interns, graduate students and professional researchers, visited BDFFP field sites at least once. We assumed that BDFFP researchers become exposed to forest environments independently of gender. They were also expected to make use of protection against vector bites more regularly than rural settlers, while the overall time of exposure is probably higher among settlers. We took advantage of gender-independent exposure to forest environment among BDFFP researchers and used this population as a control group to test our hypothesis about the role of differential exposure in generating sex-biased CL incidence.

image

Figure 1. Study area map. (1) Location of the study areas in South America, north of Manaus, Brazil. (2) Biological Dynamics of Forest Fragments Project (BDFFP) field research sites. (3) The rural settlement of Rio Pardo. Dark grey represents primary tropical upland forest; light grey represents areas of forest disturbance such as second-growth forests, deforested areas and cropland; the continuous line represents the BR-174 highway; dashed lines represent unpaved roads; and the dotted line in panel (2) is the Pardo River.

Download figure to PowerPoint

CL epidemiological survey

Our epidemiological data emphasised self-reports of CL clinical manifestations. We also recorded the history of medical treatment against CL, which is unlikely to have been administered in the absence of a confirmed diagnosis and highly likely to be remembered by patients. The CL treatment adopted by the Brazilian Ministry of Health is the only treatment in the clinical-epidemiological context of our study involving intravenous or intramuscular injections of pentavalent antimony for at least 20 days (Departamento de Vigilância Epidemiológica 2006). Additionally, in Rio Pardo all participants were examined by a trained physician, who recorded the presence of characteristic CL scars or active lesions.

We collected Rio Pardo data between November 2008 and June 2009. Semi-structured questionnaires were given to family heads, with questions about all individuals living in the household. Besides time of residence in Rio Pardo, questions focused on disease history, including occurrence of CL cases, year of development of CL skin lesions and type of treatment used against the disease. Our incidence estimates were based on autochthonous cases of CL only.

As many BDFFP researchers no longer live in the study region, we used the BDFFP database and its technical series to list all scientists, graduate students, interns and field assistants who probably worked in the project sites between 1979 and 2009. We sent a pre-survey email to each of these individuals and identified those that had indeed conducted fieldwork in the area; then, we sent an electronic questionnaire to each of them. Because researchers that had CL could have been more prone to answer the survey, we tackled this possible response-bias issue by sending out the electronic questionnaires over three rounds, each covering the list of non-responding potential participants. The questionnaire recorded (i) total time spent in the field and (ii) disease history (occurrence of CL, year of development of skin lesions, type of treatment). To control for gender-specific behavioural risk factors, the BDFFP questionnaire also asked about (iii) the period of the day spent working inside the forest and (iv) the use of any protection against insect bites. The Institutional Review Board of the Brazilian Instituto Nacional de Pesquisas da Amazônia approved all the procedures herein described (Protocol 206/09). Prior to interviews, all subjects signed an informed consent form.

Data analysis

We tested predictions about CL incidence in the Rio Pardo and BDFFP populations using linear models implemented in a Bayesian framework (Royle & Dorazio 2008). Our analysis was based on two logistic models that predict CL incidence, expressed as the probability that one individual reports a new case of CL over a given period of exposure. Model 1 compared CL incidence between Rio Pardo and BDFFP individuals of both sexes with the purpose of testing our hypothesis that sex bias in CL incidence was due solely to differences in exposure, and thus should not appear in the BDFFP population. As this model indirectly investigated possible effects of sex hormones on disease incidence, we fitted it only to data on individuals over 12-years old. Model 1 can be summarised by the following equations, where index i designates individuals from either Rio Pardo or BDFFP:

  • display math(1a)
  • display math(1b)

In equation (1), logit (θi) is the natural logarithm of the odds of any individual i reporting CL, which is represented by the formula ln(θi/1- θi). We use logit (θi) as the link function to estimate CL prevalence using a logistic model of three additive terms, sex (βS), exposure time (βT) and population (βP), as well as one interaction term between sex and population (βSP). The data vector yi contained all binary observations of CL (yi = 1 if individual i reported CL, yi = 0 otherwise); we modelled yi as a random draw from a Bernoulli distribution with incidence parameter θi Eq. (1a). In turn, θi was given by a logistic function of sex (Si = 1 if i is male and 0 if female), time of exposure measured in years (Ti), population from which individual i was sampled (Pi = 1 for Rio Pardo and 0 otherwise) and an interaction term that expressed a possible difference between the effect of sex across populations (SPi = 1 if i is a man from Rio Pardo and 0 in all other cases; Eq. (1b). Time Ti was residence time for Rio Pardo and time spent in the field sites for the BDFFP. We standardised T by subtracting the mean and dividing by the standard deviation.

Model 2 also examined the effects of exposure and sex, asking whether the effects of exposure were the same among males and females, but was focused on BDFFP individuals only. This choice reflected the fact that we are more certain that ‘exposure time’ meant exposure to sandfly habitats for the BDFFP than for Rio Pardo data. This model included the same effects of sex and exposure as in Model 1 plus an interaction term representing the possibility that males and females respond differently to the same time of exposure (STi). In Model 2, we modelled the data as a Bernoulli draw just like in Eq. (1a) but used the following logistic function for the incidence parameter:

  • display math(2)

In equation (2), logit (θi) is the logistic link function estimating CL prevalence using two additive terms, sex (βS) and exposure time (βT), as well as one interaction term between sex and population (βSP). Parameter estimates where derived from posterior distributions generated by a Markov chain Monte Carlo (MCMC) algorithm with non-informative normally distributed priors (mean zero, variance 100). We fitted models using a combination of R and WinBUGS, via the R package R2WinBUGS (Spiegelhalter et al. 2003; Sturtz et al. 2005).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

CL epidemiological overview

The Rio Pardo survey included 646 settlers from 225 households. We identified 56 autochthonous CL cases in the settlement; a physician confirmed all these reports based on the presence of characteristic scars. Intravenous or intramuscular injections were the main reported treatment; three suspected CL cases reportedly treated with skin ointments were not considered in the analyses. In total, 47 (7%) male settlers reported having had a case of CL, whereas only nine (1.4%) females did so (Table 1). The mean age at which rural settlers reported having had CL was 29-years old (SD = 14).

Table 1. Cutaneous leishmaniasis reports among Rio Pardo settlers and Biological Dynamics of Forest Fragments Project (BDFFP) researchers
PopulationCutaneous leishmaniasisTotalUnadjusted odds ratio
Yes (%)NoEstimate95% CI
  1. Yes, past clinical evidence of cutaneous leishmaniasis; No, no such evidence; CI, confidence interval.

Rio Pardo
Female9 (3.39)2562651  
Male47 (12.34)3343814.001.938.32
Sub-total56 (8.67)590646   
BDFFP
Female8 (11.11)64721  
Male19 (20.21)75942.030.834.94
Sub-total27 (16.27)139166   
Overall
Female17 (5.09)3203341  
Male66 (13.90)4094753.041.755.28
Total83 (10.22)729812   

Our compilation of subjects associated to the BDFFP resulted in a list of 376 people. Of these, 237 had worked in the field research sites, and 166 responded to our questionnaire; 19 men and eight women reported CL (Table 1). The mean age at which field researchers reported having had CL was 30 years (SD = 1.7). Exposure time, measured as the number of years in the field, was comparable among male (inline image  = 1.87; SD = 3.10) and female researchers (inline image = 1.15; SD = 0.88) (Mann–Whitney U = 1390; P = 0.88), which supports the use of the BDFFP population as a control group for the population of rural settlers. Additionally, 40 (43%) men and 23 (32%) women in the BDFFP reported that most of their field duties were carried out during nocturnal or crepuscular periods. Among female researchers, 68 (95%) reported using some form of protection against mosquito bites when working in the forest, whereas only 70 (74%) of men reportedly did so (Fisher's exact test, P = 0.005). In order to rule out the influence of behavioural biases on the probability of reporting CL, we ran an additional model (results available upon request from the authors) estimating the effects of using protection against insect bites, such as the use of insect repellents and protective clothing, and its interaction with sex. This model showed that although the use of protection against insect bites is more common among females than males, this behavioural bias had a negligible relation with CL incidence.

CL incidence and sex

Table 2 shows parameter estimates for Model 1. While the effect of being male was moderate (95% confidence bounds of the βS posterior distribution overlap zero at the lower end), its mean estimate was positive (βS = 0.70 ± 0.47), and the effect of sex did not show an obvious interaction with site (βSP = 1.19 ± 0.68, with 95% confidence bounds overlapping zero). Time of exposure had a clearly positive effect on incidence (βT = 0.76 ± 0.15). The odds of reporting CL were larger for males than for females in both populations and particularly among settlers (Table 2). As shown by the negative estimate of the effect of being from Rio Pardo (βP = −2.37 ± 0.66), rural settlers were less likely to report CL than individuals from the BDFFP population. Figure 2 shows CL incidence estimates (θ) for men and women in both Rio Pardo and the BDFFP, obtained from Model 1, with an exposure time of 1 year. In both populations, incidence estimates were higher for males than for females.

Table 2. Model 1 (Equation 1) parameters. Model 1 compares cutaneous leishmaniasis (CL) incidence among Rio Pardo settlers and researchers from the Biological Dynamics of Forest Fragments Project; it tests the hypothesis that exposure time drives sex bias in CL incidence
ParameterDescriptionMeanSE95% CBOR95% CI
  1. SE, standard error; CB, confidence bounds from the posterior distribution; OR, odds ratio [estimated as OR = exp (β)]; CI, confidence interval.

β 0 Intercept−1.520.41−2.35−0.76
β S Effect of being male0.700.47−0.191.652.010.835.21
β T Effect of 1 year of exposure0.760.150.471.062.131.592.89
β P Effect of being from Rio Pardo−2.370.66−3.66−1.140.090.030.32
β SP Interaction sex × population1.190.68−0.102.513.280.9112.27
image

Figure 2. Model 1 estimates of cutaneous leishmaniasis incidence in men and women for both populations. Points represent the mean estimate of disease incidence for 1 year of individual exposure for males and females of the Rio Pardo settlement and the Biological Dynamics of Forest Fragments Project (BDFFP). Vertical lines are 95% confidence bounds of the posterior distribution.

Download figure to PowerPoint

A closer look at the BDFFP data, based on Model 2, confirmed the tendency of a weak effect of sex on incidence and revealed an interaction between sex and exposure time (Table 3). The effect of being male was moderately positive (βS = 0.96 ± 0.59), the effect of exposure time was obviously positive (βT = 2.51 ± 0.87). Furthermore, the negative interaction between sex and exposure time (βST = −1.62 ± 0.92) indicated that while males were more likely to report the disease at short exposures, such sex bias tended to disappear with sustained exposure.

Table 3. Model 2 (Equation (2)) parameters. Model 2 examines the incidence of cutaneous leishmaniais (CL) in the population of researchers of the Biological Dynamics of Forest Fragments Project (BDFFP), estimating whether the effect of exposure is sex-biased
ParameterDescriptionMeanSE95% CBOR95% CI
  1. SE, standard error; CB, confidence bounds from the posterior distribution; OR, odds ratio [estimated as OR = exp (β)]; CI, confidence interval.

β 0 Intercept−2.500.51−3.61−1.61
β S Effect of being male0.960.590.122.182.611.138.85
β T Effect of 1 year of exposure2.510.871.034.4012.302.8081.45
β ST Interaction sex × exposure−1.620.92−3.590.010.180.031.01

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Our results added new information to previous clinical-epidemiological evidence of sex-biased CL incidence, with typically higher case frequency among males than among females (Weigle et al. 1993; Davies et al. 2000; Wasserberg et al. 2002; Calvopiña & Hashiguchi 2004; Guerra-Silveira & Abad-Franch 2013). After comparing two populations that differ in gender-related exposure, we suggested that incidence is male-biased in both, not just because the researchers showed a measurable, albeit relatively weak, effect of sex on incidence, but also because there was a strong interaction between sex and exposure time in this population. Therefore, both gender-related differential exposure and sex-related physiological factors likely contributed to male–female differences in the risk of clinical CL caused by Le. (V.) guyanensis.

The interpretation of our epidemiological data required consideration of two main issues. First, infection with Leishmania parasites does not always result in clinical CL symptoms; and second, geographic variation in the sandfly fauna could lead to differences in disease transmission between sites. As some Leishmania infections are asymptomatic, exposure to the parasites was likely higher than we perceived. Serological tests could provide a more precise assessment of exposure; however, we felt justified in using epidemiological information for two main reasons: 1) serology tests sometimes detect antibodies to Leishmania sp. in only 85% of patients clinically diagnosed with Le. (V.) guyanensis (Romero et al. 2005), and 2) serology tests can fail to detect an existing bias in clinical CL because they may respond to exposure to parasites even in the absence of clinical manifestation of the disease (Coimbra et al.1996; Guerra-Silveira & Abad-Franch 2013). We were particularly interested in male–female differences in the incidence of clinically evident CL, with cases defined by disease signs and/or symptoms.

As for geographic variation in vector fauna, the sandfly fauna of the BDFFP is likely to resemble that of Rio Pardo. Both sites are only about 75 km from each other and lie within the same eco-region – the Uatumã-Trombetas moist forests. Entomological surveys in Rio Pardo (W. Ramos, J.F. Medeiros, G.R. Julião, C.M. Rios - Velásquez, E.F. Marialva, S.J.M Desmouliere, S.L.B. Luz, F.A.C. Pessoa, unpublished data) showed that sandfly species composition was similar to what has typically been reported from the central Amazon region (Alencar 2007; Barbosa et al. 2008), particularly as regards the presence of the main local CL vector, Lu. umbratilis (W. Ramos, J.F. Medeiros, G.R. Julião, C.M. Rios - Velásquez, E.F. Marialva, S.J.M Desmouliere, S.L.B. Luz, F.A.C. Pessoa, unpublished data). Further research is needed to describe the sandfly fauna of BDFFP sites, but we had no reason to think that a local peculiarity would lead to a false impression of sex-biased incidence among researchers.

Our modelling results provided support for a combined effect of sex and exposure time as CL risk factors: first, the average odds of reporting CL were more than two times higher for males than for females in both populations (Table 1), and, second, our models estimated a relatively strong, negative effect of the interaction between male sex and time of exposure to sand fly habitats (Table 3). Sex-related differences in the clinical manifestations of Leishmania spp. infections can be at least partially explained by endocrine–immune interactions (Roberts et al. 2001; Snider et al. 2009; Klein & Roberts 2010). Protective immunity against parasites of the genus Leishmania critically depends on effective T helper (Th)1-type cellular responses (Alexander et al. 1999; Roberts et al. 2001; Alexander & Bryson 2005). As oestrogen upregulates Th1-type responses, females can potentially mount more efficient immunity against these parasites (Snider et al. 2009). Male hamsters experimentally infected with Le. (V.) guyanensis presented increased occurrence and severity of cutaneous lesions, supporting the role of sex hormones in immune modulation (Travi 2002).

Based on evidence implying an enhanced protective immune response in female hosts, we suggested that sex and vector exposure determined the higher risk of clinical development of CL among men. Our BDFFP results suggested that, at low levels of exposure, endocrine–immune interactions resulted in more effective responses in women; yet, this extra protection tended to wane under long-lasting exposure, perhaps because repeated infection events overwhelm cellular immunity. This would explain why male BDFFP researchers had higher probability of reporting CL than their female counterparts particularly at short exposure. On the other hand, the stronger male bias among settlers was likely the result of male-biased exposure and, to a lesser extent, higher overall male susceptibility.

Another important finding of our study was that according to the odds ratio obtained from our model parameters, Rio Pardo settlers were about 11 times less likely to report CL as BDFFP researchers (odds ratio = 0.09; see Table 2). Most settlers had a long history of exposure to sandfly bites, whereas most researchers were probably first bitten by sandflies during fieldwork. It has been shown experimentally that mice treated with proteins found in vector saliva were less vulnerable to Leishmania (Viannia) braziliensis infection, suggesting a protective effect of sandfly bites (Oliveira et al. 2008). Prior exposure of mice to bites of uninfected Phlebotomus papatasi, an Old World CL vector, resulted in some degree of immunity against Leishmania major (Kamhawi et al. 2000). However, the immune response induced by sandfly saliva can lead to either infection aggravation or resistance, depending on the parasite species and on the salivary proteins to which the hosts are exposed (Gomes et al. 2008; Oliveira et al. 2008; Collin et al. 2009). Whatever the precise mechanism, there is evidence that previously non-exposed human populations in Amazonia had a higher probability of developing CL than long-term residents of endemic areas (Lainson & Rangel 2005). Additionally, researchers may be more intensely exposed to sandfly habitats than rural settlers because of the contrasting dwelling conditions in the BDFFP sites vs. the settlement area. BDFPP researchers slept in open-walled premises immediately adjacent to forest environments that support sandfly populations, whereas most rural settlers lived in wooden-walled houses surrounded by crops or deforested areas where Lu. umbratilis is rarer than in nearby forest or forest edges (W. Ramos, J.F. Medeiros, G.R. Julião, C.M. Rios - Velásquez, E.F. Marialva, S.J.M Desmouliere, S.L.B. Luz, F.A.C. Pessoa, unpublished data).

Finally, we believe that a deeper understanding of genetic variability among parasites, vectors, and hosts would help foster understanding of CL epidemiology. The diversity of Le. (V.) guyanensis genotypes in the study area is unknown, yet different parasite lineages can result in specific disease manifestations (Nolder et al. 2007; Doudi et al. 2010). Genetically divergent sandfly populations can differ in their vector competence (Scarpassa & Alencar 2012), contributing to inter-population differences in CL incidence. We did not take into account genetic variation in our study populations, which included individuals from different geographic and ethnic origins who may have differed in immune competence against Leishmania infections, but we hope that our results will motivate further work on the mechanisms underlying CL epidemiological patterns. A consistent identification of higher-risk population groups and an improved understanding of the mechanisms driving CL incidence will provide a robust background for planning effective prevention strategies and reducing disease burden.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

We provided epidemiological evidence for sex bias in CL incidence: men were moderately more likely to report the disease than women. This sex effect was particularly strong at low levels of exposure, and appeared also when exposure to vector habitats was sex-unbiased. This suggested that sex-specific physiological factors were involved in the clinical epidemiology of CL in the central Amazon. Women seemed better equipped than men to contain the infection and its clinical outcomes, but sex bias tended to shrink at high levels of exposure. We also suggested that people who come into contact with forested habitats for the first time are at higher risk of developing CL; this is the case not only of field researchers, but also of military personnel, tourists or bird-watchers. These groups should strive to comply with strict personal protection measures to avoid infection from vector bites while in the forest.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Special thanks are due to Rio Pardo settlers and BDFFP field researchers who contributed to our CL epidemiological survey. Fieldwork was supported by grants from IDRC (Canada), CNPq (Brazil), FAPEAM (Brazil) and the Research Program on Infectious Disease Ecology in the Amazon of the Instituto Leônidas e Maria Deane (RP-IDEA, ILMD – Fiocruz (Brazil). Letícia Soares received a fellowship from the Brazilian Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for MSc studies at the graduate program in Ecology of the Instituto Nacional de Pesquisas da Amazônia (INPA). We are very thankful to Eduardo Venticinque and Toby V Barrett, who provided fundamental ideas for this study. We are also grateful to Felipe AC Pessoa, Luis Fernando Chaves, Jacob Koella, Richard Reithinger, Aloísio Falqueto and Ana Rabello who kindly read and commented on the written version of LS's MSc project. This is contribution number 20 of the Research Program on Infectious Disease Ecology in the Amazon (RP-IDEA) of the Instituto Leônidas e Maria Deane, and contribution number 642 of the Biological Dynamics of Forest Fragments Project (BDFFP).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  • Alencar RB (2007) Emergência de flebotomíneos (Diptera: Psychodidae) em chão de floresta de terra firme na Amazônia Central do Brasil: uso de um modelo modificado de armadilha de emergência. Acta Amazonica 37, 287292.
  • Alexander J & Bryson K (2005) T Helper (h)1/Th2: paradox rather than paradigm. Immunology Letters 99, 1723.
  • Alexander J, Satoskar AR & Russell DG (1999) Leishmania species: models of intracellular parasitism. Journal of Cell Science 112, 29933002.
  • Armijos RX, Weigel MM, Izurieta R et al. (1997) The epidemiology of cutaneous leishmaniasis in subtropical Ecuador. Tropical Medicine & International Health 2, 140152.
  • Barbosa MGV, Fé NF, Marcião AHR et al. (2008) Fauna de flebotomíneos (Diptera: Psychodidae) em um foco de leishmaniose tegumentar americana na área periurbana de Manaus, Estado do Amazonas. Revista da Sociedade Brasileira de Medicina Tropical 41, 485491.
  • Bierregaard R, Gascon C & Lovejoy TE (2001) Lessons from Amazonia: The Ecology and Conservation of a Fragmented Forest (1st edn.) Yale University Press, New Haven, CT, USA, pp. 478.
  • Calvopiña M & Hashiguchi RAY (2004) Epidemiology of leishmaniasis in Ecuador: current status of knowledge - a review. Memórias do Instituto Oswaldo Cruz 99, 663672.
  • Coimbra CEA, Santos RV & Valle ACF (1996) Cutaneous leishmaniasis in Tupí-Mondé Amerindians from the Brazilian Amazonia. Acta Tropica 61, 201211.
  • Collin N, Gomes R, Teixeira C et al. (2009) Sand fly salivary proteins induce strong cellular immunity in a natural reservoir of visceral leishmaniasis with adverse consequences for Leishmania. PLoS Pathogens 5, e1000441.
  • Davies CR, Reithinger R, Campbell-Lendrum D et al. (2000) The epidemiology and control of leishmaniasis in Andean countries. Cadernos de Saúde Pública 16, 925950.
  • Departamento de Vigilância Epidemiológica (2006) Atlas de Leishmaniose Tegumentar Americana: Diagnóstico Clínico e Diferencial (1st edn.) Ministério da Saúde do Brasil, Brasília, Brazil, pp. 136
  • Doudi M, Hejazi SH, Razavi MR et al. (2010) Comparative molecular epidemiology of Leishmania major and Leishmania tropica by PCR-RFLP technique in hyper endemic cities of Isfahan and Bam, Iran. Medical Science Monitor 16, 530535.
  • Gomes R, Teixeira C, Teixeira MJ et al. (2008) Immunity to a salivary protein of a sand fly vector protects against the fatal outcome of Visceral Leishmaniasis in a hamster model. Proceedings of the National Academy of Sciences of the U.S.A. 105, 78457850.
  • Guerra JAO, Talhari S, Paes MG et al. (2003) Aspectos clínicos e diagnósticos da leishmaniose tegumentar americana em militares simultaneamente expostos à infecção na Amazônia. Revista da Sociedade Brasileira de Medicina Tropical 36, 587590.
  • Guerra JAO, Ribeiro JAS, Coelho LIARC et al. (2006) Epidemiologia da leishmaniose tegumentar na Comunidade São João, Manaus, Amazonas, Brasil. Cadernos de Saúde Pública 22, 23192327.
  • Guerra-Silveira F & Abad-Franch F (2013) Sex bias in infectious disease epidemiology: patterns and processes. PLoS One 8, e62390.
  • Kamhawi S, Belkaid Y, Modi G et al. (2000) Protection against cutaneous leishmaniasis resulting from bites of uninfected sand flies. Science 290, 13511354.
  • Killick-Kendrick R (1990) Phlebotomine vectors of the leishmaniases: a review. Medical and Veterinary Entomology 4, 124.
  • Klein SL & Roberts CW editors (2010) Sex Hormones and Immunity to Infection (1st edn.) Springer Verlag, Berlin, Germany, pp.x+319.
  • Lainson R (1988) Ecological interactions in the transmission of the leishmaniases. Philosophical Transactions of the Royal Society of London B: Biological Sciences 321, 389404.
  • Lainson R & Rangel EF (2005) Lutzomyia longipalpis and the eco-epidemiology of American visceral leishmaniasis, with particular reference to Brazil: a review. Memórias do Instituto Oswaldo Cruz 100, 811827.
  • Lainson R, Shaw JJ, Silveira FT et al. (1994) The dermal leishmaniases of Brazil, with special reference to the eco-epidemiology of the disease in Amazonia. Memórias do Instituto Oswaldo Cruz 89, 435443.
  • Nolder D, Norma R, Davies CR et al. (2007) Multiple hybrid genotypes of Leishmania (Viannia) in a focus of mucocutaneous leishmaniasis. The American Journal of Tropical Medicine and Hygiene 76, 573578.
  • Oliveira F, Lawyer PG, Kamhawi S & Valenzuela JG (2008) Immunity to distinct sand fly salivary proteins primes the anti-Leishmania immune response towards protection or exacerbation of disease. PLoS Neglected Tropical Diseases 2, e226.
  • Owens IPF (2002). Sex differences in mortality rate. Science 297, 20082009.
  • Reithinger R, Dujardin JC, Louzir H et al. (2007) Cutaneous leishmaniasis. The Lancet Infectious Diseases 7, 581596.
  • Roberts CW, Walker W & Alexander J (2001) Sex-associated hormones and immunity to protozoan parasites. Clinical Microbiology Reviews 14, 476488.
  • Romero GAS, Orge MG, Guerra MVF et al. (2005) Antibody response in patients with cutaneous leishmaniasis infected by Leishmania (Viannia) braziliensis or Leishmania (Viannia) guyanensis in Brazil. Acta Tropica 93, 4956.
  • Royle JA & Dorazio RM (2008) Hierarchical Modeling and Inference in Ecology: the Analysis of Data on Populations, Metapopulations and Communities (1st edn.) Elsevier, London, UK, pp. 464.
  • Scarpassa VM & Alencar RB (2012) Lutzomyia umbratilis, the main vector of Leishmania guyanensis, represents a novel species complex? PLoS One 7, e37341.
  • Snider H, Lezama-Davila C, Alexander J & Satoskar AR (2009) Sex hormones and modulation of immunity against leishmaniasis. NeuroImmunoModulation 16, 106113.
  • Spiegelhalter D, Thomas A, Best N & Lunn D (2003). WinBUGS Version 14. MRC Biostatistics Unit, Cambridge, UK.
  • Sturtz S, Ligges U & Gelman A (2005) R2WinBUGS: a package for running WinBUGS from R. Journal of Statistical Software 12, 117.
  • Travi BL (2002) Gender is a major determinant of the clinical evolution and immune response in hamsters infected with Leishmania spp. Infection and Immunity 70, 22882296.
  • Wasserberg G, Abramsky Z, Anders G et al. (2002) The ecology of cutaneous leishmaniasis in Nizzana, Israel: infection patterns in the reservoir host, and epidemiological implications. International Journal for Parasitology 32, 133143.
  • Weigle KA, Santrich C, Martinez F et al. (1993) Epidemiology of cutaneous leishmaniasis in Colombia: environmental and behavioral risk factors for infection, clinical manifestations, and pathogenicity. The Journal of Infectious Diseases 168, 709714.