Asymptomatic Leishmania infection in HIV-positive outpatients on antiretroviral therapy in Pernambuco, Brazil

Background Visceral leishmaniasis (VL) in HIV-positive individuals is a global health problem. HIV-Leishmania coinfection worsens prognosis and mortality risk, and HIV-Leishmania coinfected individuals are more susceptible to VL relapses. Early initiation of antiretroviral therapy can protect against Leishmania infection in individuals living in VL-endemic areas, and regular use of antiretrovirals might prevent VL relapses in these individuals. We conducted a cross-sectional study in Petrolina, Brazil, an VL-endemic area, to estimate the prevalence of asymptomatic Leishmania cases among HIV-positive outpatients. Methods We invited any HIV-positive patients, aged ≥ 18-years-old, under antiretroviral therapy, and who were asymptomatic for VL. Patients were tested for Leishmania with enzyme-linked immunosorbent assays (ELISA)-rK39, immunochromatographic test (ICT)-rK39, direct agglutination test (DAT), latex agglutination test (KAtex), and conventional polymerase chain reaction (PCR). HIV-Leishmania coinfection was diagnosed when at least one VL test was positive. Results A total of 483 patients were included. The sample was predominantly composed of single, < 48-years-old, black/pardo, heterosexual males, with fewer than 8 years of schooling. The prevalence of asymptomatic HIV-Leishmania coinfection was 9.11% (44/483). HIV mono-infected and HIV-Leishmania coinfected groups differed statistically significantly in terms of race (p = 0.045), marital status (p = 0.030), and HIV viral load (p = 0.046). Black/pardo patients, married patients, and those with an HIV viral load up to 100,000 copies/ml presented higher odds for HIV-Leishmania coinfection. Conclusions A considerable number of asymptomatic Leishmania cases were observed among HIV-positive individuals in a VL-endemic area. Given the potential impact on transmission and health costs, as well as the impact on these coinfected individuals, studies of asymptomatic Leishmania carriers can be useful for guiding public health policies in VL-endemic areas aiming to control and eliminate the disease.


Introduction
One of the most neglected diseases [1], visceral leishmaniasis (VL) is a parasitic disease that mostly affects tropical and subtropical regions [2,3]. Endemic to more than 60 countries, it is estimated that 50,000-90,000 cases occur annually worldwide [4]. Brazil and six other countries account for about 90% of all cases [3]. In the American continent, the majority of VL-positive individuals reside in Brazil [5]. Most of VL cases in Brazil are notified in the Northeast region, and Pernambuco is a VL-endemic area [6].
VL is considered to be an opportunistic infection for those living with the human immunodeficiency virus (HIV) [7]. HIV-positive individuals who live in VL-endemic areas have an increased risk of Leishmania infection as compared to HIV-negative individuals [8]. When VL occurs in HIV-infected individuals, the prognosis is typically poor and the mortality rate is high [9]. Consequently, several countries have performed studies to estimate the prevalence of Leishmania infection in HIV-positive individuals [10][11][12][13][14][15].
Asymptomatic Leishmania infected individuals, despite the typically low parasite load, might contribute to maintaining the transmission cycle of Leishmania parasites in endemic regions during episodes of increased parasite load and disease relapse [16,17]. In Brazil, due to the high prevalence of HIV-Leishmania coinfection observed in previous studies [10,15,18], and due to the other potential implications (e.g. frequent relapses, mother-to-child transmission [19]), testing for Leishmania should be strongly recommended for all HIV-positive individuals. In addition, once HIV-Leishmania coinfection is diagnosed, early start of highly active antiretroviral therapy (HAART) should be recommended as a protective factor against VL relapses [20]. Moreover, for those HIV-positive individuals who are negative for leishmaniasis and are living in VL-endemic areas, HAART could decrease the risk of Leishmania infection [21]. The Brazilian national HIV/AIDS program assists all persons who live with HIV with HAART, free of charge, and in the same way it provides treatment for all individuals affected by VL. Despite the impact on the public health system, there have been few studies following up these HIV-Leishmania coinfection cases. Most of these studies have focused on hospitalized patients. However, testing for Leishmania in HIV-positive outpatients in VL-endemic areas may be useful for guiding health policies aiming to control and eliminate leishmaniasis, and it could improve treatment and outcomes for those affected by the disease.

PLOS NEGLECTED TROPICAL DISEASES
In this study, we aimed to estimate the prevalence of asymptomatic Leishmania cases in HIV-positive outpatients under continuous use of HAART, in a VL-endemic area, Pernambuco, in Northeast Brazil.

Ethics statement
The study was approved by the research ethics committee of Instituto Aggeu Magalhães, Fiocruz Pernambuco (approval number 51235815.0.0000.5190). All subjects were adults and provided written, informed consent. This study was conducted in accordance with the Declaration of Helsinki.

Study design and sample
We performed a cross-sectional study aiming to estimate the prevalence of HIV-Leishmania coinfection in HIV-positive outpatients from the municipality of Petrolina, in the state of Pernambuco, Brazil. Petrolina is an VL-endemic area from where most VL cases in the state of Pernambuco are reported [22]. The study was conducted in a public HIV outpatient clinic, which serviced about 600 individuals regularly using HAART at the time of this investigation.
The study population included any HIV-positive patients, using HAART, aged 18-yearsold or older. Individuals under treatment for VL and those with current VL symptoms were excluded. Based on a previous prevalence study in Pernambuco [18], the minimum sample size was calculated to be 159 individuals for a 95% confidence interval (Epi Info 7.2.3.0 software, https://www.cdc.gov/epiinfo/index.html). We publicly invited patients at this outpatient clinic to participate in the study. Many of them demonstrated interest in knowing their serological status for Leishmania, as they resided in an endemic area. Due to the high demand, we decided to include anyone who requested to be tested. Therefore, there was a higher number of participants recruited compared to the originally calculated sample size. Each participant received the results of the tests done in the study.

Data collection and laboratory procedures
After the interview and physical examination, peripheral venous blood and urine samples were collected from the patients at the same time when the samples were taken for analyzing the lymphocyte T CD4+ (LTCD4+) count or HIV viral load. The samples were stored, processed, and analyzed at Fiocruz Pernambuco, a referral public research center. Participants were tested for VL with enzyme-linked immunosorbent assays (ELISA)-rK39, immunochromatographic test (ICT)-rK39, direct agglutination test (DAT), latex agglutination test (KAtex) and polymerase chain reaction (PCR) test. All HIV-positive individuals with at least one positive test for VL were considered HIV-Leishmania coinfection cases. Hemogram, biochemistry, LTCD4+ count, and HIV viral load data were obtained from the medical records.
For the ELISA-rK39 assays, the commercial recombinant rK39 antigen was purchased from Rekom Biotech (Granada, Spain) and the assays were essentially carried out as previously described by Scalone et al [23] and Abass et al [24]. For ICT-rK39, we used the OnSite Leishmania IgM/IgG Combo test (CTK Biotech, Inc., Poway, CA, USA) following the manufacturer's instructions.
For DAT, we used a freeze-dried antigen from Biomedical Research (Amsterdam, The Netherlands) and titers of 1:3,200 or higher were considered to indicate a positive test [25]. Leishmania antigen was detected in urine by means of the KAtex kit (Kalon Biological Ltd., Guildford, UK) according to the manufacturer's instructions.

Statistical analysis
Data were entered and stored on spreadsheets using Microsoft Excel Professional Plus 2016 software (Microsoft Corp., Redmond, WA, USA). Data analysis was performed in Stata SE 12.0 software for Windows (StataCorp, College Station, TX, USA).
Frequencies and means with 95% confidence interval of the variables of interest were obtained. We compared an HIV mono-infected and an HIV-Leishmania coinfected group. For binary/categorical variables, the chi-square test was used (significance level p < 0.05). For continuous variables, we used the two-sample Wilcoxon rank-sum (Mann-Whitney) test. Variables that yielded a p-value � 0.1 in univariate analysis were included in a multivariate analysis using logistic regression.

Results
Of 487 HIV-positive individuals who agreed to enroll in the study, three individuals were excluded as they presented typical symptoms of VL and one individual was excluded for not using HAART. The study population was composed mainly of single (45.5%), black/pardo (83.4%), heterosexual (58.2%) men (61.3%) with 8 or fewer years of schooling (48.24%). About 3% reported using intravenous drugs ( Table 1).
The prevalence of HIV-Leishmania coinfection was 9.11% (44/483). The highest positivity by VL test was seen with DAT (3.53%), followed by ELISA-rK39 (2.48%), and PCR kDNA (2.28%) ( Table 2). Two individuals tested positive with DAT and KAtex, and one person tested positive with DAT and rK39. In addition, of the individuals who tested positive for Leishmania spp, three (6.8%) reported having had previous diagnoses of VL, and all of whom reported having been treated.
Regarding general laboratory findings, all results were compatible with the reference standards (Table 3). When the two groups were compared, we observed statistically significant  (Table 5). Black/pardo individuals were at an increased risk (odds ratio, OR: 7.85; p = 0.044) of being HIV-Leishmania co-infected, as compared with white individuals. Marriage/stable unions and a detectable HIV viral load up to 100,000 copies/mL were also associated with an increased risk (OR: 5.12, p = 0.029 and OR: 2.01, p = 0.047, respectively).

Discussion
This study focused on HIV-Leishmania coinfection in outpatients in Pernambuco, which had not been reported previously. The prevalence of asymptomatic HIV-Leishmania coinfection was 9.11% (44/483). There were statistically significant differences between the HIV monoinfected and HIV-Leishmania coinfected groups in term of race (p = 0.045), marital status (p = 0.030), and HIV viral load (p = 0.046). Black/pardo patients, married patients, and those Table 3 with an HIV viral load up to 100,000 copies/mL presented higher odds for HIV-Leishmania coinfection.

. Comparison of general laboratory characteristics between HIV-Leishmania coinfected and HIV mono-infected groups (univariate analysis) among HIVpositive individuals tested for
The prevalence of asymptomatic HIV-Leishmania coinfection in this study (9.1%) was lower than that in a previous study in Pernambuco of hospitalized HIV-positive patients who were tested for VL (16.9%) [18]. This previous study involved three referral hospitals for infectious diseases that serviced the entire state. Despite the higher percentage, compared with the current study, there were fewer cases (35 vs 44) and we tested more individuals (483 vs 207). A similar study in Minas Gerais, Southeastern Brazil, observed a prevalence of asymptomatic Leishmania infection in HIV-positive individuals of 20% [15]. Minas Gerais is a Brazilian state with high VL endemicity, and it would be expected to have a higher prevalence than that observed in our study. In the Metema district of Northwestern Ethiopia, a pilot study in HIVinfected adults identified a prevalence of 12.8% in males and 4.2% in females for asymptomatic HIV-Leishmania coinfection cases, based on the same tests used in the present study [28].
In the VL tests used, we observed a low coincidence of results. Based on previous studies [15,28,29], a large variation in the test results could be expected. Furthermore, the sample consisted of HIV-positive individuals with no suspicion of VL. The highest prevalence rate was observed with the DAT, a test with high sensitivity that is considered to be a good diagnostic tool for immunocompromised individuals [30,31]. The DAT has also showed the highest prevalence rate in a previous prevalence study in Pernambuco with symptomatic patients [18]. We observed a low prevalence with the ICT-rK39 rapid test (1%). Apparently, this is not the best serological screening test for individuals living in L. infantum-endemic regions [15,18,29,30]. Despite the reduced sensitivity of serological tests for VL in HIV-positive individuals [8,30], this type of test should not be excluded, as a positive result should be considered when associated with clinical features [32]. Serological tests are important as screening tests for VL, particularly in VL-endemic areas and for detection of asymptomatic HIV-Leishmania coinfected persons, as they normally have higher LTCD4+ counts. In our study, most participants had an LTCD4+ count exceeding 350 cells/mm 3 , and thus they had potentially similar humoral response as HIV-negative persons.
In terms of epidemiological aspects, we observed statistically significant differences for race and marital status. Indeed, in Brazil race and poverty are strongly connected, and most people affected by neglected tropical diseases live in low income regions [33], such as the region in which this study was performed. The mean LTCD4+ count in HIV-Leishmania co-infected individuals was similar to that observed in the HIV mono-infected group. In the present study, almost three-quarters of all individuals presented with LTCD4+ exceeding 350 cells/ mm 3 . Most participants from the previous study of hospitalized patients in Pernambuco had LTCD4+ counts lower than 200 cells/ mm 3 in both groups (VL-HIV and HIV) [18]. New cohort studies with paired samples may better explain the behavior of Leishmania infection in our population. In the present study, all HIV-positive individuals were on HAART, which explains the higher LTCD4+ count, and which might be a protective factor against developing VL.
We observed that, a detectable viral load, although lower than 100,000 copies/ml, was associated with Leishmania infection. As all patients were asymptomatic, this could indicate an initial HAART failure or irregular use of the treatment. In the previous study of hospitalized VL-HIV coinfected patients, only 16% had an undetectable viral load [18], while in the present study the viral load was undetectable in 63% of the coinfected individuals. Since regular use of HAART usually increases the LTCD4+ count, which, in turn, is a protective factor against VL relapses, campaigns to encourage the regular use of antiretrovirals should be intensified among individuals living with HIV in VL-endemic areas. In Brazil, to date, only secondary prophylaxis for VL is recommended, and the only one marker used to guide this prophylaxis is the LTCD4+ count.
The factors determining maintenance of an asymptomatic VL state have not yet been established. This balance between the parasite infection and the host's immune response, such as, for example, in blood donors or in HIV-AIDS patients, probably extends beyond nutritional status and genetic factors. Due to the increased risk of relapses and the poor prognosis, it is important for HIV-positive persons, particularly those living in VL-endemic areas, to know about a previous Leishmania infections. Identification of new markers or tests that might delimit active disease, suggest cure, and predict relapses is urgent. It may be challenging to distinguish active VL cases from another opportunistic infection, and new and less-invasive markers for VL could help health professionals in making more accurate diagnoses and consequently avoiding unnecessary treatments.

Conclusions
Visceral leishmaniasis remains an important problem in Brazil, particularly in the Northeastern region. We observed a considerable number of asymptomatic Leishmania cases in HIVpositive individuals. Studies focusing on health care of asymptomatic individuals could be useful for public health policies in VL-endemic areas, facilitating monitoring of the progress of leishmaniasis control. In addition, at the individual level, it is important to follow up all these HIV-Leishmania coinfected persons in terms of VL prophylaxis and treatment, and to predict relapses. Consequently, we strongly recommend testing for Leishmania in all HIV-positive individuals in VL-endemic areas.