HIV / AIDS-related visceral leishmaniasis : a clinical and epidemiological description of visceral leishmaniasis in northern Brazil

introduction: This study aimed to describe the main features of visceral leishmaniasis (VL), both related to and independent of human immunodefi ciency virus (HIV) infection, in patients who were registered in Tocantins, Brazil. Methods: Data from 1,779 new patients with VL, 33 of whom were also infected with HIV, were reviewed. Results: The incidence of VL/HIV coinfection increased from 0.32/100,000 inhabitants in 2007 to 1.08/100,000 inhabitants in 2010. VL occurred predominantly in children aged 10 years or younger, while VL/HIV was more common in patients aged between 18 and 50 years. There were more male patients in the VL/HIV group than in the VL group. Relapse rates were also considerably higher in the VL/HIV (9.1%) group than in the VL group (1.5%). Despite a similar clinical presentation, VL/HIV patients exhibited a higher proportion (24.2%) of concomitant infectious diseases and jaundice. Pentavalent antimonials were used for the initial treatment of VL and VL/HIV infections. However, amphotericin B deoxycholate and liposomal amphotericin B were also widely used in the treatment of VL/HIV coinfection. The mortality rate was higher in the VL/HIV coinfection group (19.4%) than in the VL group (5.4%). Furthermore, the mortality rate due to other causes was signifi cantly higher in the VL/HIV group (12.9%) than in the VL group (0.7%). Conclusions: The study showed that the incidence, clinical characteristics and outcomes among the VL and VL/HIV patients in this state are similar to those from other endemic regions, indicating that both infections are emerging with increasing frequency in Brazil.


MeTHODs
in this study, we aimed to report the results of a comparative study (based on data from the Information System for Notifi able Diseases) on the epidemiology, clinical presentation, drugs used and outcome of both VL and VL/HIV patients in Tocantins, the newest Brazilian state, between 2007 and 2010.

Study design and data collection
A descriptive study of confi rmed cases with VL and VL/HIV infections was conducted between 2007 and 2010 in Tocantins, Brazil.Tocantins is the newest state of the country (created in 1988) and is located in the central part of Northern Brazil.The 2010 census estimated that the state had a population of 1,383,445 inhabitants, distributed over 277,720.520km 2 , and a population density of 4.98 inhabitants per square kilometer; moreover, it comprises 139 municipalities.
The study was based on the data from the Information System for Notifi able Diseases (Sistema de Informação de Agravos de Notifi cação -SINAN).The patients were divided into two groups: the VL group, consisting of patients without HIV infection, and the VL/HIV coinfection group, consisting of patients with VL and HIV (with and without AIDS).An epidemiological description was produced from both groups.However, reports in which the "HIV coinfection", "drug initially administered" and "developments in the case" fi elds were left blank or incomplete were excluded from both groups.In this database, the diagnosis of VL was made by the presence of symptoms, a blood test for confi rmation (immunological methods) and a parasitological diagnosis.

inclusion and exclusion criteria
The study included patients diagnosed with VL due to the presence of amastigotes in bone marrow smears or indirect immunofl uorescence, which were considered positive with results higher than 1:80, and clinical manifestations that included hepatomegaly or splenomegaly, anemia and thrombocytopenia.The study excluded patients undergoing prior treatment with antileishmanial drugs.
The "LV Investigation Form" included an "HIV coinfection" fi eld.Cases were selected for the study using this information.Patients with a confi rmed diagnosis of VL and positive HIV serum tests were included, while those who left the "HIV coinfection" fi eld blank or marked it as negative were excluded.Co-infected patients were diagnosed with HIV by ELISA, and the result was confi rmed by either indirect immunofl uorescence (IIF) or western blot analysis.
Associated causes were considered for patients who, regardless of the effi cacy of the treatment, eventually died from cytopenia or thrombocytopenia, and data were discarded in cases where the patient underwent an inadequate treatment course (e.g., someone who should have been treated with amphotericin B because they had cardiovascular involvement but was treated with Glucantime instead), suffered adverse reactions to the medication, such as renal or hepatic insuffi ciency, or developed associated infections such as pneumonia or severe sepsis.
Subjects from 0-65 years of age of either sex were included in the study.Females who were pregnant or lactating were not excluded.Exclusion criteria included forms with insuffi cient information regarding the initial drug used in the treatment, HIV coinfection status and the progression of the case.

data analysis
Data were codified and analyzed using the Statistical Package for the Social Sciences (SPSS for Windows, version 17.0).The variables used for comparative analyses were gender, age, scholarship, clinical manifestations, initial drug use and development of the disease.All variables are presented as numbers and proportions, and age and scholarship are also presented as means and standard deviations (SDs).The indicators were considered to be the following: the lethality of the VL and VL/HIV infections; the mortality rate directly related to VL and linked to other causes in both groups; the relapse rate in each group; and the annual rate of VL/HIV coinfection.Only new cases were used for the analysis, except for the analysis of relapses.The chi-squared test was used to calculate the difference in the nominal variables between the two groups.The differences in the means between the two groups due to age and scholarship were calculated using Student's t-test.A signifi cance level (α) of 0.05 was considered statistically relevant.The variables with signifi cant results were inserted in the stepwise multivariate model to determine the odds ratios (ORs) with 95% confi dence intervals (CIs) for the association between the HIV/VL coinfection.We used HIV positivity as a dependent variable, each factor as an independent variable and age, gender and scholarship as covariates.Alpha was set at p < 0.05.

Ethical considerations
The study was approved by the Committee of Ethics in Human Research (Comitê de Ética em Pesquisas com Seres Humanos) of the Federal University of Tocantins (Universidade Federal do Tocantins -UFT) (number 030/2009) and was conducted in accordance with the ethical guidelines of the Declaration of Helsinki (created in 1964 and revised in 2002).Permission to conduct the study was obtained from the Health Department of the State of Tocantins (Secretaria da Saúde do Estado do Tocantins -SESAU).
According to SINAN, 1,779 VL patients were registered in Tocantins between 2007 and 2010.Of those, 33 (2.1%) were coinfected with HIV.There were no data for VL/HIV coinfection status, the initial drug used or developments in the case in 210 reports; thus, these patients were excluded from the comparative analysis.During the same period, there were 23 relapses in the VL group and 3 relapses in the VL/HIV group.Among these 26 cases, fi ve were excluded due to the lack of data regarding the initial drug used for treatment and developments in the case.Figure 1 depicts the incidence rate per year for LV/HIV and LV-only cases.The incidence of LV/HIV coinfections rose from 0.32/100,000 inhabitants in 2007, to 0.47/100,000 FIGURE 1 -Incidence of VL/HIV-confi rmed cases per 100,000 habitants per year in Tocantins, Brazil, from 2007 to 2010.VL/HIV: visceral leishmaniasis and human immunodefi ciency virus coinfection group; VL: visceral leishmaniasis single infection group; *Incidence: number of new cases per 100,000 habitants per year.Details regarding the demographic characteristics of the VL and VL/HIV groups are provided in table 1.The mean age of the VL/HIV group (27.9±15.1)was higher than that of the VL group (18.0±14.9);t=4.134, p<0.01.The more frequent age range for VL/HIV patients was 18-50 years, which is considered the economically active population in Brazil.In the VL group, there were more patients aged 10 years and younger (χ 2 =35.873, p<0.01).The mean scholarship was higher in the VL group (6.9±3.7)than in the VL/HIV group (5.2±3.4), but the proportion between study stages was not signifi cant (χ 2 =10.461, p=0.06).The proportion of male patients was higher in the LV/HIV group (78.8%) than in the VL group (57.5%); χ 2 =5.878, p=0.01.The relapse rate was also considerably higher in the VL/HIV group (9.1%) than in the VL group (1.5%); χ 2 =11.328, p=0.01.
Despite the similar clinical presentation among the VL patients with or without HIV coinfection (table 2), a higher proportion (24.2%) of patients in the VL/HIV group had The drugs most frequently used in the treatment of patients in both groups are shown in table 3. The drug of choice for the initial treatment of VL (84.2%) and VL/HIV coinfection (62.5%) was pentavalent antimonials (χ 2 =19.935; p<0.01).However, amphotericin B deoxycholate and liposomal amphotericin B were also widely used for the treatment of VL/HIV.
The mortality rate (table 3) was higher in the VL/HIV coinfection group (19.4%) than in the VL group (5.4%).The VL-related mortality rate was 4.7%, and the VL/HIV-related mortality rate was 6.5%.Furthermore, the mortality attributed to other causes was signifi cantly higher in the VL/HIV group (12.9%) than in the VL group (0.7%); χ 2 =47.973, p<0.01.
For the multiple logistic regression analysis, we selected HIV positivity as a dependent variable and the following factors as independent variables: gender, age, scholarship, concomitant infections, jaundice, drugs, outcome and the occurrence of relapse.Gender, scholarship and age were selected as covariates.It was observed that being HIV positive was a risk factor for concomitant infections (OR=0.35  The transmission of VL has gradually spread to various Brazilian regions.It is believed that rural-urban migration, agroindustrial and man-made projects such as dams, irrigation systems and wells as well as deforestation contribute to the dissemination of this disease 14 .In Tocantins, located in northern Brazil, new urban VL cases have been reported since 2000 11 .This suggests that environmental changes such as the destruction of the cerrado vegetation, the construction of new cities and the rapid and intense migration of rural populations to urban peripheries could have played a role in the transmission of VL.HIV/AIDS is also endemic in Brazil 15 , and an increase in the number of cases of VL/HIV coinfections has been observed since the early 1990s.The number of cases is projected to increase due to the geographical overlap of the two infections as a result of the urbanization of leishmaniasis and the internalization of HIV infection.In VL-endemic regions, HIV infection increases the risk of developing clinically evident leishmaniasis.Leishmaniasis is endemic to Tocantins and is in the process of expanding both geographically and in magnitude.However, information on coinfection by VL and HIV in the northern region of Brazil is still scarce.Thus, our objective was to describe the main epidemiological and clinical features of VL/HIV-infected and VL-infected patients in Tocantins.

DIscUssIOn
During the four years of this study (2007-2010), the signifi cant number of VL cases when compared to the number of VL/HIV cases could have been overestimated because no HIV test was performed preceding the completion of the "LV Investigation Form", which resulted in only 33 patients completing the "HIV infection" fi eld.
Additionally, of the total number of reports for VL infection, 210 records had incomplete "HIV infection" fi elds.This reinforces the importance of fi lling out the form in full, as the information contained in it is extremely important for the planning and execution of surveillance and disease control.
The highest prevalence rate (47.7%) of VL infection in children ≤ 5 years old was reported in other region of Brazil 16,17 , with similarly high prevalence rates reported in other countries [18][19][20] , and it could refl ect increased exposure to sandfl ies 20 .In our study, there were two cases of VL in pregnant patients who transmitted the disease congenitally to their infants 21 .The diagnosis of VL among pregnant women was parasitological due to the presence of Leishmania in the bone marrow aspirate.PCR amplifi cation of parasite kDNA from newborn bone marrow samples suggested that the leishmaniasis was transmitted vertically because the newborns developed signs of the disease shortly after birth 21 .
VL/HIV coinfections were most commonly observed in the age range of 18-50 years (72.7%), and the proportion of male patients was higher in the LV/HIV group (78.8%) than in the VL group (57.5%).The same scenario was found in the central-west region of Brazil between 2000 and 2006 12 , the southeast region between 2000 and 2005 8 and the northeast region between 2001 and 2005 22 .This shows that cases of coinfection have spread throughout Brazil.
Although it has been proposed that VL/HIV coinfections show different clinical manifestations -a lack of visceromegaly or fever, for instance 23,24 -our study is in accordance with other reports 25,7 that show that the initial clinical presentation of VL/HIV patients is similar to that of HIV-negative individuals, with fever, splenomegaly and hepatomegaly associated with weakness, weight loss, mucocutaneous pallor, airway infections and/or diarrhea, edema and jaundice.The only difference in symptoms between the two groups was an increased prevalence of infection in HIV-infected patients (25.9% versus 10.7%).However, our study confi rms the fi ndings of previous reports showing that HIV-induced immunosuppression impairs the body's protective mechanisms against intracellular parasites, such as Leishmania species.Additionally, VL could induce intracellular HIV replication, thus speeding up the clinical course of the HIV infection [26][27][28] .Consequently, it also demonstrates that VL has emerged as a serious opportunistic infection in HIV-infected patients 29 .
Rev Soc Bras Med Trop 47(1):38-46, Jan-Feb, 2014 The treatment for VL in HIV-infected patients is limited to pentavalent antimonials (sodium stibogluconate (SSG) and meglumine antimoniate) and amphotericin B (AmB, typically in the liposomal formulation AmBisome, Gilead Sciences Inc., Astellas Pharma, North Deerfi eld, IL) 25 .Although the use of pentavalent antimonials is no longer recommended in HIV-infected patients by most experts in the fi eld, due to their unacceptable toxicity and high rates of treatment failure and mortality 30 , we used pentavalent antimonials as the fi rst choice of treatment for VL and VL/HIV infections in our study.In 2005, the World Health Organization stressed the need for multicenter trials of fi rst-line treatment and secondary prophylaxis for patients with VL infected with HIV and the need to include treatment regimens with liposomal amphotericin B 31 .Nevertheless, the number of patients who received amphotericin B deoxycholate and liposomal amphotericin B was much higher in the VL/HIV group (31.3%) than in the VL group (11.7%).This indicates that the WHO recommendations for liposomal AmB as the best treatment choice for coinfected patients because of its success rate in the HIV-positive population 2 were followed.
In 2011, the Brazilian Health Ministry (Ministério da Saúde do Brasil) published a new guideline 32 recommending the use of AmB deoxycholate as the fi rst-choice drug in the treatment of VL/HIV coinfection.However, some studies have demonstrated a high frequency of relapses during follow-up 2,33 , further supporting the need for combined therapies.
Furthermore, a VL/HIV-coinfected patient's severely immunocompromised state could hinder the destruction of the parasite and contribute to an increased predisposition to frequent relapses 34 .This tendency toward relapses was observed in our studies (data not shown).
The immune deficiency caused by HIV facilitates the multiplication of the Leishmania parasite and further reduces the rates of cure through conventional treatments 2,7,35,36 .The lethality among the VL/HIV patients in this study was substantially higher than that in the VL patients (data not shown).Generally, infectious complications and bleeding are the main risk factors for death in children with LV.Furthermore, the involvement of the liver in children, which is generally not pronounced and is reversible after treatment, could be quite severe, causing a fatal outcome 37 .Lethality in the VL/HIV group prevailed in individuals older than 40 years, confi rming the observation that adults older than 45 years have a higher risk of dying, possibly due to immune decline at this age.The progression of the disease symptoms with a consequent delay in diagnosis and treatment of patients with VL or VL/HIV has been identifi ed as a risk factor for death 38 .Adverse reactions, including renal and liver failure, cardiac abnormalities and pancreatitis, and hepatitis drugs were also death-precipitating factors.The effi cacy of the therapeutic approach could not be ascertained, and our results are in accordance with other reports 7,34,37 .Considering that other opportunistic diseases frequently develop during VL episodes in VL/HIV patients 33 , there was a highly accentuated death rate due to other causes in the VL/HIV group (12.9%) in comparison to the VL group (0.7%).It has been suggested that knowledge , to 0.62/100,000 inhabitants in 2009 and, fi nally, to 1.08/100,000 inhabitants in 2010.The incidence of LV alone decreased from 30.96/100,000 inhabitants in 2007, to 31.08/100,000 inhabitants in 2008, to 24.92/100,000 inhabitants in 2009 and, fi nally, to 28.88/100,000 inhabitants in 2010.
VL/HIV: visceral leishmaniasis and human immunodefi ciency virus coinfection group; VL: visceral leishmaniasis single infection group.χ 2 : chi-square.t: T-student test.a Age is presented as the mean and standard deviation (SD) in the fi rst line and divided into groups by number and proportion below.b Scholarship is presented as the mean and SD in the fi rst line and divided into groups by number and proportion below.c Gender is presented as a number and proportion.concomitant

TABLE 3 -
Drugs used in the treatment of patients with visceral leishmaniasis and patients with visceral leishmaniasis associated HIV and their clinical outcomes in Tocantins, Brazil from 2007 to 2010.

TABLE 4 -
Results the multiple logistic regression analysis.Controlled by age and scholarship; b Controlled by gender and scholarship; c Controlled by gender and age; d Controlled by gender, age and scholarship; e Controlled by gender, age, scholarship and case type; -: Parameter set to zero or there were not enough parameters to be calculated.Result close to zero.
OR: odds ratio; CI: confi dence interval.a