INTRODUCTION
Visceral leishmaniasis (VL) is a vector-borne parasitic disease caused by several Leishmania spp., and if untreated, death occurs in 90% of cases1. VL is endemic to 70 countries2, and approximately 90% of the VL cases documented on the American continent occur in Brazil3. It is considered an underreported disease, and some studies have demonstrated a high frequency of relapses during follow-up. VL symptoms include fever, weight loss, hepatosplenomegaly, lymphadenopathy, pancytopenia and hypergammaglobulinemia4. Some cases of VL present atypically, as they involve the lungs, pleura, oral mucosa, larynx, esophagus, stomach, small intestine, skin and bone marrow5. There is regional variation in the response to antileishmanial drugs, and recommendations for treatments vary by region6 as a result.
Since 1980, more than 608,230 cases of AIDS have been registered in Brazil, with an incidence of 17.9/100,000 inhabitants in 20107, and unlike VL, the propagation of HIV transmission to regions with lower urbanization rates and to small and medium-sized cities has been observed2. HIV infections increase the risk of developing VL by 100- to 2,320-fold in endemic areas, reduce the likelihood of a therapeutic response and greatly increase the probability of relapse2. However, VL promotes the clinical progression of HIV and the development of AIDS-defining conditions8.
In general, patients with coinfections show very similar clinical features to classical VL9, although the usual clinical features associated with VL (prolonged fever, spleen and liver enlargement, wasting and pancytopenia) are not always present, and atypical presentations occur more often in immunocompromised patients. Furthermore, a clinical diagnosis can also be more difficult to establish due to other concomitant opportunistic diseases10.
Considering that the endemicity of VL in many Brazilian regions is high and that some reports have addressed the clinical features, drug response and outcomes of HIV-related VL9,11–13, in this study, we aimed to report the results of a comparative study (based on data from the Information System for Notifiable 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.
METHODS
Study design and data collection
A descriptive study of confirmed 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 277,720.520km2, 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 Notifiable Diseases (Sistema de Informação de Agravos de Notificação – 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” fields 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 confirmation (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 immunofluorescence, 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” field. Cases were selected for the study using this information. Patients with a confirmed diagnosis of VL and positive HIV serum tests were included, while those who left the “HIV coinfection” field blank or marked it as negative were excluded. Co-infected patients were diagnosed with HIV by ELISA, and the result was confirmed by either indirect immunofluorescence (IIF) or western blot analysis.
Associated causes were considered for patients who, regardless of the efficacy 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 insufficiency, 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 insufficient 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 significance level (α) of 0.05 was considered statistically relevant. The variables with significant results were inserted in the stepwise multivariate model to determine the odds ratios (ORs) with 95% confidence 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).
RESULTS
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, five 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 inhabitants in 2008, to 0.62/100,000 inhabitants in 2009 and, finally, 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, finally, to 28.88/100,000 inhabitants in 2010.
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 (χ2=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 significant (χ2=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%); χ2=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%); χ2=11.328, p=0.01.
Demographic characteristics | VL/HIV patients | VL patients | χ2/t | p | ||
---|---|---|---|---|---|---|
Age in years (Mean, SD)a | 27.9±15.1 | 18.0±14.9 | 1.456 | <0.01 | ||
0-10 | 8 | 24.2 | 900 | 58.6 | 35.873 | <0.01 |
11-17 | 0 | 0.0 | 136 | 8.9 | ||
18-50 | 24 | 72.7 | 402 | 26.2 | ||
>50 | 1 | 3.0 | 98 | 6.4 | ||
Years of scholarship (Mean, SD)b | 5.2±33.4 | 6.9±3.7 | 4,736 | 0.02 | ||
none | 1 | 3.0 | 24 | 1.6 | 10.461 | 0.06 |
primary school | 12 | 36.4 | 418 | 27.2 | ||
secondary school | 7 | 21.2 | 189 | 12.3 | ||
university | 0 | 0.0 | 16 | 1.0 | ||
unknown | 13 | 39.4 | 902 | 57.5 | ||
Genderc | ||||||
male | 26 | 78.8 | 887 | 57.7 | 5.878 | 0.01 |
female | 7 | 21.2 | 649 | 42.3 | ||
Case type | ||||||
new case | 30 | 90.9 | 1,503 | 98.5 | 11.328 | 0.01 |
relapse | 3 | 9.1 | 23 | 1.5 |
VL/HIV: visceral leishmaniasis and human immunodeficiency virus coinfection group; VL: visceral leishmaniasis single infection group. χ2: chi-square. t: T-student test.
aAge is presented as the mean and standard deviation (SD) in the first line and divided into groups by number and proportion below.
bScholarship is presented as the mean and SD in the first line and divided into groups by number and proportion below.
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 concomitant infectious diseases (χ2=6.023, p=0.02) and jaundice (χ2=4.599, p=0.04).
Characteristics | VL/HIV patients | VL patients | χ2 | p | ||
---|---|---|---|---|---|---|
Fever | 31 | 93.9 | 1,488 | 96.9 | 0.902 | 0.34 |
Weakness | 28 | 84.8 | 1,354 | 88.2 | 0.336 | 0.357 |
Weight loss | 25 | 75.8 | 1,119 | 72.9 | 0.138 | 0.44 |
Mucocutaneous pallor | 22 | 66.7 | 1,117 | 72.7 | 0.636 | 0.27 |
Splenomegaly | 20 | 60.6 | 1,082 | 70.4 | 0.227 | 0.38 |
Hepatomegaly | 17 | 51.5 | 969 | 63.1 | 0.952 | 0.21 |
Cough | 17 | 51.5 | 830 | 54.0 | 0.085 | 0.45 |
Concomitant infectious diseases | 8 | 24.2 | 175 | 11.4 | 6.023 | 0.02 |
Edema | 5 | 15.2 | 214 | 13.9 | 0.072 | 0.47 |
Jaundice | 9 | 27.3 | 215 | 14.0 | 4.599 | 0.04 |
Hemorrhage | 0 | 0.0 | 50 | 3.3 | 1.079 | 0.35 |
VL/HIV: visceral leishmaniasis and human immunodeficiency virus coinfection group; VL: visceral leishmaniasis single infection group; χ2: chi-square.
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 (χ2=19.935; p<0.01). However, amphotericin B deoxycholate and liposomal amphotericin B were also widely used for the treatment of VL/HIV.
Characteristics | VL/HIV patients | VL patients | χ2 | p | ||
---|---|---|---|---|---|---|
Drugs | ||||||
pentavalent antimony | 20 | 62.5 | 1,272 | 84.2 | 19.935 | <0.01 |
amphotericin B deoxycholate | 7 | 21.9 | 154 | 10.2 | ||
pentamidine | 0 | 0.0 | 4 | 0.3 | ||
liposomal amphotericin B | 3 | 9.4 | 22 | 1.5 | ||
other | 0 | 0.0 | 18 | 1.2 | ||
not used | 2 | 6.2 | 40 | 2.6 | ||
Outcome | ||||||
cured | 25 | 80.0 | 1,349 | 89.8 | 47.973 | <0.01 |
abandoned | 0 | 0.0 | 7 | 0.5 | ||
deaths from VL | 2 | 6.5 | 71 | 4.7 | ||
deaths from other causes | 4 | 12.9 | 11 | 0.7 | ||
transferred to another state | 0 | 0.0 | 65 | 4.2 |
VL/HIV: visceral leishmaniasis and HIV coinfection group; VL: visceral leishmaniasis single infection group. χ2: chi-square.
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 significantly higher in the VL/HIV group (12.9%) than in the VL group (0.7%); χ2=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, CI=1.14-0.85, p=0.02) and relapse (OR=0.23, CI=0.05-1.05, p=0.05). For more details, see Table 4.
Variable | Uncontrolled | Controlled | ||||
---|---|---|---|---|---|---|
OR | CI (95%) | p | OR | CI (95%) | p | |
Gender Malea | 0.37 | 0.16 – 0.85 | 0.02 | 0.44 | 0.19 – 1.02 | 0.06 |
Ageb | ||||||
0-10 | 1.14 | 0.14 – 0.92 | 0.90 | 0.70 | 0.08 – 6.00 | 0.25 |
11-17 | – | – | – | – | – | – |
18-50 | 0.17 | 0.02 – 1.28 | 0.08 | 5.66 | 0.76 – 42.50 | 0.09 |
>50 | – | – | – | – | – | – |
Scholarshipc | ||||||
none | 1.41 | 0.14 – 14.30 | 0.77 | 2.14 | 0.20 – 22.56 | 0.53 |
primary school | 2.05 | 0.56 – 7.50 | 0.28 | 2.10 | 0.55 – 0.77 | 0.28 |
secondary school | 1.59 | 0.40 – 6.36 | 0.51 | 1.43 | 0.35 – 5.83 | 0.62 |
university | – | – | – | – | – | – |
unknown | 4.92 | 1.32 – 18.50 | 0.02 | 2.15 | 0.47 – 9.31 | 0.31 |
Infectiousd | 0.37 | 0.16 – 0.84 | 0.02 | 0.35 | 0.14-0.85 | 0.02 |
Jaundiced | 0.44 | 0.20 – 0.95 | 0.04 | 0.53 | 0.24 – 1.16 | 0.11 |
Drugse | ||||||
pentavalent antimony | 3.18 | 0.72 – 14.07 | 0.13 | 2.58 | 0.56 – 11.80 | 0.22 |
amphotericin B deoxycholate | 1.10 | 0.22 – 5.50 | 0.91 | 0.79 | 0.15 – 4.10 | 0.78 |
pentamidine | – | – | – | – | – | – |
liposomal amphotericin B | 0.38 | 0.06 – 2.36 | 0.29 | 0.37 | 0.05 – 2.52 | 0.31 |
other | – | – | – | – | – | – |
not used | – | – | – | – | – | – |
Outcomee | ||||||
cured | 3.48 | – | 0.99 | – | – | – |
abandoned | – | – | 0.99 | 1.36 | 1.36 – 1.40 | 0.99 |
deaths from VL | 2.29 | – | 0.98 | – | – | 0.99 |
deaths from other causes | 1.77 | – | 0.99 | – | – | 0.96 |
transferred to another state | – | – | 0.97 | – | – | – |
recidived | 6.53 | 1.86 – 22.95 | <0.01 | 0.231 | 0.050-1.060 | 0.05 |
OR: odds ratio; CI: confidence interval.
aControlled by age and scholarship;
bControlled by gender and scholarship;
cControlled by gender and age;
dControlled by gender, age and scholarship;
eControlled by gender, age, scholarship and case type; -: Parameter set to zero or there were not enough parameters to be calculated. Result close to zero.
DISCUSSION
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 disease14. In Tocantins, located in northern Brazil, new urban VL cases have been reported since 200011. 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 Brazil15, 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.
During the four years of this study (2007-2010), the significant 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” field.
Additionally, of the total number of reports for VL infection, 210 records had incomplete “HIV infection” fields. This reinforces the importance of filling 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 Brazil16,17, with similarly high prevalence rates reported in other countries18–20, and it could reflect increased exposure to sandflies20. In our study, there were two cases of VL in pregnant patients who transmitted the disease congenitally to their infants21. The diagnosis of VL among pregnant women was parasitological due to the presence of Leishmania in the bone marrow aspirate. PCR amplification 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 birth21.
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 200612, the southeast region between 2000 and 20058 and the northeast region between 2001 and 200522. 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 instance23,24 – our study is in accordance with other reports25,7 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 confirms the findings 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 infection26–28. Consequently, it also demonstrates that VL has emerged as a serious opportunistic infection in HIV-infected patients29.
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 Deerfield, IL)25. Although the use of pentavalent antimonials is no longer recommended in HIV-infected patients by most experts in the field, due to their unacceptable toxicity and high rates of treatment failure and mortality30, we used pentavalent antimonials as the first 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 first-line treatment and secondary prophylaxis for patients with VL infected with HIV and the need to include treatment regimens with liposomal amphotericin B31. 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 population2 were followed.
In 2011, the Brazilian Health Ministry (Ministério da Saúde do Brasil) published a new guideline32recommending the use of AmB deoxycholate as the first-choice drug in the treatment of VL/HIV coinfection. However, some studies have demonstrated a high frequency of relapses during follow-up2,33, 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 relapses34. 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 treatments2,7,35,36. 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 outcome37. Lethality in the VL/HIV group prevailed in individuals older than 40 years, confirming 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 identified as a risk factor for death38. Adverse reactions, including renal and liver failure, cardiac abnormalities and pancreatitis, and hepatitis drugs were also death-precipitating factors. The efficacy of the therapeutic approach could not be ascertained, and our results are in accordance with other reports7,34,37. Considering that other opportunistic diseases frequently develop during VL episodes in VL/HIV patients33, 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 regarding the laboratory and clinical profiles of patients and their association with death from VL could assist in clinical management and reduce lethality39–41. Therefore, since 2011, the Brazilian Health Ministry (Ministério da Saúde do Brasil) has recommended the differential diagnosis of opportunistic infections and the use of serological tests for HIV in patients with VL in VL-endemic regions such as Tocantins.
There are some limitations to our study. Although the sample size was large, the number of HIV patients was small, which restricted our multivariate analysis. Because it was only possible to analyze the variables that were included on the notification form (data from the Information System for Notifiable Diseases), other associations could not be ascertained. However, this is the first comparative study on the epidemiologic and clinical features of confirmed cases of VL and VL/HIV infections in Tocantins, Brazil. Thus, our results contribute to increasing our knowledge of VL and VL/HIV infections, which can be used to implement strategies for the prevention and control of these diseases.
In conclusion, the VL and VL/HIV infections that occur in northern Brazil are comparable to those reported by other studies performed in other endemic areas throughout the world, attesting the fact that VL and VL/HIV infections have been emerging with increased frequency in Brazil. Therefore, there is a distinct need to develop and implement methods to identify new forms of treatment and to improve the diagnostic accuracy and therapeutic management of health services that function to control these diseases.