Slow clinical improvement after treatment initiation in Leishmania / HIV coinfected patients

Introduction: In Brazil there is a large area of overlap of visceral leishmaniasis (VL) and HIV infection, which favored a increased incidence of coinfection Leishmania/HIV. Methods: This study evaluated 65 consecutive patients with VL and their clinical response to treatment in two health care settings in Belo Horizonte, Brazil. Results: At baseline, the clinical picture was similar between both groups, although diarrhea and peripheral lymphadenomegaly were more frequent in HIV-infected subjects. HIV-positive patients had lower median blood lymphocyte counts (686/mm3 versus 948/mm3p = 0.004) and lower values of alanine aminotransferase (ALT) (48IU/L versus 75.6IU/L p = 0.016) than HIV-negative patients. HIV-positive status (hazard ratio = 0.423, p = 0.023) and anemia (HR = 0.205, p = 0.002) were independent negative predictors of complete clinical response following antileishmanial treatment initiation. Conclusions: This study reinforces that all patients with VL should be tested for HIV infection, regardless of their clinical picture. This practice would allow early recognition of coinfection with initiation of antiretroviral therapy and, possibly, reduction in treatment failure.

The outbreak of human immunodeficiency virus/acquired immunode ciency syndrome (HIV/ AIDS) during the past 30 years changed the clinical and epidemiological pa ern of visceral leishmaniasis (VL) in many parts of the world.Since the mid-1980's, when the rst case of Leishmania/HIV coinfection was reported, the incidence of coinfection has steadily increased and, at the end of the 90s, it had been reported in 35 countries 1,2 .
e majority of the cases in South America are reported in Brazil, where there is a large endemic area for VL with a strong trend to urbanization.At the same time, HIV infection has spread into back country and rural areas, where most cases of VL are reported 3 .In general, patients with coinfection show clinical features very similar to classical VL, although atypical clinical presentations are more frequent in HIV-positive patients.Typically, coinfected patients present with intense parasitism and have low antibody response.This favors parasitological diagnosis through bone marrow or spleen aspirates and turn serological tests less sensitive than in HIV-negative patients 4,5 .CD4+ (cluster of di erentiation)T lymphocytes counts are lower than 200 cells/mm 3 in 62-92% of coinfected patients and lower than 50 cells/mm 3 in 42% [6][7][8] .
Classically, Leishmania/HIV coinfection is treated with the same drugs prescribed for classical VL (i.e., pentavalent antimonials and amphotericin B, desoxicolate or lipidic formulations), but the optimal dosage and the duration of treatment remains to be determined.Only 17-56% of patients show complete clinical response a er the rst course of therapy 9,10 .Pentavalent antimonials exhibit more toxicity in HIV-infected patients, and 11-28% of patients may need to interrupt therapy due to adverse events 10,11 .
In this study, the clinical and laboratorial features of HIV-negative and HIV-positive patients with visceral leishmaniasis in Brazil were compared.Factors associated with early clinical response a er antileishmanial therapy initiation were analyzed.

METHODS RESULTS
Souza GF et al -Slow improvement of Leishmania/HIV coinfected A retrospective cohort study was conducted using medical records of consecutive patients diagnosed with the rst episode of con rmed VL between January 2000 and December 2005 at two referral centers for leishmaniasis in Belo Horizonte, Brazil: Centro de Pesquisas René Rachou and Hospital Eduardo de Menezes.e following data were gathered from clinical records using standardized forms: age, gender, HIV status, CD4+ Tcell count, clinical signs/ symptoms, blood cells counts and renal and liver function tests.
Patients with clinical suspicion of VL and laboratory con rmation were included.Laboratory-con rmed VL was deemed in the presence of a positive (i.e, ≥ 1:80) polymerase chain reaction.Patients with a past history of VL were excluded.
Human immunode ciency virus infection was diagnosed at the moment of the rst VL episode in 55.6% of the patients.Sixty-one patients initiated treatment with meglumine antimoniate; nine individuals discontinued the drug prematurely due to adverse events [pancreatitis (6) e primary endpoint for this study was the time until the occurrence of complete clinical response a er antileishmanial treatment initiation.Complete clinical response was de ned as the resolution of all clinical signs and symptoms present before treatment initiation.Patients who died before treatment completion were considered nonresponders.The unadjusted association between baseline patients' characteristics and the outcome variable was initially assessed by means of Cox proportional hazards regression analysis.Afterward, a multivariate Cox proportional hazards model was t using all variables significant at p≤0.25 in univariate analysis.The likelihood ratio test with forward and backward stepwise elimination was used for variable selection.Results are reported as hazard ratios (HR) and 95% confidence intervals (CI).The proportional hazards assumption was verified using baseline hazard function plots and the log-minus-log test of proportionality.e value of p<0.05 was considered statistically significant.Statistical Package for the Social Sciences (SPSS) for Windows (version 10.0, SPSS Inc., Chicago, IL) was used for statistical analysis.

Ethical considerations
e study was approved by the Universidade Federal de Minas Gerais and Centro de Pesquisas René Rachou of the Fundação Oswaldo Cruz -Institutional Review Boards.assess the e ect of choosing alternative cuto points and of treating these variables as continuous instead of categorical; c hemoglobin<11g/dl; d leukocytes<4,000/mm 3 ; e lymphocytes<1,000/mm 3; f eosinophils<50/mm 3 ; g platelets<150,000/mm 3 .
treatment with amphotericin B, but this drug was switched to meglumine antimoniate in one patient due to acute renal failure a er total recovery of renal function.Median follow-up was 20 days (IQR = 13-28) for HIV-positive patients and 21 days (IQR = 13.5-30) for HIV-negative patients.Eight patients (six HIV-positive and two HIV-negative) died at a median follow-up of 13.5 days.Univariate associations between patients' baseline factors and time to complete clinical cure following antileishmanial treatment initiation are described in Table 2.A er adjusting for confounders, HIV-positive status (HR= 0.423, 95%CI= 0.202-0.888;p= 0.023) and anemia (HR= 0.205, 95%CI= 0.074-0.568;p = 0.002) remained as independent (negative) predictors of complete clinical cure.

DISCUSSION
e clinical presentation of VL in HIV-infected patients did not di er from that classically described for VL in HIV-negative subjects, as most patients in both groups showed the classical triad of fever, splenomegaly and citopenia.In fact, clinical manifestations of VL caused by L. infantum and L. donovani in HIV-infected patients are not signi cantly di erent from those in non-HIVinfected patients 1,6,7,8 .Classical features commonly present are fever, weight loss, hepatosplenomegaly, and pancytopenia.Some authors, however, reported atypical clinical manifestations in coinfected patients, lacking visceromegalies or fever 9,12,13 .Visceral leishmaniasis is a broad clinical spectrum disease, so atypical forms of presentation are relatively frequent.
In this study, higher frequency of diarrhea in HIV positive patients, which is not a clinical manifestation commonly reported in patients with VL.Diarrhea that could not be explained by other condition was found in 7.7% of coinfected patients 14 .coinfected patients treated with miltefosine (failure occurred in 18% of HIV-coinfected patients, compared with 5% of non-HIV-infected patients) 17 .High frequency of adverse events due to antileishmanial treatment has also been reported.Adverse e ects were observed in 56% of coinfected patients when antimonials were used without the upper limit of dosage, and in 28% of patients, treatment with meglumine antimoniate was permanently discontinued due to serious adverse e ects 10,11 .In a study from Ethiopia, 1/3 of coinfected patients died due to treatment-related pancreatitis 17 .In our study, 9 (14.8%)out of 61 individuals who started treatment with antimonials discontinued the drug due to adverse events.In this study, anemia was also a predictor of worse clinical outcome a er treatment.It is possible that lower hemoglobin levels may serve as a proxy for the severity of the patients' illness or advanced immunossupression.In HIV patients coinfected with Leishmania spp.thrombocytopenia was identi ed as independent predict or of mortality, with each reduction of 10,000 platelets/mm 3 increasing the mortality risk ratio by 6% 14 .e platelet count was not associated with increased mortality in the present study (data not shown).This study has a number of limitations.Registration bias precluded the analysis of a number of variables originally included in our study protocol (e.g., quanti cation of weight loss).e extent to which early clinical response predicts long-term response or even parasitological cure cannot be ascertained by our data.e small In coinfected patients, intestinal infections and concomitant use of antimicrobials and/or antiretrovirals may explain the higher frequency of diarrhea.
In this group, low total lymphocyte count was associated with HIV infection.CD4+ T cell count was lower than 200 cells/mm 3 in 75% of coinfected patients.e low lymphocytes counts observed in HIV-positive patients, due to depletion of CD4+ T cells seen in HIV infection, was largely reported in coinfected patients in the Mediterranean basin 9,10,11,15 .At the time of clinical presentation of VL, 79 to 100% of HIV-positive patients coinfected with L. infantum have CD4+ cell counts lower than 200 cells/mm 3 .Low CD4+ T cell counts may prevent a full clinical recovery and is associated with multiple relapses 8,11 .
Human immunode ciency virus infection was an independent predictor of poorer clinical response to treatment in this studygroup.Using the treatment regimens recommended by the Brazilian Ministry of Health 16 most HIV-negative patients with VL show clinical improvement, very often before the end of therapy.In coinfected patients, however, a number of studies reported higher treatment failure and higher lethality rates, regardless of the treatment schedule used 10,11,[13][14][15] .Lower response rate in HIV-infected compared to HIV-negative individuals (54.8% versus 89.7%) was reported 11 .More recently, a high proportion of treatment failure was observed in
, severe arthralgia (2), cardiotoxicity (1)] and received full courses of amphotericin B. Four patients initiated (IQR; interquartile range).Continuous variables were categorized based on clinically relevant criteria.Univariate comparisons of clinical and laboratory characteristics between HIV-positive and HIV-negative patients at baseline (i.e, before initiation of antileishmanial treatment) were conducted by means of Student's t test or Mann-Whitney's U test for continuous variables, and by means of Pearson's qui-square test or Fisher's exact test for categorical variables, as appropriate.