Home SVCC                                                  Area: English - Español - Português

Efficacy of Trypanocidal Therapy for Chronic
Asymptomatic Trypanosoma Cruzi Infection.
A Meta-Analysis

Villar, Juan C.; Villar, Luis A.; Marin-Neto, José A.;
Ebrahim, Shah; Yusuf, Salim

Population Health Research Institute, McMaster University, Hamilton, Canada;
Universidad Autónoma de Bucaramanga, Bucaramanga, Colombia;
Universidad Industrial de Santander, Bucaramanga, Colombia;
Universidad de São Paulo, São Paulo, Brazil
and Bristol University, Bristol, United Kingdom

Background: Recently published randomized clinical trials (RCTs) have challenged the prevailing guideline not to use Trypanocidal therapy (TT) for chronic asymptomatic Trypanosoma cruzi infection based on results from earlier studies.
Objective: To systematically review the existing evidence on the effects of TT for chronic asymptomatic T. cruzi infection
Methods: We included published RCTs testing TT that include asymptomatic T. cruzi-chronically infected individuals and at least one clinical or parasite-related outcome. Search for references was performed in MEDLINE, The Cochrane Controlled trials database, EMBASE, BIREME and the Tropical diseases research division of WHO database, using English, Spanish and Portuguese terms. Abstracts were first screened and then papers appraised for inclusion, evaluation and data extraction by two independent reviewers.
Data synthesis: Five papers (total randomized population=756) met the inclusion criteria. Two RCTs tested Benznidazole in school children and three tested different agents in adults. The pooled odds ratios and their 95%CI (Fixed models) were: Incidence of ECG abnormalities: 0.41 (0.09, 1.85); Negative seroconversion (AT ELISA): 10.91 (6.07, 19.58); Positive xenodiagnosis during the follow up: 0.24 (0.15, 0.37); antibody titres standardized mean difference: -0.54 (-0.84, -0.31). Nitroimidazolic derivatives, but no other agents significantly modify parasite-related outcomes compared to placebo.
Conclusions: TT for chronic asymptomatic T. cruzi infection has been tested in few, small size RCTs, which have failed to provide data on clinical outcomes. Nitroimidazolic derivatives improve parasite-related outcomes, especially in children. Given its importance for public health, more RCTs including clinical endpoints, testing these or newer agents in this population are needed.


   Chagas' heart disease (CHAD), the result of human infection by Trypanosoma cruzi, remains as the leading cause of burden due to parasitic diseases, as it accounts for 2,740,000 disability-adjusted life years lost in the Americas. Virtually all the burden of CHAD comes from the 30% estimate of asymptomatic, chronically infected individuals progressing to the chronic phase (CP). In addition to vector control, a policy aimed to stop the transmission of infection, preventing the incidence of CP among the 16-18 million people already infected is still an unsatisfied public health priority. Although Trypanocidal Therapy (TT) had not been considered effective for chronic T. cruzi infection, more recent observational and experimental studies have suggested that TT might be useful for asymptomatic, T. cruzi-infected population. This fact, and the correlation found between T. cruzi inoculate, amount of parasitic antigens and inflammatory lesions in the heart have re-opened the debate on whether T. cruzi infection is sufficient to cause CHAD or TT is potentially useful to prevent it.

   To provide a synthesis of the best evidence available on the effects of TT in the chronic asymptomatic T. cruzi infection.

   We included published trials that randomly allocated participants with chronic T. cruzi infection without symptomatic Chagas' heart disease to one or more of the trypanocidal drugs or to control treatment with or without placebo. Interventions selected were oral trypanocidal drugs in any dose given for at least 30 days and compared against control or placebo. Included studies had to evaluate at least one of the following outcomes: All-cause mortality, sudden death, incidence of heart failure, side effects of treatment (collectively named here as "clinical outcomes") or reduction in parasite load, reduction in antibody titers to T. cruzi or negative seroconversion (collectively named here as "parasite-related outcomes"). For searching references, we consulted: MEDLINE, EMBASE, LILACS, The Cochrane Controlled Trials Database (CCTD) and Tropical diseases research division (TDR) at the WHO and hand search of references cited in the review articles retrieved. Except in the case of the CCTD, for consistency, we used the same search strategy criteria in all databases, with the equivalent English, Spanish and Portuguese terms for ["Trypanosomiasis" or "Chagas"] and ["Chemotherapy" or "Treatment"] in the title or abstract.

   Abstracts were first screened first for pre-selection and then evaluated for inclusion using the whole report. Agreement for selection was evaluated and differences in opinion were resolved by consensus and a third reviewer opinion if necessary. A Data abstraction and quality assessment forms were filled independently by two reviewers. A descriptive review of the trials was performed. When appropriate, pooled effect estimates and their 95% confidence intervals of the individual clinical endpoints were obtained by computing odds ratios (OR) using the method proposed by Yusuf and Peto and the random models statistical approach. Changes in serology were examined quantitatively by pooling the standardized, weighted mean differences after intervention. Individual differences were computed by subtracting the mean of antibodies after the treatment from the baseline. For each study the variance of the mean differences was inferred using the methodology suggested by Follman. When available, for parasite-related outcomes, both symptomatic and asymptomatic T. cruzi chronically infected patients were included in the analysis.

   1306 abstracts were screened for eligibility. Of these, 87 (6.7%) were considered potentially relevant. 43 of these studies were pre-selected for further appraisal. Of these papers evaluated in full, 5 (11.7%) met the inclusion criteria. General characteristics of the included studies are shown in Table 1 All five studies were phase-three RCTs. Although all papers were reported as randomized, double blind and were placebo-controlled, no paper reported the methods for randomization, the methods for concealing the treatment allocation, or the blinding for outcome assessment. The characteristics of the placebo were not described in any included study either. Only Andrade and Sosa-Estani described the details of the withdrawals during the trial. Andrade performed an intention-to-treat analysis and reported a sample size calculation on beforehand. The total number of randomized participants in the five studies was 756. Control of re-infection was reported in three studies (Sosa-Estani, Andrade and Coura). At least two positive serological tests were required to diagnose chronic T. cruzi infection in all studies. Overall follow-up varied from 12 months (Coura, Gianella) to 48 months (Apt, Sosa-Estani). Andrade reported that 86% of participants finished the protocol. Only one of seventeen dropouts was attributable to side effect of the medication (BZD). Participants who dropped out of this study had no clinical or serological differences to the population remaining in the study. Sosa-Estani reported that 6 of 55 (11%) BZD-treated children dropped out because of side effects (defined in the study protocol as the need to stop study medication, but no need for hospitalization). Apt reported 62/217 (28.6%) dropouts in those who received Itraconazole compared to 2/187 (1.1%) subjects (symptomatic and asymptomatic) allocated to ALLO and attributed that to the length of treatment required with ITRA as the main reason. Gianella reported 10 dropouts, five for each arm of treatment.

   The only outcome other than parasitic-related evaluated was the incidence of ECG changes in those participants in IP of CHAD in two studies (Andrade and Sosa-Estani, n=198). Another report (Apt) showed data on electrocardiogram, but not by allocated treatment. Conversely, all studies reported parasitic-related outcomes before and after the TT, using study-specific definitions. The pooled ORs and their 95% confidence intervals are shown in Figure 1. After pooling the data for these outcomes, heterogeneity among studies was found for xenodiagnosis (p<0.001). The analysis of this outcome by different categories is shown in Table 2 Individually, both BZD and NFTMX significantly decreased the rate of positive xenodianosis but neither ALLOP nor ITRA showed a significant effect compared to placebo. However, within the population randomized to no nitrosamine derivatives (ITRA and ALLOP), Gianella's RCT yielded significantly different results from Apt (p=0.013 and p=0.005 when included participants allocated to both drugs and only to ALLOP respectively). This author performed five sets of xenodiagnosis during one year of follow-up, excluding for the new set those participants who had already had a positive xenodiagnosis (i.e. therapeutic failures). Indeed, at the end of the follow up, all patients treated with placebo and all but one of thirteen patients treated with ALLOP had yielded positive xenodiagnosis. The other authors did not report on what basis the xenodiagnosis was carried out.

   Of two studies testing BZD in children, 7 of 115 treated dropped out of the study because of side effects. However, the criteria for discontinuation were different in both studies. Andrade reported that less than 5% of participants complained of a variety of minor symptoms, but rash and pruritus were reported as higher in those receiving the drug (8/64 with BZD versus 2/65 with Placebo, p=0.094). Sosa-Estani stated that BZD was well tolerated in children and reported no severe side effects, but a 11% incidence of moderate side effects in BZD-treated children, who were dropped out as stated above. Overall, this study estimated the incidence of side effects as less than 20%. The only study testing BZD in adults reported a non-quantified variety of mild side effects (skin reactions, peripheral neuropathy, digestive disturbances), but stated that they were less intense than those seen with NFTMX. Side effects were also reported in studies testing ALLO or ITRA. Apt stratified tolerance into satisfactory, moderate and unsatisfactory and reported no significant differences between placebo and active treatment. A proportion of 10.3%, 9.7% and 6.7% of those receiving ALLO, ITRA and Placebo respectively reported moderate side effects (p=0.412). However, one case of Steven-Johnson syndrome was observed among 155 patients treated with ALLO in this study. Gianella also reported one case of severe skin reaction requiring hospitalization and steroid treatment among 13 patients evaluated who were allocated to ALLO. Three studies (Apt, Andrade and Sosa-Estani) evaluated blood chemistry (liver enzymes, blood cell counts) but did not report abnormalities in children under TT. Apt reported a transient increase in aspartate aminotransferase and alkaline phosphatase in three patients of 185 receiving ITRA.

   Despite major public health importance, until now all forms of TT for chronic asymptomatic T. cruzi infection have been tested in only 756 participants, randomized into five small studies, all of them published after 1995. These trials were not designed to assess clinical outcomes and failed to report key methodological issues. In addition, because of the scarcity in the data reported, none of the pooled outcomes included all randomized participants. Thus, all observations on the effects of these agents for chronic asymptomatic T. cruzi infection should be interpreted in the light of the small number of participants in studies intended to evaluate parasite-related, but not clinical outcomes. Hence, at present, no experimental evidence is available to support the use of TT for improving clinical outcomes in chronic asymptomatic T. cruzi infection.

   The most important finding in this review was that TT, in particular Nitroimidazolic derivatives, improved parasite-related outcomes. The effect of these agents on parasite-related outcomes was demonstrated in this review as both large and consistent across trials, populations and outcomes measured. The strongest effect was found for BZD, an agent that significantly reduced the proportion of positive xenodiagnosis in both children and adults by about 90% and lead to a 11-fold increase of the rate of negative seroconversion of treated children, when serology was tested using the AT ELISA technique. Conversely, studies testing ALLO or ITRA failed to demonstrate a significant effect on these outcomes. In addition to the absence of effect on parasite load, all the severe side effects reported in the trials included in this review occurred in participants treated with these agents (1.1%). Although these results are in favors of the use of BZD in children and either NFTMX or BZD for adults for reducing antibodies or the parasite load respectively, whether this effect will result in clinical benefit remains to be proven. Moreover, although effective for parasite clearance, a variety of side effects were reported in up to 20% of participants treated with nitroimidazolic derivatives. While these results provide a spur to reduce the uncertainty about the clinical effect of these agents, they also make a case for the development of more and safer trypanocidal agents available for clinical practice, as suggested by others.

   The data gathered by this systematic review also make some points about the future of the assessment of the efficacy of TT in asymptomatic, T. cruzi-infected population. First, the need to improve the tools to assess the parasitic-related outcomes in population under TT. Xenodiagnosis, the most sensible test available to isolate parasites was positive in only 70 of 296,l (23.6%) of seropositive participants randomized to placebo. In addition, the effect of TT in terms of serology significantly varied from technique to technique in the same population under similar treatment. Likewise, the extent to which antibodies (in terms of either serological status or antibody titers) and parasitic load in (terms of xenodiagnosis, haemoculture or PCR) correlates remains also unclear and adds complexity to the evaluation of chemotherapy for CHAD. Secondly, the need to improve the risk stratification of asymptomatic participants at baseline. The only "clinical outcome" recorded in the included trials was the incidence of ECG abnormalities in participants with indeterminate phase. Only seven abnormal ECGs out of 198 (3.5%) were found after the follow up. Such a four-year incidence of events should be taken into account in the design of further clinical trials, especially considering that this is a surrogate, not a clinical outcome. Thus, important contributions to the understanding of both the natural history and the clinical assessment of parasite-related outcomes in trials testing TT. Recording clinical outcomes in future trials might be not only more reliable but clinically applicable, provided a sufficient number of events derived form a large number of participants is collected.

Based on the current evidence available, recommendations about TT for asymptomatic T. cruzi infection should be considered as valid only for parasite-related outcomes and based on relative sparse pieces of evidence.



Your questions, contributions and commentaries will be
answered by the authors in the Clinical Cardiology mailing list.
Please fill in the form (in Spanish, Portuguese or English) and press the "Send" button.

or commentary
Name and Surname:
E-Mail address:


2nd Virtual Congress of Cardiology

Dr. Florencio Garófalo
Steering Committee
Dr. Raúl Bretal
Scientific Committee
Dr. Armando Pacher
Technical Committee - CETIFAC

Copyright© 1999-2001 Argentine Federation of Cardiology
All rights reserved


This company contributed to the Congress: