Use of benznidazole to treat chronic Chagas disease: An updated systematic review with a meta-analysis

Background Approximately 6 million people worldwide are affected by Chagas disease, with many in the chronic phase of the disease (CCD). It is crucial to evaluate the effectiveness of benznidazole for CCD treatment. Methods/Principal findings We updated a meta-analysis published in 2009 up to February 2021, including controlled trials (RCT) and prospective observational studies (OBS) that compared benznidazole vs placebo/no-treatment (P/nT). Main outcomes evaluated were clinical progression (CP) and seroreversion with subgroup analysis performed according to study design and participants’ age. Parasitological response and safety were also described. We identified 879 articles and selected nine for inclusion (corresponding to eight studies). After adding the nine articles from the previous meta-analysis, 17 studies were analyzed corresponding to 6640 patients. The odds ratio (OR) for seroreversion in children treated with benznidazole vs P/nT was 38.3 (95%CI: 10.7–137) and 34.9 (95%CI: 1.96–624.09) in RCT and OBS, respectively. In adults the OR for seroreversion in OBS was 17.1 (95%CI: 2.3–129.1). CP was only evaluated in adults, where benznidazole did not demonstrate a beneficial effect: OR 0.93 (95%CI: 0.8–1.1) and OR 0.49 (95%CI:0.2–1.2) for RCT and OBS, respectively. Most outcomes were deemed to have a low level of certainty, except for the beneficial effect in children and the low efficacy in adults (moderate certainty). Conclusions Benznidazole should be recommended for CCD in children, though this is only based on serological response and a moderate grade of evidence, while in adults benznidazole efficacy remains uncertain. More data on clinical efficacy of benznidazole in CCD is needed in both children and adults.


Introduction
Chagas disease, a parasitic infection caused by Trypanosoma cruzi, affects at least 6 million people worldwide [1]. It is endemic from the south of the United States to the south of Argentina and Chile, where people are mainly infected by several species of blood-sucking triatomine insects [2]. Chagas disease can be also transmitted through blood and blood products, transplantation, laboratory accidents, oral transmission, and vertical transmission during pregnancy or childbirth [3][4][5][6]. Thus, it can be transmitted outside of endemic areas.
Currently, 70.2 million people are at risk of infection and there are still 38,593 new infections and 12,000 deaths per year in endemic areas [1]. As a consequence of migration movements, Chagas disease has spread beyond its traditional boundaries affecting other territories, especially the United States, where 347,000 persons are estimated to be infected [7,8] and some European countries where an estimated 123,078 infected persons reside (86,618 of them in Spain) [9,10]. Chagas disease is also a concern in other countries, such as Canada (with an estimated 156,960 number of migrants from which 3.5% [5,553 persons] were estimate to be infected in 2006) and, to a lesser extent, it is also present in African, Eastern Mediterranean and Western Pacific countries [11,12].
Treatment is always indicated in the acute phase of the disease, as it has been demonstrated to improve clinical outcomes in addition to achieving parasitological clearance and seroreversion in most patients (76% to 100%) [13][14][15]. Without treatment, patients will develop chronic Chagas disease, and 30-40% of them will develop visceral involvement in about 20 years, with cardiomyopathy being the most common complication (14-45%) followed by megacolon and/or megaesophagus (10-20%) [16][17][18][19]. Treatment effectiveness seems to be lower in chronic Chagas disease in adult patients (2-40%) followed-up and treated in endemic areas [15,20]. In patients with established cardiomyopathy, chronic Chagas disease does not significantly reduce clinical progression of the disease [18,21].
Currently there are only two available drugs for Chagas disease treatment: benznidazole and nifurtimox. Both drugs dates from the 1960s-70s and are frequently associated with treatment discontinuation secondary to adverse effects [22][23][24]. Not surprisingly, it was not until August 2017 and August 2020, respectively, that the FDA has approved their use for the treatment of Chagas disease, both for pediatric use and off-label use in adults, [25,26] while in Europe is not formally approved. However, benznidazole has a better safety profile and tolerance and it is generally the first choice drug. It is administered orally in daily doses of 5 mg/kg per day in adults and 7.5 mg/kg in children, in two or three daily doses for 60 days, although in some settings only 30 days are used for adult patients [15,18,27].
The potential benefits of benznidazole in chronic Chagas disease, compared with placebo or no treatment (P/nT), was previously analysed in a systematic review and meta-analysis published in 2009 [28]. Most of the evidence in that analysis came from observational studies, especially in adult patients. Since then, there have been several well-designed, randomized, controlled trials of the treatment of chronic Chagas disease comparing benznidazole mostly with antifungal agents [21,[29][30][31].
The availability of high quality randomized controlled trial data provides an opportunity to update the previous systematic review with evidence published during the last 11 years in order to evaluate the balance between risks and benefits of benznidazole at standard dose as compared with placebo/no-treatment in chronic Chagas disease [28]. Specifically, we aimed to assess the effectiveness of benznidazole to prevent clinical progression (cardiovascular, digestive or other clinical event or death) and to promote seroreversion after treatment.

Materials and methods
We performed a systematic literature review, updating the previous meta-analysis performed by Pérez-Molina et al [28]. We followed the PRISMA statement to report our findings [32]. The protocol of the study is available at: DOI 10.17605/OSF.IO/TY836.

Eligibility criteria
We included randomized controlled trials and prospective comparative observational studies using benznidazole at the standard dose compared with placebo/no-treatment for the treatment of chronic Chagas disease. We included studies with patients treated for at least 30 days or a daily fixed dose of 300 mg of benznidazole. Studies involving drugs other than benznidazole were included provided they had a benznidazole and a placebo arm. All studies using benznidazole without a placebo/no-treatment arm were excluded.

Patient population
We included information from chronic Chagas disease patients (both adults and children) with and without visceral involvement, with a confirmed Chagas disease infection by at least two serological tests or by parasitological techniques. Studies focused on acute infections, pregnant or breast-feeding women, or immunocompromised patients (haematological malignancies, cancer, bone marrow and solid organ transplants, hypogammaglobulinemia, or HIV infection) were excluded.

Literature search, data collection, and reporting of results
We searched Medline, EMBASE, LILACS, and Cochrane Central Register of Controlled Trials (CENTRAL) for articles published between January 2008 and February 17, 2021 with no language restrictions. A medical librarian (N.A-D) conducted the search strategies using a combination of keywords and standardized index terms (S1 Text).
We also searched on ClinicalTrials.gov to identify unpublished clinical trials. We performed a secondary search by consulting the references of the articles included and the abstracts of the most important scientific meetings on the field (CISTM, ECMID, ECTMIH, Taller de Chagas, Chagas Disease Clinical Research Platform). All scientific information (abstract and others) not published during the 3 years following their presentation at scientific congresses were excluded.
Two reviewers independently identified eligible studies (BC and CC-A) by applying the inclusion and exclusion criteria using an eligibility form. A third investigator (JAP-M) resolved disagreements. When there were some unclear aspects of the studies, we contacted the authors for clarification. BC and CC-A performed data extraction using a pre-specified data collection form. For each study arm, we collected the primary and secondary endpoints, if available: number of patients with clinical progression, seroreversion, and parasitological clearance (PCR, blood culture, and xenodiagnosis). We also collected variables related to study design and rates of treatment discontinuation.
Two reviewers (BC and CC-A) independently assessed the risk of bias for each study using the Cochrane Risk of Bias Tool for Randomized Controlled Trials [33] and the Newcastle-Ottawa scale for prospective cohort studies [34]. Reviewers judged each criterion for bias risk and resolved any disagreements in discussion with a third reviewer (JAP-M).
We summarized our findings using Summary of Findings (SoF) tables [35] and include, in a per outcome basis, ratings of certainty of evidence using the GRADE approach [36,37]. We analyzed separately the evidence from randomized controlled trials and observational comparative studies.

Statistical analysis
We pooled the effect estimates of individual studies for the benznidazole and placebo/no-treatment groups. The meta-analysis was based on a random effect model. We used odds ratios (OR) as the measure for the pooled effect. Missing patients were considered as benznidazole treatment failure for further analysis. We performed a sensitivity analysis of the primary and secondary endpoints including only those studies with more than 60 months of follow-up taking into account the importance of the time elapsed since treatment for the cure criteria.
Heterogeneity was evaluated using the I 2 statistic and Chi 2 test. We regarded heterogeneity as substantial if the I 2 statistic was > 50% or there was a low p-value (<0.10) in the Chi 2 test for heterogeneity. If more than ten studies were available in a meta-analysis, we performed funnel plots to assess the risk of publication bias or small studies effect and complemented them with additional tests of funnel asymmetry (Egger and Begg's tests).
We anticipated the following sources of heterogeneity to drive subgroup analyses for the primary outcomes: study design (observational versus randomized clinical trial), presence of visceral involvement versus indeterminate form, year for study initiation, participant age (children versus adults), and the country where patients were treated (endemic versus nonendemic region). The statistical significance of the between groups differences were obtained using meta-regression analyses weighted by the inverse of the standard error of the logarithm of the OR. The dependent variable in this analysis was the logarithm of the OR and independent variables, one at each time, were the a priori defined sources of heterogeneity.
All analyses were conducted using Stata version 16 statistical software (StataCorp, College Station, TX, USA).

Results
We screened 879 articles for eligibility. After reading titles and abstracts, we excluded 626 references as they were not performed in humans, were performed on patients with acute or congenital infection, were based on other drugs than benznidazole, did not have placebo or no treatment arm, or examined diagnostic tools in Chagas disease. After a full-text reading of the remaining 38 articles, we finally selected nine articles reporting on eight studies (Fig 1).
Therefore, in this update we included data from 17 studies, including nine from the original systematic review [28] ( Table 1).
The final set of studies included eight randomized controlled trials and nine prospective observational studies including information from 6640 patients (2938 patients received benznidazole and 3702 patients taking placebo/no-treatment). Three studies were performed in children 1-14 years old, 13 studies were performed in adults, and one study did not specify the participants' age [49]. Table 1 shows general characteristics of the studies included in the review. Most of the studies were performed in Argentina (n = 7/17; 41.2%), followed by Brazil (n = 4/17; 23.5%) and Bolivia or in Spain, with more than 90% of patients of Bolivian origin (n = 3/17; 17.6%). Information about participants' sex was available in 13 studies with an average of 54.5% (3558/ 6530) female. Follow-up period and Chagas disease clinical form (cardiomyopathy, digestive involvement, neuropathy, or mixed clinical form) were available in 16 studies. Half of the studies (8/16) included only patients with the indeterminate clinical form, one (6.3%) only with cardiomyopathy and, in seven studies (43.8%), patients with and without visceral involvement. The follow-up period ranged from 3 to 552 months with a median of 36 months (p25-p75:13.5-76.2). Patients were managed with benznidazole at dose of 7.5 mg/kg/day for children and from 5-10 mg/kg/day for adult patients. A fixed dose of 300 mg per day during a variable treatment period (between 40 and 80 days) on the basis of the patient's weight was used in two studies [21,51].
The results of the risk of bias assessment of prospective observational studies using Newcastle-Ottawa scale are shown in S2 Text. As per patient selection, most of the studies had an adequate representation of the general population exposed to Chagas disease with reasonable diagnostic and clinical evaluation prior to the inclusion of participants. Regarding comparability domain, five studies [42,[45][46][47]49] were considered at an overall high risk of bias, while the comparability was considered adequate in only one study [18]. As per the outcome evaluation, most of the studies showed a high risk of bias because the high rates of lost to follow up. There was one study with no information of the follow-up length [49], and two with a suboptimal assessment of outcomes during the study period [42,43].
The risk of bias of randomized controlled trials was generally considered low or unclear (S1 Fig). Blinding of participants and allocation concealment was unclear in several studies. Two studies showed an unclear risk of bias in two or more dimensions [41,50]. Only one study was deemed to have an overall high risk of bias [40].
We did not find significant heterogeneity among studied variables in the metaregression analysis. Thus, we finally performed a subgroup analysis according to the type of study and the age of the population, based on methodological and clinical decisions.
When only studies with more than 60 months of follow-up were considered, the OR decreased to 12.8 (95% CI: 1.   We did not find any studies evaluating the effect of benznidazole treatment on the clinical progression of chronic Chagas disease in children. On the other hand, benznidazole efficacy in adults has only been evaluated in one randomized controlled trials which included patients diagnosed with mild to moderate cardiomyopathy [21]. The OR for progression of cardiac disease in patients treated with benznidazole versus placebo was 0.93 (95% CI: 0.8-1.1) (one RCT n = 2854 patients) (Fig 4). This lack of effect was similar when we consider data from OBS with an OR 0.49 (95% CI: 0.2-1.2) (four studies, n = 1331 patients) (Fig 4). The sensitivity analysis restricted to studies with more than 60 months of follow-up did not change overall estimations that were virtually the same (OR = 0.48; 95% CI: 0.17-1.35) (S3 Fig). Only one prospective observational studies [22], performed in a mixed group of adult patients with and without cardiomyopathy, found a significant reduction of the risk of clinical progression in those treated with benznidazole (OR: 0.27; 95% CI: 0.1-0.5) (n = 566) (Fig 4).
Certainty of evidence has been very variable due to several factors such as methodological issues, studies' length of follow-up, representativeness of the studied population, rate of losses to follow-up and imprecision in the meta-analysis.

Discussion
Even though more than 110 years have passed since the discovery of Chagas disease and more than 50 years since the availability of active drugs against T. cruzi, some controversy remains. Even in the absence of a formal randomized controlled trials, there is a general agreement on the efficacy of benznidazole in acute infection and reactivation disease based on evidence from studies performed in the late 1960s and 1970s, and the accumulated experience so far [15,52].
Regarding chronic Chagas disease treatment with benznidazole in children, the certainty of evidence is moderate, with the indication established based mainly on the results of randomized controlled trials conducted in the 1990s, which used seroreversion as a surrogate marker of therapeutic response. The high rates of seroreversion and the clinical practice during decades had supported this indication, as reflected the WHO Guidelines for the diagnostic and References: [21], [42], [18], [47], [48].
https://doi.org/10.1371/journal.pntd.0010386.g004 treatment of Chagas disease [15,52,53]. Furthermore, children are currently the only patients for which the US FDA has approved benznidazole for the treatment of Chagas disease in the USA [25]. Nevertheless, the certainty of the evidence is moderate to low, whether randomized controlled trials [38,41] or prospective observational studies studies [45] are considered (Table 2). Unfortunately, our estimations for seroreversion in children treated with benznidazole vs placebo/no-treatment are imprecise (wide confidence interval ranges) and have not changed from our previous meta-analysis as there has been no new evidence to include. Thus, to increase the degree of certainty concerning this indication, it would be desirable to conduct studies (as for example, prospective cohorts compared with historical controls and healthy children) with clinical outcomes in children treated with benznidazole.
In adult populations with chronic Chagas disease, there is a significant debate regarding indications for benznidazole as its benefits depends on individual factors (as visceral involvement), adding this to a poor tolerability in this population [21,22]. The time needed to achieve seroreversion is challenging and thus, conducting studies with such extended follow-up is the main limitation. Adults may have a lower frequency of seroreversion than children because of the long time-lapse from infection to treatment, leading to lower treatment efficacy. The WHO Guidelines provides a conditional recommendation in adults with chronic Chagas disease without specific organ damage, based on low certainty regarding the effects of the intervention [15]. Nevertheless, after benznidazole approval by the US FDA, the monthly average of treated patients increase from five persons to 13 (comparing all the previous period to the 9 months following its commercialization), with 90% of these patients being adults, despite that the FDA's indication was only for paediatric use [54]. In addition, it is expected that chronic Chagas disease would progressively increase its relevance in endemic and non-endemic areas as a consequence of the empowerment of affected populations [55], better knowledge of the disease by health workers [56], and better access to diagnosis and treatment [57]. Therefore, it is mandatory to establish a common position on the indication for treatment in this population.
The efficacy of benznidazole in avoiding clinical progression has only been formally evaluated in adults. In the case of children with chronic Chagas disease, the WHO Guidelines establish a low certainty of evidence for the treatment regarding clinical outcomes and a strong recommendation based on the experts' consensus that serological negativization is equivalent to a therapeutic response [15]. Benznidazole demonstrated no effect in reducing the progression of heart disease in a large, randomized placebo-controlled clinical trial performed in adult population with mild to severe cardiomyopathy (Fig 4) [21]. Certainty of evidence was considered moderate for benznidazole's effect on this outcome because this study did not include the entire clinical spectrum of Chagas disease in adults (Table 2) [21]. However, this study provided important evidence that benznidazole was not valuable for patients with heart disease. Similarly, WHO Guidelines do not recommend the treatment for adult patients with chronic Chagas disease with specific organ damage based on moderate certainty regarding the effects of the intervention [15]. Regarding observational studies, benznidazole demonstrated only a marginal effect (Fig 4) with a very low degree of certainty. The significant risk of bias limited the degree of certainty resulting from the different proportion of visceral involvement among study participants, non-homogeneous length of follow-up, and inconsistency of results between studies (Table 2). Thus, randomized controlled trials in adults without visceral involvement are needed to determine whether benznidazole is helpful in this indication.
This meta-analysis's main limitations are a consequence of the particular characteristics of the studies included and the different behaviour of the disease in children and adults. The fact that Chagas disease is a neglected disease has limited the generation of high-quality evidence over time. The long latency period of chronic Chagas disease until the development of visceral involvement, competing risks in long periods of follow-up, and lack of surrogate markers are also barriers that make research challenging. In addition, parasitological outcomes cannot be considered a reliable surrogate marker for benznidazole's therapeutic efficacy. Blood cultures and xenodiagnosis have very low sensitivity [58] and have fallen into disuse. On the other hand, PCR techniques are much more sensitive, but, for patients with visceral involvement they have not been demonstrated to be predictive of clinical response [21].
In recent years the trypanocidal effect of other drugs, such as E1224 (ravuconazole prodrug) [30], posaconazole [29,31], and fexinidazole (ClinicalTrials.gov Identifier: NCT03587766), has been tested in clinical trials. The impact on the tolerability of various benznidazole administration schemes has also been analyzed [59][60][61]. Unfortunately, these randomized controlled trials have short follow-up periods and use PCR as the primary outcome, whose role after treatment is mainly as a marker of therapeutic failure, until more information becomes available. The authors of a meta-analysis comparing the efficacy and safety of a fixed vs an adjusted dose of benznidazole, did not find direct evidence comparing both schemes. Nevertheless, through indirect comparisons, they find a low to very low certainty of evidence for clinical outcomes and moderate for the efficacy (positive PCR) and safety outcomes (drug discontinuation, peripheral neuropathy and mild rash) [61]. As in our meta-analysis, they find serological response that favors the use of benznidazole vs placebo (RR 0.88; IC95% 0.84-0.93) [61]. So far, no new drug has demonstrated sufficient efficacy, nor the alternative benznidazole schemes improvement of tolerability, to be included in current therapeutics.
Our meta-analysis shows a benefit of treating chronic Chagas disease with benznidazole (with a moderate degree of certainty) in achieving seroreversion in children. It would be helpful to have studies based on clinical outcomes to better understand the drug's efficacy profile given the lack of information on this context. As for adults, benznidazole has not demonstrated efficacy for patients with cardiomyopathy (moderate certainty). On the other hand, the quality of evidence for patients without visceral involvement is poor, with benznidazole demonstrating a marginally positive effect in the best case. Longer follow-up RCTs with adequate designs in adults without visceral involvement are needed to know whether benznidazole is helpful in this indication and to identify which patients would benefit from treatment. Considering that there will be no new drugs in the short term, early diagnosis, treatment in the early chronic phase and of women of childbearing potential are essential to maximizing the current drugs' benefits.  [40], [38], [21], [29], [41], [30], [51], [50]. prospective observational study. References: [42], [18], [47], [21].  [40], [21], [50], [29], [30], [51], [43], [46], [47]. (TIF) S6 Fig. Evaluation of the effect of benznidazole compared with placebo or no treatment on the response to therapy in adult patients according to secondary parasitological endpoint by type of study (only in those with >60 months follow-up period). RCT: Randomized Clinical Trial; OBS: prospective observational study. References: [43], [47], [21]. (TIF)