Benznidazole therapy for Chagas disease in asymptomatic Trypanosoma cruzi-seropositive former blood donors : evaluation of the effi cacy of diff erent treatment regimens

Introduction: Chagas disease currently affects 5.7 million people in Latin America and is emerging in non-endemic countries. There is no consensus concerning the effi cacy of trypanocidal therapy for patients with the chronic form of the disease. We evaluated cardiac function and sociodemographic, clinical, and serologic characteristics of a group of asymptomatic Trypanosoma cruzi-seropositive former blood donors, and compared the effects of benznidazole treatment applied for different lengths of time. Methods: Blood donors who screened positive for T. cruzi between 1998 and 2002 were recruited 10 years later for follow-up (n = 244); 46 individuals had received treatment. Three subjects had terminated treatment prematurely. The remaining 43 individuals were divided into two groups: individuals who had received benznidazole therapy for 50-60 days (n = 28; BT ≤60 group) or more than 60 days (n = 15; BT >60). Serologic assays, biochemical tests, electrocardiographic, echocardiographic, and clinical examinations were performed on all participants. Parasite loads were determined by qualitative and quantitative polymerase chain reaction. Results: Parasitemia was signifi cantly reduced in the BT ≤60 and BT >60 groups compared with the untreated group. There were no differences in epidemiologic profi les or clinical, biochemical, electrocardiographic, or echocardiographic data between any of the groups. Conclusions: Despite elimination or signifi cant reduction in parasitemia in patients with chronic Chagas disease who received benznidazole, there was no clinical difference between those who were treated for >60 days and those treated for a shorter duration. Furthermore, the adverse effects of benznidazole appear to be less severe than previous reports would suggest.


INTRODUCTION
Chagas disease, also known as American trypanosomiasis, is an infl ammatory disease caused by the fl agellate protozoan Trypanosoma cruzi (Trypanosomatidae).The disease is endemic in Latin America and is estimated to affect some 5.7 million people (1) , a large proportion (20-40%) of whom will gradually evolve to the chronic phase of the disease and exhibit myocardiopathy, megaesophagus, megacolon, or other associated conditions in later life (2) .Although the incidence of Chagas disease is declining in most Latin American countries following the establishment of public policies of surveillance, prevention, and control (3) (4) , the number of cases is increasing in non-endemic regions such as North America, Europe, Asia, and Oceania as a result of the migration of infected individuals.Despite its detrimental impact on the world economy, trypanosomiasis is considered a neglected disease (5) .
The classical form of transmission of Chagas disease is vectorial via the feces of triatomine bugs, although nonvectorial modes of transmission (oral, vertical or congenital, blood transfusion, organ transplantation, and laboratory accidents) have also been reported.In Brazil, for example, the epidemiology of trypanosomiasis has changed substantially over the years such that it now resembles that of other infectious diseases that are not vector-borne (1) .Indeed, oral transmission through the ingestion of contaminated food is currently the most frequent mode of infection of Chagas disease in Brazil (68.9%), while congenital transmission and transfusion-related transmission are a problem in non-endemic regions (6) .Hence, safeguards relating to all aspects of blood transfusion have been strengthened in many countries.As a result, donated blood is screened for known blood-borne infections, including Chagas disease.In the United States (US) for example, the Retrovirus Epidemiology Donor Study-II (REDS-II; 2004 -2012) was established with the aim of improving blood product safety and availability (7) (8) .
The two antiparasitic drugs that could be employed in the treatment of T. cruzi infection (i.e.nifurtimox and benznidazole) have not been approved by the US Food and Drug Administration (FDA) but are available at the request of a physician for use under investigational protocols (9) .It is important to emphasize, however, that prolonged treatment with benznidazole is contraindicated because of the associated side-effects and safety risks.The Benznidazole Evaluation for Interrupting Trypanosomiasis (BENEFIT) project, described recently by Morillo et al. (10) , was a multicenter, doubleblinded, randomized, placebo-controlled trial involving the administration of benznidazole or matching placebo to 2,854 patients presenting with chronic Chagas cardiomyopathy.The results showed that benznidazole therapy did not reduce the clinical decline in cardiac function despite signifi cant reductions in parasite load in the blood over the 5 years of follow-up.The authors concluded that there is a need to research the effects of benznidazole in patients with the indeterminate form of Chagas disease without cardiomyopathy and also the effects of prolonged benznidazole therapy (10) .
A retrospective cohort study, conducted in collaboration with the international component of the REDS-II program, evaluated a population of supposedly healthy blood donors in the Brazilian cities of São Paulo and Montes Claros, and showed that 24% of 499 T. cruzi-seropositive individuals presented with cardiomyopathy within a 10-year follow-up period, while only 5% of the 488 seronegative subjects exhibited this condition within the same time period (11) .Interestingly, only 46 seropositive subjects reported prior administration of benznidazole and treatment periods ranged from 6 to 700 days.In the present study we investigated T. cruzi-seropositive blood donors identifi ed in Montes Claros with the objectives of: I) evaluating the sociodemographic, clinical, serologic, electrocardiographic (ECG) and echocardiographic (ECHO) characteristics of the individuals, and II) comparing the effects of benznidazole treatment applied for different lengths of time, namely 50-60 days and >60 days.This is the fi rst report describing treating blood donors with benznidazole; hence, it contributes to a better evidence-based judgment of the appropriateness of prolonged treatment with this drug.

Study design
This cross-sectional study involved 244 supposedly healthy volunteers who were registered at a blood donation center in Montes Claros, Brazil, during the period 1998-2002 and had been selected 10 years later for follow-up examinations in a previous investigation (11) .The inclusion criteria were: I) positive serologic test results for Chagas disease based on enzyme-linked immunosorbent (ELISA) and hemagglutination assays for T. cruzi antibodies performed at the time of blood donation, II) absence of comorbidities such as diabetes and renal insuffi ciency, and III) a signed document giving informed consent for participation.Subjects were excluded if their serologic test results were not completed.During the period 2008-2010, potential participants in the follow-up study were recruited by either mail or telephone.New serologic assays were performed on blood samples of all participants selected.These tests were repeated in the US REDS-II Central Laboratory at the Blood Systems Research Institute [(BSRI); San Francisco, CA, US] in order to confi rm the diagnosis using methodology and reagent kits (Ortho-Clinical Diagnostics, Raritan, NJ, US) approved by the FDA.The parasite load in the blood of each subject was determined by qualitative and quantitative polymerase chain reaction (PCR) assays.Participants also underwent standard biochemical testing, ECG and ECHO investigations, and clinical examination by a physician (nonblinded test), and were subsequently offered counseling if appropriate.

Collection of data
Selected participants were called for an interview and were asked to complete a standardized questionnaire relating to sociodemographic data (age, sex, skin color, civil status, educational level, smoking and drinking habits, and physical activity), cardiovascular and digestive symptoms, and use of medicines.On the day of the interview, blood samples (20mL) for molecular tests were collected from participants and transferred to vials containing the anticoagulant ethylenediaminetetraacetic acid (EDTA).Each sample was mixed immediately with an equal volume of a solution containing 6M guanidine hydrochloride in 0.2M EDTA and stored at room temperature.Samples were subsequently boiled for 15 min (to shear T. cruzi nuclear and kinetoplast deoxyribonucleic acid (kDNA), cooled, submitted to vortex mixing and distributed as 1mL aliquots into appropriately labeled cryovials.The vials were maintained in Brazil at -20°C until required for blind analysis by PCR or shipment to BSIR in containers cooled with dry ice (-70°C) for repeat blind testing.
The qualitative and quantitative analysis of blood parasite load was performed by targeted-capture real time polymerase chain reaction (TC-RT-PCR) following the methodology described by Virreira et al. (12) in which the target was kinetoplast DNA minicircles from T. cruzi.Specifi c DNA extraction was improved by the TC step employing a T. cruzi-specifi c 20-mer capture oligonucleotide bound to magnetic beads.TC-RT-PCR analysis was repeated four times for each sample and the result was considered positive when two or more tests detected the specifi c amplicon, or negative when none or only one of the tests detected the amplicon.According to Sabino et al. (13) , this assay is able to detect a single parasite in 20mL of blood.The results of the PCR assays were sent to the US REDS-II Coordination Center (Westat; Rockville, MD, US) for decoding and analysis.
Blood samples (20mL) were also collected for the preparation of serum and plasma required for assessment of high density lipoprotein (HDL) cholesterol, low density lipoprotein (LDL) cholesterol, total cholesterol, triglycerides, glycemia, troponin, myoglobin, N-terminal pro B-type natriuretic peptide (NT-proBNP), creatine kinase fraction MB (CK-MB), and tumor necrosis factor-alpha (TNFα).The levels of troponin, CK-MB and NT-proBNP were determined using the VITROS® System (Ortho Clinical Diagnostics, Raritan, NJ, US), as approved by the FDA (11) .
A resting ECG was performed using a General Electric Healthcare (Waukesha, WI, US) model MAC 1200 ECG instrument.ECG traces were analyzed by a local cardiologist and blind re-evaluations were carried out at the Epidemiological Cardiology Research Center in the Wake Forest University (Winston-Salem, NC, US).The Minnesota Code classifi cation system was employed in the interpretation of ECG traces (14) (15) considering the following parameters: QRS complex interval, corrected QT (QTc) interval, heart rate, PR interval, atrioventricular block (AVB), complete right bundle branch block (RBBB), complete left bundle branch block (LBBB), left anterior fascicular block (LAFB), ventricular premature beats (VPBs), Q wave abnormalities, and ST and T waves abnormalities.QRS ≥120ms and QTc >440ms were considered prolonged.
Echocardiographic assessment was performed using a General Electric Healthcare Vivid 3 ultrasound unit following the recommendations of the American Society of Echocardiography (16) .Images were recorded in digital format and analyzed at the Cardiovascular Branch of the Echocardiography Laboratory in the National Heart, Lung, and Blood Institute (NHLBI; Bethesda, MD, US).The parameters analyzed were: left ventricular ejection fraction (LVEF), left ventricular mass index (LVMI), wall motion score index (WMSI), left ventricular hypertrophy (LVH), Doppler-derived septal E/E' >15, and left ventricular diastolic diameter (LVDD).LVEF was calculated using the modifi ed biplane Simpson's method (16) .LVEF <50% characterized left ventricular systolic dysfunction.
The results of all of the examinations and assays performed were sent to the US REDS-II centralized reading centers for blind evaluation.

Experimental groups
The initial study population (n = 244) was divided into two groups comprising subjects who had received treatment with benznidazole (BT group; n = 46) and non-treated subjects (NT group; n = 198).Subsequently, three members of the BT group were removed from the study because of premature termination of treatment (one after 6 days and two after 30 days).The remaining 43 members of the BT group were subdivided according to the duration of treatment, namely, 50-60 days (BT ≤60; n = 28) and more than 60 days (BT >60; n = 15).

Statistical analysis
The Kolmogorov-Smirnov test was used to test the normality of the distributions.Normally distributed variables were expressed in terms of the mean ± standard deviation (SD) while continuous variables that were not normally distributed were expressed in terms of the median and interquartile range (IQR).The statistical significance of differences between groups was assessed using the Mann-Whitney or Kruskal-Wallis tests.Categorical variables were expressed as numbers and percentages, and differences were determined using the χ 2 or Fisher's exact tests.For all tests, the level of signifi cance was set at p <0.05.The strength of the relationship between parasite load and duration of treatment was evaluated using the Spearman correlation coeffi cient.All statistical analyses were performed with the aid of Statistical Package for the Social Sciences (SPSS) version 18 software (Chicago: SPSS Inc.).

Ethical considerations
The study was submitted to and approved by the National Research Ethics Commission (protocol no.CONEP 3 1312/2006) and all procedures followed the ethical principles for medical research involving human subjects proposed by the Declaration of Helsinki.Written informed consent was obtained from all participants.
The BT group comprised only 18.9% of the study population (n = 46).There were no differences between the BT and NT groups (Not Treated) with respect to sociodemographic variables.When the duration of treatment was taken into consideration, the sociodemographic characteristics of the NT, BT 60, and BT >60 groups were similar, except for a signifi cantly higher frequency of patients who reported having drunk alcohol in the previous month among individuals in the BT ≤60 group (Table 1).
The three groups were also similar with respect to the biochemical variables (Table 2) and clinical, ECG, and ECHO characteristics (Table 3).However, quantitative PCR revealed that the parasite load of the NT group was signifi cantly higher than that of the BT groups (p <0.05) (Table 4).Qualitative PCR showed that the parasite load of the NT group was signifi cantly higher (p <0.05) than that of the BT ≤60 group but was not signifi cantly higher than that of the BT >60 group (Table 4).Qualitative and quantitative PCR demonstrated that there were no signifi cant differences in parasite load between the BT ≤60 and BT >60 groups.

DISCUSSION
Chagas disease, described more than 100 years ago (17) , remains a potentially life-threatening condition that negatively impacts on the quality of life of millions of people in Latin America, with attendant economic consequences.Infection is mediated mainly by vectors, but the epidemiologic profi le of the disease has changed markedly in Brazil over recent years such that ingestion of contaminated food is currently the most important mode of transmission.Moreover, while the disease was originally confi ned to Latin American countries, its prevalence in the US, Canada, Europe, and some Western Pacifi c countries is on the increase owing mainly to population mobility (3) .
Treatment of adults with the indeterminate form of Chagas disease remains controversial chiefl y because of the absence of appropriate blinded randomized clinical trials.In the past, parasite persistence was not thought to be relevant to the physiology of complications in Chagas disease.Many other reasons also motivated the paradigm of not treating Chagas disease (18) .Despite the current recommendations, treatment is not implemented in >99% of infected individuals (2) .In the present study, the proportion of seropositive blood donors who had been treated with benznidazole was also modest (18.9%; 46/244) while the duration of treatment varied markedly and ranged between 6 and 700 days.
Few studies have focused on the efficacy of different durations of benznidazole therapy in the treatment of Chagas disease.Thirty years ago, Coura et al. (19) compared the effectiveness of 30-and 60-day treatments, but observed no statistical difference between the two regimens regarding the control of parasitemia, thus suggesting that there was no advantage in opting for a longer therapy.However, in a more recent report, Coura et al. (20) suggested a treatment regimen comprising 30 days of therapy repeated every two months over a 6 to 12 month period.In contrast, the study conducted by Morillo et al. (10) for the BENEFIT project involved benznidazole treatment administered for a single period of 40 to 80 days.
In a systematic review and meta-analysis of benznidazole therapy applied to patients with chronic Chagas disease who were in the indeterminate phase or had visceral involvement, Pérez-Molina et al. (21) considered 696 reports and selected for detailed analysis three clinical trials and six observational studies of which six involved adults.In these studies, the duration of treatment varied widely encompassing regimens of 30 (22) , 45-60 (23) and 60 (24) days, while the follow-up periods were also somewhat disparate, ranging from 12 months to 24 years (median 4 years) (21) .Regarding the effi cacy of benznidazole administered in the late chronic phase of the disease, the authors of the meta-analysis concluded that the effi cacy of therapy was doubtful and that the benefi cial effect was marginal (21) .Uncertainty appeared to be even greater in asymptomatic individuals and those older than 50 years; this was largely due to differences in the study populations, endpoints, and durations of follow-up.However, the non-randomized and non-blinded prospective study conducted by Viotti et al. (22) confi rmed clinical improvement and increased seroreversion rates of benznidazole-treated subjects compared with their non-treated counterparts (n = 283 in each group).Moreover, Fabbro et al. (23) reported improvement of the ECG profi les of benznidazoletreated patients, while Fragata-Filho et al. (25) found that 79.1% (263/310) of benznidazole-treated patients and 46.8% (47/310) of those untreated had normal ECG fi ndings after a follow-up period of around 20 years.The occurrence of ECG abnormalities and clinical events (heart failure, stroke, mortality, and death) was less prevalent in treated patients.In contrast, Morillo et al. (10) concluded that therapy with benznidazole in patients with established Chagas cardiomyopathy reduced signifi cantly serum parasite load but did not signifi cantly reduce clinical cardiac deterioration through fi ve years of follow-up.
In the present study, although the length of treatment varied considerably, two groups with different benznidazole treatment regimens could be defi ned.While parasite load was signifi cantly reduced in the BT group compared with the NT group, no signifi cant differences were detected between BT ≤60 and BT >60 groups.Quantitative and qualitative PCR revealed a signifi cant reduction in the parasite load of the BT 60 group compared with NT.However, an analogous reduction in the BT >60 group could only be confi rmed by quantitative PCR, probably because this group was small and the statistical power to detect differences by qualitative PCR was not so strong.Regarding the clinical cardiac conditions, no alterations were detected in the study population, regardless of group.It is important to emphasize, however, that the frequencies of use of cardiovascular drugs by the study population at the time of evaluation were 23.2, 25.0 and 6.7% for the NT, BT ≤60, and BT >60 groups, respectively, although the differences were not statistically signifi cant (p = 0.313).
It is of interest to note that in the study conducted by Morillo et al. (10) , which involved only patients with established Chagas cardiomyopathy, the effi cacy of treatment as assessed by conversion to negative PCR results was 66.2% at the end of treatment and 46.7% after 5 years or more of follow-up.This fi nding suggests that the standard regimen of benznidazole may be less effective in patients with cardiomyopathy.In contrast, the preliminary results of a randomized placebo-controlled evaluation of benznidazole treatment on long-term disease progression in adults with chronic Chagas disease (the TRAENA study) (26) (27) showed that rates of negative parasitemia increased with time after treatment from 55.97% after two months, to 62.59% after 8-16 months, and 72.81% after 9-11 years.
In our study, side effects (allergic reactions) to the drug were detected in 19.5% of individuals, and therapy was discontinued in only 6.5% (3/46) of cases.Morillo et al. (10) reported that 14.4% of patients permanently discontinued treatment with benznidazole because of adverse events, a rate that is lower than that previously reported in observational and small randomized trials.These authors recommended that a dose of 300mg/day should not be surpassed since low doses appeared to increase tolerance to the drug.
New evidence supports the idea that the level of circulating parasites is relevant to the pathogenesis of chronic Chagas disease (11) .Furthermore, the association between parasite load (as measured by PCR) and the severity of clinical conditions has been confi rmed (12) .In addition, various blinded randomized studies have demonstrated the effi cacy of benznidazole therapy and that earlier treatment is associated with a more successful outcome (9) (18) (28) (29) (30) .Such findings have encouraged some agencies, including members of the network Nuevas Herramientas para el Diagnóstico y la Evaluación del Paciente con Enfermedad de Chagas (NHEPACHA; New Tools for the Diagnosis and Evaluation of Chagas Disease Patients), the World Health Organization and the Centers for Disease Control and Prevention, to recommend antiparasitic treatment for all T. cruzi-infected individuals in the chronic phase, even though this remains controversial in current medical practice.The need for a paradigm shift becomes stronger when one considers that delay in treatment while awaiting further studies leaves most of the infected population without any treatment.
Limitations of the present study include recall bias and the absence of information in the study population relating to dosage and the length of time after treatment completion.Moreover, employing positive ELISA results in the inclusion criteria may have led to the exclusion from the study of serologically cured patients, thereby reducing the ability to detect differences favoring treatment.In addition, the small sample population probably reduced the statistical power to detect differences.A strength of the study is that sampling a population of blood donors allowed inclusion of asymptomatic individuals, thus minimizing the effect of referral bias.
In conclusion, our study revealed that parasite load was signifi cantly lower in the benznidazole-treated group than the non-treated group.However, comparison between individuals treated for 50-60 days and those treated for a more extended period (>60 days) showed that prolonged treatment conferred no advantage.Since the majority of blood donors are asymptomatic they are ideal targets for etiologic treatment in specialized health facilities.However, treatment regimens are currently inconsistent and this problem needs to be addressed with dosage, duration, and treatment intervals specifi ed according to age and clinical groups.Although our results indicate that the duration of treatment does not infl uence the parasite load, the study was not designed to draw conclusions about the most appropriate therapeutic regimen for individuals with chronic asymptomatic or indeterminate Chagas disease.

FIGURE 1 .
FIGURE 1. Correlation between parasite load as estimated by quantitative PCR and duration of benznidazole treatment.PCR: polymerase chain reaction.Spearman correlation coeffi cient r = 0.13; p = 0.401.

TABLE 1
Sociodemographic characteristics of former blood donors registered in Montes Claros, State of Minas Gerais, Brazil.
NT: non-treated; BT: benznidazole-treated; SD: standard deviation.Data expressed in number and percentage, and used χ2 or Fisher's exact tests.Values followed by different lowercase superscript letters ( a, b ) are signifi cantly different.*Statistically signifi cant ( χ 2 test).Age expressed in mean and SD, and used Kruskal-Wallis test.

TABLE 2
Biochemical profi le of former blood donors registered in Montes Claros, State of Minas Gerais, Brazil.

TABLE 3
Clinical, ECG and ECHO alterations of former blood donors registered in Montes Claros, State of Minas Gerais, Brazil.

TABLE 4
Parasite load as determined by quantitative and qualitative PCR.