Quantifying anti-trypanosomal treatment effects in chronic indeterminate Chagas disease: a secondary analysis of individual patient data from two proof-of-concept trials

Background

Determining parasitological cure in chronic Chagas disease is compromised by very low blood trypomastigote densities, which fluctuate close to or below the limit of quantitative PCR (qPCR) detection (approximately one parasite per 10 mL). We aimed to improve the statistical methodology used to analyse serial qPCR data to estimate treatment efficacy.

Methods

In this secondary analysis, we pooled clinical and laboratory data from two prospective randomised controlled trials (E1224 [NCT01489228] and BENDITA [NCT03378661]) in Bolivian adults (aged 18–50 years) with chronic indeterminate Chagas disease. Both trials included positive and negative control groups, consisting of placebo and standard of care benznidazole (300 mg per day for 8 weeks), respectively. In E1224, participants were enrolled between July 19, 2011, and July 26, 2012, and the experimental groups were fosravuconazole monotherapies (400 mg per week for 4 weeks or 8 weeks, or 200 mg per week for 8 weeks); in BENDITA, participants were enrolled between Nov 30, 2016, and July 27, 2017, and the experimental groups were shorter or lower-dose benznidazole regimens (300 mg per day for 2 weeks or 4 weeks, or 150 mg per day for 4 weeks), or combinations of fosravuconazole 300 mg weekly for 8 weeks with either benznidazole 150 mg per day for 4 weeks or benznidazole 300 mg per week for 8 weeks. Triplicate qPCRs were done on one to three blood samples taken at eight to 12 follow-up visits over 1 year. The primary analysis included patients randomly assigned to placebo or patients who took an active treatment for more than 80% of the allocated treatment duration. We estimated treatment efficacy under a probabilistic hierarchical Bayesian model fitted to the serial blood qPCR data.

Findings

441 patients (231 from E1224; 210 from BENDITA; 320 [73%] female and 121 [27%] male) provided 34 804 individual qPCR cycle threshold values over 5402 unique visits, comprising 449 patient-years of follow-up. In the per-protocol population (n=424), an estimated 81% (70–89) of participants had parasitological cure following the standard of care 8-week benznidazole regimen. In comparison, spontaneous self-cure occurred in only 4% of patients allocated to placebo (95% credible interval [CrI] 1–9). All benznidazole regimens had similar estimated cure proportions (95% CrIs >63%) except the 2-week regimen (63% cured [43–81]; posterior probability of inferiority relative to standard of care 8 weeks was 0·95). Fosravuconazole showed dose dependency in both efficacy and risk of increased liver aminotransferases but overall was relatively ineffective (all regimens had <40% estimated cure rates). Parasite densities in recurrences after fosravuconazole were only slightly lower than before treatment, whereas recurrent parasitaemias after benznidazole were substantially lower.

Interpretation

Therapeutic assessments in Chagas disease must account probabilistically for qPCR test performance and low post-treatment parasite densities. In chronic Chagas disease in Bolivia, once-weekly benznidazole dosing for 8 weeks or daily dosing over 4 weeks have similar efficacies as the current 8 weeks daily regimen. These results suggest that the total benznidazole dose in the standard of care regimen is excessive.