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

BackgroundThe current antiparasitic treatment for chronic Chagas disease of 8 weeks daily benznidazole or nifurtimox is poorly tolerated and reaches only a small minority of those with chronic infections. Defining parasitological cure is compromised by the low blood trypomastigote densities, which fluctuate close to or below the limit of qPCR detection. MethodsTo address this limitation and improve the assessment of parasitological cure we developed a probabilistic model of therapeutic efficacy based on serial qPCR data. We pooled clinical and laboratory data from two prospective trials in Bolivian adults with chronic indeterminate Chagas disease. In both trials randomised arms included placebo or standard of care benznida-zole (300mg/day for 8 weeks). In the first trial, the experimental arms were fosravuconazole monotherapies (400mg/week for 4 or 8 weeks, or 200mg/week for 8 weeks); in the second trial the experimental arms were shorter or lower dose benznidazole regimens (300mg/day for 2 or 4, or 150mg/day for 4 weeks), or combinations of fosravuconazole 300mg weekly for 8 weeks with either benznidazole 150mg/day for 4 weeks or benznidazole 300mg/week for 8 weeks. Serial parasite densities were estimated from triplicate qPCRs targeting T. cruzi satellite DNA taken from one to three 5 or 10ml blood samples at 8-12 visits over one year. Treatment efficacies were estimated under a hierarchical Bayesian model, taking as input serial cycle threshold (Ct) data grouped by time point, blood draw and technical replicate. The primary analysis was done in a per-protocol population defined as patients randomised to placebo or patients who took an active treatment >80% of the allocated treatment duration. ResultsThe two trials randomised 441 patients. 34,804 qPCR Ct values were recorded over 5,402 unique visits, comprising 449 participant years follow-up. In a per-protocol population (n=424), an estimated 81% (95% Credible Interval [CrI] 70 to 89%, n=69) had parasitological cure following benznidazole 300mg/day for 8 weeks. All other benznidazole regimens had similar estimated cure proportions (95% CrIs >63%) except the 2-week regimen (63% cured [95%CrI 43-81%], n=27, probability of inferiority relative 8-week: 0.95). Recurrent parasitaemias following benznidazole were at substantially lower densities than at baseline. In comparison, only 3.9% of patients allocated to placebo were cured (95%CrI 1 to 9%, n=77). Fosravuconazole was relatively ineffective: 23% cured following 400mg for 8 weeks (95%CrI, 10 to 40%, n=45); 9% following 400mg for 4 weeks (95%CrI 3 to 21%, n=46); and 2% following 200mg for 8 weeks (95%CrI 0 to 11%, n=48). Recurrent parasitaemias one year after fosravuconazole treatment were only slightly lower than at baseline. Fosravuconazole caused dose-dependent increases in liver transaminases. ConclusionsTherapeutic assessments in Chagas disease must account probabilistically for qPCR test performance and low density post treatment parasitaemias. In Bolivian chronic Chagas disease, weekly dosing for eight weeks or daily dosing over four weeks both appear as effective as the current eight weeks daily regimen. The total dose of benznidazole in the current standard of care regimen is excessive.