A randomized ten-year follow-up study involving 91 Chagas patients and 41 uninfected controls was undertaken to determine the effectiveness of nitroderivative therapy. Anti-Trypanosoma cruzi antibodies were consistently lower one year after treatment than 10 years thereafter (P < 0.001). The blood of all treated and 93.7% of untreated Chagas patients yielded polymerase chain reaction (PCR) product from probes annealing to T. cruzi nuclear DNA, indicating active infection. Competitive PCR showed means +/- standard deviations of 20.1+/-22.6 T. cruzi/ml of blood from untreated and 13.8+/-14.9 from treated Chagas patients, but the differences between means were not statistically significant (P > 0.05). Electrocardiograms recorded a gamut of alterations several-fold more frequent in Chagas patients, regardless of treatment, than in uninfected controls (P < 0.001). These results show that nitroderivative therapy for T. cruzi infections is unsatisfactory and cannot be recommended since it fails to eradicate the parasite or change the progression of heart disease in chronic Chagas patients.