The objective of the present study was to evaluate the ability of the gradual effort test (GET) to detect ventricular extrasystoles (VE), and, in particular, to identify patients with high arrhythmic risk in Chagas' heart disease, as well as to compare the results to those obtained by 24 h ambulatory electrocardiographic monitoring (Holter monitoring). A total of 103 Chagasic patients aged 20 to 55 years who exhibited no clinical signs of heart failure or complex arrhythmias under resting conditions were submitted to 24 h Holter monitoring and to a GET with discontinuous loads first applied to reach 120 bpm and then submaximal or maximal heart rate. The results of 24 h ambulatory electrocardiographic monitoring demonstrated that 52.4% of patients had no VE, 20.4% had few isolated VE (Grade I), 1.9% had frequent monomorphic VE (grade II), 4.8% very frequent polymorphic VE (grade III), and the remaining 20.4% had repetitive forms (doubles or bursts of 3 or more) classified as grade IV. In 68% of these patients, the GET did not detect VE. These were the patients who showed no arrhythmias or isolated VE during the Holter test. The 32% of patients who exhibited VE when submitted to the GET included all patients shown to have grade III and grade IV arrhythmia (27 cases) and 5 patients shown to have grade I arrhythmia by the Holter test. The GET showed 100% sensitivity in the detection of pathological (greater than 100 VE daily) and high-risk arrhythmias (greater than 1000 VE per 24 h, polymorphism or repetitive forms), with a specificity of 93 and 92%, respectively.(ABSTRACT TRUNCATED AT 250 WORDS)