Eight hundred and forty three patients with confirmed positive serology for Trypanosoma cruzi
infection were seen between 1984 and 2003 at the Hospital Eva Perón, Buenos Aires, Argentina. All patients were evaluated with ECG, chest radiography, bidimensional echocardiography, and stress testing under basal conditions. Patients were grouped into four stages according to a modified Kuschnir classification.2
Patients in groups 0 (normal ECG and chest radiograph) and I (abnormal ECG and normal chest radiograph) and aged 20–60 years were enrolled for a longitudinal 10 year follow up.
Patients in groups II (left ventricular (LV) enlargement) and III (heart failure) (79 patients) were excluded. Other exclusion criteria were reduced LV systolic function (26 patients), associated pathological conditions (92 patients), incomplete screening period (21 patients), and incomplete stress testing (38 patients). Patients were assessed longitudinally by ECG every six months (group 0) and four months (group I) and by evaluation of the cardiothoracic index from annual chest radiography.
ECG abnormalities were defined as those specific for Chagas disease as previously reported.3
Cardiomegaly was diagnosed by a cardiothoracic index > 0.50. The clinic variables evaluated were age at admission, sex, specific ECG abnormalities, clinical group, and baseline LV diastolic diameter.
The primary outcome of the study was progression to a more severe stage of the disease. The appearance of new ECG abnormalities indicated a shift from group 0 to group I; LV dilatation indicated a shift from group I to group II (confirmed by echocardiography); and the development of heart failure indicated progression from group II to group III.
Stress testing was performed on a bicycle ergometer, in both continuous and stepwise modes, with progressive workload increments of 150 kilogrameters every three minutes. Leads DII, V1, and V5 were monitored during the test and recorded throughout at three minute intervals during exercise and for nine minutes in the recovery stage. Abnormalities of the stress testing were failure to increase in systolic blood pressure, chronotropic incompetence, and the presence of ventricular extrasystoles during both the exercise and the recovery periods. Ventricular arrhythmias were further divided into complex (non‐sustained ventricular tachycardia and ventricular couples) and non‐complex (monomorphic and polymorphic ventricular extrasystoles).
Cox regression analysis was used to assess the relation between the evaluated variables and clinical group changes (dependent variable). Multivariate analysis considered all the variables significant at p < 0.05 in the univariate analysis.