Electrocardiographic Ventricular Repolarization Parameters in Chronic Chagas’ Disease as Predictors of Asymptomatic Left Ventricular Systolic Dysfunction GIL F. SALLES,* CLAUDIA R.L. CARDOSO,* SERGIO S. XAVIER,* , † ANDREA S. SOUSA,* , † and ALEJANDRO HASSLOCHER-MORENO† From *Department of Internal Medicine, Clementino Fraga Filho University Hospital, Federal University of Rio de Janeiro, Rio de Janeiro and the †Evandro Chagas Hospital, Oswaldo Cruz Foundation, Rio de Janeiro, Brazil SALLES, G.F., ET AL.: Electrocardiographic Ventricular Repolarization Parameters in Chronic Chagas’ Disease as Predictors of Asymptomatic Left Ventricular Systolic Dysfunction. Electrocardiographic re- polarization parameters are potential markers of arrhythmogenic risk and have not been evaluated in Chagas’ disease. The aim of this report was to investigate their associations with LV systolic function as- sessed by two-dimensional echocardiography. In a cross-sectional study involving 738 adult outpatients in the chronic phase of Chagas’ disease, maximal QTc and T wave peak-to-end (TpTe) intervals, and QT, QTapex (QTa), JT and TpTe interval dispersions, and variation coefficients were measured and calculated from 12-lead standard ECGs. Clinical, radiological, ECG, and echocardiographic data were recorded. In bivariate statistical analysis, all repolarization parameters were significantly increased in patients with moderate or severe LV systolic dysfunction, and these patients showed more clinical, radiologic, and ECG abnormalities. Receiver operating characteristic curve analysis demonstrated that isolatedly QTd had the best predictive performance for LV dysfunction, with an 80% specificity and 67% sensitivity for values >60 ms in the subgroup of chagasic patients with abnormal ECGs and no heart failure. Multivariate lo- gistic regression selected, as the best predictive model for LV dysfunction in this subgroup of patients, the presence of cardiomegaly on chest X ray (OR 14.06, 95% CI, 5.54–35.71), QTd > 60 ms (OR 9.35, 95% CI, 4.01–21.81), male gender (OR 7.70, 95% CI, 2.98–19.91) and the presence of frequent premature ventric- ular contractions (PVCs) on ECG (OR 4.06, 95% CI, 1.65–9.97). This model showed 90% specificity and 71% sensitivity. In conclusion, QTd was associated to LV systolic function and could be used to predict asymptomatic dysfunction in chronic Chagas’ disease. The presence of cardiomegaly, frequent PVCs, and male sex refined LV function stratification in these patients. (PACE 2003; 26:1–10) Chagas’ disease, QT intervals, QT dispersion, systolic dysfunction, ventricular repolarization, electrocardiology Introduction Chagas’ heart disease remains an important public health problem in Latin American coun- tries, where it is estimated that nearly 20 million are infected and 25–30% will develop symptoms of congestive heart failure, ventricular arrhyth- mias, or thromboembolism. 1 It is the leading cause of cardiovascular death in endemic areas 2 from sudden arrhythmic or progressive heart failure. Since the original description of Day et al., 3 QT dispersion (QTd), defined as the greatest in- terlead variability of QT intervals, is presumed Address for reprints: Gil F. Salles, M.D., Rua Croton, 72, Jacarepagua, Rio de Janeiro, Brazil. Fax: 55-21-25622759; e-mail: gilsalles@hucff.ufrj.br Received July 10, 2002; revised July 29, 2002; accepted August 7, 2002. to represent a noninvasive measurement of ven- tricular repolarization inhomogeneity and a poten- tial marker of arrhythmogenic risk. Within the last decade, various repolarization parameters have been evaluated in several clinical conditions like long QT syndromes, 4 coronary artery disease, 5,6 heart failure of different etiologies, 7,8 other car- diopathies, 9 land primary noncardiac diseases like diabetes mellitus 10 or in population-based stud- ies, 11 with rather inconsistent results. 12 Although Chagas’ heart disease, because of its main characteristics (a chronically evolving my- ocarditis with fibrosis, hypertrophy, and dilata- tion, accompanying autonomic dysfunction, and a high prevalence of serious ventricular arrhyth- mias and sudden arrhythmic death), 13 seems a per- fect candidate for assessment of ventricular repo- larization dispersion parameters; however, such a study has never been reported. As left ventricular PACE, Vol. 26 May 2003 1