[A comparison of three electrocardiogram coding systems for chronic Chagas' disease].

OBJECTIVE To compare three electrocardiogram coding systems (Buenos Aires, New York Heart Association, and modified Minnesota) for use with Chagas' disease. METHODS Between January 1976 and December 1978 persons older than 2 years of age living in the town of Agua Comprida, Minas Gerais, Brazil, underwent a patient history, physical examination, conventional electrocardiography, and serology for Chagas' disease. Those records were archived, and in 1999 from this sample, 100 seropositive individuals were matched by sex and age with 100 seronegative controls. The electrocardiograms of these 200 patients were read using the three coding systems. Each system was assessed in terms of its gradient, that is, the difference between the number of abnormalities found in the electrocardiograms of individuals with and without Chagas' disease. RESULTS The gradient was 7% for the New York Heart Association coding system, 15% for the modified Minnesota system, and 17% for the Buenos Aires one. The ability of the Buenos Aires method to detect electrocardiographic abnormalities was significantly greater in the group with Chagas' disease than in controls (P = 0.012), followed by the modified Minnesota system (P = 0.028) and the New York Heart Association system (P = 0.31). The Buenos Aires method was also easier to use, in terms of both the analysis of electrocardiograms and queries to a database. CONCLUSIONS Although the Buenos Aires method had advantages over the two other systems, some changes are still needed in it to make it more useful. It would be worth carrying out further studies to evaluate this method with the changes that we have suggested in this study.