The electrocardiographic diagnosis of myocardial infarction in the presence of ventricular conduction defects. A new attempt to solve an old problem

Abstract

Ventricular conduction defect (VCD) in certain cases may obscure the electrocardiographic (ECG) signs of myocardial infarction (MI). While many previous studies have dealt with this problem, criteria derived from such studies have usually been limited in their clinical usefulness because their sensitivity and specificity were not indicated. Realizing the need for such information, we have gathered 816 records with VCD, and analyzed their Prank-lead ECGs to determine optimal diagnostic criteria for MI. The data used in the study were collected in a prospective fashion from six Veterans Administration Medical Centers, thereby including a fairly representative hospital population of adult males. Two thirds of these patients were white and one third black. Cases were randomly grouped into a training set of 547 and a test set of 269 subjects. The training set was analyzed first. In the presence of RVCD, 72% of cases with MI manifested a Q-wave abnormality in at least one lead. We defined a Q-wave abnormality as any one of the following: a Q/R amplitude ratio in lead x of more than 0.20, in lead y of greater than 0.22, and in lead z of less than 0.10. Among cases without MI, 29% met at least one of the same three criteria (false positives). We did not find R- or S- wave parameters useful to differentiate cases with MI from those without. In contrast to RVCD, a Q-wave abnormality in the presence of LVCD was significant for MI only when present in lead x or lead y. Forty-one percent of cases with MI had a Q/R amplitude ratio in lead x of more than 0.20, and/or a Q/R ratio in lead y exceeding 0.22. Only 11% of cases without MI manifested either one or both Q-wave abnormalities (false positives). A Q/R ratio in lead z of less than 0.10 did not differentiate between cases with MI and those without. In the absence of LVCD, this finding denotes loss of anterior forces, and thus probable anterior MI. Neither did we find unique R- or S- wave measurements to distinguish between LVCD cases with and without MI. There was good repeatability of results when the same criteria were applied to the 269 test cases. On this set, the three criteria for MI in the presence of RVCD had 79% sensitivity and 72% specificity; for MI in the presence of LVCD, sensitivity was 44% and specificity 91%.