Evaluation of electrocardiographic criteria for left ventricular hypertrophy based on anatomical comparison.

The reliability of precordial electrocardiographic criteria for left ventricular hypertrophy (LVH) was evaluated by comparing it with the anatomical findings of 421 postmortem hearts examined by the chamber dissection technique. Of these 421 postmortem hearts, 136 without cardiovascular disease served as normal controls. The weights of the left ventricular free wall with interventricular septum (LV + S) and of the whole heart in normal controls were 153 +/- 36 grams (mean +/- SD) and 289 +/- 67 grams, respectively. Correlation coefficients between LV + S and whole heart were 0.89 for men and 0.90 for women. Anatomical LVH was defined as the weight of LV + S exceeding the average plus one standard deviation of that of a normal control heart. Anatomically established LVH was found in 99 (28%) of 353 cases having precordial electrocardiograms available for evaluation. The sensitivity, specificity, and predictive accuracy of electrocardiographic criteria for LVH were as follows: SV1 + RV5 greater than 3.5 mV, 56.6%, 76.0% and 47.9%; SV1 + RV5 greater than 4.0 mV, 45.5%, 87.4% and 58.4%; SV1 + RV5 greater than 5.0 mV, 21.2%, 94.5% and 60.0%; SV1 + RV5 greater than 3.5 mV with strain type ST-T, 35.4%, 93.3% and 67.3%; SV1 + RV5 greater than 4.0 mV with strain type ST-T, 29.3%, 94.9% and 69.0%; SV1 + RV5 greater than 5.0 mV with strain type ST-T, 14.1%, 98.0% and 73.7%; strain type ST-T, 36.4%, 92.5% and 65.5%; nonspecific ST-T, 26.3%, 76.4% and 30.2%; SV1 + RV5 greater than 3.5 mV with nonspecific ST-T, 12.1%, 93.3% and 41.4%; and right anteriorly directed T vector, 39.4%, 89.4% and 59.1%.(ABSTRACT TRUNCATED AT 250 WORDS)