We examined the relations of gender differences in electrocardiographic (ECG) voltages and QRS duration to differences in cardiac dimensions and body size between men and women and gender differences in test performance of ECG criteria for the detection of echocardiographic left ventricular hypertrophy in 389 subjects (112 women and 277 men). ECG voltage-duration products were calculated as the product of QRS duration and voltages. Among subjects with normal left ventricular mass and also among subjects with left ventricular hypertrophy, men had longer QRS duration, higher Cornell voltage, higher 12-lead sum of QRS voltage, and higher Cornell and 12-lead voltage-duration products than did women. Significant gender differences in QRS duration, Cornell voltage, the 12-lead sum of voltage and their voltage-duration products remained after adjusting for the greater left ventricular mass, height, and weight in men than women. Comparison of areas under receiver operating characteristic curves using gender-specific criteria demonstrated higher performance of QRS duration, Cornell voltage, the 12-lead sum of QRS voltage, and the respective voltage-duration products for the identification of left ventricular hypertrophy in men than women. Thus, gender differences in body size and left ventricular mass do not completely account for gender differences in QRS duration and voltage measurements, and ECG criteria for left ventricular hypertrophy have lower accuracy in women even when gender differences in partition value selection are taken into account.(ABSTRACT TRUNCATED AT 250 WORDS)