Race- and sex-specific ECG models for left ventricular mass in older populations. Factors influencing overestimation of left ventricular hypertrophy prevalence by ECG criteria in African-Americans

The validity of the reported high prevalence of left ventricular hypertroph y (LVH) among African-American men and women has been questioned owing to c onflicting echocardiographic evidence. We used echocardiographic left ventr icular mass (LVM) from M-mode measurements to evaluate associations between LVM, body size, and electrocardiographic (ECG) variables in 3,627 white an d African-American men and women 65 years of age and older who were partici pants of the Cardiovascular Health Study (CHS), a multicenter cohort study of risk factors for coronary heart disease and stroke. ECG amplitudes used in LVH criteria were substantially higher in African-Americans, with appare nt LVH prevalence 2 to 3 times higher in African American men and women tha n in white men and women, although there was no significant racial differen ce in echocardiographic LVM. The higher apparent LVH prevalence by Sokolow- Lyon criteria in African-American men is in part owing to smaller lateral c hest diameter. In women, reasons for racial differences in ECG LVH prevalen ce remain largely unexplained although a small part of the excess LVH in Af rican-American women by the Sokolow-Lyon criteria appears to be owing to a larger lateral chest semidiameter in white women. ECG variables alone were too inaccurate for LVM prediction, and it was necessary to incorporate in a ll ECG models body weight that was properly adjusted for race and sex. This resulted in modest LVM prediction accuracy, with R-square values ranging f rom .22 to.36. Race- and sex-specific ECG models introduced for LVM estimat ion with an appropriate adjustment for body size differences are expected t o facilitate evaluation of LVH status in contrasting racial population grou ps.