Ethnic differences in the identification of left ventricular hypertrophy in the hypertensive patient

Left ventricular hypertrophy (LVH) is more prevalent in black than white hypertensives, but this difference is greater when identified by electrocardiography (ECG) than by echocardiography. We evaluated the proposal that current ECG criteria for LVH are less specific, and therefore, less useful, in blacks than whites. In a retrospective cross-sectional study, 408 subjects (271 white, 137 black) referred to a hypertension clinic for assessment of hypertension underwent measurement of blood pressure, ECG voltages (Sokolow-Lyon and Cornell sex-specific), and echocardiographic left ventricular mass index (LVMI). Black subjects had greater ECG voltages than whites, even when closely matched for LVMI. In black subjects, current ECG criteria were twice as sensitive as in whites (Sokolow-Lyon: 44.9% v 22.5%, P = .003, Cornell: 30.4% v 15.7%, P = .03). They were less specific in blacks using the Sokolow-Lyon criteria (73.5% v 86.8%, P = .02) but this failed to reach significance using the Cornell criteria (83.8% v 91.8%, P = .07). When voltage criteria were adjusted to give matched sensitivities and specificities, respectively, differences in specificity and sensitivity were no longer apparent. Receiver operating characteristic curve analyses confirmed no significant differences in overall performance of either ECG criteria between blacks and whites. In conclusion, ECG detection of LVH is insensitive in both ethnic groups. Sensitivity is higher in blacks due to higher LVMI in those with LVH. Apparent differences in specificity are due to ethnic differences in ECG voltages that are unrelated to differences in LVMI. When these differences are taken into account, there are no overall differences in test accuracy. However, given the prognostic importance of the detection of LVH, currently accepted ECG voltage criteria for the detection of LVH remain of equal or greater value in black hypertensives compared with whites. (C) 1999 American Journal of Hypertension, Ltd.