Determination of Sensitivity and Specificity of Electrocardiography for Left Ventricular Hypertrophy in a Large, Diverse Patient Population

BACKGROUND: Electrocardiography (ECG) is poorly sensitive, but highly specific for the diagnosis of left ventricular hypertrophy. However, previous studies documenting this were small and lacked patient diversity. Furthermore, little is known about the impact of patient characteristics on the sensitivity and specificity of ECG for left ventricular hypertrophy. To address this issue, the present study was conducted to ascertain the sensitivity and specificity of ECG for left ventricular hypertrophy in a large, diverse patient population. METHODS: We performed a retrospective cohort study using ECG and echocardiography (ECHO) data from a large metropolitan health system. All patients had one ECG and ECHO on file, obtained within 1 week of each other. Sensitivity and specificity of ECG for left ventricular hypertrophy were determined by comparing results from the MUSE (R) 12-SL (GE Healthcare, Chicago, IL) computer-generated algorithm for ECG to ECHO left ventricular mass index. Subgroup analyses of individual patient characteristics were performed with corresponding chi-squared analyses to determine significance. RESULTS: A total of 13,960 subjects were included in the study. The typical subject was 60 years of age or older, female, overweight, and hypertensive, and demonstrated low socioeconomic status. The sensitivity and specificity of ECG for left ventricular hypertrophy in the overall cohort were 30.7% and 84.4%, respectively, with multiple patient characteristics influencing these results. CONCLUSIONS: This is the first study to confirm the sensitivity and specificity of ECG for left ventricular hypertrophy in a large, diverse patient population with significant minority representation. Furthermore, although these statistical measures are influenced by patient characteristics, such differences are likely not clinically significant. (C) 2020 Elsevier Inc. All rights reserved.