Short‐term repeatability of electrocardiographic criteria of left ventricular hypertrophy

Background Left ventricular hypertrophy (LVH) is a marker of cardiac end‐organ damage and a risk factor for cardiovascular morbidity and mortality. Although clinical trials and cohort studies commonly use the electrocardiogram (ECG) for LVH assessment, the repeatability of ECG‐LVH criteria has not been sufficiently examined. Therefore, we evaluated the repeatability of ECG‐LVH criteria. Methods Participants ( n  = 63) underwent two standard ECGs at each of two visits, two weeks apart. The ECGs were processed centrally to calculate Cornell voltage (CV) LVH, Cornell voltage product (CVP) LVH, Sokolow–Lyon (SL) LVH, and Sokolow–Lyon product (SLP) LVH. We also used the waveforms measurements contributing to these LVH criteria as continuous variables, referred to here as CV‐index, CVP‐index, and SL‐index. We calculated the intraclass correlation coefficient (ICC), minimal detectable change (95% confidence), and the prevalence‐adjusted bias‐adjusted kappa (PABAK). Results ICCs (95% confidence intervals (CI)) were 0.97 (0.96, 0.98) for CV‐index, 0.97 (0.95, 0.98) for CVP‐index, and 0.93 (0.90, 0.96) for log of SL‐index. Minimal detectable change between repeat measures of CV‐index, CVP‐index, and log of SL‐index were ≥236.7 mV, ≥26.7 mV, and ≥0.09 mV, respectively. The within‐visit PABAK was 1 for all ECG‐LVH criteria, except for the first visit SLP‐LVH (PABAK = 0.93). Between‐visit PABAK ranged from 0.83 to 0.97 across LVH criteria. Conclusions CV, CVP, and SL ECG‐LVH as continuous variables have excellent repeatability, and as binary variables have excellent within‐visit agreement and good between‐visit agreement. These results alleviate concerns about the repeatability the ECG‐LVH use in clinical trials and epidemiologic studies.