Reproducibility of electrocardiographic criteria for left ventricular hypertrophy in hypertensive patients in general practice

Before changes in ECG voltage criteria can be accepted as evidence for changes in LVH, the variability of ECG measurements must be known. Here we report on the results of a study, on the variability of electrocardiographic single lead voltage parameters and (voltage) criteria for left ventricular hypertrophy in hypertensive patients in general practice. Two electrocardiograms were recorded, from 64 patients at an interval of 2 min, to measure the minute-to-minute variability. From 77 patients, two electrocardiograms were recorded at an interval of one week to measure the day-to-day variability. The coefficient of variation of voltage parameters for single leads ranged in the day-to-day group from 9.2% in R/I to 42% in T/V1, and the coefficient of variation for voltage combinations in this group ranged from 10.0% for the Sokolow-Lyon criteria to 13.7% for Gubner-Ungerleider criteria. The reclassification percentages in the day-to-day group ranged from 0% for Gubner-Ungerleider to 17% for Minnesota code criteria. A factor analysis showed that studies which use the Romhilt-Estes score, the Sokolow-Lyon or the Minnesota criteria to detect ECG-LVH are not comparable with studies which use the Cornell or Gubner-Ungerleider criteria.