Variability of the electrocardiogram in normal young men

Abstract

Electrocardiograms were taken on twelve normal young men in the basal state on eleven different occasions over a period of two months. Three standard and three chest leads were used, and thirty-five electrocardiographic items were measured. In addition, one set of electrocardiograms was measured by the same assistant on three different occasions. The amount of variation contributed by interindividual differences, by day-to-day physiologic fluctuation, and by inaccuracy in measuring the records was determined statistically. The ratio of the intraindividual to the interindividual variance was used as the criterion of the consistency of repeated electrocardiographic determinations. Out of thirty-five electrocardiographic items, the consistency was very high (consistency coefficient ≧ 0.90) in fourteen, high (0.89 to 0.80) in seven, moderate (0.79 to 0.70) in three, low (0.69 to 0.60) in three, and very low (< 0.60) in eight. The consistency measures were corrected for the homogeneity of the sample of subjects by substituting estimates of interindividual variance based on much larger samples, and for the inaccuracy of measurement of the electrocardiographic records. The 90 per cent expectancy range of intraindividual variability was calculated. These data may be applied for interpretation of borderline electrocardiograms and for evaluation of degrees of deterioration or improvement in serial electrocardiograms.