Background: P-wave indices of maximum P-wave duration and P-wave dispersion have been examined in a broad array of cardiovascular and noncardiovascular disease states. The P-wave indices literature has been highly heterogeneous in measurement methodologies, described quality control metrics, and distribution of values. We therefore sought to determine the reproducibility of P-wave indices in a community-based cohort.
Methods: P-wave indices were measured in sequential subjects enrolled in the Framingham Heart Study. Electrocardiograms were obtained at the 11th biennial visit of the Original Cohort (n = 250) and the initial visit of the Offspring Cohort (n = 252). We determined the mean P-wave durations, interlead correlations, and P-wave indices. We then chose 20 ECGs, 10 from each cohort, and assessed intrarater and interrater variability.
Results: The maximum P-wave duration ranged from 71 to 162 ms with mean of 112 + or - 12 ms. The minimum P-wave duration ranged from 35 to 103 ms with mean of 65 + or - 10 ms. P-wave dispersion ranged from 12 to 82 ms. The mean P-wave dispersion was 48 + or - 12 ms (40-56). The intrarater intraclass correlation coefficient (ICC) was r = 0.80 for maximum P-wave duration and r = 0.82 for P-wave dispersion. The interrater ICC was 0.56 for maximum P-wave duration and 0.70 for P-wave dispersion.
Conclusions: We demonstrated excellent intrarater reproducibility and fair interrater reproducibility for calculating P-wave indices. Reproducibility is frequently lacking in studies of P-wave indices, but is an essential component for the field's growth and epidemiologic contribution.