Introduction: Electrocardiography is an essential tool to assess the liability of test articles to produce torsade de pointes. The number of cardiac cycles that must be measured from a dog to accurately characterize the relationship between RR and QT intervals, and thus assess this liability, is unknown.
Methods: In this study, electrocardiograms (ECGs) were obtained from 12 conscious dogs with sinus rhythm. In each dog, RR and QT intervals were measured for 12 cardiac cycles. Measurements for each were then averaged over all 12 cycles, and those results compared to the average of both the initial 6 and 3 cycles, as well as to the middle cycle alone, for 12, 6, and 4 of the dogs. QTc was calculated by dividing each QT by the cube root of the preceding RR interval.
Results: We found no significant differences in the results of measurements of RR, QT, or QTc obtained from 12, 6, 3, or 1 cycle, whether from 12, 6, or 4 dogs. Intraobserver variability of ECG measurements was tested by having a single observer measure 10 copies of 12 different ECGs. The greatest coefficient of variation (S.D./mean) for the measurement of any ECG parameter was less than 2.5%.
Discussion: We conclude that measurements of RR and QT intervals made by a trained observer from 1 cardiac cycle accurately reflect those that are averaged from 3, 6, or 12 cycles whether the number of dogs per group is 12, 6, or 4.