Interobserver and inter-cycle variability in the electrocardiographic assessment of dogs

Abstract

Abstract: The electrocardiogram, which graphically records the phases of the cardiac cycle, plays a crucial role in diagnosing arrhythmias and conduction abnormalities. However, the reliability of results can be significantly affected by the level of expertise of the evaluator. Acknowledging potential discrepancies in interpretations of the same electrocardiographic readings across different cardiac cycles and between analysts of varying experience, this study aimed to assess the interobserver and inter-cycle variability in electrocardiographic exams of dogs. This retrospective research was carried out at the Veterinary Hospital Superintendency Unit of the Federal University of Fronteira Sul, Realeza Campus, PR, Brazil. The study involved a random selection of electrocardiographic records from 50 dogs, collected between September 2018 and December 2019, without discrimination based on breed, sex, age, or weight. Various parameters were analyzed, including heart rate and rhythm; durations of P waves; QRS complexes; PR and QT intervals; amplitudes of P, R, and T waves; and both atrial and ventricular electrical axes. The data underwent Shapiro-Wilk normality testing, with parametric data analyzed via analysis of variance and Tukey's test, and non-parametric data assessed using the Kruskal-Wallis followed by Dunn's tests. The final step involved applying the Kappa Agreement Coefficient Test to the data. The findings revealed variability in heart rhythm interpretation, with disagreement between evaluators of low to intermediate experience levels. Specifically, low-level evaluators tended to underestimate maximum heart rates, recording a median rate of 105, compared to 122 by high-level evaluators (p=0.0100). Additionally, P wave durations were consistently overestimated by low- and intermediate-level evaluators, who reported median durations of 52 and 51, respectively, against 48 by their highly experienced counterparts (p=0.0064). No significant differences emerged in analyses comparing a single cardiac cycle to either three cycles or the average of three cycles. The study highlighted the impact of evaluator expertise on the variability of electrocardiographic interpretations, particularly concerning P wave duration, for which the evaluators tended to overestimate measurements, and the degrees of agreement, which did not represent their experience level. The findings stress the importance of training and routine practice in electrocardiographic examinations to mitigate the risk of mistaken interpretations and analyses.