Comparison of QT Correction Methods in the Pediatric Population of a Community Hospital: A Retrospective Study

Accurate measurement of QT interval, the ventricular action potential from depolarization to repolarization, is important for the early detection of Long QT syndrome. The most effective QT correction (QTc) formula has yet to be determined in the pediatric population, although it has intrinsically greater extremes in heart rate (HR) and is more susceptible to errors in measurement. The authors of this study compare six different QTc methods (Bazett, Fridericia, Framingham, Hodges, Rautaharju, and a computer algorithm utilizing the Bazett formula) for consistency against variations in HR and RR interval.

Descriptive Retrospective Study. We included participants from a pediatric cardiology practice of a community hospital who had an ECG performed in 2017. All participants were healthy patients with no past medical history and no regular medications.

ECGs from 95 participants from one month to 21 years of age (mean 9.7 years) were included with a mean HR of 91 beats per minute (bpm). The two-sample paired t-test or Wilcoxon signed-rank test assessed for any difference between QTc methods. A statistically significant difference was observed between every combination of two QTc formulae. The Spearman’s rank correlation analysis explored the QTc/HR and QTc/RR relationships for each formula. Fridericia method was most independent of HR and RR with the lowest absolute value of correlation coefficients. Bazett and Computer had moderate correlations, while Framingham and Rautaharju exhibited strong correlations. Correlations were positive for Bazett and Computer, reflecting results from prior studies demonstrating an over-correction of Bazett at higher HRs. In the linear QTc/HR regression analysis, Bazett had the slope closest to zero, although Computer, Hodges, and Fridericia had comparable values. Alternatively, Fridericia had the linear QTc/RR regression coefficient closest to zero. The Bland-Altman method assessed for bias and the limits of agreement between correction formulae. Bazett and Computer exhibited good agreement with minimal bias along with Framingham and Rautaharju. To account for a possible skewed distribution of QT, all the above analyses were also performed excluding the top and bottom 2% of data as sorted by heart rate ranges (N = 90). Results from this data set were consistent with those derived from all participants (N = 95).

Overall, the Fridericia correction method provided the best rate correction in our pediatric study cohort.