P2A098: Comparison of QT Correction Methods in the Pediatric Population of a Community Hospital: A Retrospective Study

Background: Accuracy of QT measurement is important for early detection of congenital long QT syndrome. As QT interval is strongly affected by heart rate (HR), several formulae have been proposed to correct for this relationship. The ideal formula would eliminate any HR dependence of QT. Although the performances of different QT correction (QTc) formulae have been well-explored in adults, they have been less studied in the pediatric population that has intrinsically greater extremes in HR and is, therefore, more susceptible to errors in measurement.

Methods: This is a descriptive retrospective study. We included patients in the pediatric cardiology practice of a community hospital who had an ECG performed in the year 2017. Patient demographics were evaluated using mean and standard deviation for continuous variables, and frequency and percent for categorical variables. Spearman’s rank correlation tests were used to evaluate for any correlation between two variables. Two-sample paired t-test or Wilcoxon signed-rank test, as appropriate, were used to test the difference in every two QTc values (p-value was adjusted with fdr method because of multiple comparisons). Scatter plots were drawn to observe relationships between two variables and corresponding univariate linear regression lines were added. Linear regression models were used to determine the association between QTc value and HR or RR. All p-values are two-sided with statistical significance evaluated at the 0.05 alpha level. Bland-Altman Plots were used to assess the agreement between two QTc formulae. All analyses were performed in R Version 4.1.2 (The R Foundation for Statistical Computing).

Results: ECGs from 95 patients from ages 1 month to 21 years (mean 9.7 years) were included. The average HR was 91 bpm. Two-sample paired t-test or Wilcoxon signed-rank test, as appropriate, showed a statistically significant difference between every two QTc formulae. Fridericia was most independent of HR and RR with the lowest absolute value of Spearman's coefficients. Bazett and Computer had moderate correlations, while Framingham and Rautaharju exhibited strong correlations. Over-correction at higher HRs were observed for Bazett and Computer. Bazett had the linear QTc/HR regression slope closest to zero; computer, Hodges, and Fridericia, however, had comparable values. Alternatively, Fridericia had the linear QTc/RR regression nearest zero. Bazett and Computer had good agreement with minimal bias in the Bland-Altman analysis, along with Framingham and Rautaharju.

Conclusion: Overall, Fridericia correction formula provided the best rate correction in our study cohort. Our findings support the use of this formula in the pediatric population, although limitations of our study, including a small sample size, must be considered. If Bazett, the most commonly implemented formula, represents a valid and reliable method for QTc calculation in a particular population, Computer algorithm may be an appropriate and alternative method to facilitate QTc measurement.