Abstract 12894: Model Selection for Pediatric Coronary Diameter Curves in Japan

Abstract

Introduction: The Z score project began in May 2010 as an effort to develop a Pediatric Coronary Diameter Z score Calculator, which could be used to easily and quickly compute coronary diameter Z scores. Participants in the project, who included children without pre-existing diseases, were asked to register the following data left main coronary (LM) artery, left anterior descending (LAD) artery, left circumflex (LCX) artery, right coronary artery (RCA), age, sex, height, and weight. Although the estimated necessary sample size for the Z score project was 4000, valid data has been collected on 4118 children as of December 2011. The objective of this study was to derive pediatric coronary diameter curves in Japan.

Methods: We analyzed the data with the LMS method, using age, height, and body surface area (BSA). BSA was calculated by the Du Bois expression and the Haycock expression (BSA and BSA h, respectively). We compared the relative goodness of fit of these models using Akaike’s information criterion (AIC), generalized AIC (GAIC), and Schwarz Bayesian Criteria (SBC).

Results: All criterions revealed the age and the height models to be the poorest in predicting coronary diameters in the pediatric population. Additionally, the goodness of fit of the BSA and BSA h models were quite similar (Table).

Conclusions: The BSA and BSA h models were better at predicting pediatric coronary diameter by comparison to the age or height models. Since BSA estimation by the Du Bois expression is quite popular in Japan, we developed the Pediatric CoronaryDiameter Z score Calculator using BSA.