The aim of our study was to determine the most appropriate mathematical model for right ventricular (RV) function normal values in children.

Normal children aged 0 to 18 years referred to the pediatric cardiology consultation were screened in this prospective cross-sectional study. Each RV parameters (S’, E’, A’ waves, TEI index, TAPSE) was modeled with four explanatory variables: age, size, weight and body surface area. Four different mathematical models were successively applied: a linear model, a polynomial model of degree 2, a linear regression model by spline with free knot, and a polynomial regression model of degree 2 by spline with a free knot. The choice of the best explanatory variable and the best modelization method was made on the basis of the model's coefficient of determination (R²).

We included 314 normal children aged 2 days to 18 years (46% female, 88 infants, 26 neonates, 226 children). The modelization of the RV variables over all ages did not follow a linear model, with the lowest coefficients of determination in the four explanatory variables. The quadratic spline model best adjusted the modelization of all RV variables. No model accurately explained the TEI index in any of the four explanatory variables. The S’ wave was best modelized by the quadratic spline model with the weight. The knot at 6.86kg corresponded to the curves’ break. The E’ wave was best modelized by the quadratic spline model with the age. The knot at 0.29 years corresponded to the curve's break. The A’ wave was best modelized by the quadratic model with the size. The knot at 81cm corresponded to the break of the curve. The TAPSE was best modelized by the quadratic spline model with the weight (Fig. 1). The knot at 9.04kg corresponded to the curve's break (Fig. 1).

The quadratic spline model best applies to echocardiographic RV reference values in pediatrics.