How should we determine normal echocardiographic right ventricle function reference values in pediatrics?

Area of Science:

• Pediatric Cardiology
• Biomedical Engineering
• Medical Imaging Analysis

Background:

• Pediatric Z-scores for echocardiographic right ventricle (RV) variables are essential for assessing cardiac health.
• Existing models for RV function reference values in children lack a standardized, optimal mathematical approach.
• Accurate RV assessment is vital for diagnosing and managing pediatric cardiac conditions.

Purpose of the Study:

• To identify the most appropriate mathematical model for establishing pediatric Z-scores of echocardiographic right ventricle (RV) function.
• To compare the efficacy of linear, polynomial, and spline regression models in parameterizing RV variables.
• To determine the optimal explanatory variables (age, height, weight, body surface area) for RV Z-score modelization.

Main Methods:

• A prospective cross-sectional study involving 314 healthy children aged 2 days to 18 years.
• Echocardiographic RV parameters (S', E', A' waves, TEI index, TAPSE) were measured.
• Four mathematical models (linear, quadratic, linear spline, quadratic spline) were applied to RV variables using age, height, weight, and BSA as predictors.
• The adjusted coefficient of determination (aR² ) was used to select the best-fitting model.

Main Results:

• RV variable modelization did not follow a simple linear pattern.
• A single explanatory variable was insufficient for all Z-scores; specific, independent variables were required per parameter.
• Quadratic spline regression models demonstrated the best fit for most RV variables.
• Specific optimal predictors were identified: weight for S' and TAPSE, age for E', and height for A' waves.

Conclusions:

• Spline regression models provide a superior fit for echocardiographic right ventricle reference values in pediatrics compared to linear or simple polynomial models.
• The selection of specific explanatory variables (weight, age, height) is critical for accurate RV Z-score calculation in children.
• These findings support the use of advanced statistical modeling for robust pediatric echocardiographic reference standards.