Z-score regression model for coronary artery diameter in healthy Chinese Han children

Abstract

Background

Coronary artery distension and aneurysm are complications of Kawasaki disease in children.

Purpose

To develop a Z-score regression model for coronary artery diameter in children that could be used as reference.

Material and Methods

This retrospective analysis included children with normal heart structure between March 2013 and April 2017. Body surface area (BSA) was calculated. The diameters of the right coronary, left main coronary, left anterior descending, and circumflex arteries were measured by echocardiography. Pearson correlation analysis was used to establish linear, exponential, logarithmic, power, and square root regression models.

Results

The analysis included 509 children (280 boys) aged 1 day to 15.2 years. Coronary artery diameters were significantly correlated with age, height, body mass, BSA, and $\sqrt{\mathrm{BSA}}$ (r = 0.663–0.826; P < 0.05), with a stronger correlation for $\sqrt{\mathrm{BSA}}$ than BSA (P < 0.05). The adjusted determination coefficients ( ${\mathrm{R}}_{\mathrm{a}}^2$ ) were higher for the exponential and square root models than for the other models (P < 0.05). The random error term variance was constant for the exponential model (P > 0.05), and processing with the weighted least-square methods eliminated heteroscedasticity in the other models. The Z-scores were normally distributed for the exponential and square root models (P > 0.05).

Conclusion

Overall, the square root model was the optimal equation for the calculation of coronary artery Z-score in Chinese Han children. This model could be used to facilitate the diagnosis of coronary artery distension in children with suspected Kawasaki disease.