Head Circumference for Growth Curves and Percentiles for 0-2-Year-Old Children in Ankara: Comparison with GAMLSS and Quantile Regression Methods.

Abstract

Objective: Growth curves are important tools for assessing the growth and development of children in the target population by age. Measuring head circumference (HC) in infants is an important tool in monitoring infant health and brain development. The aims of this study are to construct current 0-2-year-old HC growth curves and percentiles and compare the methods used in the construction of growth curves for HC measurements by gender. Materials and Methods: The study is a retrospective research that includes the HC measurements of a total of 2832 (n = 1438 girls, n = 1394 boys) children examined between 2018 and 2021 in Başkent University hospital, Ankara. Lambda-Mu-Sigma, LMSP, and LMST methods based on Box-Cox Cole and Green (BCCG), Box-Cox power exponential (BCPE), and Box-Cox t (BCT) distributions respectively, and quantile regression (QR) method were used to construct the fitted growth curves. Model performances were evaluated using the generalized Akaike information criterion. The analyses were conducted using R 4.1.2 version. Results: According to the LMS and QR methods, percentile values of HC measurements by gender were calculated and the results were compared. Smoothed HC growth curves were constructed and compared for both methods by gender. The present study, Neyzi et al study, and World Health Organization (WHO) standards were compared for the third, 50th and 97th percentiles of HC. Conclusion: This study showed that the third-97th HC percentiles calculated by LMS and QR methods are very close to each other. Additionally, this study showed that the HC percentiles of Turkish children were slightly different compared to WHO standards.