Identifying children at risk of high myopia using population centile curves of refraction

Purpose: To construct reference centile curves of refraction based on population-based data as an age-specific severity scale to evaluate their efficacy as a tool for identifying children at risk of developing high myopia in a longitudinal study. Methods: Data of 4218 children aged 5-15 years from the Guangzhou Refractive Error Study in Children (RESC) study, and 354 first-born twins from the Guangzhou Twin Eye Study (GTES) with annual visit were included in the analysis. Reference centile curves for refraction were constructed using a quantile regression model based on the cycloplegic refraction data from the RESC. The risk of developing high myopia (spherical equivalent ≤ -6 diopters [D]) was evaluated as a diagnostic test using the twin follow-up data. Results: The centile curves suggested that the 3^rd, 5^th, and 10^th percentile decreased from -0.25 D, 0.00 D and 0.25 D in 5 year-olds to -6.00 D, -5.65D and -4.63 D in 15 year-olds in the population- based data from RESC. In the GTES cohort, the 5th centile showed the most effective diagnostic value with a sensitivity of 92.9%, a specificity of 97.9% and a positive predictive value (PPV) of 65.0% in predicting high myopia onset (≤-6.00D) before the age of 15 years. The PPV was highest (87.5%) in 3^rd centile but with only 50.0% sensitivity. The Mathew's correlation coefficient of 5^th centile in predicting myopia of -6.0D/-5.0D/-4.0D by age of 15 was 0.77/0.51/0.30 respectively. Conclusions: Reference centile curves provide an age-specific estimation on a severity scale of refractive error in school-aged children. Children located under lower percentiles at young age were more likely to have high myopia at 15 years or probably in adulthood.