Factors predicting myopia incidence in China and Europe

To develop a predictive model for myopia incidence using population-based data from Chinese and European children.

Analysis of four longitudinal studies from the UK, Sweden and China was conducted. Data from 4405 non-myopic children aged 6 to 16.8 years with spherical equivalent (SE) refraction from -0.49 D to +10.00 D were analysed. Kaplan-Meier and Cox proportional hazard models were used to evaluate the probability of myopia incidence by 12 and 24 months based on age, sex, parental myopia, cycloplegic SE, axial length and axial length/corneal radius of curvature (AL/CR). Hyperopic reserve was defined as the minimum level of hyperopia required to provide <10% probability of developing myopia by 24 months.

The cumulative incidence of myopia by the 24th month was 18.8% for Chinese and 6.7% for European (p < 0.001) populations. Based on multivariate Cox regression, a greater risk of myopia incidence was seen in Chinese children, younger ages, females, children with parental myopia, less hyperopic SE and higher AL/CR. SE had the highest predictive accuracy (C statistic 0.90). The AL/CR ratio had a greater predictive accuracy (C statistic 0.75) than axial length alone (C statistic 0.63). Predictive accuracy of all variables was similar between ethnicities (p > 0.50) apart from axial length, which was higher in Chinese children (C statistic 0.65) versus European children (C statistic 0.55, p = 0.04). To avert the risk of myopia, Chinese eyes required a greater hyperopic reserve ranging from 0.5 to 1.5 D compared to European eyes ranging from 0 to 0.5 D, depending on age and other predictive factors.

SE is the strongest predictive factor for both ethnicities. The influence of predictive factors is similar between ethnicities/regions though Chinese children have a greater risk of developing myopia and require a higher hyperopic reserve. These data could be useful for developing a predictive tool of myopia incidence.