Age-continuous centile curves reveal upper-tail divergence of anisometropia in 389 056 schoolchildren: implications for risk stratification in myopia control

Objective To characterise the age-continuous distribution of anisometropia in a large school-based cohort, derive age-specific centile curves and determine whether age-related increases in anisometropia reflect a generalised distributional shift or selective divergence in the upper tail. Methods and analysis This school-based cross-sectional study included 389 056 children aged 6–18 years from five prefectures of Xinjiang, China, examined from March to June 2025 after prespecified exclusions. Spherical-equivalent anisometropia (Aniso_SE) was the absolute interocular difference in spherical equivalent; clinically significant anisometropia was defined as Aniso_SE≥1.00 dioptre (D). Age-specific prevalence was modelled using restricted cubic splines; centile curves were generated by quantile regression; and school-level clustering was accounted for using mixed-effects models. Results Mean age was 11.79 years and 50.2% were male. Mean Aniso_SE was 0.44 D and 42 917 participants (11.03%) had clinically significant anisometropia. Prevalence increased non-linearly with age, peaking at approximately 17%–19% around age 16 years. The median remained stable at about 0.25 D, whereas the 90th percentile exceeded 1.00 D by age 12 years and reached 1.4 D by age 16 years. Age and myopia severity were independently associated with clinically significant anisometropia. Conclusions Age-related worsening of anisometropia was driven by selective upper-tail divergence rather than a generalised distributional shift. Age-continuous centile curves may help identify children requiring closer monitoring in school-based myopia control programmes.