Factors predicting the rate of eye growth in Chinese children at different ages

Abstract

Purpose: To explore the prediction of myopia development from eye shape obtained by MRI in school children.

Methods: Healthy students were recruited from primary school (children, n=39, 11.5 ± 1.0y) or high school (adolescents, n=28, 17.7 ± 1.1y) in Wenzhou. Axial lengths (AL) were measured by IOLMaster, and the axial and cycloplegic peripheral spherical equivalents (SEs) were measured by autorefraction. MRI was used to measure transverse and sagittal planes through the center of the eyes. AL measured using MRIs were compared with that using IOLMaster to verify best sections for analysis of VC depth and measures of ocular shape (EQ/AL). The children were remeasured after 3.7 ± 1.3 ys and the adolescents after 4.6 ± 1.9 ys. The refraction rate and the AL rate was used as the dependent variable in linear regression models where the baseline data were used as independent covariates.

Results: In both age groups, regardless of refractive state, IOLMaster and MR images of AL were significantly correlated (R²=0.90), and the MR scanning localization in transverse and sagittal sections were in good agreement (R²=0.86). Though the adolescents had higher myopia (-0.66 ± 1.10 D) and longer AL (24.15 ± 0.91 mm) than the children (-0.41 ± 0.76 D; 23.66 ± 0.63 mm), children had more active eye growth (dAL/D/yr: 0.13 ± 0.46 vs. 0.47 ± 0.50, p=0.006). The eye shape index from MRIs in both groups did not differ in transverse plane but in the sagittal plane the index in children was higher than that in adolescents (1.01 ± 0.04 vs. 0.98 ± 0.03, p<0.001). The best single baseline predictor of the change in SEs in children and the change in AL in adolescent was the on-axis SE at baseline, but it did not contribute to the multiple models. The significant predictors to the multiple models included baseline AL, the asymmetry from nasal to temporal 30^o off-axis, sagittal eye shape index, the difference of the VCD from temporal 30^o to on-axis, and the relative defocus on the tangential image shell at temporal off-axis 30° in both groups when the AL change as dependent variable (children: R²=0.327, p=0.029; adolescent: R²=0.376, p=0.016); and only in children (R²=0.312, p=0.038), not in adolescents (R²=0.267, p=0.103) all those parameters contributed to the multiple models when the SE change as dependent variable.

Conclusions: The eye shape index in sagittal section had contributed more to predicting eye growth in children compared to adolescents.