A multiple regression approach to study optical components of myopia in Singapore school children

Previous studies on the optical components of childhood myopia adjusted only for age and gender or reported only the change of individual components with time. We perform a study to assess the association of the optical components of the eye contributing to the degree of myopia by building a statistical model which can better assess the individual effects of each optical component.

Children 6-12 years old with myopia of -1 to -6 D were used in this study. Autokeratometry was performed to determine the mean corneal power in two perpendicular meridia. Ultrasonography was used to determine the vitreous chamber depth (VCD). Ultrasonography and Scheimflug biometry were used to determine lens thickness (LT) and anterior chamber depth. A stepwise multiple linear regression was used to determine statistically significant independent variables that explain the variation in the mean spherical equivalent of the left eye. The independent variables investigated were anterior chamber length, LT, VCD, mean corneal power, age and gender. The first regression model used measurements of anterior chamber depth and LT from ultrasonography, while the second model used measurements from Scheimflug photography. Statistical significance was set at p < 0.05.

Both statistical models have the same significant independent variables - VCD, LT and gender, with similar estimates. In both models, VCD was found to have the strongest partial association with mean spherical equivalent, followed by gender and LT. A more myopic refractive error was associated with (1) a longer eye, (2) an increased LT, and (3) female gender.

Myopia was associated with a longer vitreous chamber. This remained the primary determinant of myopia in Chinese children after consideration of other optical components. The effect of gender and LT on myopia requires further study. These results are in agreement with multiple regression analyses in Singapore adults.