Abstract
Purpose :
Research using conventional linear models concluded that corneal biomechanics were compromised in high myopia. We investigated whether the linearity assumption was appropriate and re-examined these associations across different levels of myopia.
Methods :
Myopic (spherical equivalent refraction, SER < -0.25D) eyes of 12,392 adults aged 40-69 years without any history of systemic and ocular conditions were identified in the UK Biobank. Data from both eyes were averaged. Ordinary least squares (OLS) regression was employed to test the linear association between corneal hysteresis (CH) or corneal resistance factor (CRF), separately, and SER while controlling for age, sex, corneal radius and intraocular pressure. Quantile regression (QR) was used to test the same set of associations across 49 equally spaced conditional quantiles of SER, controlling for the same covariates.
Results :
In OLS regression, every standard deviation (SD) decrease in CH and CRF was associated with 0.07D (95% CI: 0.03-0.11; p=0.001) and 0.08D (95% CI: 0.03-0.13; p=0.001) higher myopia. However, residual analyses strongly indicated that the linearity assumption was violated (Figure 1). QR revealed no evidence of a significant association between CH/CRF and SER towards the upper end of the SER distribution (lower myopia), but a significant (p<0.05) positive association became evident from -3D, i.e., every SD decrease in CH and CRF was associated with 0.07D and 0.08D higher myopia (Figure 2). The magnitude of association increased exponentially with increasing myopia: in the -5D quantile, every SD decrease in CH and CRF was associated with 0.19D (95% CI: 0.10-0.28; p<0.001) and 0.24D (95% CI: 0.12-0.36; p<0.001) higher myopia; in the -8D quantile, this further increased to 0.45D (95% CI: 0.26-0.64; p<0.001) and 0.57D (95% CI: 0.33-0.81; p<0.001) higher myopia per SD decrease in CH and CRF.
Conclusions :
Corneal biomechanics appeared compromised at a much lower level of myopia than previously thought. Of note, these changes were observed to be exponential (not linear) with increasing myopia.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.