Abstract
Purpose:
Intraocular pressure (IOP) and central corneal thickness (CCT) are important risk factors for glaucoma progression. However, it is uncertain whether CCT risk is independent or mediated through IOP measurement error. The Ocular Response Analyzer (ORA) and Dynamic Contour Tonometer (DCT) make IOP measurements less affected by corneal biomechanics than the Goldmann applanation tonometer (GAT). Our hypothesis was that the correlation between the rate of visual field (VF) loss and ORA IOP and DCT IOP is stronger than with GAT IOP.
Methods:
516 previously untreated open angle glaucoma patients were monitored over 2 years in the United Kingdom Glaucoma Treatment Study with VF testing, and tonometry with the GAT, DCT and ORA at 11 visits. The eye with the most negative baseline mean deviation (MD) was selected and only eyes with VFs over at least 5 visits were included. Pearson’s correlation coefficient (R) was used to measure the strength of the relationship between median follow-up IOP [GAT, DCT, corneal compensated IOP (IOPcc) and Goldmann correlated IOP (IOPg)] and MD slope. IOP models, with CCT, corneal hysteresis (CH), corneal resistance factor (CRF) and axial length (AL) included as covariates, were compared using the Akaike Information Criterion (AIC) and Akaike weights (AWs). Data analysis was performed with the statistical software R.
Results:
342 eyes of 342 patients were included. Pearson’s R of the MD slope against median GAT IOP, DCT IOP, IOPcc and IOPg was 0.263, 0.230, 0.294 and 0.244, respectively (P<0.001 in all cases). The AIC values (AWs) for the GAT, DCT, IOPcc and IOPg models were 1006 (0.039), 1012 (0.002), 1000 (0.953) and 1010 (0.007), respectively. IOPcc had the highest AW, suggesting it had the highest probability (95.3%) of being the best predictor of VF progression. Following inclusion of CCT, CH, CRF and AL in each IOP model, a backward stepwise AIC approach was applied to remove non-significant predictors; the best two models were IOPcc alone (AIC=1000, AW=0.69) and GAT IOP with CH (AIC=1001, AW=0.31). CCT did not contribute to any model and, on its own, was not correlated to the MD slope (R=-0.0047).
Conclusions:
IOP is a relatively poor predictor of glaucomatous progression, but IOPcc and GAT IOP with CH measured by the ORA, were the best models of VF progression in this study.
Keywords: 568 intraocular pressure •
466 clinical (human) or epidemiologic studies: treatment/prevention assessment/controlled clinical trials •
629 optic nerve