Abstract
Purpose: :
The SCORE Study primary outcome papers demonstrated only weak and negative correlation between changes from baseline in center point thickness (CPT) by optical coherence tomography and contemporaneous changes in visual acuity letter score (VA) by the electronic Early Treatment Diabetic Retinopathy Study (E-ETDRS) . We now investigate associations between CPT and subsequent VA.
Methods: :
The SCORE Study evaluated intravitreal triamcinolone injection(s) (1mg or 4mg) versus standard care for vision loss associated with macular edema secondary to central or branch retinal vein occlusion (CRVO or BRVO). VA and CPT were measured at baseline and at 4-month intervals through Month 36. Auto-and cross-correlations and ordinary least squares were used to identify significant relationships between non-contemporaneous CPT and VA. Family-wide type I error was controlled by Hochberg’s method.
Results: :
Over 90% of the correlations between non-contemporaneous CPT and VA changes are negative, many significantly so even after Hochberg adjustment. In BRVO, there are more Hochberg-significant correlations between VA and prior CPT changes than between contemporaneous VA and CPT changes. The correlation between early CPT and later VA changes typically exceeds in magnitude the corresponding correlation between early VA and later CPT changes. However, autocorrelations of changes in CPT and changes in VA are much larger. In a regression to predict VA, lagged VA measurements were important predictors, but not lagged CPT. Using prior VA reduces root mean squared prediction error of current VA by 20-60%, while adding prior CPT reduces prediction error by less than 5-6%.
Conclusions: :
There is a weakly predictive relation between current CPT and VA four months later, but a far better predictor is current VA.
Clinical Trial: :
www.clinicaltrials.gov NCT00105027
Keywords: retina • visual acuity • imaging/image analysis: clinical