First Author: G.Menon UK
Co Author(s): Y. Yang B. Reeves F. Ghanchi U. Chakravarthy 0 0 0 0 0 0 0 0 0
Back to previous
Among 297 participants with wet AMD assigned to discontinuous anti-VEGF treatment using ranibizumab or bevacizumab in the IVAN trial, we sought to quantify associations between (1) lesion area and activity with baseline visual function, and lesion characteristics and visual function with (2) reading 68 ETDRS or more at the last visit, (3) time to lesion reactivation after initial 3 injections and (4) the need for 3 or fewer injections/year.
This was a multicentre trial conducted in the UK. IVAN Study group.
Wet AMD lesions at baseline line were classified as more (presence of haemorrhage or classic CNV) or less active and by quartile of lesion area. Predictors investigated (all available after initial 3 injections) were: lesion characteristics and ETDRS distance acuity, logMAR near visual acuity (NVA) and Pelli-Robson contrast sensitivity (CS) at baseline and 3 months. Data were analysed by multiple linear, logistic or Cox regression, using the likelihood ratio test to optimize model fit.
More active (-9.4 letters, 95%CI -12.7 to -6.1, p<0.0001) and larger lesions (-1.6 letter per quartile, 95%CI -3.0 to -0.1, p=0.03) were independently associated with reading fewer ETDRS letters at baseline; similar associations were observed for NVA and CS. ETDRS letters read at baseline and 3 months independently predicted reading 68 ETDRS or more at the last visit (pseudo r2 = 0.34, p<0.0001). Median (95% CI 44% to 56%) time to lesion reactivation was 78 days. Larger lesion area (but not activity) and worst quartile of ETDRS letters at baseline predicted longer time to lesion reactivation (p<0.002). Lesion area and NVA at baseline were the best independent predictors of needing 3 or fewer injections/year but the model fitted poorly (pseudo r2 = 0.04, p=0.04) and neither association was linear.
Participants presenting with larger and more active lesions had worse baseline visual function. ETDRS letters read at baseline and 3 months, but not lesion characteristics, predicted good ETDRS outcome. The factors investigated were unable to predict the need for few injections.