Background/aims To evaluate baseline low-luminance visual acuity (LLVA) as a predictor of visual acuity improvement in patients with neovascular (wet) age-related macular degeneration (wAMD) receiving antivascular endothelial growth factor A (anti-VEGF) therapy.

Methods In the HARBOR trial, 1084 treatment-naïve patients ≥50 years of age with subfoveal wAMD received intravitreal ranibizumab 0.5 or 2.0 mg monthly or as needed. To measure LLVA, patients read a normally illuminated ETDRS (Early Treatment Diabetic Retinopathy Study) chart with a neutral density filter placed in front of the study eye. Patients were assigned into quartiles based on the magnitude of the difference between best-corrected visual acuity under optimal luminance (BCVA) and LLVA (BCVA–LLVA gap). The association between mean change in BCVA from baseline and BCVA–LLVA gap at baseline was analysed using a general linear model.

Results A smaller baseline BCVA–LLVA gap predicted significantly higher BCVA gains over 24 months (p<0.0001 at each month; Pearson correlation), even after controlling for baseline BCVA or stratifying by treatment arm. Patients in the smallest baseline BCVA–LLVA gap quartile gained an average of +13.4 letters compared with +2.4 letters for patients in the widest baseline BCVA–LLVA gap quartile. At months 12 and 24, the smallest baseline BCVA–LLVA gap quartile had the highest proportion of ≥15−≥30-letter gain, and the widest baseline BCVA–LLVA gap quartile had the highest proportion of ≥15-/≥30-letter loss (p<0.0001; Fisher's exact test).

Conclusions The baseline BCVA–LLVA gap is a significant predictor of visual acuity response to anti-VEGF treatment in patients with wAMD.

Trial registration number NCT00891735; Post-results.