Abstract
BACKGROUND: Thiazolidinediones, commonly used antidiabetic medications, have been associated with an increased risk of development of diabetic macular oedema and increased vascular endothelial cell permeability. Macular neovascularisation in age-related macular degeneration (AMD) and associated fluid leakage may be influenced by thiazolidinediones. This study aims to determine the association between thiazolidinedione usage and retinal morphological outcomes or visual acuity (VA) in patients treated with bevacizumab or ranibizumab for neovascular AMD (nAMD). METHODS: Secondary analysis of data from the Comparison of Age-related Macular Degeneration Treatments Trials. Participant self-reported diabetes status and thiazolidinedione usage at baseline. VA, intraretinal, subretinal and subretinal pigment epithelium fluid, and foveal thickness of retinal layers were evaluated at baseline and during 2-year follow-up. Comparisons of outcomes between thiazolidinedione usage groups were adjusted by macular neovascularisation lesion type in multivariable regression models. RESULTS: Patients taking thiazolidinedione (n=30) had lower adjusted mean VA score at baseline (difference -6.2 letters; p=0.02), greater proportion with intraretinal fluid (IRF) at year 2 (75% vs 50%, adjusted OR 2.8; p=0.04), greater mean decrease in subretinal tissue complex thickness from baseline at year 1 (difference -75.1 um; p=0.02) and greater mean decrease in subretinal thickness at year 1 (difference -41.9 um; p=0.001) and year 2 (difference -43.3 um; p=0.001). CONCLUSIONS: In this exploratory analysis, patients with diabetes taking thiazolidinediones and treated with bevacizumab or ranibizumab for nAMD had worse baseline mean VA, greater reductions in subretinal and subretinal tissue complex thickness from baseline, and greater proportions with IRF comparing to patients not taking thiazolidinediones. TRIAL REGISTRATION NUMBER: ClinicalTrials.gov NCT00593450.