This editorial commentary evaluates a large OHDSI network study (Cai et al.) examining whether semaglutide use in adults with type 2 diabetes is associated with a changed risk of neovascular age-related macular degeneration (NVAMD), compared with other GLP-1RAs (dulaglutide, exenatide) and non–GLP-1RA diabetes drugs (empagliflozin, sitagliptin, glipizide).
No statistically significant difference in NVAMD risk was found between semaglutide and any comparator; however, despite ~228,000 semaglutide new users, only a maximum of 79 patients contributed to any single comparison, yielding hazard ratios ranging from 0.25 to 2.08 with 95% CIs as wide as 0.05–5.9—making clinically meaningful effects impossible to rule out.
- Study population skewed young (mean age 55–65 years), far below the ~75-year average onset of NVAMD, limiting event accrual and generalizability. - Up to 3 of 4 meta-analyses relied almost exclusively on a single database after others failed diagnostic criteria, despite 12 databases enrolled. - Dose-response analysis and baseline non-neovascular AMD status were not assessable from administrative data.
This study does not establish a large increase or decrease in NVAMD risk with semaglutide in diabetic patients, but the wide confidence intervals mean a doubling or halving of risk cannot be excluded. Clinicians should await studies in older populations (≥70 years) before drawing firm conclusions, especially for patients using semaglutide for weight loss or non-diabetic indications.
Explore related topics
