RESUMO
AIMS: The effect of Glucagon-like peptide 1 receptor agonists (GLP1 RA) on diabetic retinopathy (DR) remains controversial. Previous reviews combined data from randomized clinical trials (RCTs) with or without cardiovascular (CV) benefits and did not address confounders, therefore may have generated misleading results. The study aimed to examine the effect of GLP1RA on DR in type 2 diabetes (T2DM) in RCTs with or without CV benefits and distinguish the effect by major confounders. METHODS: We conducted electronic searches of multiple databases and a manual search using references lists. We included 13 RCTs examining the effect of GLP1 RA on health outcomes/adverse events including DR or DR complications in T2DM. We performed a random-effects model meta-analysis. RESULTS: GLP1RA was associated with an elevated risk of rapidly worsening DR in four major RCTs with CV benefits in T2DM (OR 1.23, 95 % CI 1.05-1.44). The association between GLP1 RA and DR was significant in subgroups of RCTs with length over 52 weeks (1.2, 1.00-1.43), using placebo as a comparator (1.22, 1.05-1.42). In subgroups with patients who had T2DM ≥10 years (1.19, 0.99-1.42) or with subjects enrolled from multiple countries (1.2, 0.99-1.46), the association appeared to be evident but did not reach statistical significance. CONCLUSIONS: GLP1 RA including liraglutide, semaglutide, and dulaglutide are associated with an increased risk of rapidly worsening DR in RCTs with CV benefits. Further data from clinical studies with longer follow-up purposefully designed for DR risk assessment, particularly including patients of established DR are warranted.