Introduction: Sodium-glucose cotransporter‑2 (SGLT‑2) inhibitors are widely used in type 2 diabetes, but their effect on the incidence of diabetic retinopathy remains uncertain.
Aim: To update the evidence on the relationship between SGLT-2 inhibitors and diabetic retinopathy incidence using a Bayesian network meta‑analysis (NMA) of randomised controlled trials (RCTs).
Method: A systematic search on Embase, PubMed, Web of Science, and ClinicalTrials.gov to 25 August 2025 identified RCTs in adults with type 2 diabetes comparing an SGLT‑2 inhibitor with placebo or active comparators and reporting diabetic retinopathy outcomes. A random‑effects Bayesian NMA estimated odds ratios (ORs) with 95% credible intervals (CrI) with ranking based on the surface under the cumulative ranking curve (SUCRA). Evidence quality was assessed using the Cochrane Risk-of-Bias tool and Confidence in Network Meta-Analysis (CINeMA). Network meta-regression explored treatment‑by‑covariate interactions including baseline diabetic retinopathy risk. Model‑based NMA (MBNMAdose) evaluated dose-response patterns.
Results: Thirty RCTs (eight gliflozins; 70,310 participants across study arms) formed a star‑shaped network without head‑to‑head trials. CINeMA indicated low confidence for most comparisons, mainly due to imprecision, with some concerns about indirectness and generally no concerns regarding heterogeneity or incoherence. No SGLT‑2 inhibitor altered diabetic retinopathy risk versus placebo (all 95% CrIs included 1.0). Probability rankings suggested empagliflozin most likely to be favourable and luseogliflozin least favourable (SUCRA values of 85.88% and 22.96%, respetively), but between‑drug differences were not statistically significant. Network meta‑regression identified effect modification by baseline diabetic retinopathy risk: higher placebo event rates were associated with lower relative odds under SGLT‑2 inhibition. No dose-response signal for diabetic retinopathy was detected, in contrast to dose‑related improvements in HbA1c and reductions in body weight and blood pressure.
Conclusion: SGLT‑2 inhibitors showed a neutral effect on the incidence of diabetic retinopathy. Baseline diabetic retinopathy risk appears to moderate relative effects, and no dose-response was evident for diabetic retinopathy. Choice among agents should be driven by overall efficacy, safety and patient factors rather than retinal risk; dedicated trials with prespecified ophthalmic endpoints are warranted.
扫码关注我们
求助内容:
应助结果提醒方式:
