Potential Efficacy of Metformin for Age-Related Macular Degeneration: A Systematic Review and Meta-Analysis

Abstract

Topic: Metformin, a widely used diabetes medication, has shown potential for treating age-related macular degeneration (AMD) due to its antioxidative, anti-inflammatory, and antiangiogenic properties. This study aims to systematically review and analyze the efficacy of metformin in reducing AMD prevalence.

Clinical relevance: Metformin's potential to serve as a treatment for AMD could significantly reduce the burden of vision loss, offering a cost-effective and widely accessible solution.

Methods: A systematic search was conducted in OVID Embase, OVID MEDLINE, Cochrane Library, and Web of Science databases on May 2, 2024. Both observational and interventional studies were included if they involved oral metformin use before AMD diagnosis. Data were extracted and analyzed using a random-effects model meta-analysis, with subgroup analyses based on study design, AMD subtype, sex, and metformin dosage.

Results: Eighteen observational studies were identified, which together included a total of 2 683 234 individuals. Nine studies had a case-control design, 7 were retrospective cohort studies, and 2 were cross-sectional studies. The meta-analysis revealed a significant reduction in the odds of AMD among metformin users (pooled odds ratio [OR] = 0.86, 95% confidence interval = 0.79-0.93, P = 0.0002, I² = 90%). The association was significant in both patients with diabetes (pooled OR = 0.89) and without diabetes (pooled OR = 0.70), although only 2 studies reported nondiabetic ORs. Dose-response analysis revealed significant protective effects at low doses. Sensitivity analysis indicated that the removal of an outlier study did not alter the overall effect. Bias analysis using the Risk of Bias in Nonrandomized Studies of Interventions tool revealed significant risks of bias, particularly due to confounding.

Conclusion: Although the current evidence suggests a potential protective role of metformin in AMD, all studies showing an effect of metformin have been observational and thus subject to bias. Randomized clinical trials are needed to determine the effectiveness of metformin for preventing the onset of AMD.

Financial disclosures: Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Keywords: Age-related macular degeneration; Meta-analysis; Metformin; Systematic review.

Figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart summarizing the study selection process. This flowchart depicts the systematic identification and screening of studies included in the meta-analysis. The initial database search yielded 800 records, of which 41 full-text articles were assessed for eligibility, and 18 studies were included in the final analysis. Each step of the selection process, including the reasons for exclusion at each stage, is outlined.

Figure 2

Risk of bias assessment using the ROBINS-I tool. The bar chart shows the proportion of included studies rated as low, moderate, serious, or critical risk of bias across 7 domains, including confounding, selection of participants, and outcome measurement. Most studies exhibited a serious risk of bias in the confounding domain due to the observational design. The findings underscore the need for cautious interpretation and highlight key areas for methodological improvement in future research. ROBINS-I = Risk of Bias in Nonrandomized Studies of Interventions.

Figure 3

Forest plot of pooled ORs for the association between metformin use and AMD. The plot shows individual study ORs and their 95% CIs, along with the overall pooled estimate (OR = 0.86, 95% CI = 0.79–0.93, P = 0.0002, I² = 90%). The diamond at the bottom represents the pooled estimate, indicating a statistically significant protective effect of metformin use against AMD. The high I² statistic reflects considerable heterogeneity among studies. AMD = age-related macular degeneration; CIs = confidence intervals; ORs = odd ratios.

Figure 4

Forest plot of pooled HRs for the progression of AMD among metformin users. Individual study HRs and their 95% CIs are displayed alongside the pooled estimate (HR = 0.82, 95% CI = 0.55–1.23, P = 0.34, I² = 99%). The wide confidence intervals and high I² value suggest variability and limited consensus across studies. The results indicate no statistically significant association between metformin use and AMD progression. AMD = age-related macular degeneration; CIs = confidence intervals; HRs = hazard ratios.

Figure 5

Funnel plot for sensitivity analysis assessing publication bias. The plot depicts the distribution of effect sizes (ORs) vs. their standard errors. Asymmetry is observed, with one study (Jiang et al) appearing to be an outlier. This study's exclusion in the sensitivity analysis resulted in a marginally higher pooled OR (0.90, 95% CI = 0.84–0.96, P = 0.003) and a reduction in heterogeneity (I² = 87%), suggesting that publication bias may partially influence the findings. CIs = confidence intervals; ORs = odd ratios.