Does metformin protect against AMD? Large retrospective study says no

26 Sep 2025
Jairia Dela Cruz
Jairia Dela Cruz
Jairia Dela Cruz
Jairia Dela Cruz
Does metformin protect against AMD? Large retrospective study says no

There is no evidence suggesting that metformin can prevent the development or progression of age-related macular degeneration (AMD) among older adults, according to an analysis of data from the TriNetX medical records database.

In a cohort of 379,450 AMD-free participants, the risk of developing the eye disease did not significantly differ between those who were prescribed metformin and those who were not (risk ratio [RR], 0.90, 95 percent confidence interval [CI], 0.86–0.94). The risk remained similar at 5 years (RR, 0.94, 95 percent CI, 0.90–0.99) and 10 years (RR, 0.91, 95 percent CI, 0.87–0.94). [JAMA Ophthalmol 2025;doi:10.1001/jamaophthalmol.2025.3070]

Likewise, the risk of progression in a cohort of 18,488 participants with mild or moderate nonexudative AMD did not significantly differ between those who were and those who were not prescribed metformin (geographic atrophy [GA]: RR, 0.87, 95 percent CI, 0.76–1.01; nonfoveal GA: RR, 0.80, 95 percent CI, 0.67–0.97; foveal GA: RR, 0.96, 95 percent CI, 0.80–1.14; neovascular AMD: RR, 1.03, 95 percent CI, 0.91–1.17). These findings were consistent at the 5-year and 10-year follow-up.

“In summary, the results of our analysis do not support a protective effect of metformin against the development of AMD or its progression to geographic atrophy or neovascular AMD,” the investigators said.

The interest in metformin for preventing or treating AMD comes from its theorized protective effects against various age-related diseases. Specifically, metformin activates AMP-activated protein kinase, which can reduce the buildup of drusen, alleviate inflammation, protect retinal pigment epithelial cells and photoreceptors from death, stop choroidal neovascularization, and stimulate clearance of dead retinal pigment epithelium debris. [Ophthalmol Retina 2020;4:1118-1119; Lancet 2023;401:1459-1472; Int J Ophthalmol 2021;14:1274-1280]

There is no consensus in the scientific community about whether metformin helps in AMD, and the existing literature provides a mixed bag of evidence, according to the investigators. “We feel that our study works toward a consensus and clarifies a lack of a role for metformin in AMD development or progression,” they said, emphasizing that all conclusions in the study are associations and not causality.

They also acknowledged the need for more research, suggesting that future prospective analyses focus on evaluating whether the dosage and duration of metformin use might influence AMD.

In a linked editorial Dr Su-Hsun Liu from the University of Colorado Anschutz Medical Campus in Aurora, Colorado, US, and two colleagues praised the investigators for their careful and honest reporting of their findings, which “serve as a valuable reminder of the complexities and limitations inherent in using observational data to explore potential drug repurposing opportunities.” [JAMA Ophthalmol 2025;doi:10.1001/jamaophthalmol.2025.3183]

In light of the growing interest in metformin’s anti-inflammatory and anti-ageing properties, Liu and colleagues highlighted that the drug, like any other medications, carries potential safety risks, including gastrointestinal side effects, vitamin B12 deficiency with long-term use, and, in rare cases, lactic acidosis, particularly in individuals with kidney impairment.

“Until high-quality causal evidence is available, the totality of the currently available evidence does not support the consideration of clinicians to prescribe metformin for the sole purpose of preventing AMD or slowing its progression,” they said.

The analyses included 189,725 matched pairs of participants in the AMD-free cohort (average age 76 years, 53.6 percent female, 59.7 percent White, 79.8 percent had diabetes) and 9,244 matched pairs of participants in the cohort with mild to moderate nonexudative AMD (average age 80.5 years, 57 percent female, 79 percent White, 85.3 percent had diabetes). Propensity score matching was performed based on age, sex, race, hypertension, diabetes, and other systemic conditions.