Metformin use was associated with a lower incidence of intermediate age-related macular degeneration (AMD) over 5 years among patients with type 2 diabetes, according to a population-based study that used systematic retinal image grading and real-world prescribing data.
Researchers analyzed data from the Individualized Screening for Diabetic Retinopathy (ISDR) study, a National Institute for Health Research–funded program in Liverpool, UK, that linked diabetic retinopathy screening photographs with general practitioner prescribing records. Unlike prior epidemiologic studies that relied on billing codes or unvalidated clinical diagnoses, this study graded AMD directly from color fundus photographs using a modified Age-Related Eye Disease Study (AREDS) system, allowing the analysis to distinguish between early, intermediate, and late AMD.
The study included 2,545 randomly selected ISDR patients aged 50 years or older with diabetes who attended screening in 2011. After excluding 456 patients with ungradable images or no images, the analysis focused on 2,089 patients with type 2 diabetes who had gradable images at five-year follow-up. Of these, 836 patients were prescribed metformin at baseline and during follow-up, while 1,253 were not.
At baseline, most patients had no AMD, but a meaningful minority already had early, intermediate, or late disease. Researchers reported that 284 patients had early AMD, 90 had intermediate AMD, and 12 had late AMD at baseline.
After adjusting for confounders including age, sex, diabetic retinopathy status, glycemic control, and diabetes duration, metformin use was associated with a 37% lower incidence of intermediate AMD over five years. The association remained consistent across regression models, including analyses that used multiple imputation to account for missing data.
In contrast, metformin was not associated with a reduced incidence of early AMD or with progression from early or intermediate AMD to late AMD. Researchers noted that the number of patients who developed late AMD during follow-up was relatively small, which limited the study’s ability to detect meaningful differences in advanced disease outcomes.
For ophthalmologists, the study adds clinically relevant nuance to a growing body of literature suggesting that metformin may have pleiotropic benefits beyond glycemic control. Multiple meta-analyses have reported modest associations between metformin exposure and reduced odds of AMD, but most prior studies relied on administrative data sources that are prone to misclassification – particularly for intermediate AMD, which may be asymptomatic and inconsistently documented in routine care.
The researchers suggested that the protective association observed at the intermediate stage is biologically plausible. Metformin has known antioxidant and anti-inflammatory effects, may promote autophagy, and has been linked to pathways involved in aging-related retinal degeneration. However, the researchers emphasized that their findings remain observational and cannot establish causality.
The study also had several limitations relevant to clinical interpretation. The dataset did not include metformin dose, prior duration of use, or adherence. Optical coherence tomography, which has better sensitivity for detecting AMD, was not available for the patient cohort, and so AMD grading relied solely on non-stereoscopic color fundus photographs. In addition, baseline differences between metformin users and nonusers – including lower AMD prevalence at baseline among metformin users – may have influenced progression patterns despite statistical adjustment.
Nevertheless, the researchers conclude that the findings support prospective clinical trials evaluating metformin as a potential intervention to reduce AMD progression prior to vision-threatening disease.
Source: BMJ Open Ophthalmology.
Researchers analyzed data from the Individualized Screening for Diabetic Retinopathy (ISDR) study, a National Institute for Health Research–funded program in Liverpool, UK, that linked diabetic retinopathy screening photographs with general practitioner prescribing records. Unlike prior epidemiologic studies that relied on billing codes or unvalidated clinical diagnoses, this study graded AMD directly from color fundus photographs using a modified Age-Related Eye Disease Study (AREDS) system, allowing the analysis to distinguish between early, intermediate, and late AMD.
The study included 2,545 randomly selected ISDR patients aged 50 years or older with diabetes who attended screening in 2011. After excluding 456 patients with ungradable images or no images, the analysis focused on 2,089 patients with type 2 diabetes who had gradable images at five-year follow-up. Of these, 836 patients were prescribed metformin at baseline and during follow-up, while 1,253 were not.
At baseline, most patients had no AMD, but a meaningful minority already had early, intermediate, or late disease. Researchers reported that 284 patients had early AMD, 90 had intermediate AMD, and 12 had late AMD at baseline.
After adjusting for confounders including age, sex, diabetic retinopathy status, glycemic control, and diabetes duration, metformin use was associated with a 37% lower incidence of intermediate AMD over five years. The association remained consistent across regression models, including analyses that used multiple imputation to account for missing data.
In contrast, metformin was not associated with a reduced incidence of early AMD or with progression from early or intermediate AMD to late AMD. Researchers noted that the number of patients who developed late AMD during follow-up was relatively small, which limited the study’s ability to detect meaningful differences in advanced disease outcomes.
For ophthalmologists, the study adds clinically relevant nuance to a growing body of literature suggesting that metformin may have pleiotropic benefits beyond glycemic control. Multiple meta-analyses have reported modest associations between metformin exposure and reduced odds of AMD, but most prior studies relied on administrative data sources that are prone to misclassification – particularly for intermediate AMD, which may be asymptomatic and inconsistently documented in routine care.
The researchers suggested that the protective association observed at the intermediate stage is biologically plausible. Metformin has known antioxidant and anti-inflammatory effects, may promote autophagy, and has been linked to pathways involved in aging-related retinal degeneration. However, the researchers emphasized that their findings remain observational and cannot establish causality.
The study also had several limitations relevant to clinical interpretation. The dataset did not include metformin dose, prior duration of use, or adherence. Optical coherence tomography, which has better sensitivity for detecting AMD, was not available for the patient cohort, and so AMD grading relied solely on non-stereoscopic color fundus photographs. In addition, baseline differences between metformin users and nonusers – including lower AMD prevalence at baseline among metformin users – may have influenced progression patterns despite statistical adjustment.
Nevertheless, the researchers conclude that the findings support prospective clinical trials evaluating metformin as a potential intervention to reduce AMD progression prior to vision-threatening disease.
Source: BMJ Open Ophthalmology.
