August 2023
Volume 64, Issue 11
Open Access
Retina  |   August 2023
Association of Metformin and Other Diabetes Medication Use and the Development of New-Onset Dry Age-Related Macular Degeneration: A Case–Control Study
Author Affiliations & Notes
  • Gabriel T. Kaufmann
    Department of Ophthalmology and Visual Science, Pritzker School of Medicine, The University of Chicago, Chicago, Illinois, United States
  • Max J. Hyman
    The Center for Health and the Social Sciences, The University of Chicago, Chicago, Illinois, United States
  • Reem Gonnah
    Department of Ophthalmology and Visual Science, Pritzker School of Medicine, The University of Chicago, Chicago, Illinois, United States
  • Seenu Hariprasad
    Department of Ophthalmology and Visual Science, Pritzker School of Medicine, The University of Chicago, Chicago, Illinois, United States
  • Dimitra Skondra
    Department of Ophthalmology and Visual Science, Pritzker School of Medicine, The University of Chicago, Chicago, Illinois, United States
  • Correspondence: Dimitra Skondra, Department of Ophthalmology and Visual Science, Pritzker School of Medicine, The University of Chicago, 5841 South Maryland Avenue, MC 2114, Chicago, IL 60637, USA; dimitraskondra@gmail.com
Investigative Ophthalmology & Visual Science August 2023, Vol.64, 22. doi:https://doi.org/10.1167/iovs.64.11.22
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Gabriel T. Kaufmann, Max J. Hyman, Reem Gonnah, Seenu Hariprasad, Dimitra Skondra; Association of Metformin and Other Diabetes Medication Use and the Development of New-Onset Dry Age-Related Macular Degeneration: A Case–Control Study. Invest. Ophthalmol. Vis. Sci. 2023;64(11):22. https://doi.org/10.1167/iovs.64.11.22.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: To investigate if metformin use is associated with decreased odds of developing new non-neovascular (“dry”) age-related macular degeneration (AMD).

Methods: Case–control study examining 194,135 cases with diagnoses of new-onset AMD between 2008 and 2017 and 193,990 matched controls in the Merative MarketScan Research Databases. The diabetic subgroup included 49,988 cases and 49,460 controls. Multivariable conditional logistic regressions identified the risks of exposures on the development of dry AMD. Main outcome measures were odds ratios (ORs) of developing dry AMD with metformin use.

Results: In multivariable conditional logistic regression, any metformin use was associated with decreased odds of developing dry AMD (OR = 0.97; 95% confidence interval [CI], 0.95–0.99). This protective effect was noted for cumulative 2-year doses of metformin of 1 to 270 g (OR = 0.93; 95% CI, 0.90–0.97) and 271 to 600 g (OR = 0.92; 95% CI, 0.89–0.96). In a diabetic subgroup, metformin use below 601 g per 2 years decreased the odds of developing dry AMD (1–270 g: OR = 0.95; 95% CI, 0.91–0.99; 271–600 g: OR = 0.92; 95% CI, 0.89–0.96). Unlike in diabetic patients with diabetic retinopathy, diabetic patients without diabetic retinopathy had decreased odds of developing dry AMD with any metformin use (OR = 0.97; 95% CI, 0.94–0.998) and cumulative two-year doses of 1 to 270 g (OR 0.96; 95% CI, 0.91–0.998) and 271 to 600 g (OR = 0.92; 95% CI, 0.88–0.96).

Conclusions: Metformin use was associated with decreased odds of developing dry AMD. The protective effect was observed for cumulative 2-year doses below 601 g. In diabetics, this association persisted, specifically in those without diabetic retinopathy. Therefore, metformin may be a strategy to prevent development of dry AMD.

Age-related macular degeneration (AMD) is the leading cause of irreversible blindness in developed countries. Although some disease is classified as neovascular (“wet”) AMD and may be treated with both U.S. Food and Drug Administration (FDA)-approved and off-label anti-vascular endothelial growth factor injections, 85% of the disease burden is non-neovascular (“dry”) AMD.1 Unfortunately, there are no established treatments to prevent the development of dry AMD, and interventions aimed at slowing down its progression are limited in number and effectiveness; they include AREDS2 vitamins,24 smoking cessation, and a healthy diet.26 In the United States, over 10 million people have AMD, and globally the number exceeds 200 million.1 Furthermore, in North America, the estimated financial burden of AMD on healthcare systems in 2008 was greater than $250 billion.1 As the incidence and prevalence of AMD, especially dry AMD, are expected to grow as the global population ages, there is an unmet need for safe and non-invasive strategies to prevent it or slow down its progression. 
Metformin, a commonly prescribed oral diabetes medication, has demonstrated protective effects in diseases other than diabetes,7 including in cancer, cognitive decline, vascular disease, and primary open-angle glaucoma.819 Metformin has also been found to have anti-angiogenic, anti-inflammatory, and neuroprotective effects in the retina, suggesting a possible method by which metformin may slow AMD progression.2022 A case–control study by our group showed any metformin use was associated with decreased odds of developing new-onset AMD of any subtype—dry, wet, or unspecified (odds ratio [OR] = 0.94; 95% confidence interval [CI], 0.92–0.96). Specifically, 2-year cumulative doses up to 1080 g reduced these odds (1–270 g: OR = 0.91; 95% CI, 0.88–0.94; 271–600 g: OR = 0.90; 95% CI, 0.87–0.93; 601–1080 g: OR = 0.95; 95% CI, 0.92–0.98).23 Metformin use has been associated with decreased odds of developing AMD in four additional case–control studies,2326 and a meta-analysis found a non-statistically significant decrease in the development of AMD in patients taking metformin, as well.27 
Despite this evidence suggesting that metformin use may decrease the development of AMD of any subtype, evidence regarding the association between metformin use and the development of dry AMD is less conclusive. In a subgroup analysis of a retrospective cross-sectional study of 3120 patients with diabetes 60 years of age and older, metformin use provided a statistically significant reduction in the risk of developing dry AMD.24 In contrast, a retrospective cohort analysis found a conflicting association between metformin and the development of dry AMD, with a statistically significant increased hazard of developing dry AMD with active metformin use and a statistically significant decreased hazard of developing dry AMD with prior metformin use only.28 Furthermore, a case–control study in South Korea with 2330 total cases found metformin use increased the odds of developing dry AMD,29 and, finally, a case–control study of 1008 patients in a single county in Minnesota found that patients with no AMD had a statistically significantly increased likelihood of any metformin use compared to patients with dry AMD.30 Considering these results and the growing prevalence of dry AMD, a large, national case–control study investigating the association between any metformin use or metformin by dose and the development of dry AMD was warranted. 
Methods
The data originated from a peer-reviewed case–control study of patients 55 years of age and older with new-onset AMD of any subtype between January 2008 and December 2017 in the Merative MarketScan Commercial and Medicare Databases (Merative, Ann Arbor, MI, USA).31 These annual databases represent the health services of approximately 19 million to 57 million employees, dependents, and retirees in the United States with primary or Medicare coverage through privately insured fee-for-service, point-of-service, or capitated health plans. Each case was matched to a control by year, age, U.S. Census Bureau region, hypertension, anemia, and Charlson Comorbidity Index group (0, 1, 2, ≥3). The sample included 312,404 cases and 312,376 controls, and it achieved statistical balance.23 Complete methodological information, including inclusion criteria and diagnostic and procedural coding, is available in the Methods and Supplemental Online Content of our previous study.23 The subset of cases with dry AMD were identified by International Classification of Diseases (ICD), Ninth Revision (ICD-9), code 362.51 and ICD, Tenth Revision (ICD-10), code H35.31. Patients who were diagnosed with wet AMD prior to or on the date of the dry AMD diagnosis were excluded. Therefore, cases had no previous diagnoses of AMD of any subtype before their dry AMD diagnosis, which included any new diagnosis of early-, intermediate-, advanced-, or unspecified-stage dry AMD. Not until 2017 (the final year in this study) were ICD-10 codes expanded to allow us to distinguish among these stages (e.g., H35.31xx).32 This sample included 194,135 cases and 193,990 matched controls. 
We performed descriptive statistics for cases and matched controls. For categorical variables, this included frequencies and percentages, and significant differences were determined by χ2 tests. For continuous variables, this included means and standard deviations, and significant differences were determined by two-sample t-tests. 
Univariate conditional logistic regression analyzed the association between the development of dry AMD and (1) the risk factors of AMD, including female sex, diabetes, hypertension, hyperlipidemia, obesity, smoking, nonproliferative diabetic retinopathy (NPDR), and proliferative diabetic retinopathy (PDR); and (2) exposure to metformin and other medications, including insulin, sulfonylureas, meglitinides, other diabetes medications (i.e., exenatide, sitagliptin, and pramlintide), and statins in the 2 years prior to the index date. Multivariable conditional logistic regression analyzed the association between the development of dry AMD and metformin use after adjusting for the risk factors of AMD and exposure to other medications. Multivariable conditional logistic regression also analyzed the association between the development of dry AMD and 2-year cumulative doses of metformin, which were categorized by quartile: 1 to 270 g, 271 to 600 g, 601 to 1080 g, and >1080 g. 
We repeated the univariate and multivariable conditional logistic regressions in the subset of dry AMD cases and matched controls with diabetes (cases, n = 49,988; controls, n = 49,460). We also performed multivariable conditional logistic regression to analyze the association between the development of dry AMD and any metformin use or metformin by dose in diabetics with and without diabetic retinopathy. Finally, we performed multivariable conditional logistic regression to analyze the association between the development of dry AMD and (1) the combination of metformin with one of the other medications compared to metformin alone, or (2) the combination of metformin with one of the other medications compared to the use of neither medication. 
Statistical significance was set at α = 0.05. Data cleaning and analysis were performed in SAS 9.4 (SAS Institute, Chicago, IL, USA). Figures were produced with the ggplot2 package (version 3.3.5) in R 3.6.3 and RStudio 1.4.1103 (R Foundation for Statistical Computing, Vienna, Austria). The University of Chicago Institutional Review Board exempted this study from full review because personal identifiable information was not available in the data. 
Results
The sample included 194,135 new dry AMD cases and 193,990 matched controls (Table 1). Of the 14,812 cases in 2017, 3760 (12.7%) had early-stage dry AMD (H35.31 × 1), 912 (3.1%) had intermediate-stage dry AMD (H35.31 × 2), 152 (0.5%) had advanced-stage dry AMD (H35.31 × 3, H35.31 × 4), and 24,782 (83.7%) had unspecified-stage dry-AMD (H35.31, H35.31x, H35.31 × 0). Compared to controls, cases were more likely female (58.2% of cases and 54.8% of controls; P < 0.001) (Table 1) and were more likely to have a history of smoking, hyperlipidemia, NPDR, and PDR (smoking: 5.8% of cases and 4.3% of controls, P < 0.001; hyperlipidemia: 50.2% of cases and 48.3% of controls, P < 0.001; NPDR: 3.4% of cases and 1.7% of controls, P < 0.001; PDR: 0.5% of cases and 0.4% of controls, P < 0.001). The prevalence of diabetes was not statistically significantly different between cases and controls (25.8% of cases and 25.5% of controls, P = 0.071). 
Table 1.
 
Characteristics of Dry AMD Cases and Matched Controls
Table 1.
 
Characteristics of Dry AMD Cases and Matched Controls
Exposure to metformin did not differ between cases and controls in the full sample or in diabetics (13.3% of cases and 13.5% of controls in the full sample, P = 0.135; 48.8% of cases and 49% of controls in diabetics, P = 0.558) (Supplementary Table S1). However, the metformin doses were different between cases and controls in the full sample (1–270 g: 3.1% of cases and 3.3% of controls; 271–600 g: 2.9% of cases and 3.1% of controls; 601–1080 g: 3.7% of cases and 3.8% of controls; >1080 g: 3.6% of cases and 3.3% of controls; P < 0.001) and in diabetics (1–270 g: 10.8% of cases and 11.2% of controls; 271–600 g: 10.5% of cases and 11.2% of controls; 601–1080 g: 13.9% of cases and 14.0% of controls; >1080 g: 13.6% of cases and 12.6% of controls; P < 0.001). 
Furthermore, cases in the full and diabetic samples had less exposure to sulfonylureas compared to controls (8.2% of cases and 8.7% of controls in the full sample, P < 0.001; 30.8% of cases and 32.5% of controls in diabetics, P < 0.001), and, in the full cohort, controls had greater exposure to glitazones and statins (glitazones: 3.3% of cases and 3.4% of controls, P = 0.021; statins: 53.0% of cases and 53.7% of controls, P < 0.001). 
Univariate conditional logistic regression in the full sample found statistically significant associations between the development of dry AMD and female sex, smoking, hyperlipidemia, NPDR, or PDR (Table 2). Univariate conditional logistic regression also found the use of any metformin did not reduce odds of developing dry AMD (OR = 0.99; 95% CI, 0.97–1.01); however, cumulative 2-year doses of metformin of 1 to 270 g and 271 to 600 g decreased the odds of developing dry AMD (1–270 g: OR = 0.95; 95% CI, 0.91–0.98; 271–600 g: OR = 0.93; 95% CI, 0.90–0.97). Cumulative doses of >1080 g increased the odds (OR = 1.07; 95% CI, 1.04–1.11). 
Table 2.
 
Univariate and Multivariable Conditional Logistic Regression Models for Dry AMD Cases and Matched Controls (N = 388,125)
Table 2.
 
Univariate and Multivariable Conditional Logistic Regression Models for Dry AMD Cases and Matched Controls (N = 388,125)
In multivariable conditional logistic regression models adjusting for risk factors of AMD and exposures to other medications, any metformin use decreased the odds of developing dry AMD (OR = 0.97; 95% CI, 0.95–0.99) (Table 2). Cumulative 2-year doses of 1 to 270 g and 271 to 600 g decreased the odds of developing dry AMD (1–270 g: OR = 0.93; 95% CI, 0.90–0.97; 271–600 g: OR = 0.92; 95% CI, 0.89–0.96). Cumulative doses of >1080 g increased the odds (OR = 1.06; 95% CI, 1.02–1.11). 
In diabetics, univariate conditional logistic regression found statistically significant associations between the development of dry AMD and female sex, smoking, hyperlipidemia, NPDR, or PDR (Table 3). Exposure to any metformin did not affect the odds of developing dry AMD (OR = 0.98; 95% CI, 0.95–1.001). Two-year cumulative doses of 1 to 270 g and 271 to 600 g decreased the odds of developing dry AMD (1–270 g: OR = 0.95; 95% CI, 0.91–0.99; 271–600 g: OR = 0.92; 95% CI, 0.89–0.96). These associations persisted in the multivariable conditional logistic regression (any: OR = 0.97; 95% CI, 0.95–1.000; 1–270 g: OR = 0.95; 95% CI, 0.91–0.99; 271–600 g: OR 0.93; 95% CI, 0.89–0.97). Cumulative doses of >1080 g were associated with increased odds of developing dry AMD in the univariate and multivariable conditional logistic regressions (univariate: OR = 1.05; 95% CI, 1.01–1.09; multivariable: OR = 1.05; 95% CI, 1.01–1.10). Furthermore, the odds of developing dry AMD were not affected by any metformin use or metformin by dose in diabetics with diabetic retinopathy; however, in diabetics without diabetic retinopathy, the odds decreased for any metformin use and cumulative 2-year doses of 1 to 270 g and 271 to 600 g (any: OR = 0.97; 95% CI, 0.94–0.998; 1–270 g: OR = 0.96; 95% CI, 0.91–0.998; 271–600 g: OR = 0.92; 95% CI, 0.88–0.96) (Fig. 1; Supplementary Table S2). 
Table 3.
 
Univariate and Multivariable Conditional Logistic Regression Models for Dry AMD Diabetic Cases and Matched Controls (N = 99,448)
Table 3.
 
Univariate and Multivariable Conditional Logistic Regression Models for Dry AMD Diabetic Cases and Matched Controls (N = 99,448)
Figure 1.
 
Multivariable conditional logistic regression odds ratios for any metformin use and metformin by dose among dry AMD diabetics with and without diabetic retinopathy. Notes: Error bars show 95% confidence intervals. Diabetic retinopathy includes nonproliferative and proliferative diabetic retinopathy.
Figure 1.
 
Multivariable conditional logistic regression odds ratios for any metformin use and metformin by dose among dry AMD diabetics with and without diabetic retinopathy. Notes: Error bars show 95% confidence intervals. Diabetic retinopathy includes nonproliferative and proliferative diabetic retinopathy.
In diabetics not exposed to insulin, exposure to metformin decreased the odds of developing dry AMD (OR = 0.95; 95% CI, 0.92–0.98) (Supplementary Table S3). This association persisted for cumulative 2-year doses of metformin under 1080 g (1–270 g: OR = 0.93; 95% CI, 0.89–0.98; 271–600 g: OR = 0.92; 95% CI, 0.87–0.96; 601–1080 g: OR = 0.95; 95% CI, 0.91–0.997) (Supplementary Table S3). The same associations between metformin use up to 1080 g per 2 years and decreased odds of new dry AMD were found in diabetics without diabetic retinopathy who were not exposed to insulin (any: OR = 0.93; 95% CI, 0.89–0.93; 1–270 g: OR = 0.93; 95% CI, 0.89–0.98; 271–600 g: OR = 0.92; 95% CI, 0.87–0.96; 601–1080 g: OR = 0.95; 95% CI, 0.91–0.998) (Supplementary Table S3). Metformin use was not significantly associated with a change in the odds of dry AMD development in diabetic patients with diabetic retinopathy who were not exposed to insulin. 
Univariate conditional logistic regression found statistically significant associations between the development of dry AMD and exposure to sulfonylureas, glitazones, and statins (Table 2). In multivariable conditional logistic regression, sulfonylureas, statins, and insulin decreased the odds of developing dry AMD (sulfonylureas: OR = 0.89; 95% CI, 0.87–0.92; statins: OR = 0.95; 95% CI, 0.93–0.96; insulin: OR = 0.89; 95% CI, 0.86–0.92). In diabetics, univariate conditional logistic regression found a statistically significant association between the development of dry AMD and exposure to sulfonylureas (Table 3). In multivariable conditional logistic regression, sulfonylureas and insulin decreased the odds of developing dry AMD (sulfonylureas: OR = 0.92; 95% CI, 0.89–0.94; insulin: OR = 0.90; 95% CI, 0.87–0.93). 
Finally, metformin in combination with sulfonylureas or statins decreased the odds of developing dry AMD compared to metformin alone (sulfonylureas: OR = 0.90; 95% CI, 0.87–0.94; statins: OR = 0.96; 95% CI, 0.91–0.998) (Fig. 2, Supplementary Table S4). In contrast, metformin in combination with other diabetes medications increased the odds compared to metformin alone (OR = 1.06; 95% CI, 1.02–1.11). Metformin in combination with insulin or sulfonylureas decreased the odds of developing dry AMD compared to the use of neither medication (insulin: OR = 0.90; 95% CI, 0.81–0.99; sulfonylureas: OR = 0.91; 95% CI, 0.82–0.997) (Supplementary Table S5). Metformin in combination with other diabetes medications did not affect the odds compared to the use of neither medication (OR = 1.09; 95% CI, 0.99–1.20). 
Figure 2.
 
Multivariable conditional logistic regression odds ratios for other medications in combination with metformin vs. metformin alone. Notes: Error bars show 95% confidence intervals.
Figure 2.
 
Multivariable conditional logistic regression odds ratios for other medications in combination with metformin vs. metformin alone. Notes: Error bars show 95% confidence intervals.
Discussion
This case–control study analyzed the association of any metformin use or metformin by dose and the development of dry AMD controlling for the risk factors of AMD and exposure to other medications. It found that any metformin use decreased the odds of developing dry AMD and, in particular, cumulative 2-year doses of 1 to 270 g and 271 to 600 g decreased the odds by 3% to 10% and 4% to 11%, respectively. This result was found in an analysis of all patients in our study, of whom 25.8% of cases and 25.5% of controls had a diagnosis of diabetes, as metformin may also be prescribed for polycystic ovarian syndrome (PCOS), prediabetes, and obesity.33 In this study, metformin exposures may represent prescriptions for known diabetes, prescriptions of metformin for PCOS, prediabetes, and obesity, or coding inaccuracies, as patients may not have seen (and thus not have a claim from) a primary care physician or endocrinologist in the year prior to the index date. In diabetics specifically, low doses of metformin (i.e., 1–270 g and 271–600 g) also decreased the odds of developing dry AMD by 1% to 9% and 4% to 11%, respectively. Finally, further stratification found that any metformin use only decreased the odds of developing dry AMD in diabetics without diabetic retinopathy. 
Our group previously demonstrated that any metformin use and cumulative 2-year metformin doses up to 1080 g decreased the odds of developing AMD.23 Other studies have found similar associations between metformin and the development of dry AMD. A retrospective cross-sectional study analyzed 3120 diabetic patients (122 of whom had dry AMD) and found decreased odds of developing dry AMD after any metformin use (OR = 0.59; 95% CI, 0.46–0.77).24 A case–control study compared 504 patients with dry AMD to 504 patients without dry AMD and found increased odds of any metformin use in patients without dry AMD (OR = 1.66; 95% CI, 1.09–2.51).30 These studies suggested a protective role of metformin in the development of dry AMD, and our group expanded on their work by analyzing the protective effects of any metformin use and metformin by dose in a large, national sample of patients with dry AMD. 
Not every study has demonstrated a protective effect of metformin on the development of dry AMD. For example, historical use of metformin demonstrated a decreased hazard of developing dry AMD (hazard ratio [HR], 0.95; 95% CI, 0.92–0.98), and current use of metformin demonstrated an increased hazard (HR = 1.08; 95% CI, 1.04–1.12).28 A case–control study in South Korea found that metformin increased the odds of dry AMD in a subgroup analysis (OR = 1.62; 95% CI, 1.02–2.60).29 Furthermore, a meta-analysis assessed three subgroup analyses of the association between metformin and dry AMD and identified no relationship (OR = 0.98; 95% CI, 0.62–1.57).34 Thus, our study adds to this literature by supporting the association between metformin and the development of dry AMD in a subtype-specific analysis. 
In this study, metformin did not support a dose–response association in the development of dry AMD, as higher doses are not more protective against the development of dry AMD. However, this conclusion is consistent with our prior work showing that metformin was not protective of AMD at the highest doses >1080 g per 2 years.23 Notably, in this study, doses of metformin > 1080 g per 2 years showed increased odds of developing dry AMD (OR = 1.06; 95% CI, 1.02–1.11). A possible explanation could be that patients who require the highest doses of metformin may have the most severe cases of diabetes and remain at an increased risk for dry AMD despite protection from metformin. Furthermore, the decreased odds of developing dry AMD in diabetics without diabetic retinopathy compared to those with diabetic retinopathy supports the conclusion that metformin may be most effective in patients with well- to moderately controlled diabetes, in which diabetic retinopathy is less likely.35 Additionally, this study found that metformin exposure < 1080 g per 2 years was protective in subgroups of both diabetic patients and diabetic patients without diabetic retinopathy, all of whom had not been exposed to insulin (Supplementary Table S3). This is further suggestive that metformin may be most effective in patients without severe diabetes, as insulin is standardly prescribed in diabetic patients with the most severe disease as measured by laboratory values and failure of other therapies.36,37 There is also some preclinical evidence to suggest that lower doses of metformin are more protective in extending healthspan and lifespan in mice compared to high doses.38 Additional preclinical and clinical studies are needed to further elucidate the relationship of different doses of metformin and dry AMD. These conclusions are also consistent with the aforementioned conflicting studies showing decreased hazard ratios only in patients with a history of metformin use rather than current metformin use, as patients with active uncontrolled diabetes did not benefit from metformin use in the short term.28 
When assessing medication interactions, metformin in combination with sulfonylureas or statins decreased the odds of developing dry AMD compared to metformin alone, suggesting a possible synergistic effect. Sulfonylureas included drugs with the generic names of glimepiride, glyburide, glipizide, tolazamide, tolbutamide, and chlorpropramide. Glyburide (i.e., glibenclamide) specifically has been shown to have retinal neuroprotective properties and to reduce the risk of diabetic retinopathy in mice models.3941 Additionally, statins have been suggested to have neuroprotective effects, having been assessed in non-retinal contexts.42 Therefore, the use of metformin in combination with a sulfonylurea, such as glyburide, or with a statin may offer greater protection from dry AMD. In contrast, metformin in combination with insulin did not affect the odds of developing dry AMD compared to metformin alone. Despite each medication independently reducing the odds of developing dry AMD (metformin: OR = 0.97; 95% CI, 0.95–0.99; insulin: OR = 0.90; 95% CI, 0.87–0.93), the addition of insulin to metformin did not offer greater protection. Still, metformin in combination with insulin or sulfonylureas decreased the odds of developing dry AMD compared to taking neither medication, suggesting that their individual protective effects persist when taken together. 
Other diabetes medications have been found to increase the odds of developing dry AMD (OR = 1.10; 95% CI, 1.06–1.14). These diabetes medications include drugs with the generic names of exenatide, sitagliptin, and pramlintide, which are second-line medications used in patients with other comorbidities.43 Again, this result suggests that more unhealthy patients, such as those with less-controlled diabetes, may not benefit from metformin use. Interestingly, metformin in combination with the other diabetes medications did not affect the odds of developing dry AMD compared to the use of neither medication (OR = 1.09, CI, 0.99–1.20). Although the reason for increased odds of developing dry AMD for other diabetes medications is not known, combining them with metformin may ameliorate risk. However, this result may also reflect differences in diabetes severity, diabetes control, and systemic status of patients requiring both medications compared to neither. 
We focused on metformin in this analysis for several reasons. Metformin has been found to have other anti-aging and neuroprotective properties beyond its antidiabetic properties.2022 Metformin also has, from our perspective, promise as a potential preventative agent, as metformin is inexpensive, widely available globally, noninvasive, and safe with low profile of adverse effects, and it has already been used safely for years for other indications besides diabetes such as obesity and PCOS. Specifically, in contrast to metformin, insulin requires injection and, while sulfonylureas and insulin have shown evidence of retinal neuroprotection in diabetic retinopathy animal models,39 they are known to cause potentially dangerous hypoglycemia as an adverse effect. Thus, metformin, rather than other medications, could be well suited for older patients without diabetes who are at risk of developing dry AMD, as it avoids the risk of the adverse effects of other diabetes medications. 
Strengths of our study include a large sample size of 312,404 cases and 312,376 controls derived from national data, allowing for robust statistical analysis of the association between the development of dry AMD and any metformin use or metformin by dose. Additionally, given the sample size, our dataset allowed for subgroup analyses, such as in diabetics with and without diabetic retinopathy (retinopathy, n = 10,273; no retinopathy, n = 89,175). Multivariable analyses were also adjusted for established AMD risk factors, as well as medication exposures, reducing confounding in the results. 
This study must be evaluated within its limitations. First, the data lack specific medical information, including hemoglobin A1c and blood sugar levels, which would have provided information on diabetes severity. Laboratory measures are the established methods to assess diabetes severity,44,45 but these data were not included in our institution's MarketScan subscription. Therefore, we were only able to determine diabetes severity from ICD codes, which specify diabetes complications rather than diabetes severity.46 A validated measure of diabetes severity using ICD-9 codes does exist, the Diabetes Complication Severity Index (DCSI); however, it also requires laboratory data.47 The DCSI has not been validated using ICD-10 codes, but it has shown similar results.48 Nonetheless, the DCSI publication found that diabetic retinopathy was consistent with more severe diabetes.47 Accordingly, in this study, we adjusted for NPDR and PDR in its multivariable analyses, which means that some variation in diabetes severity was potentially captured in the study design. Furthermore, cases were matched to controls on Charlson Comorbidity Index score, which incorporates known complications of diabetes such as kidney disease, cerebrovascular disease, and congestive heart failure. Finally, our study found that metformin exposure was protective in the subgroup of diabetic patients who had not been exposed to insulin, providing further support that the protective effects of metformin that we are seeing are not simply due to diabetes severity. Thus, although we were unable to measure diabetes severity directly via laboratory measures and DCSI, our study design incorporated to some degree measures of diabetes severity in an attempt to mitigate its role as a potential confounder. Given the limitation regarding measures of diabetes severity in this study, future studies, both retrospective and prospective, are needed to shed more light on the complicated and important association between diabetes severity and the development of AMD. 
Second, the study also relied on diagnostic and procedural coding. The absence of a medical review of patients’ charts made us unable to determine the appropriateness of diagnoses. The large sample size should have helped to mitigate the effects of coding inaccuracies; however, systematic biases in the MarketScan data may still exist. Additionally, given the number of analyses performed, false positives are possible at our significance level of α = 0.05. Finally, we were unable to directly explore the association of metformin use and the development of dry AMD by disease severity. Although ICD-10 codes contain information on the severity of dry AMD, including the presence of geographic atrophy, ICD-9 codes do not allow for this determination. As the relevant ICD-10 codes started to be used in 2017, the sample size of patients with geographic atrophy was too small for analysis. 
This work adds to the existing literature on metformin use and the development of AMD. We analyzed the odds of developing dry AMD in patients from a large, national database, controlling for confounding factors and also assessing the protective effects of different metformin doses. We concluded that low doses of metformin up to 600 g over 2 years may help prevent the development of dry AMD. This association maintained statistical significance in diabetics, specifically in those without diabetic retinopathy. Furthermore, sulfonylureas and statins showed a possible synergistic effect with metformin, as they decreased the odds of developing dry AMD compared to metformin alone. 
Ultimately, these data suggest a potential role for metformin in preventing dry AMD, and follow-up studies should expand on this work by analyzing the association between metformin use and the progression of dry AMD, specifically to advanced dry AMD or geographic atrophy. As our study analyzed all stages of dry AMD (early, intermediate, and advanced) together, we are unable to assess relationships between metformin use and specific stages of AMD, an important analysis that ought to be pursued in future studies, as these associations may differ among early, intermediate, and advanced dry AMD. ICD-10 coding and the availability of more timely data in health insurance claims databases now permit preclinical and clinical investigations into this association. 
Although our study found that metformin use was associated with decreased odds of developing new-onset dry AMD, these results will have to be further validated in prospective trials. Database studies often reach similar conclusions as prospective trials, but randomized controlled trials are still needed for more conclusive evidence.49 We believe that this is feasible given the safety profile of metformin and its use in indications other than diabetes; furthermore, a phase II clinical trial (NCT02684578) is currently ongoing to assess whether metformin use decreases geographic atrophy progression, and metformin is being actively considered for a prospective trial in attenuating aging.50,51 There are also ongoing registered trials for metformin as an intervention for primary open-angle glaucoma and for ABCA4 retinopathy/Stargardt disease (NCT05426044, NCT04545736). Thus, we believe that our work adds to evidence suggesting the potential of metformin in the prevention of dry AMD, but more preclinical and clinical studies are needed. To definitively determine the efficacy of metformin in preventing dry AMD, metformin would have to be evaluated in patients without diabetes in a prospective clinical trial. 
Acknowledgments
Supported by the University of Chicago Institute for Translational Medicine and the Bucksbaum Institute for Clinical Excellence. 
Certain data used in this study were supplied by Merative as part of one or more MarketScan Research Databases. Any analysis, interpretation, or conclusion based on these data is solely that of the authors and not Merative. 
Disclosure: G.T. Kaufmann, None; M.J. Hyman, None; R. Gonnah, None; S. Hariprasad, None; D. Skondra, None 
References
Thomas CJ, Mirza RG, Gill MK. Age-related macular degeneration. Med Clin North Am. 2021; 105(3): 473–491. [CrossRef] [PubMed]
Gorusupudi A, Nelson K, Bernstein PS. The Age-Related Eye Disease 2 Study: micronutrients in the treatment of macular degeneration. Adv Nutr. 2017; 8(1): 40–53. [CrossRef] [PubMed]
Mukhtar S, Ambati BK. The value of nutritional supplements in treating Age-related macular degeneration: a review of the literature. Int Ophthalmol. 2019; 39(12): 2975–2983. [CrossRef] [PubMed]
Age-Related Eye Disease Study Research Group. A randomized, placebo-controlled, clinical trial of high-dose supplementation with vitamins C and E, beta carotene, and zinc for age-related macular degeneration and vision loss: AREDS Report No. 8. Arch Ophthalmol. 2001; 119(10): 1417–1436. [CrossRef] [PubMed]
Merle BMJ, Colijn JM, Cougnard-Grégoire A, et al. Mediterranean Diet and incidence of advanced age-related macular degeneration: the EYE-RISK Consortium. Ophthalmology. 2019; 126(3): 381–390. [CrossRef] [PubMed]
Merle BM, Silver RE, Rosner B, Seddon JM. Adherence to a Mediterranean diet, genetic susceptibility, and progression to advanced macular degeneration: a prospective cohort study123. Am J Clin Nutr. 2015; 102(5): 1196–1206. [CrossRef] [PubMed]
Campbell JM, Bellman SM, Stephenson MD, Lisy. K. Metformin reduces all-cause mortality and diseases of ageing independent of its effect on diabetes control: a systematic review and meta-analysis. Ageing Res Rev. 2017; 40: 31–44. [CrossRef] [PubMed]
Sehdev A, Shih YCT, Vekhter B, Bissonnette MB, Olopade OI, Polite BN. Metformin for primary colorectal cancer prevention in patients with diabetes: a case–control study in a US population. Cancer. 2015; 121(7): 1071–1078. [CrossRef] [PubMed]
Tseng CH. Metformin significantly reduces incident prostate cancer risk in Taiwanese men with type 2 diabetes mellitus. Eur J Cancer. 2014; 50(16): 2831–2837. [CrossRef] [PubMed]
Tseng CH. Use of metformin and risk of kidney cancer in patients with type 2 diabetes. Eur J Cancer. 2016; 52: 19–25. [CrossRef] [PubMed]
Mazzone PJ, Rai H, Beukemann M, Xu M, Jain A, Sasidhar M. The effect of metformin and thiazolidinedione use on lung cancer in diabetics. BMC Cancer. 2012; 12(1): 410. [CrossRef] [PubMed]
Bosco JLF, Antonsen S, Sørensen HT, Pedersen L, Lash TL. Metformin and incident breast cancer among diabetic women: a population-based case–control study in Denmark. Cancer Epidemiol Biomarkers Prev. 2011; 20(1): 101–111. [CrossRef] [PubMed]
Jiralerspong S, Palla SL, Giordano SH, et al. Metformin and pathologic complete responses to neoadjuvant chemotherapy in diabetic patients with breast cancer. J Clin Oncol. 2009; 27(20): 3297–3302. [CrossRef] [PubMed]
Luchsinger JA, Perez T, Chang H, et al. Metformin in amnestic mild cognitive impairment: results of a pilot randomized placebo controlled clinical trial. J Alzheimers Dis. 2016; 51(2): 501–514. [CrossRef] [PubMed]
Ng TP, Feng L, Yap KB, Lee TS, Tan CH, Winblad B. Long-term metformin usage and cognitive function among older adults with diabetes. J Alzheimers Dis. 2014; 41(1): 61–68. [CrossRef] [PubMed]
Herath PM, Cherbuin N, Eramudugolla R, Anstey KJ. The effect of diabetes medication on cognitive function: evidence from the PATH Through Life Study. BioMed Res Int. 2016; 2016: e7208429. [CrossRef]
Holman RR, Paul SK, Bethel MA, Matthews DR, Neil HAW. 10-Year follow-up of intensive glucose control in type 2 diabetes. N Engl J Med. 2008; 359(15): 1577–1589. [CrossRef] [PubMed]
Zhou L, Liu H, Wen X, Peng Y, Tian Y, Zhao L. Effects of metformin on blood pressure in nondiabetic patients: a meta-analysis of randomized controlled trials. J Hypertens. 2017; 35(1): 18–26. [CrossRef] [PubMed]
Lin HC, Stein JD, Nan B, et al. Association of geroprotective effects of metformin and risk of open-angle glaucoma in persons with diabetes mellitus. JAMA Ophthalmol. 2015; 133(8): 915–923. [CrossRef] [PubMed]
Joe SG, Yoon YH, Choi JA, Koh JY. Anti-angiogenic effect of metformin in mouse oxygen-induced retinopathy is mediated by reducing levels of the vascular endothelial growth factor receptor Flk-1. PLoS One. 2015; 10(3): e0119708. [CrossRef] [PubMed]
Han J, Li Y, Liu X, et al. Metformin suppresses retinal angiogenesis and inflammation in vitro and in vivo. PLoS One. 2018; 13(3): e0193031. [CrossRef] [PubMed]
Xu L, Kong L, Wang J, Ash JD. Stimulation of AMPK prevents degeneration of photoreceptors and the retinal pigment epithelium. Proc Natl Acad Sci USA. 2018; 115(41): 10475–10480. [CrossRef] [PubMed]
Blitzer AL, Ham SA, Colby KA, Skondra D. Association of metformin use with age-related macular degeneration: a case–control study. JAMA Ophthalmol. 2021; 139(3): 302–309. [CrossRef] [PubMed]
Stewart JM, Lamy R, Wu F, Keenan JD. Relationship between oral metformin use and age-related macular degeneration. Ophthalmol Retina. 2020; 4(11): 1118–1119. [CrossRef] [PubMed]
Brown EE, Ball JD, Chen Z, Khurshid GS, Prosperi M, Ash JD. The common antidiabetic drug metformin reduces odds of developing age-related macular degeneration. Invest Ophthalmol Vis Sci. 2019; 60(5): 1470–1477. [CrossRef] [PubMed]
Chen YY, Shen YC, Lai YJ, et al. Association between metformin and a lower risk of age-related macular degeneration in patients with type 2 diabetes. J Ophthalmol. 2019; 2019: e1649156.
Romdhoniyyah DF, Harding SP, Cheyne CP, Beare NAV. Metformin, a potential role in age-related macular degeneration: a systematic review and meta-analysis. Ophthalmol Ther. 2021; 10(2): 245–260. [CrossRef] [PubMed]
Eton EA, Wubben TJ, Besirli CG, Hua P, McGeehan B, VanderBeek BL. Association of metformin and development of dry age-related macular degeneration in a U.S. insurance claims database. Eur J Ophthalmol. 2022; 32(1): 417–423. [CrossRef] [PubMed]
Lee H, Jeon HL, Park SJ, Shin JY. Effect of statins, metformin, angiotensin-converting enzyme inhibitors, and angiotensin II receptor blockers on age-related macular degeneration. Yonsei Med J. 2019; 60(7): 679–686. [CrossRef] [PubMed]
Starr MR, Dalvin LA, AbouChehade JE, et al. Incidence of metformin use in patients with age-related macular degeneration versus normal controls: a population-based study in Olmsted County, Minnesota. Eye (Lond). 2023; 37(9): 1861–1865. [CrossRef] [PubMed]
Kulaylat AS, Schaefer EW, Messaris E, Hollenbeak CS. Truven Health Analytics MarketScan databases for clinical research in colon and rectal surgery. Clin Colon Rectal Surg. 2019; 32(1): 54–60. [PubMed]
Lum F, Repka M, Vicchrilli S. How to use the ICD-10 codes for age-related macular degeneration. EyeNet Mag. 2017; 9: 61–62.
Lv Z, Guo Y. Metformin and its benefits for various diseases. Front Endocrinol. 2020; 11: 191. [CrossRef]
Liang KH, Chen CH, Tsai HR, Chang CY, Chen TL, Hsu WC. Association between oral metformin use and the development of age-related macular degeneration in diabetic patients: a systematic review and meta-analysis. Invest Ophthalmol Vis Sci. 2022; 63(13): 10. [CrossRef] [PubMed]
Lee R, Wong TY, Sabanayagam C. Epidemiology of diabetic retinopathy, diabetic macular edema and related vision loss. Eye Vis (Lond). 2015; 2: 17. [CrossRef] [PubMed]
Chun J, Strong J, Urquhart S. Insulin initiation and titration in patients with type 2 diabetes. Diabetes Spectr. 2019; 32(2): 104–111. [CrossRef] [PubMed]
ElSayed NA, Aleppo G, Aroda VR, et al. 9. Pharmacologic approaches to glycemic treatment: Standards of Care in Diabetes–2023. Diabetes Care. 2022; 46(suppl 1): S140–S157.
Martin-Montalvo A, Mercken EM, Mitchell SJ, et al. Metformin improves healthspan and lifespan in mice. Nat Commun. 2013; 4: 2192. [CrossRef] [PubMed]
Berdugo M, Delaunay K, Naud MC, et al. The antidiabetic drug glibenclamide exerts direct retinal neuroprotection. Transl Res. 2021; 229: 83–99. [CrossRef] [PubMed]
Berdugo M, Delaunay K, Lebon C, et al. Long-term oral treatment with non-hypoglycemic dose of glibenclamide reduces diabetic retinopathy damage in the Goto–KakizakiRat model. Pharmaceutics. 2021; 13(7): 1095. [CrossRef] [PubMed]
Hussien NR, Al-Naimi MS, Rasheed HA, Al-Kuraishy HM, Al-Gareeb AI. Sulfonylurea and neuroprotection: the bright side of the moon. J Adv Pharm Technol Res. 2018; 9(4): 120–123. [PubMed]
Chen CJ, Ding D, Ironside N, et al. Statins for neuroprotection in spontaneous intracerebral hemorrhage. Neurology. 2019; 93(24): 1056–1066. [CrossRef] [PubMed]
American Diabetes Association. 9. Pharmacologic approaches to glycemic treatment: Standards of Medical Care in Diabetes–2020. Diabetes Care. 2020; 43(suppl 1): S98–S110. [PubMed]
American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2011; 34(suppl 1): S62–S69. [PubMed]
Lee CS, Krakauer C, Su YR, et al. Diabetic retinopathy and dementia association, beyond diabetes severity. Am J Ophthalmol. 2023; 249: 90–98. [CrossRef] [PubMed]
Dugan J, Shubrook J. International Classification of Diseases, 10th Revision, Coding for Diabetes. Clin Diabetes. 2017; 35(4): 232–238. [CrossRef] [PubMed]
Young BA, Lin E, Von Korff M, et al. Diabetes complications severity index and risk of mortality, hospitalization, and healthcare utilization. Am J Manag Care. 2008; 14(1): 15–23. [PubMed]
Glasheen WP, Renda A, Dong Y. Diabetes Complications Severity Index (DCSI)—update and ICD-10 translation. J Diabetes Complications. 2017; 31(6): 1007–1013. [CrossRef] [PubMed]
Wang SV, Schneeweiss S, RCT-DUPLICATE Initiative, et al. Emulation of randomized clinical trials with nonrandomized database analyses: results of 32 clinical trials. JAMA. 2023; 329(16): 1376–1385. [CrossRef] [PubMed]
Barzilai N, Crandall JP, Kritchevsky SB, Espeland MA. Metformin as a tool to target aging. Cell Metab. 2016; 23(6): 1060–1065. [CrossRef] [PubMed]
Kulkarni AS, Gubbi S, Barzilai N. Benefits of metformin in attenuating the hallmarks of aging. Cell Metab. 2020; 32(1): 15–30. [CrossRef] [PubMed]
Figure 1.
 
Multivariable conditional logistic regression odds ratios for any metformin use and metformin by dose among dry AMD diabetics with and without diabetic retinopathy. Notes: Error bars show 95% confidence intervals. Diabetic retinopathy includes nonproliferative and proliferative diabetic retinopathy.
Figure 1.
 
Multivariable conditional logistic regression odds ratios for any metformin use and metformin by dose among dry AMD diabetics with and without diabetic retinopathy. Notes: Error bars show 95% confidence intervals. Diabetic retinopathy includes nonproliferative and proliferative diabetic retinopathy.
Figure 2.
 
Multivariable conditional logistic regression odds ratios for other medications in combination with metformin vs. metformin alone. Notes: Error bars show 95% confidence intervals.
Figure 2.
 
Multivariable conditional logistic regression odds ratios for other medications in combination with metformin vs. metformin alone. Notes: Error bars show 95% confidence intervals.
Table 1.
 
Characteristics of Dry AMD Cases and Matched Controls
Table 1.
 
Characteristics of Dry AMD Cases and Matched Controls
Table 2.
 
Univariate and Multivariable Conditional Logistic Regression Models for Dry AMD Cases and Matched Controls (N = 388,125)
Table 2.
 
Univariate and Multivariable Conditional Logistic Regression Models for Dry AMD Cases and Matched Controls (N = 388,125)
Table 3.
 
Univariate and Multivariable Conditional Logistic Regression Models for Dry AMD Diabetic Cases and Matched Controls (N = 99,448)
Table 3.
 
Univariate and Multivariable Conditional Logistic Regression Models for Dry AMD Diabetic Cases and Matched Controls (N = 99,448)
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×