Glucagon-like peptide 1 receptor agonists (GLP-1RA), a class of medication originally approved for diabetes1 and now also approved for weight management in patients with obesity,2 have grown exponentially in popularity over the past few years. GLP-1RAs mimic the endogenous GLP-1, which belongs to the incretin family of metabolic hormones involved in insulin production, delayed gastric emptying, and enhanced satiety.1 These latter effects contribute to the weight-loss benefits of GLP-1RAs, which have since revolutionized the management of obesity.3
Although well established as an extremely effective medication for glycemic control in diabetes, GLP-1RAs have sparked controversy due to their effect on diabetic retinopathy (DR). Initial concerns were raised after findings from the cardiovascular outcome trial (CVOT), Semaglutide Unabated Sustainability in Treatment of Type 2 Diabetes 6 (SUSTAIN 6), revealed a greater risk of DR-related complications, defined as the need for laser therapy, intravitreal injections, development of vitreous hemorrhage, and blindness, in semaglutide users compared to placebo (3.0% vs 1.8%, HR=1.76, P=.02).4 This led to the FDA-mandated warning label for GLP-1RAs to include the risk of DR progression, raising questions for patients and ophthalmologists alike about the safety of this class of medication.
Potential Role of Rapid Improvement in Glycemic Control in DR Progression
Much of the earlier literature investigating whether GLP-1RAs are associated with DR progression used meta-analysis of randomized control trials (RCTs) with DR complications as a secondary characteristic. Bethel et al included the trials LEADER, SUSTAIN 6, EXSCEL, HARMONY, REWIND, and PIONEER-6 in their meta-analysis and found that GLP-1RAs were associated with DR; further meta-regression analysis revealed that mean reduction in HbA1c percentage was a key risk factor in this relationship.5 Similarly, Vilsbøll found in a post hoc mediation analysis of SUSTAIN 6 that the large and rapid reduction of HbA1c percentage in patients on semaglutide seemed to drive the DR complications observed in that group.6 The study also did not find an association between DR and semaglutide in their meta-analysis across SUSTAIN 1-5 and 2 Japanese CVOTs.6
In contrast, Yoshida and colleagues found through a meta-analysis of 4 CVOTs—including trials that specifically showed cardiovascular benefit of GLP-1RAs and excluding trials testing off-market GLP-1RAs—that GLP-1RA was associated with retinopathy complications after 52 weeks on the drug.7 However, a letter to the editor showed that, had the meta-analysis included trials for off-market GLP-1RAs, which still represent valuable data, the association would not have been significant.8
If rapid and dramatic glycemic control on GLP-1RAs is the underlying mechanism of the association with DR, this phenomenon has already been described in the literature in association with insulin use. This phenomenon, also known as a paradoxical “early worsening” of DR, has been described in insulin users since the 1980s.9,10 The mechanism of the relationship is poorly understood, with hypotheses including that a drop in blood sugar can lead to a rise in VEGF expression11,12 and that systemic glycemic changes can correspond with glucose concentration in the vitreous, and in turn lead to retinal injury.13 One of the strengths of GLP-1RA in diabetic management is the efficacy of HbA1c reduction,14 which could explain the newfound association between this powerful class of blood sugar–lowering agents and retinopathy.
Despite the strength in the large cumulative sample size and compilation of data from the statistically robust design of RCTs, one of the major limitations of meta-analysis studies for retinopathy is the lack of standardization in reporting and detecting retinopathy across the included trials. Careful inspection of study protocols revealed that several of the CVOTs thus mentioned, including the landmark SUSTAIN 6 trial, did not require dilated fundus exams at baseline.4,15-18 The baseline HbA1c, indicating severity of diabetic disease, and DR severity included in the trials also varied widely. The inclusion and exclusion criteria also varied. Although SUSTAIN 6 included patients with proliferative DR (PDR), who would be expected to require more interventions for retinopathy complications, some other CVOTs did not. Given that any of these factors can dramatically affect the rates of accurately detected retinopathy complications, the heterogeneity in quality of ophthalmic data from CVOTs warrant cautious interpretation of such meta-analysis studies.
Conflicting Results on GLP-1RA and DR Progression
An alternative study design that enables more consistency in ophthalmic-specific data is retrospective studies using real-world data. The results of such studies, including those of large sample size using national cohort data, is still conflicting. A multi-institutional study in Taiwan by Lin and colleagues found that GLP-1RA users were more likely than sodium-glucose cotransporter 2 inhibitors (SGLT-2I) users to develop PDR or require interventions for DR complications.19 Another study by Wai and colleagues using a multi-institutional US database found greater risk of progression to PDR and DME in GLP-1RA users compared to SGLT-2I controls after propensity score matching for baseline characteristics.20
Some limitations to these studies, however, were highlighted in responses to these publications. A letter to the editor to the Lin et al study noted the different definitions of DR progression used as the primary outcome between their study and the original CVOT studies.21 A comment on Wai’s study questioned the use of a database platform with limited ophthalmic data to answer ophthalmic questions.22 The comments made on these studies reflect the need for greater dialogue in a broader sense of the challenges of using large databases for addressing ophthalmic, and DR-specific, questions.
Meanwhile, other large database studies indicate a lack of association between GLP-1RA and DR. Barkmeier and colleagues compared the effect of multiple antihyperglycemic agents—namely SGLT-2I, dipeptidyl peptidase-4 (DPP-4), GLP-1RA, and sulfonylureas—on the need for treatment of diabetic macular edema or PDR in a cohort of 371,698 patients using US claims data. They found that although SGLT-2I was associated with lower rates of treatment compared to the other 3 drug classes, these latter groups had comparable outcomes compared to each other.23 This highlights how outcomes can differ depending on the control groups GLP-1RAs are compared to. Still yet, other studies have found no association with DR and GLP-1RA compared to various control groups, including SGLT-2Is.24-27
One limitation of many big data studies is the reliance on ICD-10 codes for diagnostic changes in DR. Studies have shown the propensity for inaccuracy in coding for DR,28 particularly after hospital systems switched from ICD-9 to ICD-10 coding,29 and a recent study highlighted the difference in coding for DR severity between retina specialists and general ophthalmologists.24 The inability to manually check these large data sets may pose a challenge in accurately detecting an event as rare as early worsening retinopathy. Another limitation in comparing these retrospective studies is the inconsistency in the outcomes by which the association of GLP-1RAs with DR was measured. Many studies emulated the definition of DR complications in the SUSTAIN 6 trial, while others measured the incidence of new DR, PDR, or DME. Moreover, another limitation of many studies was the lack of adjustment within the analytic model for the degree of HbA1c reduction while on the drug, because this is the leading hypothesis of the mechanism of this relationship.
Nuanced Investigation of the Relationship Between GLP-1RA and DR
Given the inherent limitations in diagnostic coding accuracy within large data sets, findings from large database studies may be best interpreted in conjunction with results from retrospective cohort studies employing manual chart review, which may provide more detailed clinical context. In one such study, Joo et al found no difference between 692 GLP-1RA and 289 SGLT-2I users in DR by clinical worsening, using ICD-10 codes, or by need for procedures after propensity score matching across baseline variables. The cases of worsening were confirmed by a retina specialist after manual review of the patient’s electronic medical record and available retinal imaging, such as OCT.24
Similarly, the AngioSafe Type 2 Diabetes study, in a case-control analysis of 3,154 patients with type 2 diabetes, found that GLP-1RA use was not associated with severe DR, although HbA1c—among other covariates such as diabetes duration—was.27 The cases of severe DR were confirmed by fundus photography.
In contrast, a case series by Varadhan et al found in 165 exenatide-treated patients that 30% of the cohort developed or experienced progression of DR.30 However, in 80% of the patients who experienced worsening, DR improved despite continued treatment with GLP-1RA, implying a temporary effect.
These retrospective studies, although smaller than the large database studies, enable more detailed analysis by virtue of including nuanced data from retinal imaging and longitudinal follow-up after retinopathy development or progression. Nevertheless, the smaller sample sizes limit the statistical power to detect a relatively uncommon event like retinopathy worsening.
A Neuroprotective Role of GLP-1RA in the Retina
Although clinical data have shown a potential for GLP-1RAs in contributing to retinopathy, animal models of DR have rather consistently indicated a neuroprotective effect of these drugs on the retina. GLP-1RAs were found to prevent progression of DR,31 showed anti-inflammatory properties in the retina of diabetic mice,32,33 and enhanced the vascular architecture of the retina.34 Additionally, a genome-wide association study (GWAS) in a Swedish population found that GLP1R, the gene encoding the receptor for GLP-1RA, had an inverse association with severe nonproliferative DR (OR=0.72; 95% CI, 0.53–0.98).35
Outside of DR, GLP-1RAs have also been shown to provide neuroprotective benefit in neurodegenerative diseases, like Alzheimer and Parkinson diseases,1 and ocular disorders, such as glaucoma and age-related macular degeneration (AMD).36 A recent study found, in a cohort of 9,669 patients, that GLP-1RA use was associated with a lower rate of primary open-angle glaucoma, wet AMD, and dry AMD compared to insulin use.36 Collectively, these data support a growing body of evidence of the promising role of GLP-1RAs in retinal health.
Recommendations on Management of GLP-1RA Users
The most definitive way to investigate the relationship between GLP-1RA and DR is through a prospective trial, and the first such trial is currently under way. The FOCUS trial will follow long-term effects on DR in 1,500 patients with type 2 diabetes randomized to semaglutide or placebo. The trial is estimated to be completed in 2027.37
Currently, the American Academy of Ophthalmology has not announced official recommendations about GLP-1RA management in the context of the potential for DR worsening. Generally, the cases of retinopathy progression have been found to be temporary and manageable with modern therapies for DR. For most patients, the cardiovascular and blood sugar control benefit of these drugs outweigh the small risk of retinopathy progression. Retinal experts do recommend obtaining a baseline eye exam and considering closer follow-up of patients on GLP-1RAs within the first 12 to 18 months of starting the drug, particularly for those who have more severe DR at baseline, high baseline HbA1c percentage, and rapid improvement in glycemic control similar to the characteristics of the patients enrolled in SUSTAIN 6.38 RP
References
1. Müller TD, Finan B, Bloom SR, et al. Glucagon-like peptide 1 (GLP-1). Mol Metab. 2019;30:72-130. doi:10.1016/j.molmet.2019.09.010
2. US Food and Drug Administration. FDA approves new drug treatment for chronic weight management, first since 2014. June 21, 2021. Accessed March 23, 2025. https://www.fda.gov/news-events/press-announcements/fda-approves-new-drug-treatment-chronic-weight-management-first-2014
3. Wilding JPH, Batterham RL, Calanna S, et al. Once-weekly semaglutide in adults with overweight or obesity. N Eng J Med. 2021;384(11):989-1002. doi:10.1056/NEJMoa2032183
4. Marso SP, Bain SC, Consoli A, et al. Semaglutide and cardiovascular outcomes in patients with type 2 diabetes. N Eng J Med. 2016;375(19):1834-1844. doi:10.1056/NEJMoa1607141
5. Bethel MA, Diaz R, Castellana N, Bhattacharya I, Gerstein HC, Lakshmanan MC. HbA1c change and diabetic retinopathy during GLP-1 receptor agonist cardiovascular outcome trials: a meta-analysis and meta-regression. Diabetes Care. 2021;44(1):290-296. doi:10.2337/dc20-1815
6. Vilsbøll T, Bain SC, Leiter LA, et al. Semaglutide, reduction in glycated haemoglobin and the risk of diabetic retinopathy. Diabetes Obes Metab. 2018;20(4):889-897. doi:10.1111/dom.13172
7. Yoshida Y, Joshi P, Barri S, et al. Progression of retinopathy with glucagon-like peptide-1 receptor agonists with cardiovascular benefits in type 2 diabetes—a systematic review and meta-analysis. J Diabetes Complications. 2022;36(8):108255. doi:10.1016/j.jdiacomp.2022.108255
8. Goldenberg RM. Progression of retinopathy with glucagon-like peptide-1 receptor agonists with cardiovascular benefits in type 2 diabetes—a systematic review and meta-analysis. J Diabetes Complications. 2022;36(9):108285. doi:10.1016/j.jdiacomp.2022.108285
9. Hooymans JM, Ballegooie EV, Schweitzer NM, Doorebos H, Reitsma WD, Slutter WJ. Worsening of diabetic retinopathy with strict control of blood sugar. Lancet. 1982;2(8295):438. doi:10.1016/s0140-6736(82)90464-0
10. Grunwald J, Riva C, Martin D, Quint A, Epstein P. Effect of an insulin-induced decrease in blood glucose on the human diabetic retinal circulation. Ophthalmology. 1987;94:1614-1620.
11. Kennedy A, Frank RN. The influence of glucose concentration and hypoxia on VEGF secretion by cultured retinal cells. Curr Eye Res. 2011;36(2):168-177. doi:10.3109/02713683.2010.521968
12. Abcouwer SF, Lin C mao, Wolpert EB, et al. Effects of ischemic preconditioning and bevacizumab on apoptosis and vascular permeability following retinal ischemia-reperfusion injury. Invest Ophthalmol Vis Sci. 2010;51(11):5920-5933. doi:10.1167/iovs.10-5264
13. Casson RJ, Wood JPM, Osborne NN. Hypoglycaemia exacerbates ischaemic retinal injury in rats. Br J Ophthalmol. 2004;88(6):816-820. doi:10.1136/bjo.2003.024661
14. Tsapas A, Avgerinos I, Karagiannis T, et al. Comparative effectiveness of glucose-lowering drugs for type 2 diabetes: a systematic review and network meta-analysis. Ann Intern Med. 2020;173(4):278-286. doi:10.7326/M20-0864
15. Marso SP, Daniels GH, Brown-Frandsen K, et al. Liraglutide and cardiovascular outcomes in type 2 diabetes. N Engl J Med. 2016;375(4):311-322. doi:10.1056/NEJMoa1603827
16. Husain M, Birkenfeld AL, Donsmark M, et al. Oral semaglutide and cardiovascular outcomes in patients with type 2 diabetes. N Engl J Med. 2019;381(9):841-851. doi:10.1056/NEJMoa1901118
17. Gerstein HC, Colhoun HM, Dagenais GR, et al. Dulaglutide and cardiovascular outcomes in type 2 diabetes (REWIND): a double-blind, randomised placebo-controlled trial. Lancet. 2019;394(10193):121-130. doi:10.1016/S0140-6736(19)31149-3
18. Holman RR, Bethel MA, Mentz RJ, et al. Effects of once-weekly exenatide on cardiovascular outcomes in type 2 diabetes. N Engl J Med. 2017;377(13):1228-1239. doi:10.1056/NEJMoa1612917
19. Lin TY, Kang EYC, Shao SC, et al. Risk of diabetic retinopathy between sodium-glucose cotransporter-2 inhibitors and glucagon-like peptide-1 receptor agonists. Diabetes Metab J. 2023;47(3):394-404. doi:10.4093/dmj.2022.0221
20. Wai KM, Mishra K, Koo E, et al. Impact of GLP-1 agonists and SGLT-2 inhibitors on diabetic retinopathy progression: an aggregated electronic health record data study. Am J Ophthalmol. 2024;265:39-47. doi:10.1016/j.ajo.2024.04.010
21. Ko J, Moon SJ. Risk of diabetic retinopathy between sodium-glucose cotransporter-2 inhibitors and glucagon-like peptide-1 receptor agonists (Diabetes Metab J 2023;47:394-404). Diabetes Metab J. 2023;47(4):571-572. doi:10.4093/dmj.2023.0164
22. Lum F. Comment on Impact of GLP-1 agonists and SGLT-2 inhibitors on diabetic retinopathy progression: an aggregated electronic health record study. Am J Ophthalmol. 2024;268:416. doi:10.1016/j.ajo.2024.08.037
23. Barkmeier AJ, Herrin J, Swarna KS, et al. Comparative effectiveness of glucagon-like peptide-1 receptor agonists, sodium-glucose cotransporter 2 inhibitors, dipeptidyl peptidase-4 inhibitors, and sulfonylureas for sight-threatening diabetic retinopathy. Ophthalmol Retina. 2024;8(10):943-952. doi:10.1016/j.oret.2024.05.003
24. Joo JH, Sharma N, Shaia J, et al. The effect of glucagon-like peptide-1 receptor agonists on diabetic retinopathy at a tertiary care center. Ophthalmol Sci. 2024;4(6):100547. doi:10.1016/j.xops.2024.100547
25. Tauqeer Z, Bracha P, Hua P, Yu Y, Cui QN, VanderBeek BL. Glucagon-like peptide-1 receptor agonists are not associated with an increased risk of progressing to vision-threatening diabetic retinopathy. Ophthalmic Epidemiol. Published online October 10, 2024:1-4. doi:10.1080/09286586.2024.2399764
26. Ueda P, Pasternak B, Eliasson B, et al. Glucagon-like peptide 1 receptor agonists and risk of diabetic retinopathy complications: cohort study in nationwide registers from two countries. Diabetes Care. 2019;42(6):e92-e94. doi:10.2337/dc18-2532
27. Gaborit B, Julla JB, Besbes S, et al. Glucagon-like peptide 1 receptor agonists, diabetic retinopathy and angiogenesis: the AngioSafe Type 2 Diabetes Study. J Clin Endocrinol Metab. 2020;105(4):dgz069. doi:10.1210/clinem/dgz069
28. Chen JS, Copado IA, Vallejos C, et al. Variations in electronic health record–based definitions of diabetic retinopathy cohorts: a literature review and quantitative analysis. Ophthalmol Sci. 2024;0(0). doi:10.1016/j.xops.2024.100468
29. Cai CX, Michalak SM, Stinnett SS, Muir KW, Fekrat S, Borkar DS. Effect of ICD-9 to ICD-10 transition on accuracy of codes for stage of diabetic retinopathy and related complications: results from the CODER study. Ophthalmol Retina. 2021;5(4):374-380. doi:10.1016/j.oret.2020.08.004
30. Varadhan L, Humphreys T, Walker AB, Varughese GI. The impact of improved glycaemic control with GLP-1 receptor agonist therapy on diabetic retinopathy. Diabetes Res Clin Pract. 2014;103(3):e37-39. doi:10.1016/j.diabres.2013.12.041
31. Liu J, Wang H, Huang C. Exendin-4, a GLP-1 receptor agonist, suppresses diabetic retinopathy in vivo and in vitro. Arch Physiol Biochem. 2025;131(1):1-10. doi:10.1080/13813455.2023.2274279
32. Oezer K, Kolibabka M, Gassenhuber J, et al. The effect of GLP-1 receptor agonist lixisenatide on experimental diabetic retinopathy. Acta Diabetol. 2023;60(11):1551-1565. doi:10.1007/s00592-023-02135-7
33. Ramos H, Bogdanov P, Sampedro J, Huerta J, Simó R, Hernández C. Beneficial effects of glucagon-like peptide-1 (GLP-1) in diabetes-induced retinal abnormalities: involvement of oxidative stress. Antioxidants (Basel). 2020;9(9):846. doi:10.3390/antiox9090846
34. Wei L, Mo W, Lan S, et al. GLP-1 RA improves diabetic retinopathy by protecting the blood-retinal barrier through GLP-1R-ROCK-p-MLC signaling pathway. J Diabetes Res. 2022;2022:1861940. doi:10.1155/2022/1861940
35. Zheng D, Li N, Hou R, et al. Glucagon-like peptide-1 receptor agonists and diabetic retinopathy: nationwide cohort and Mendelian randomization studies. BMC Medicine. 2023;21(1):40. doi:10.1186/s12916-023-02753-6
36. Allan KC, Joo JH, Kim S, et al. Glucagon-like peptide-1 receptor agonist impact on chronic ocular disease including age-related macular degeneration. Ophthalmology. Published online January 23, 2025. doi:10.1016/j.ophtha.2025.01.016
37. Long-term effects of semaglutide on diabetic retinopathy in subjects with type 2 diabetes. Clinicaltrials.gov identifier: NCT03811561. Updated April 16, 2025. Accessed April 18, 2025. https://clinicaltrials.gov/study/NCT03811561
38. Taylor R. Update on semaglutide risks. EyeNet. November 1, 2021. Accessed April 18, 2025. https://www.aao.org/eyenet/article/update-on-semaglutide-risks