Glucagon-like peptide-1 receptor agonists and diabetic retinopathy: nationwide cohort and Mendelian randomization studies

Abstract

Background: The ability of glucagon-like peptide-1 receptor agonists (GLP-1 RAs) to decrease certain microvascular events has called for the investigation of GLP-1 RAs against diabetic retinopathy (DR), but the evidence is limited. By combining data from observational and Mendelian randomization (MR) studies, we aimed to investigate whether GLP-1 RAs decrease the risk of DR.

Methods: We combined data from several Swedish Registers and identified patients with incident type 2 diabetes being treated with GLP-1 RAs between 2006 and 2015, and matched them to diabetic patients who did not use GLP-1 RAs as the comparisons. The Cox proportional hazards models were applied to assess the risk of DR. We further performed the summary-data-based MR (SMR) analyses based on the Genotype-Tissue Expression databases and the Genome-Wide Association Study of DR from the FinnGen consortium.

Results: A total of 2390 diabetic patients were treated with GLP-1 RAs and the incidence of DR was 5.97 per 1000 person-years. Compared with diabetic patients who did not use GLP-1 RAs having an incidence of 12.85 per 1000 person-years, the adjusted hazard ratio (HR) of DR was 0.42 [95% confidence interval (CI), 0.29-0.61]. Genetically-predicted GLP1R expression (the target of GLP-1 RAs) showed an inverse association with background [odds ratio (OR)=0.83, 95% CI, 0.71-0.97] and severe nonproliferative DR (OR=0.72, 95% CI, 0.53-0.98), and a non-significant association with overall (OR=0.97, 95% CI, 0.92-1.03) and proliferative DR (OR=0.98, 95% CI, 0.91-1.05).

Conclusions: Both observational and mendelian randomization analyses showed a significantly lower risk of DR for patients treated with GLP-1 RAs, which calls for further studies to validate these findings.

Keywords: Cohort study; Diabetic retinopathy; GLP-1 RAs; Mendelian randomization.

Fig. 1

Identification of eligible patients and development of cohorts in the study and the design of summary-data-based Mendelian randomization (SMR) model. A Identification of eligible patients and development of cohorts in the observational study. GLP-1 RAs, Glucagon-like peptide-1 receptor agonists; DDD, Defined daily dose. B The design of summary-data-based Mendelian randomization (SMR) model. ×: there are no associations between genetic variants and exposure (or outcome). GWAS, Genome-Wide Association Study; eQTL, Expression quantitative trait loci; SMR, Summary-data-based Mendelian randomization; SNP, Single nucleotide polymorphism; MAF, Minor allele frequency

Fig. 2

Cumulative incidence estimates (Kaplan-Meier) for DR and the forest plot about the association between GLP-1 RAs and DR using Cox regression. DR, Diabetic retinopathy; GLP-1 RAs, Glucagon-like peptide-1 receptor agonists. A Cumulative incidence estimates (Kaplan-Meier) for DR in non-GLP-1 RAs and GLP-1 RAs groups. The blue and red lines indicate the cumulative incidence rate of the non-GLP-1 RAs group and the GLP-1 RAs group, separately. DR, Diabetic retinopathy; GLP-1 RAs, Glucagon-like peptide-1 receptor agonists. B The forest plot about the association between GLP-1 RAs and DR using Cox regression for four models. The association between GLP-1 RAs and DR was significant in all models. HR, Hazard ratio; CI, Confidence interval; GLP-1 RAs, Glucagon-like peptide-1 receptor agonists; DR, Diabetic retinopathy

Fig. 3

Association of GLP-1 RAs and DR using Cox regression in the sensitivity and subgroup analyses. The treatment of GLP-1 RAs was also associated with the risk of DR after excluding the patients with less than 6 months of follow-up. Subgroup analyses indicated GLP-1 RAs were associated with DR in men but not in women. GLP-1 RAs, Glucagon-like peptide-1 receptor agonists; HR, Hazard ratio; CI, Confidence interval; DR, Diabetic retinopathy