Glucagon-Like Peptide 1 Receptor Agonists and Risk of Diabetic Retinopathy Complications: Cohort Study in Nationwide Registers From Two Countries

Post hoc analyses of the Trial to Evaluate Cardiovascular and Other Long-term Outcomes With Semaglutide in Subjects With Type 2 Diabetes (SUSTAIN-6) showed an increased risk of diabetic retinopathy complications associated with semaglutide versus placebo among patients with a history of retinopathy, especially among those with insulin use (1). Although concerns have been raised regarding the safety of glucagon-like peptide 1 (GLP-1) receptor agonists in patients with existing retinopathy, subgroup analyses by retinopathy status at baseline have not been presented for other trials of GLP-1 receptor agonists; the two observational studies on GLP-1 receptor agonists and diabetic retinopathy performed to date have excluded patients with a history of retinopathy (2) or treatment of retinopathy (3).

We conducted a cohort study (January 2010 to December 2016) of patients with existing retinopathy using nationwide data in Sweden and Denmark from the prescription registers, patient registers, population registers, the national bureaus of statistics, and the Swedish National Diabetes Register. Data sources and general methods used have been described in detail elsewhere (4,5). We included patients, aged 35–84 years, who filled their first prescription for either a GLP-1 receptor agonist or a dipeptidyl peptidase 4 (DPP-4) inhibitor (active comparator) during the study period and had a history of diabetic retinopathy (retinal photocoagulation, treatment with intravitreal agents, vitreous hemorrhage, vitrectomy, retinal detachment, retinal bleeding, and diabetic eye complications/retinopathy). We excluded patients who had previously filled prescriptions for any of the study drugs; had no hospital contact or prescription drug in the past year; had dialysis, renal transplantation, severe pancreatic disorders, end-stage illness, or drug misuse; or were hospitalized for any reason or had a hospital contact for diabetic retinopathy within 30 days before cohort entry. Using logistic regression, we estimated propensity scores for the probability of starting a GLP-1 receptor agonist versus a DPP-4 inhibitor given the status of 60 covariates, including sociodemographic characteristics, comorbidities, comedications, and health care utilization, at cohort entry; two-way interaction terms between country and each covariate were included in the model. Patients with a propensity score outside the overlapping area of the distribution for the two study drug groups were excluded.

Patients were considered exposed to the study drugs as long as prescriptions were refilled before the estimated end date of the most recent prescription, including a grace period of 180 days. The primary outcome, retinopathy complications, was a composite of retinal photocoagulation, treatment with intravitreal agents, vitreous hemorrhage, vitrectomy, retinal detachment, or retinal bleeding, which were identified by procedure and diagnostic codes. Patients were followed from cohort entry to end of exposure to the study drug, crossover to the other study drug, the outcome event, death, emigration, 5 years since cohort entry, or end of study period. Cox proportional hazards regression was used to calculate hazard ratios (HRs). Adjusted HRs were calculated using propensity score weighting. The study was approved by the Regional Ethical Review Board in Stockholm, Sweden, and the Danish Data Protection Agency.

The study population included 6,650 new users of GLP-1 receptor agonists (liraglutide, 91.6%; exenatide, 4.3%; dulaglutide, 2.7%; and lixisenatide, 1.4%) and 11,630 new users of DPP-4 inhibitors with history of retinopathy.

After propensity score weighting, the treatment groups were well balanced on baseline characteristics (data on file). Use of GLP-1 receptor agonists was not associated with a statistically significant increase in risk of diabetic retinopathy complications (adjusted HR 1.07 [95% CI 0.95–1.20]). Adjusted HRs in subgroup analyses by country, sex, age-group, insulin treatment, and glycated hemoglobin level at cohort entry (for patients in Sweden) were similar to those observed in the primary analyses (Table 1). The findings were similar in an additional analysis in which the follow-up time was truncated at 6 months after cohort entry and in sensitivity analyses applying an intention-to-treat exposure definition, truncation of large weights, or additional adjustments for glycated hemoglobin, blood pressure, albuminuria, estimated glomerular filtration rate, BMI, and smoking in the Swedish part of the cohort (Table 1).

In this nationwide cohort study of patients with existing retinopathy from two countries, we found no association between use of GLP-1 receptor agonists and diabetic retinopathy complications in analyses using DPP-4 inhibitors as the active comparator. Liraglutide comprised the majority (91.6%) of GLP-1 receptor agonist use in our study, and examination of individual GLP-1 receptor agonists remains a topic for future studies.

Data from the placebo-controlled SUSTAIN-6 trial suggest that retinopathy complications associated with semaglutide may occur due to early worsening of existing retinopathy following rapid improvement of glycemic control (1). We found no evidence of a risk increase associated with use of GLP-1 receptor agonists versus DPP-4 inhibitors among patients who might have a higher likelihood of experiencing a rapid improvement of glycemic control, including those with a higher glycated hemoglobin and those with insulin treatment at cohort entry. In addition, findings remained similar when follow-up time was truncated at 6 months after cohort entry, although incidence rates in both groups were higher in this time period.