GLP-1 medicines and diabetic eye disease: what six large trials show

Why this study matters

Diabetic retinopathy is damage to the back of the eye caused by diabetes. It is a leading cause of sight loss. Doctors have long known that when blood sugar improves quickly, retinopathy can briefly get worse before it gets better. GLP-1 medicines lower blood sugar, so researchers wanted to know whether they affect eye disease over the long term. This analysis matters for Canadian readers with type 2 diabetes who take, or are thinking about, one of these medicines.

Who was in the study

This was not a single trial. It was a meta-analysis, which pools results from several trials to get a bigger picture. The researchers found six placebo-controlled cardiovascular outcome trials of GLP-1 receptor agonists that had planned in advance to record retinopathy. Cardiovascular outcome trials are large, long trials designed to check whether a medicine is safe for the heart. According to the extracted data, the six trials included 49936 participants. All of them were adults with type 2 diabetes and added heart risk. The median follow-up was 3.4 years.

What the researchers did

The researchers searched the published literature for GLP-1 heart-outcome trials that reported retinopathy. They used the published trial reports as their data. From each trial they extracted HbA1c (a blood test showing average blood sugar over about three months), systolic blood pressure, and body weight over the course of follow-up, by treatment group.

They then did two things. First, a random-effects meta-analysis, which combines the retinopathy results from all six trials into one estimate. Second, a meta-regression, which asks whether trials with larger changes in HbA1c, blood pressure or weight also had more retinopathy. A meta-regression works at the level of whole trials, not individual people.

What they found

The pooled odds ratio for retinopathy with a GLP-1 medicine versus placebo was 1.10, with a 95% confidence interval of 0.93 to 1.30 (R32-C121). An odds ratio of 1 means no difference. A confidence interval is the range in which the true value is likely to sit. Because this range includes 1, the analysis did not show a clear increase or decrease in retinopathy with the medicines. The results of the six trials differed from one another more than expected by chance (heterogeneity of 52.2%).

The meta-regression told a different story. Trials with larger average HbA1c reductions during follow-up had more retinopathy (slope 0.77). No relationship was found for blood pressure or weight. Looking at HbA1c at set times, the link was present at 3 months and at 1 year. Each extra 0.1% (1.09 mmol/mol) of HbA1c reduction was associated with a 6%, 14% or 8% increase in the log odds of retinopathy at 3 months, 1 year and overall follow-up respectively.

The authors conclude that the size of the HbA1c drop was correlated with retinopathy risk in this group, but that the long-term effect of better glucose control on eye disease was not measured in these trials. They suggest that eye health should be checked when glucose-lowering treatment is being stepped up.

Side effects and people who stopped

This analysis looked only at retinopathy. It does not report other side effects or how many people stopped treatment in the trials.

What this study does not tell you

• None of the six trials was designed to measure eye outcomes, and they used different definitions and methods for recording retinopathy.
• The meta-regression is based on only six data points (one per trial), which limits how much weight it can carry.
• The meta-regression cannot show that lower HbA1c caused eye problems in any individual person. It shows a pattern across trials.
• A median follow-up of 3.4 years is short compared with the time retinopathy takes to develop.
• All participants had type 2 diabetes and added heart risk. The findings do not apply to people without diabetes using these medicines for weight management.
• The summary we reviewed does not state who funded the analysis.

What it means in Canada

For Canadian readers with type 2 diabetes, this analysis is one reason clinicians often talk about eye checks when blood sugar treatment changes. It did not find a clear direct effect of GLP-1 medicines on retinopathy across the six trials. For how these medicines are used in Canada, see semaglutide in Canada and tirzepatide in Canada. For coverage questions, see public drug coverage by province.

Related reading

Other studies in this library that bear on the same question.

Source