Stopping and restarting GLP-1 medicines: a study of 125,474 adults

Why this study matters

GLP-1 medicines work while people take them. Knowing how often people stop, what is linked to stopping, and whether they come back matters for anyone considering long-term treatment. This is the largest study we reviewed on this question. It used electronic health records, so the researchers could also look at how weight change and side effects related to stopping and restarting. It is a US study, but the patterns are relevant to Canadian readers thinking about what treatment might look like over time.

Who was in the study

The study included 125,474 adults with overweight or obesity who newly started liraglutide, semaglutide or tirzepatide between January 1, 2018 and December 31, 2023. Everyone had a body mass index (BMI, a measure of weight against height) of 27 or more, a weight recorded within 60 days before starting, and regular care in the previous year. The mean age was 54.4 years (standard deviation 13.1), 82,063 (65.4%) were women, and 76,524 (61.0%) had type 2 diabetes. The records came from a collective of US health care systems.

What the researchers did

This was a retrospective cohort study. Retrospective means it looked back at existing records. Cohort means a defined group was followed over time. Nobody was assigned a treatment.

People were followed for up to two years to see whether they stopped the medicine, and then for up to two more years to see whether they restarted. All analyses were done separately for people with and without type 2 diabetes. Rates were estimated with Kaplan-Meier methods, which account for people being followed for different lengths of time. Links between personal factors, weight change, side effects and stopping or restarting were examined with a model that allowed those factors to change over time.

What they found

One-year discontinuation was 64.8% (95% confidence interval 64.4% to 65.2%) for people without type 2 diabetes and 46.5% (95% CI 46.2% to 46.9%) for people with it (R32-C131). A confidence interval is the range in which the true value most likely sits; here the intervals are very narrow.

The restart figures come from 41,792 people who stopped and had a weight measurement recorded at the time.

Several factors were linked to stopping. For each 1% of starting weight lost, the hazard of stopping was 3.1% lower (95% CI 2.9% to 3.2%) with type 2 diabetes and 3.3% lower (95% CI 3.2% to 3.5%) without. Among people with type 2 diabetes, a household income above $80,000 US was linked to less stopping (hazard ratio 0.72, 95% CI 0.69 to 0.76). Moderate or severe stomach and bowel side effects during treatment were linked to more stopping: hazard ratio 1.38 (95% CI 1.31 to 1.45) with type 2 diabetes and 1.19 (95% CI 1.12 to 1.27) without. A hazard ratio above 1.0 means the event happened more often.

Weight regain was linked to restarting. Each 1% of weight regained after stopping was associated with a 2.3% (95% CI 1.9% to 2.8%) higher hazard of restarting for people with type 2 diabetes and 2.8% (95% CI 2.4% to 3.2%) for people without.

Side effects and people who stopped

Stopping was the main outcome, reported above. The study found that moderate or severe gastrointestinal side effects during treatment were linked to a higher rate of stopping. The summary we reviewed does not report how many people had such side effects.

What this study does not tell you

• Health records show what was prescribed, not why people stopped. Cost, supply shortages, coverage rules, reaching a goal and clinician decisions cannot be separated.
• The comparison between people with and without diabetes is not a fair test. The extracted data note that differences in access and in clinical circumstances confound it.
• This is a US population. Canadian coverage, prices and supply differ.
• The links between weight change, income, side effects and stopping are associations. They do not prove cause.
• Which medicine, and at what dose, each person used is not broken down in the summary we reviewed.
• The summary we reviewed does not state the funding source.

What it means in Canada

In a large US population, stopping a GLP-1 medicine within a year was common, and restarting was also common. This does not tell any Canadian reader what to expect personally, and it does not say whether a medicine is right for anyone. Coverage rules can affect whether people keep filling prescriptions; see public drug coverage by province. For information about specific medicines, see semaglutide in Canada, tirzepatide in Canada and liraglutide in Canada.

Related reading

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

Source