Cleveland Clinic finds most people who stop GLP-1s switch or restart rather than simply regain

People who stop taking injectable semaglutide or tirzepatide often don't simply drift back to their old weight. A new retrospective study of 7,938 adults treated in a large health system in Ohio and Florida found that average weight one year after stopping changed by just +0.5% among those treated for obesity — in part because more than half of patients started something else or went back on the same drug ^[1].

The study, published in Diabetes, Obesity and Metabolism, used electronic health records from 1 January 2021 through 30 June 2025 ^[1]. Researchers included adults with overweight or obesity who started injectable semaglutide or tirzepatide for obesity or type 2 diabetes between 2021 and 2023, then stopped within 3 to 12 months. Stopping was defined as a gap of more than 90 days between running out of the previous supply and the next fill ^[1]. Patients averaged 55.7 years old (standard deviation 13.4), and 5,061 (63.8%) were female ^[1].

What happened after people stopped

In the year after discontinuation, 19.6% restarted the same medication and 35.2% received some other obesity treatment ^[1]. That included starting a different medication (27.4%), a lifestyle modification visit with a health professional (13.7%), and metabolic and bariatric surgery (0.6%) ^[1]. Those categories were not mutually exclusive, so a single patient could appear in more than one ^[1].

Weight loss while on the drug was moderate compared with what clinical trials report. From starting the medication to stopping it, mean weight change was −8.4% (95% CI, −8.7% to −8.1%) among people treated for obesity and −4.4% (95% CI, −4.7% to −4.2%) among those treated for type 2 diabetes ^[1].

After stopping, the group treated for obesity gained an average of 0.5% (95% CI, 0.0% to 1.0%) over the following year, while the group treated for diabetes actually lost a little more, −1.3% (95% CI, −1.6% to −1.0%) ^[1]. The authors are explicit that these averages conceal wide swings: there was "considerable individual-level variability" in what happened to any given person's weight ^[1].

The fine print

Not everyone in the cohort had usable weight data. The trajectory analysis drew on 6,634 patients who had a baseline plus a discontinuation or post-discontinuation weight and no censoring events ^[1]. The figure showing individual variation was based on 3,810 patients who had both a discontinuation weight and a one-year weight ^[1]. The median patient had weights recorded in 7 separate months (interquartile range, 4–13) from the start of treatment through one year after stopping ^[1].

This is observational data from routine care, not a randomized trial, and there was no comparison group of people who stayed on treatment. The abstract does not report why people stopped — cost, side effects, supply problems or reaching a goal — so that remains unknown from these sources ^[1]. It also does not break out results separately for semaglutide versus tirzepatide, or name the specific "other medications" people switched to ^[1]. Only injectable forms were studied, so oral semaglutide (Rybelsus) is not covered ^[1].

Several authors reported industry ties. One is currently employed by Novo Nordisk A/S and has reported advisory board fees from Eli Lilly, Boehringer Ingelheim and Abbott; others reported research funding or consulting relationships with Novo Nordisk, Eli Lilly, Amgen, Ethicon and Medtronic ^[1].

Why it matters for patients

Much of the public conversation about GLP-1 drugs assumes that stopping means rapid, near-total weight regain. This study suggests the real-world picture is messier. Roughly one in five people restarted the original drug and about one in three got a different obesity treatment within a year ^[1] — so the flat average weight curve reflects continued treatment, not evidence that weight stays off on its own after stopping.

The variability point is the practical one. An average of +0.5% does not mean most people gained half a percent; some people gained substantially and some lost more, and the study does not report how many fell into each group ^[1]. Real-world weight loss on treatment also ran lower than trial figures, at −8.4% for obesity ^[1], which is useful context for anyone comparing their own results with headline trial numbers.

For people weighing insurance changes, cost, or a planned pause, the finding that switching and restarting are common in routine care is worth discussing with a clinician who knows the full picture.