Why patients stop GLP-1 drugs, and restart

Most people who start a GLP-1 receptor agonist for weight loss stop within a year. A new narrative review in Diabetes, Obesity & Metabolism (Heisey et al., June 2026) pulls together the real-world evidence on how often patients discontinue, why they do, and what happens next^[1]. The short version: discontinuation is common, often temporary, and driven by a mix of side effects, cost, and unmet expectations. The authors also flag a less obvious problem: researchers have not agreed on what "discontinuation" even means, making cross-study comparisons unreliable.

How many patients stop, and how quickly

The largest US dataset on this question comes from Rodriguez et al. (JAMA Network Open, January 2025), which tracked 125,474 adults initiating a GLP-1 RA for overweight or obesity. Among those with type 2 diabetes, 46.5% had discontinued within one year; among those without diabetes, the figure was 64.8%^[2]. By two years the rates climbed to 64.1% and 84.4% respectively. That pattern held across subgroups: patients prescribed these drugs specifically for weight loss stopped sooner and more often than those prescribed for blood-sugar control.

Heisey et al. note that rates vary widely across studies, in part because some define discontinuation as a 60-day gap in refills while others use 90 or 180 days. That inconsistency makes pooled estimates unreliable, and the review calls for standardised definitions before the field can claim a firm number.

Why patients stop

Three reasons dominate the literature. First, gastrointestinal side effects: nausea, vomiting, diarrhoea, and constipation are dose-dependent and well documented across the class. In the Rodriguez cohort, moderate-to-severe GI adverse events were significantly associated with early discontinuation^[2]. Second, cost. US patients without adequate insurance coverage face list prices that can exceed $1,000 per month, and prior-authorisation hurdles add friction even for those with coverage. Third, unmet expectations: some patients stop because the weight loss plateaus or because the side effects outweigh the benefit they perceive.

Heisey et al. argue that qualitative research on patient decision-making is thin. Most studies report claims-level data (prescription fills and gaps) rather than asking patients directly why they stopped. The few interview-based studies that exist point to a more tangled picture involving body image, social pressure, and the psychological adjustment to long-term injectable therapy^[1].

What happens after stopping

Weight regain is the headline consequence, and it is proportional to the original loss. A meta-analysis by Berg et al. (Obesity Reviews, 2025) pooled data from eight RCTs covering 2,372 participants and found a mean weight regain of 9.69 kg after discontinuing semaglutide or tirzepatide, compared with 2.20 kg after stopping liraglutide^[3]. The larger regain with the newer, more potent agents reflects the larger initial loss: patients who lost more regained more in absolute terms.

The SURMOUNT-4 trial of tirzepatide showed the same pattern in a controlled setting: participants switched to placebo at week 36 regained 14.0% of body weight by week 88, while those who stayed on tirzepatide lost an additional 5.5%^[4]. That 19.4-percentage-point gap is the clearest single-trial illustration of why regulators and clinicians increasingly frame GLP-1 RAs as chronic therapy rather than a course of treatment.

Reinitiation is common. In the Rodriguez cohort, 47.3% of patients with type 2 diabetes and 36.3% without restarted a GLP-1 RA within one year of stopping. Weight regain was the strongest predictor of reinitiation: each 1% gain in body weight since discontinuation raised the hazard of restarting by roughly 2.3% to 2.8%^[2].

What the review does not resolve

Heisey et al. are upfront about the gaps. The review does not deliver a single consensus discontinuation rate because the underlying studies use different definitions, populations, and follow-up windows. It does not cover the newer agents that have not yet accumulated enough real-world data (retatrutide, orforglipron, survodutide). And it does not address the question most patients care about: whether structured lifestyle support during and after treatment can reduce regain. That question is being tested in ongoing trials, but the evidence is not in yet.

Where this fits for readers

If you are on a GLP-1 RA or considering one, the review offers a useful frame: stopping is normal, often temporary, and does not mean the drug failed. But weight regain after cessation is also real and well documented. The decision to continue, pause, or stop belongs with a clinician who knows your case. For background on specific drugs in this class, see the tirzepatide and semaglutide pages on this site.