What Happens When People Stop GLP-1s
Weight regain data from every trial withdrawal study and real-world dataset, plus metabolic parameter rebound. The ~67% regain from RCTs is a floor, not a ceiling — real-world evidence points to 74–100% regain at 12–18 months for patients who truly discontinue.
Last updated: April 2026
Context
The question the persistence data forces
The persistence data shows that the vast majority of patients stop GLP-1 therapy within three years. The question this page addresses is what happens next.
The answer from clinical trial withdrawal studies is that approximately two-thirds of weight lost is regained within one year of stopping — a figure remarkably consistent across semaglutide, tirzepatide, and liraglutide. But this ~67% figure, derived from ideal conditions with selected responders, structured withdrawal, and ongoing monitoring, is almost certainly a floor, not a ceiling.
Real-world data paint a more complex picture. The largest observational dataset (Optum, n=18,228) shows 74% of lost weight regained at one year. The most granular study (Twin Cities, n=130) found that 49% of patients who had lost weight ended up heavier than their pre-treatment baseline. A 2026 BMJ meta-analysis projects complete return to baseline weight within 18 months. One notable exception — Cleveland Clinic's 0.5% regain — appears driven by the fact that over half the cohort transitioned to alternative obesity treatment after discontinuation, demonstrating that treatment continuity can meaningfully attenuate regain.
The combination of discontinuation rates and regain rates means that the durable weight-loss benefit at the population level may be substantially smaller than headline trial numbers suggest.
Trial evidence
What RCT withdrawal studies showed
Every published withdrawal study shows substantial regain. The ~67% figure replicates across different drugs and treatment durations — but these trials measured regain under ideal conditions that systematically favor lower estimates.
| Study ↓ | Drug ↓ | N ↓ | Tx duration ↓ | Off-tx follow-up ↓ | Wt loss on tx ↓ | % of loss regained ↓ | Net from baseline ↓ |
|---|---|---|---|---|---|---|---|
| STEP 1 Extension Wilding et al., Diabetes Obes Metab, 2022 | Semaglutide 2.4 mg Injection | 228 | 68 wk | 52 wk | −17.3% | ~67% | −5.6% |
| STEP 4 Rubino et al., JAMA, 2021 | Semaglutide 2.4 mg Injection | 268 | 20 wk run-in | 48 wk | −10.6% | ~65% | ~−3.7% |
| STEP 10 McGowan et al., Lancet D&E, 2024 | Semaglutide 2.4 mg Injection | ~138 | 52 wk | 28 wk | −13.9% | ~43% | −7.9% |
| SURMOUNT-4 Aronne et al., JAMA, 2024 | Tirzepatide 10/15 mg Injection | 335 | 36 wk lead-in | 52 wk | −20.9% | ~67% | −9.9% |
| SURMOUNT-CN Chen et al., Life Metabolism, 2025 | Tirzepatide 10/15 mg Injection | 152 | 52 wk | 26 wk | −17.8 to −22.9% | ~51–54% | −8.7 to −10.6% |
Weight regain was still continuing at the final study visit in the STEP 1 Extension — the trajectory had not plateaued at 52 weeks off-treatment. A SURMOUNT-4 post-hoc showed 82.5% of withdrawn participants regained ≥25% of their loss, and 24% regained ≥75%. The 67% figure at 12 months is a snapshot of an ongoing process, not an endpoint.
Real-world evidence
What real-world data actually show
Multiple independent real-world datasets converge: regain in clinical practice exceeds the RCT benchmark of ~67%. The most likely figure is 74–100% regain at 12–18 months for patients who truly discontinue without alternative treatment.
| Source | N | Setting | Regain at ~1 year | Key finding |
|---|---|---|---|---|
| West et al. BMJ, 2026 | 9,341 | 37-study meta-analysis | 100% projected at 18 mo | Sema/tirz discontinuers regain 0.8 kg/month (9.9 kg in first year). Projected full return to baseline at 1.5 years. |
| Weintraub et al. Obesity Week, 2025 | 18,228 | Optum EHR-claims | ~74% of loss regained | Progressive regain (+4.5% at 3 mo, +7.5% at 12 mo). No plateau evident. 58% regained at least some weight. |
| Abdel-Bary et al. Obesity Medicine, 2025 | 130 | Allina Health (Twin Cities) | 65.4% gained weight | 49% of patients who had lost weight exceeded their pre-treatment baseline. Mean loss at discontinuation was only −2.26%. |
| Shah et al. Diabetes Obes Metab, 2025 | 289,000+ | Narrative review synthesis | 60–90% regain | Glycemic, BP, and lipid parameters return to baseline within 12–18 months. |
| Gasoyan et al. Diabetes Obes Metab, 2026 | 7,938 | Cleveland Clinic EHR | 0.5% | ~55% received alternative treatment after stopping. See Cleveland Clinic analysis below. |
If 50–68% of patients discontinue within a year (see persistence data), and most regain 74–100% of lost weight within 18 months, the durable weight-loss benefit at the population level may be substantially smaller than headline trial numbers suggest.
Trajectory
Rate and trajectory of regain
The RCT trajectory shows the V-shape under ideal conditions. The real-world trajectory — shown below it — is steeper and does not plateau within the measurement window.
The two charts illustrate a crucial divergence. The RCT trajectory shows patients losing 17% and retaining −5.6% at one year off treatment. The real-world trajectory shows patients losing ~10%, regaining faster without monitoring, and projecting back to baseline within 18 months. The exponential model from RCTs may describe the shape of regain, but the magnitude is worse in practice.
Structural bias
Why RCTs systematically underestimate real-world regain
The gap is not random noise. Seven structural features of withdrawal studies all push in the same direction — toward making RCTs look more favorable than reality.
Treatment continuity
How Cleveland Clinic achieved only 0.5% regain
One study reported substantially less regain than other datasets. The difference appears to reflect what happened after patients stopped their initial GLP-1 — and offers a model for how health systems can mitigate regain through structured transitional care.
Gasoyan et al. (Cleveland Clinic, 2026): 0.5% regain at 1 year
Among 7,938 patients who discontinued injectable semaglutide or tirzepatide, the obesity cohort regained only 0.5% at one year. The key to understanding this result is that a majority of patients transitioned to alternative treatments rather than simply stopping all therapy.
Gasoyan acknowledged that many patients "restart the medication or transition to another obesity treatment, which may explain why they regain less weight than patients in randomized trials." The Cleveland Clinic data demonstrates that treatment continuity — medication switching, lifestyle visits, GLP-1 restart — can meaningfully attenuate regain. The study did not report weight outcomes separately for the ~45% who received no further treatment, so the regain rate for true discontinuers remains unknown. Most employer benefit designs do not currently replicate this level of transitional care, but the data suggests they could benefit from doing so.
Beyond weight
Metabolic parameter rebound
Metabolic improvements do not all reverse at the same rate. Blood pressure reverts fastest, glycemic markers track weight, and lipid/inflammatory improvements show partial durability.
Rebounds fastest (weeks)
Tracks weight regain (months)
Most durable improvements
Liver fat/MASH: No major RCT has reported liver fat data after GLP-1 cessation. ESSENCE demonstrated MASH resolution on treatment but had no off-treatment extension.
Cardiovascular impact
Cardiovascular protection and discontinuation
SELECT established a 20% MACE reduction with continuous semaglutide. Observational data now provide initial evidence on what happens to cardiovascular protection after stopping.
Al-Aly et al., BMJ Medicine, March 2026
Target trial emulation: 333,687 U.S. veterans with T2D (132,551 GLP-1 users vs. 201,136 sulfonylurea users).
GLP-1 use for less than 18 months followed by discontinuation showed no significant residual CV protection. Interrupted-and-resumed use yielded only a 12% risk reduction vs. 18% with continuous therapy. The investigators described this pattern as "metabolic whiplash." These findings suggest that cardiovascular benefits may require sustained treatment to maintain, though RCT-level evidence on post-discontinuation CV outcomes remains absent.
Cross-drug comparison
Regain is similar across drugs when scaled to initial loss
More potent agents cause more absolute regain, but proportional regain converges around ~65–67% in RCTs. Real-world data push this higher.
Who regains more
Predictors of regain
Clinical consensus
The chronic disease argument
The consensus is remarkably uniform. The AMA, Obesity Society, World Obesity Federation, Endocrine Society, AACE/ACE, AAP, EASO, AGA, and WHO all classify obesity as a chronic, relapsing disease requiring chronic management. The World Obesity Federation states that "short-term treatments do not change the underlying biology."
Dissenting positions exist. The AMA's own Council on Science recommended against the disease classification in 2013 — the House of Delegates overrode this. Academic critics argue the model promotes pharmaceutical over structural solutions. A 2024 JAMA Viewpoint proposed "short-term, intermittent GLP-1 therapy paired with sustained lifestyle interventions."
If the medical consensus is indefinite pharmacotherapy, and most patients stop within 3 years, there is a significant gap between the clinical model and real-world usage patterns. The regain data helps explain why this gap matters: without ongoing treatment or structured transitional care, most of the clinical benefit appears to be temporary.
The restart cycle
Over one-third restart within a year
Outcomes on reinitiation remain a critical evidence gap. No study reports whether patients achieve equivalent weight loss on a second course. The Al-Aly data suggest interrupted use yields inferior CV protection (12% vs. 18%), arguing against cycling. The pattern — treatment, discontinuation, regain, reinitiation — creates a recurring cost cycle that complicates actuarial modeling.
Source: Rodriguez et al., JAMA Network Open, 2025 (125,474 adults, Truveta). DOI: 10.1001/jamanetworkopen.2024.57349
Model it
What happens to your population?
Start with a cohort on GLP-1 therapy. Apply persistence and regain rates. See how many maintain ≥5% weight loss over time. Default uses real-world regain assumptions (75%).
GLP-1 population outcome model
Connects the persistence data with the regain data from this page.
References
Sources
Questions about this data? Corrections or updates?
Get in touch