The data you need to know before you stop
Two randomized withdrawal trials define what happens after you discontinue GLP-1 weight-loss pharmacotherapy.
STEP 4 (Rubino et al., JAMA 2021): 803 adults with BMI ≥ 30 (or ≥ 27 with comorbidity) completed 20 weeks of semaglutide 2.4 mg run-in, achieving about 10.6% body weight reduction. They were then randomized 2:1 to either continue semaglutide or switch to placebo for an additional 48 weeks. At week 68:
- Continuers: lost an additional 7.9% body weight. Total reduction from baseline: ~17.4%.
- Placebo switchers: regained 6.9% body weight. Total reduction from baseline: ~5.0%.
Translation: over the 48 weeks after stopping, placebo switchers regained roughly two-thirds of the weight they had lost during the run-in phase. The regain curve was asymptotic — sharpest in months 2–6 post-stop, tapering toward week 48.
SURMOUNT-4 (Aronne et al., 2024): 670 adults with obesity completed 36 weeks of open-label tirzepatide up to 15 mg, losing about 20.9% body weight, then were randomized to continue tirzepatide or switch to placebo for 52 weeks. At week 88:
- Continuers: lost an additional 5.5%. Total from baseline: ~25.3%.
- Placebo switchers: regained 14%. Total from baseline: ~9.9%.
The magnitude of regain tracked the magnitude of the original loss. Tirzepatide patients lost more and regained more in absolute pounds. The percent-of-loss regained was similar to STEP 4 — roughly two-thirds.
Why the weight comes back
Three mechanisms, in decreasing order of how much they matter:
- Appetite rebounds immediately. GLP-1 receptor agonism stops within 2–4 weeks (half-life ~1 week for semaglutide, ~5 days for tirzepatide). Ghrelin rises, PYY and GLP-1 endogenous signaling normalize. Hunger returns to pre-treatment level within 4–6 weeks. For patients who experienced 50–60% appetite suppression, the subjective rebound is dramatic.
- Set-point biology pulls hard. The hypothalamic body-weight regulation system treats the weight loss as a threat and increases orexigenic drive (Rosenbaum et al. metabolic adaptation data). Resting metabolic rate adjusts downward roughly 8–15% below predicted at the new lower weight — this adaptation persists for years.
- Behavioral bandwidth returns to baseline. Patients who lost weight on GLP-1 without learning new food and activity habits have no behavioral scaffolding when appetite returns. Real-world studies of patients who did not add diet and training during therapy showed the fastest regain rates.
Four stopping strategies and what to expect from each
1. Cold turkey. Stop entirely. This is the scenario modeled by STEP 4 and SURMOUNT-4. Expect 60–70% regain over 12–18 months if no behavioral intervention; 40–55% if you maintain aggressive protein + resistance training + food tracking.
2. Taper over 3–6 months. Step down dose (semaglutide 2.4 → 1.7 → 1.0 → 0.5 → off; tirzepatide 15 → 10 → 7.5 → 5 → off). No RCT data on whether this reduces regain, but physiologically it gives the hypothalamic system more time to adapt and lets behavioral habits lock in. Anecdotal patient reports suggest ~30–50% regain over 12 months versus 60–70% for cold turkey, but this is uncontrolled observational data.
3. Maintenance dose. Instead of stopping, reduce to the lowest dose that prevents regain. STEP 5 tested semaglutide 2.4 mg for 104 weeks and showed sustained loss. Some clinicians now recommend maintenance doses like semaglutide 0.5–1.0 mg weekly or tirzepatide 5–7.5 mg after goal weight, balancing cost, side effects, and regain prevention. Expected regain on maintenance: 10–25% over a year.
4. Bridge to bariatric surgery or next-gen therapy. Some patients stop GLP-1 because of side effects, cost, or pregnancy planning and bridge to alternative interventions. Bariatric surgery (sleeve, bypass) produces durable 25–30% loss; future therapies (orforglipron oral GLP-1, retatrutide triple agonist, cagrilintide-sema combos) may produce better outcomes.
The medication vs chronic-disease framing
GLP-1 discontinuation looks like a moral or willpower problem from the outside but is not. Obesity is a chronic disease with a biological set point that resists voluntary weight reduction. Stopping antihypertensives causes blood pressure to rise; stopping statins causes LDL to rise; stopping GLP-1 causes weight to rise. We do not tell hypertension patients to “just stop lisinopril and maintain a healthy blood pressure,” and the analogous expectation for GLP-1 is unrealistic. The standard-of-care framing, increasingly endorsed by the Obesity Medicine Association and the Endocrine Society, is: GLP-1 is chronic pharmacotherapy for a chronic disease.
That framing changes the cost calculation entirely. See the cost-per-pound and maintenance dose cost tools.
STEP 5, maintenance durability, and the two-year read-out
STEP 5 (Garvey et al., Nature Medicine 2022, N=304) extended the STEP 1 design to 104 weeks to answer the durability question. Semaglutide 2.4 mg continuers maintained 15.2% TBWL at week 104 vs 2.6% on placebo. This is the strongest evidence that weight loss is preserved while on drug at the top dose. It also implicitly tells you what happens if you never stop: the curve flattens but does not reverse. SURMOUNT-3 (ILI run-in + tirzepatide) reached 26.6% TBWL at 88 weeks on continuing drug. The durable-while-on-therapy pattern has held across every long-duration GLP-1 trial published.
The contrasting SELECT data matters here too. Patients in the semaglutide arm of SELECT maintained ~8–10% TBWL over 39.8 months of follow-up, and the 20% MACE reduction required that continued exposure. Stopping and regaining implies losing both the weight and the cardiovascular protection — a consideration that should weigh heavily for patients with pre-existing ASCVD who are on Wegovy specifically for the SELECT indication.
Who does better on discontinuation
Real-world data and subgroup analyses suggest lower regain rates in:
- Patients who lost < 10% total body weight on therapy (less set-point disruption to recover from).
- Patients who added significant structured exercise (3+ resistance sessions/week) during therapy.
- Patients who achieved and maintained protein 0.8–1.0 g/lb through therapy — preserves metabolic rate. See the diet on GLP-1 guide.
- Younger patients (set-point plasticity is better).
- Patients without T2D (insulin-resistance biology is a separate regain driver).
When stopping is the right decision
- Pregnancy or planning pregnancy within 2 months: mandatory discontinuation.
- Intolerable GI side effects despite dose reduction, antiemetics, and dietary modification: after 2–3 months of troubleshooting.
- Cost becomes prohibitive and no alternative access path exists: stopping is better than skipping doses or stretching pens.
- Acute pancreatitis, gallstones requiring surgery, medullary thyroid cancer diagnosis: clinical contraindications.
- Goal achieved and no comorbidity driving ongoing use, paired with an aggressive maintenance plan: the hardest case, requires individual discussion with prescriber.
The biology of regain in plain terms
Why does stopping a weight-loss drug cause the weight to come back when stopping a statin does not cause LDL to return to the original level as quickly? The answer is the body-weight regulatory system, which defends a “set point” through coordinated hormonal and neural signaling. Rosenbaum, Leibel, and colleagues have spent two decades documenting that after deliberate weight loss, ghrelin rises, leptin drops more than predicted for the new body mass, PYY and CCK responses blunt, resting metabolic rate drops ~8–15% below predicted, and non-exercise activity thermogenesis falls. The set-point system treats weight loss as an existential threat and pulls hard to restore the prior weight. GLP-1s work by pharmacologically overriding those signals; remove the drug and the system reasserts. This is not a willpower failure; it is biology. The clinical implication is that obesity, like hypertension and hyperlipidemia, is a chronic condition best managed with chronic therapy.
Practical protocol for stopping
- Weeks 1–4 post-stop: keep food logging (MyFitnessPal, Cronometer). Track weight weekly, not daily. Expect 2–4 lb fast water regain — not fat.
- Weeks 4–12: this is when appetite bounces hardest. Pre-plan meals. Keep protein 0.8–1.0 g/lb. 3+ resistance sessions/week non-negotiable.
- Months 3–6: the regain curve steepens. If you are gaining faster than 1 lb/week, call your prescriber — restart low-dose therapy is an option.
- Months 6–12: regain decelerates. If you have gained back less than 30% of what you lost, your strategy is working. Above 50%, consider resuming therapy.
FAQ
If I resume, will I lose the weight again?
Yes. Limited data on restart show similar response to original treatment. A 2024 analysis of stop-and-restart patients showed 85–90% recovered their previous nadir within 6 months of restarting.
Can I take a 2–3 month “drug holiday”?
Physiologically possible. You will regain some weight during the holiday and lose it again on restart. Why you would do this is the real question — side effect break, cost gap, or life event. Discuss with prescriber; there is no trial evidence this is helpful.
What about transitioning from Wegovy to Zepbound to lose more?
Different question entirely — that is a switch, not a stop. SURMOUNT-5 head-to-head showed tirzepatide 15 mg produces roughly 7 percentage points more TBWL than semaglutide 2.4 mg. Switch protocols typically involve a 1-week washout and starting tirzepatide at 2.5 mg regardless of prior semaglutide dose. See the Wegovy vs Zepbound tool.
Does stopping affect the SELECT cardiovascular benefit?
Yes. The MACE reduction in SELECT required continued semaglutide through 39.8 months median follow-up. Stopping almost certainly reverses the benefit over time, though the durability post-stop is not formally studied.
Are there non-GLP-1 drugs that help prevent regain?
Phentermine-topiramate (Qsymia) and bupropion-naltrexone (Contrave) have been tried as post-GLP-1 maintenance with limited but non-zero evidence. Metformin may have a small role in patients with insulin resistance. None of these match the mechanism of a GLP-1 RA.