How to Stay Off GLP-1s: The 2026 Maintenance Playbook

8.1%. That's the share of patients still on an anti-obesity GLP-1 three years after starting, per a Prime Therapeutics commercial-claims analysis cited in the May 2026 Managed Healthcare Executive cover story. Six-month Medicaid retention is better at about 61%. After tirzepatide withdrawal, the Horn et al. SURMOUNT-4 post-hoc analysis in JAMA Internal Medicine (February 2026) showed 25%+ of lost weight regained within 12 months for most patients, with cardiometabolic improvements reversing proportionally. The Oxford-led January 2026 cohort study put it more bluntly: stopping GLP-1s drives weight regain faster than stopping any diet program.

If you take those three numbers at face value, the field's biggest practical problem isn't access or affordability anymore. It's that the chronic-disease drugs we have work great while you take them, and almost everyone stops taking them. Insurance turnover, cost step-edits, GI tolerance, and 'I lost the weight, why am I still on this' all push toward discontinuation; the drugs are pharmacologically designed for indefinite use. That gap is where most real-world outcomes go to die.

On May 12, 2026, Eli Lilly published SURMOUNT-MAINTAIN in The Lancet alongside concurrent ECO 2026 presentation. The trial enrolled adults from SURMOUNT-5 who had reached their maximum tolerated dose of tirzepatide (Zepbound) and lost meaningful weight, then re-randomized them to continue at MTD, drop to 5 mg, or move to placebo for 52 additional weeks of treatment plus 60 weeks of follow-up.

The headline at week 112: patients who continued tirzepatide MTD preserved all of their initial weight loss. Patients who dropped to the 5 mg dose preserved their loss except for an average 5.6 kg of regain — a partial maintenance result that's substantially better than placebo but not zero. Patients on placebo lost most of what they had gained.

The practical implication: there's no Phase 3 evidence yet for fully stopping tirzepatide and keeping the loss. But there's now Phase 3 evidence that you can taper to a maintenance dose at less than half the MTD and keep most of the result. For a patient whose insurance is denying continued MTD coverage, or who can't justify the cost of the higher dose long-term, the 5 mg-or-equivalent step-down is no longer guesswork. The kg-of-regain number gives prescribers something to negotiate against.

The companion Lilly publication in Nature Medicine on May 12, 2026, took the same SURMOUNT-5 patient pool and asked a different question: what if patients switch from injectable to oral? ATTAIN-MAINTAIN re-randomized participants who had reached their MTD on semaglutide (Wegovy) or tirzepatide (Zepbound) to either Foundayo (orforglipron) or placebo for 52 weeks.

At week 52, patients who switched from Wegovy MTD to Foundayo regained an average of just 0.9 kg — effectively preserving the injectable-phase result. Patients who switched from Zepbound MTD to Foundayo regained 5.0 kg, a slightly larger but still partial regain. Across both arms, 79.3% of the originally lost weight was preserved on Foundayo vs only 37.6% on placebo.

The Wegovy-to-Foundayo result is the surprising part. A daily oral pill held nearly all the weight loss of the maximum-tolerated weekly injectable. Mechanistically that makes sense — orforglipron is a GLP-1 agonist at the same receptor as semaglutide, just with a different molecular scaffold. Practically it means that the path 'inject Wegovy hard, switch to Foundayo pill for chronic maintenance' has Phase 3 evidence behind it for the first time. Same goes for the tirzepatide-to-Foundayo route, though with a slightly larger regain expectation.

A third piece of the maintenance puzzle landed on May 12 in a Wegovy STEP UP post-hoc analysis: 27% of patients on the higher-dose 7.2 mg formulation were 'early responders' who lost at least 15% of body weight within the first 24 weeks. Those patients went on to lose a mean 27.7% of body weight at week 72.

Why that matters for maintenance: early-response status is a cheap, observable signal for which patients are likely to do best on long-term therapy and tolerate the cost-benefit tradeoffs of staying on the drug indefinitely. A patient who hits 15% by week 24 has a clear case for continuing. A patient who hits 5% by the same milestone is having a more modest response, and the maintenance conversation looks different — possibly a switch to a different molecule, possibly a step-up to a higher dose, possibly stopping and trying an alternative approach.

The practical use: an early-response check at weeks 16-24 is a reasonable decision point for the long-term plan. Hit the threshold? The maintenance economics work out. Miss it? Have a different conversation before locking into multi-year therapy.

Taking SURMOUNT-MAINTAIN, ATTAIN-MAINTAIN, the Horn JAMA IM withdrawal data, and the early-responder lens together, the 2026 maintenance protocol looks like this:

Phase 1 (months 0-12): titrate to maximum tolerated dose. Lose what you're going to lose. Don't try to maintain on a half-dose during this phase — the trials show the loss curve is steepest at MTD and tapering early leaves results on the table.

Phase 2 (months 12-24): weight plateau plus 6+ months of stability. Run the labs (ferritin, B12, vitamin D, A1c, lipids, comprehensive metabolic panel, TSH). Confirm the cardiometabolic improvements are real and stable. Reassess body composition — DEXA if available — to make sure muscle loss isn't running ahead of fat loss.

Phase 3 (months 24+): the maintenance decision. SURMOUNT-MAINTAIN-style dose reduction to ~5 mg tirzepatide (or proportional) is the trial-validated path. ATTAIN-MAINTAIN-style switch from injectable to Foundayo is the trial-validated alternative. Full discontinuation isn't trial-validated — the Horn JAMA IM and Oxford cohort data both point to fast regain in the absence of any pharmacologic anchor.

The fourth option, increasingly cited but not yet Phase 3 validated: very-low-dose continuation. The community on the obesity and biohacking forums has been experimenting with semaglutide doses of 0.25-0.5 mg/week (vs the standard 2.4 mg) for maintenance, with anecdotal reports of preserved weight loss at quarter-dose. The mechanistic story is reasonable — GLP-1 receptor saturation may not require the full therapeutic dose for appetite-regulation effects — but the data is observational and the dosing is below trial-tested ranges. Treat it as an experiment, not a default.

Behind the trial data, pharmacy benefit managers are quietly building bridge programs that operationalize maintenance protocols. Optum Rx Weight Engage, Rightway Care Complete, and MedImpact's GLP-1 Benefit 360 all embed multidisciplinary support — typically a registered dietitian, a behavioral health touchpoint, and structured dose-step-down protocols — into the GLP-1 benefit. The programs are designed to lift the 8.1% three-year persistence number by changing what patients do at the 12-month and 24-month inflection points where most drop off.

These aren't widely advertised. Most patients don't know their PBM has a maintenance arm; most prescribers haven't built referrals to them into the standard workflow. But the programs exist, and for a patient who's reached their target weight and is wondering what to do next, asking the pharmacy explicitly about a GLP-1 transition program is worth doing.

The broader policy question — whether Medicare and Medicaid coverage will catch up to the maintenance model — sits unresolved. The April 22, 2026 cancellation of CMS's BALANCE pilot pushed it back. The Medicare GLP-1 Bridge (scheduled to start July 1, 2026 for some Part D beneficiaries at a $50/month patient cost) is the most concrete coverage development, but the eligibility rules are restrictive.

The mechanistic picture has become clearer over the last 18 months. When patients discontinue a GLP-1, three things happen on a predictable schedule:

Appetite returns within days. The pharmacologic effect on hypothalamic appetite centers reverses quickly once the drug is cleared — semaglutide's half-life is about a week, tirzepatide's similar, orforglipron shorter still. By week 2-4 post-discontinuation, most patients report appetite is back to pre-treatment baseline or higher.

The set-point hypothesis kicks in. The body defends a higher fat-mass baseline through several pathways: lower resting metabolic rate (the 'metabolic adaptation' adipose-tissue-aware homeostatic response), elevated ghrelin (the appetite-stimulating peptide that GLP-1s suppress), and reduced leptin sensitivity. The combined effect is that calorie balance shifts toward weight gain at a level of food intake that previously held the loss.

The behavioral and contextual scaffolding fades. People who lost weight on a GLP-1 often did so without the diet and exercise habits that would maintain it independently. Once the appetite suppression is gone, the absence of those habits becomes the rate-limiting variable. The Omada Health behavioral-support data from April 2026 showed that structured programs alongside GLP-1s roughly double the muscle-preservation and fat-loss-specificity of the drug-only protocol — and presumably also preserve more of the result on discontinuation. The data here is still building.

The upshot: maintenance is pharmacology and behavior in combination. The trial protocols (SURMOUNT-MAINTAIN, ATTAIN-MAINTAIN) test the pharmacology arm. The behavioral arm has to be assembled separately, and most patients don't get help with it.

If you're on a GLP-1 and considering coming off, the order I'd run through:

First: are you actually at your target? If your weight loss is still trending downward week-to-week, you're not yet at the maintenance phase. SURMOUNT-MAINTAIN and ATTAIN-MAINTAIN both enrolled patients who had reached their plateau, not patients mid-loss. Discontinuing during active loss likely produces faster and larger regain.

Second: what does your maintenance budget look like? If you can afford continued therapy at maintenance dose ($300-500/month most-common range for cash pay, less with savings cards), the trial-evidence-based path is to step down rather than stop. SURMOUNT-MAINTAIN shows ~95% loss preservation on 5 mg tirzepatide and ATTAIN-MAINTAIN shows ~99% on the Wegovy-to-Foundayo path. The maintenance economics are favorable if the absolute cost works.

Third: if cost or insurance forces a full discontinuation, plan for it. Slow taper over 8-12 weeks rather than abrupt stop. Lock in the muscle-preservation habits (resistance training 2-3×/week, 1.2-1.6 g/kg protein) for 6+ months before stopping. Schedule a follow-up at the 3-month post-stop mark to assess regain trajectory and re-engage if needed.

Fourth: if regain happens, restart sooner rather than later. The 'wait and see if I bounce back' approach generally costs more, not less. Patients who restart at the 5-10% regain threshold tend to do better on re-initiation than patients who wait until they've regained 80%+ of their lost weight.

The trial data finally supports a clear, evidence-based maintenance protocol. The behavioral half remains the bottleneck. Both halves matter.