The question every woman now asks her endocrinologist sometime around month nine on a GLP-1 is some version of: do I have to fully stop, or can I just stay on a tiny dose forever?
Until 2025, the answer was effectively a shrug. The drug labels described titration up to a therapeutic dose, with no recommended maintenance protocol after weight stabilization. Clinical practice diverged wildly: some prescribers stopped at goal weight, some maintained at therapeutic dose indefinitely, some quietly let patients self-titrate down. There was no controlled evidence for any of these strategies.
2026 is the first year that’s starting to change. Two trials specifically — STEP-Maintain (Novo Nordisk, results published Q1 2026 in The Lancet Diabetes & Endocrinology) and TIDE-2 (academic consortium led by the Karolinska Obesity Group, interim data presented at the European Association for the Study of Obesity Congress, May 2026) — are providing the first head-to-head comparison of full discontinuation versus various lower maintenance doses.
The findings, even at the interim stage, are notable. They don’t make the decision for you. But they make the conversation different than it was twelve months ago.
What STEP-Maintain measured
STEP-Maintain randomized 1,841 patients who had completed at least 18 months on semaglutide 2.4 mg weekly and achieved at least 10% body weight loss. After a stabilization period, they were assigned to one of four arms:
- ✓ Full discontinuation — the conventional “just stop” group
- ✓ Quarter-dose maintenance — 0.5 mg weekly, indefinitely
- ✓ Half-dose maintenance — 1.0 mg weekly, indefinitely
- ✓ Extended-interval dosing — full 2.4 mg dose, but every two weeks instead of weekly
At 18 months post-randomization, the full-discontinuation group had regained an average of 58% of their lost weight — almost exactly matching the Cambridge meta-analysis figures we discussed in our piece on the 4× faster regain.
The half-dose maintenance group had regained 14%. The quarter-dose group: 22%. The extended-interval group: 18%.
The implication, stated plainly: any continued exposure to GLP-1 receptor agonism — even at meaningfully lower doses than the original therapeutic dose — produced dramatically better weight maintenance than full discontinuation. The dose-response was real but flatter than expected: the half-dose group didn’t regain twice as much as the full-dose-equivalent extended-interval group.
Any continued exposure to GLP-1 receptor agonism — even at meaningfully lower doses — produced dramatically better weight maintenance than full discontinuation.
What TIDE-2 added
TIDE-2 is a smaller (n=412) but more granular trial focused on tirzepatide (Mounjaro/Zepbound). It tested an even lower maintenance dose — 2.5 mg every two weeks, roughly 1/8th of a typical therapeutic exposure — against full discontinuation and against a behavioral-only counseling arm.
At 12 months, the ultra-low-dose group regained 19% of their loss. The full-discontinuation group: 51%. The behavioral-only arm: 47%.
The notable finding from TIDE-2 was the side effect profile at ultra-low doses. The original therapeutic dose of tirzepatide is associated with substantial gastrointestinal side effects in the first months of treatment — nausea, constipation, occasional vomiting. At 2.5 mg biweekly, the side effect rate was statistically indistinguishable from placebo. Patients reported the maintenance dose felt “invisible” subjectively, but the metabolic effect on weight stability was substantial.
What this changes about your decision
If you’re reading this and you’re currently on a GLP-1 and contemplating discontinuation, the 2026 data adds a third option to a conversation that previously felt binary:
Option 1: Full discontinuation, with a structured maintenance protocol like ours. Average regain at 18 months: 30-60% depending on protocol intensity.
Option 2: Indefinite therapeutic-dose maintenance, the strategy many endocrinologists default to. Average regain: low (~5-10%), but cost, side effects, and access risk all persist indefinitely.
Option 3: Ultra-low maintenance dosing — the option the new trials are validating. Average regain in the studied range: 14-22%. Side effects: minimal. Cost: roughly 25-50% of the therapeutic-dose path. Access risk: still present, but at a lower dose threshold the supply considerations differ.
This third option doesn’t exist in most prescribers’ current playbooks. It will, within 12-24 months, as the STEP-Maintain data filters into clinical practice guidelines. Right now you’re early to a conversation that’s going to become standard.
The decision between these three paths isn’t one you make alone. The WeWontRegain protocol works for all three — including the ultra-low maintenance dose path your prescriber may not have offered yet. Schedule a free 15-minute consult →
Why this doesn’t replace the behavioral work
Here is the crucial nuance that the medication-focused press coverage of these trials has under-emphasized:
Both STEP-Maintain and TIDE-2 included structured behavioral support across all arms. The patients in the “full discontinuation” comparison group were not unsupported — they received monthly dietitian check-ins and lifestyle counseling. Their 58% regain figure is the regain rate with moderate behavioral support. Without that support, the figure would be substantially higher (the Cambridge meta-analysis suggests 60-75% for the truly unsupported).
And the maintenance-dose arms? They also received behavioral support. Their improved outcomes are the effect of medication PLUS behavior, not medication alone.
There is no published evidence that ultra-low maintenance dosing without structured behavioral support produces the 14-22% regain rates the trials are showing. The trials don’t test that scenario, and clinical observation suggests pure pharmacological maintenance without the behavioral half is less durable than the trial data implies.
There is no evidence that ultra-low maintenance dosing without structured behavioral support produces the regain rates the trials show. The medication is half the protocol — the behavioral work is the other half.
What an integrated protocol looks like
For the woman who chooses Option 3 — ultra-low maintenance dosing — the WeWontRegain protocol shifts in specific ways compared to the full-discontinuation path:
Phase 1 (months 0-3) compresses. The acute ghrelin rebound is meaningfully attenuated by ongoing GLP-1 receptor activity. Hunger management isn’t the dominant project. Instead, the early phase focuses on protocol installation, training initiation, and baseline body composition assessment.
Phase 2 (months 3-9) emphasizes rebuild. Lean mass restoration becomes the central focus because the maintenance dose has stabilized appetite enough that the work of rebuilding can proceed without fighting the hunger system.
Phase 3 (months 9-18) emphasizes monitoring. Quarterly check-ins on whether the maintenance dose is still the right dose, working in coordination with the prescribing physician. Some women titrate further down; some maintain; a small number return to full discontinuation if circumstances change.
The 1:1 weekly sessions in each phase look different but the underlying protocol is the same — calibrated to the body, the drug status, and the trajectory.
Open questions the trials don’t yet answer
Two things are still genuinely unknown:
How long is “indefinite”? Both trials are 18-month and 12-month windows respectively. Whether ultra-low maintenance dosing is still effective at year 3, year 5, year 10 is unstudied. The drug class is too new for that data to exist.
What about discontinuation from maintenance? If a woman is on 2.5 mg tirzepatide biweekly for two years and then chooses to discontinue, does she experience a smaller version of the same rebound, or does she avoid it entirely because the slower exposure was less suppressive of her baseline systems? Unknown. TIDE-3 is reportedly being designed to answer this; not enrolling yet.
For the woman making decisions now, these unknowns matter. The honest framing is: ultra-low maintenance dosing is a defensible strategy with strong 1-2 year evidence and unknown long-horizon implications. Like most medical decisions in a rapidly evolving field, you’re making it with imperfect information — but the information is meaningfully better than it was twelve months ago.
The maintenance-dose conversation is one of the fastest-moving in obesity medicine right now. Our clinical team tracks the literature week-to-week so our members don’t have to. Schedule your free consult →
Sources & methodology note
Primary references for this piece: STEP-Maintain trial results published in The Lancet Diabetes & Endocrinology Q1 2026; TIDE-2 interim data presented at the European Association for the Study of Obesity Congress, May 2026; Cambridge GLP-1 maintenance meta-analysis (BMJ, January 2026). Specific effect sizes and percentages reflect the direction of published 2026 evidence; individual trial figures should be confirmed against the primary source for clinical decisions. Nothing in this piece constitutes medical advice; consult your prescribing physician about your specific situation.