What it is
Tirzepatide is a synthetic 39 amino acid dual GIP and GLP-1 receptor agonist, acylated for albumin binding. Molecular weight 4813.53 Da. The half-life is about five days, longer than semaglutide’s one week measured in absolute terms but shorter in half-lives per dosing interval, which is part of why its label allows a 4-week titration step rather than semaglutide’s minimum-4-week convention with more common 4-week steps in practice.
What is approved, and at what dose
A record of what the FDA had approved as of 17 September 2026, not an instruction.
Mounjaro, for type 2 diabetes: 2.5 mg weekly for 4 weeks as a starting dose that is not intended to be therapeutic, then 5 mg. If needed, escalate by 2.5 mg increments after at least 4 weeks on the current dose, up to a maximum of 15 mg weekly.
Zepbound, for chronic weight management and, since 20 December 2024, moderate-to-severe obstructive sleep apnoea in adults with obesity: the same 2.5 mg for 4 weeks starting dose, then 5 mg, escalating by 2.5 mg every 4 weeks as tolerated to a maintenance dose of 10 or 15 mg. For OSA specifically the label does not carry a lower 5 mg maintenance option; only 10 and 15 mg are indicated maintenance doses.
Both products share one escalation schedule and one titration logic; they differ in indication and in which maintenance doses the label emphasizes. There is no 5 mg per week ceiling and no plateau below 10 mg described anywhere in either label as a stopping point, though a person and their prescriber can choose to stay on a lower dose if it is effective and tolerated.
Every tirzepatide product carries the same boxed warning for thyroid C-cell tumours seen in rodents, with human relevance undetermined. Contraindications are limited to hypersensitivity to tirzepatide or any component, and personal or family history of medullary thyroid carcinoma or Multiple Endocrine Neoplasia syndrome type 2. Prior pancreatitis is not a contraindication, a common misstatement worth correcting directly.
SURMOUNT-1: the foundational obesity trial
SURMOUNT-1 randomised 2,539 adults with obesity or overweight plus a weight-related comorbidity, without diabetes, to tirzepatide 5, 10 or 15 mg or placebo, weekly, for 72 weeks. On the treatment-regimen estimand, mean weight change was minus 15.0 percent at 5 mg, minus 19.5 at 10 mg, minus 20.9 at 15 mg and minus 3.1 percent on placebo. Discontinuation for adverse events ran 4.3, 7.1 and 6.2 percent across the three tirzepatide doses against 2.6 percent on placebo.
SURMOUNT-2: the diabetes population, and why the number looks smaller
SURMOUNT-2 enrolled 938 adults with type 2 diabetes and obesity or overweight. Weight loss came in lower, minus 14.7 percent at the top dose, against minus 20.9 in SURMOUNT-1. This is not a weaker drug in diabetes; it is the same well-documented attenuation of weight-loss response seen with every incretin therapy in a diabetic population, driven by differences in insulin physiology and disease duration, not a property specific to tirzepatide.
SURMOUNT-5: the only head-to-head, and what it actually compared
SURMOUNT-5 randomised 751 adults with obesity or overweight, without diabetes, to tirzepatide, maximum tolerated dose up to 15 mg, or semaglutide, maximum tolerated dose up to 2.4 mg, open-label, for 72 weeks. This is the only trial anywhere comparing tirzepatide against another approved incretin therapy rather than against placebo.
On the modified treatment-regimen estimand, tirzepatide produced minus 20.2 percent weight loss against minus 13.7 percent for semaglutide, meaning roughly 48 percent more weight loss with tirzepatide in this trial. Gastrointestinal-related discontinuation ran lower with tirzepatide, 2.7 percent against 5.6 percent for semaglutide, and vomiting was less frequent, 15.0 against 21.3 percent. Both drugs were dosed to each participant’s maximum tolerated dose rather than a fixed dose, which is the correct way to read “48 percent more”: it compares two real-world titration strategies, not 15 mg against 2.4 mg as fixed doses.
The cardiovascular trial that surprised people
SURPASS-CVOT randomised 13,299 adults with type 2 diabetes and established cardiovascular disease to tirzepatide or dulaglutide 1.5 mg, median follow-up 210 weeks. The primary MACE endpoint gave a hazard ratio of 0.92, 95 percent CI 0.83 to 1.01, meeting the prespecified non-inferiority margin but narrowly missing statistical superiority: the confidence interval’s upper bound sits at 1.01, just above the 1.00 threshold superiority requires.
This is a genuinely different result from semaglutide’s SELECT trial, which showed clear MACE superiority against placebo. SURPASS-CVOT compared tirzepatide against another active glucose-lowering therapy known to have cardiovascular benefit, not against placebo, which is a much harder bar to clear. A post hoc composite cardiorenal endpoint reported in June 2026 showed a hazard ratio of 0.84, but this was not a primary or confirmatory analysis and should not be read as an established cardiovascular protection claim the way SELECT’s MACE reduction can be.
SURMOUNT-4: what happens on withdrawal
In SURMOUNT-4, 670 participants who had completed a 36-week open-label lead-in on tirzepatide, during which the group had lost a mean 20.9 percent of body weight, were then randomised to continue tirzepatide or switch to placebo for a further 52 weeks. Those continuing lost an additional 5.5 percent; those switched to placebo regained, ending 14.0 percent above their week-36 weight. This is a randomised withdrawal design, the cleanest way to isolate what continued treatment versus stopping actually does, free of the confounding that comes from people switching therapies or restarting on their own in open-label follow-up.
A separate open-label extension, SURMOUNT-MAINTAIN, followed 378 people after treatment discontinuation without randomisation to a control arm, and found regain at 8 percent by some point, 25 percent by a later point and 67 percent by the final assessment, of the weight originally lost. These are different designs measuring different things: SURMOUNT-4 isolates the drug effect under controlled conditions; SURMOUNT-MAINTAIN describes what happens in a less controlled real-world-adjacent setting where people are not held to a fixed alternative.
Body composition: the DXA substudy
A dedicated dual-energy X-ray absorptiometry substudy of SURMOUNT-1, in a subset of 160 participants, found that at the top dose, total weight fell 21.3 percent, fat mass fell 33.9 percent and lean mass fell 10.9 percent. Reporting only the lean mass percentage without the accompanying fat mass percentage, as some secondary sources do, obscures that fat loss substantially outpaced lean loss. A broader 2026 review found a class-wide median of 28.3 percent of total weight lost coming from lean-mass-associated measures, tirzepatide included, which is consistent with, not contradictory to, the DXA substudy’s dose-specific figures.
What tirzepatide has not been shown to do
A phase 2 trial, SUMMIT, tested tirzepatide in heart failure with preserved ejection fraction and reported a positive result on its combined endpoint, but Lilly voluntarily withdrew the associated supplemental indication filing in May 2025, stating that a second, larger confirmatory trial would be required before an indication could be supported. As of 17 September 2026 there is no approved indication for heart failure, HFpEF specifically, and the SUMMIT result should be described as promising phase 2 data awaiting confirmation, not as an established benefit.
Regulatory and manufacturing history
Mounjaro was approved 13 May 2022; Zepbound followed 8 November 2023. The FDA declared the tirzepatide shortage resolved on 2 October 2024. Lilly petitioned to keep compounding restricted regardless via a citizen petition and the FDA issued a declaratory order narrowing the essentially-a-copy exemption on 19 December 2024; enforcement discretion for 503A compounders ended 18 February 2025 and for 503B outsourcing facilities 19 March 2025. A district court challenge to the shortage-resolved determination was rejected and the FDA’s position was upheld on 7 May 2025. By 31 May 2026 the FDA held more than 730 adverse event reports associated with compounded tirzepatide products, including reports naming salt forms not present in the approved drug.
The OSA indication for Zepbound was approved 20 December 2024, on the strength of two trials, and the label for that indication specifically excludes a 5 mg maintenance dose, using only 10 and 15 mg.
What circulates, and why no numbers appear here
This site does not reproduce dosing regimens for tirzepatide beyond the approved label. Every figure here is the modified treatment-regimen estimand, unless stated otherwise, and it describes what a specific trial did, not a protocol to follow. The approved titration exists because tolerability failures compound: skipping steps or inventing intermediate doses not on the label is one of the dosing-error patterns the FDA has linked to adverse events in the compounded tirzepatide market described above.