Tirzepatide: what a finished evidence base actually looks like

In plain terms

Tirzepatide, sold as Mounjaro and Zepbound, is a lab-made copy of a gut hormone, altered so it lasts about a week in the body and so it switches on two hormone receptors instead of one. Those receptors help the body release insulin when blood sugar rises, slow the stomach down, and reduce appetite.

Unlike the other three compounds here, tirzepatide has been through the full sequence of testing. Thousands of people have taken it in randomised trials where a comparison group got a dummy injection or a rival drug, and the results were published with the methods attached ^[16][17][13].

That makes this page unusual in one respect. It is not about a missing study. It is about what the evidence looks like once the main studies exist, and about the fact that new design questions appear as soon as the old ones are answered.

What it is

Tirzepatide is a linear 39-amino-acid synthetic peptide built on the native sequence of glucose-dependent insulinotropic polypeptide, with a C20 fatty diacid attached through a glutamic acid linker and two spacer units to a lysine side chain. The fatty-diacid arm gives it high affinity for albumin, which is what produces a half-life long enough for once-weekly administration. It appears in the literature as LY3298176 and is described as a twincretin or dual incretin mimetic.

It is the only member of this site's four with regulatory approval. The FDA approved it in May 2022 for type 2 diabetes mellitus ^[14]; a second approval in November 2023 covers chronic weight management in adults with obesity or with overweight plus a weight-related condition, and a later approval covers moderate-to-severe obstructive sleep apnoea in adults with obesity. Approved formulations are prescription-only. It is not approved for type 1 diabetes ^[14].

That status is why it belongs on a page about research fundamentals. Approval is not a marketing fact; it is the observable trace of a particular set of study designs having been run and having survived review.

How it works

Tirzepatide is the first approved dual incretin agonist. A single peptide activates both the GIP receptor and the GLP-1 receptor.

Engaging both enhances glucose-dependent insulin secretion, suppresses glucagon, slows gastric emptying, and reduces appetite and food intake. Glucose dependence is the mechanistically important detail: insulin release is stimulated in proportion to circulating glucose, which is why the agent on its own carries a low risk of driving blood sugar too low, and why that risk reappears when it is combined with agents that do not share the property.

The design consequence of dual agonism is that it created a specific, answerable comparison. If engaging two receptors is better than engaging one, that is a testable proposition against a selective GLP-1 receptor agonist, and it has been tested directly rather than argued from mechanism ^[17][13]. Mechanistic stories about combination effects are cheap; the head-to-head trial is what converts one into a finding.

What the research shows

This is the one page on the site where the rows fill from the bottom to the top.

Placebo-controlled randomised trial. SURMOUNT-1 enrolled 2539 adults with obesity, or with a body mass index of at least 27 plus a weight-related complication, and without diabetes, for 72 weeks. Mean weight change at week 72 was -15.0% at 5 mg, -19.5% at 10 mg and -20.9% at 15 mg, against -3.1% with placebo. The most common adverse events were gastrointestinal, mostly mild to moderate, and concentrated during the escalation period ^[16].

Active-comparator trial in diabetes. SURPASS-2, an open-label 40-week phase 3 trial in 1879 adults with type 2 diabetes, reported reductions in glycated haemoglobin of 2.01, 2.24 and 2.30 percentage points at 5, 10 and 15 mg against 1.86 percentage points for semaglutide (Ozempic, Wegovy) at 1 mg, with tirzepatide noninferior and superior at every dose. Weight reductions were also greater, with treatment differences of -1.9, -3.6 and -5.5 kg ^[17].

Active-comparator trial in obesity. SURMOUNT-5, a phase 3b open-label head-to-head trial in 751 adults with obesity and without type 2 diabetes, randomised participants to the maximum tolerated dose of either agent for 72 weeks. Least-squares mean weight change at week 72 was -20.2% with tirzepatide against -13.7% with semaglutide, with a p-value below 0.001, plus a greater reduction in waist circumference and higher proportions reaching each weight-loss threshold ^[13].

Pooled safety analysis. A systematic review and meta-analysis of nine randomised controlled trials covering 9871 participants examined two specific signals. Tirzepatide was not associated with a statistically significant increase in pancreatitis, with a relative risk of 1.46 and a 95% confidence interval from 0.59 to 3.61. It was associated with a significantly increased risk of the composite of gallbladder or biliary disease, relative risk 1.97 with a 95% confidence interval from 1.14 to 3.42, while no individual component reached significance on its own ^[15].

That last study is the most instructive one on the site for a reader thinking about design. The gallbladder signal is invisible in any single trial and visible in nine pooled. Rare events are a sample-size problem, not an attentiveness problem, and the design that answers them is a different design from the one that establishes efficacy.

Reported effects, cautions and safety

What follows is anecdotal, not clinical evidence: it comes from research-use and patient communities, is uncontrolled and unmeasured, and is included because the shape of the reports is worth comparing against the trial record. No quantities accompany it.

On the benefit side, appetite suppression and quieter preoccupation with food are frequently reported. Increased energy, reduced fatigue and improved mood or confidence are commonly reported. Better sleep and easier breathing at night, reduced joint pain, and self-reported improvements in blood sugar are reported sometimes. On the adverse side, nausea after each increase in exposure is frequently reported; constipation and diarrhoea alternating, and injection-site reactions, are commonly reported. Sulfur-tasting burps, taste changes and food aversions, hair thinning, and a stall in weight loss after an initial period are reported sometimes, and concern about muscle loss appears often enough to count as a recurring theme rather than an isolated report.

The striking thing is the overlap. Unlike the other three compounds here, the community reports and the trial record largely describe the same phenomena — gastrointestinal effects concentrated during escalation, lean-mass concern, hair shedding, plateau and regain. That agreement is not a coincidence, and it is not evidence that community reports are reliable in general. It is what community reports look like when a controlled record exists to check them against.

The cited cautions are as follows. The prescribing information carries a boxed warning regarding thyroid C-cell tumours derived from rodent studies, with contraindication in people who have a personal or family history of medullary thyroid carcinoma or multiple endocrine neoplasia type 2 ^[14]. Acute pancreatitis is a recognised class concern that is monitored on the label; the pooled analysis did not find a statistically significant increase ^[15]. Gallbladder and biliary disease was significantly increased in that same pooled analysis ^[15].

The remaining cautions are carried in this digest without an attached citation and are reported as such: dose-related gastrointestinal intolerance during escalation; low blood sugar when combined with insulin or sulfonylureas; delayed gastric emptying with its perioperative implications; loss of lean mass alongside fat; dehydration and acute kidney injury following severe gastrointestinal fluid loss; reduced reliability of oral hormonal contraception; weight regain after stopping; a higher discontinuation rate reflecting the trade-off between potency and tolerability; and reversible hair shedding during rapid weight reduction.

Where it sits in the fundamentals

Every row of the matrix is occupied, and the top two are occupied more than once. Placebo-controlled efficacy ^[16], active comparison in two different populations ^[17][13], and a pooled safety analysis at a scale no single trial reaches ^[15].

Its function on this site is comparative. Set against BPC-157, MOTS-c and CJC-1295, tirzepatide shows what the missing rows are worth. The difference between it and the other three is not that its molecule is better understood — MOTS-c arguably has the more precisely characterised molecular target ^[6] — but that somebody ran the trials.

That is the uncomfortable lesson of the four together. Mechanistic elegance and evidentiary strength are close to independent of one another, and a reader who ranks compounds by how good the mechanism story sounds will get the ordering exactly wrong.

Design brief: what is still open, and it is still design

No missing efficacy trial is specified here, because the efficacy question has been answered by the designs capable of answering it. What remains is a different set of questions, each with its own design requirement.

• Durability after stopping. Weight regain after discontinuation has been observed in trial extensions. The design that settles it is a randomised withdrawal with long follow-up and regain as a prespecified primary endpoint, rather than a secondary observation inside a trial built to measure something else.
• Body composition. Lean-mass change during rapid weight reduction is an active question. The design is a trial with a directly measured body-composition primary endpoint and a resistance-training co-intervention arm, not a substudy appended to an efficacy trial.
• Rare adverse events. The gallbladder and biliary signal emerged only at the scale of nine pooled trials ^[15]. Settling it requires either a trial powered for the event itself or prospective registry follow-up, and no further 72-week efficacy trial will resolve it however large.
• Long-term outcomes. The published endpoints are weight and glycated haemoglobin. Those are excellent intermediate measures and they are still intermediates. The design that closes the gap is a long-duration trial with hard clinical outcomes.
• Independence. Much of the strongest evidence is sponsor-funded, which is ordinary for a novel drug and worth stating when weighing the base. The design remedy is an independently funded replication, not a change to the trials already run.

The general point transfers to the other three pages. Answering the question a design was built for does not end the programme. It advances it to the next question, which needs a different design again.