Tirzepatide has large human trials. Those trials do not validate a research vial.

Why this note exists

pepmg indexes 87 compounds. For most of them the honest summary is that the human record is thin, old, or absent, and the note writes itself as an inventory of what is missing. Tirzepatide inverts that. Its file is bigger than the files of every other compound on these shelves combined, it contains multiple large randomized outcome trials, and the results are not ambiguous.

Which makes the interesting question a different one. When the evidence for a molecule is overwhelming, the weak link stops being does it work and becomes is the evidence about this. That question is not answered by another trial, and it is not answered by a price index. It is answered, so far as anything answers it, by regulatory documents — and in the twenty months to September 2026 those documents changed in a specific, datable way.[15][16]

In the index generated for this note, 36 of the 108 vendors pepmg tracks carried a tirzepatide listing, 165 listings in total.[23] Four of those 108 vendor domains match, exactly, the websites named in published FDA warning letters about unapproved GLP-1 products.[17][18][19][20]

A note on what kind of evidence this is: every efficacy figure below is human data and carries its design, phase and randomized total. One paragraph reports rodent data and is labeled as such, because FDA's own boxed warning is built on it. pepmg reports doses only as their sources published them, and does not convert a label dose into a protocol or a vial size into a schedule.

The evidence

Five randomized trials, and what each one actually established

Tirzepatide is described in the literature as a dual agonist of the glucose-dependent insulinotropic polypeptide and glucagon-like peptide-1 receptors.[3] The obesity result that anchors everything else is SURMOUNT-1: a phase 3, double-blind, randomized trial in which 2,539 adults with a BMI of 30 or more — or 27 or more with a weight-related complication, excluding diabetes — were assigned 1:1:1:1 to once-weekly subcutaneous tirzepatide at 5, 10 or 15 mg, or placebo, for 72 weeks.[1]

Mean percentage change in weight at week 72 as reported in the trial publication. Phase 3, double-blind, randomized, 2,539 adults, mean baseline weight 104.8 kg.[1]

SURMOUNT-5 is the head-to-head. A phase 3b open-label trial — the design detail that belongs in the same sentence as the result — randomized 751 adults with obesity but without type 2 diabetes 1:1 to the maximum tolerated dose of tirzepatide, 10 or 15 mg, or of semaglutide, 1.7 or 2.4 mg, once weekly for 72 weeks.[2] The least-squares mean percent weight change at week 72 was −20.2 percent with tirzepatide against −13.7 percent with semaglutide (P<0.001), and waist circumference fell 18.4 cm against 13.0 cm (P<0.001).[2] The most common adverse events in both arms were gastrointestinal, mostly mild to moderate, mostly during dose escalation.[2]

SURMOUNT-OSA is two phase 3 trials reported together, in adults with moderate-to-severe obstructive sleep apnea and obesity — trial 1 in people not on positive airway pressure, trial 2 in people who were.[5] Mean baseline apnea-hypopnea index was 51.5 and 49.5 events per hour. At week 52 the mean change in AHI was −25.3 events per hour on tirzepatide against −5.3 on placebo in trial 1, a treatment difference of −20.0 (95% CI −25.8 to −14.2, P<0.001), and −29.3 against a placebo comparator in trial 2.[5] That is the trial behind an indication, and it is the reason the OSA claim on the label is not marketing.[11]

SUMMIT went to outcomes. 731 patients with heart failure, an ejection fraction of at least 50 percent and a BMI of at least 30 were randomized 1:1 to tirzepatide up to 15 mg weekly or placebo for at least 52 weeks, with a median follow-up of 104 weeks.[4] Adjudicated cardiovascular death or a worsening heart-failure event occurred in 36 patients (9.9 percent) on tirzepatide against 56 (15.3 percent) on placebo, hazard ratio 0.62 (95% CI 0.41 to 0.95, P = 0.026).[4] Worth reading closely: the composite was driven by worsening heart-failure events, 29 against 52, while adjudicated cardiovascular deaths were 8 against 5 — a small number of events in the direction that gets least attention.[4] There is no FDA-approved heart-failure indication for tirzepatide on the labels pepmg checked.[11][12]

The qualification

The largest trial met noninferiority and was tested for superiority, and missed

SURPASS-CVOT is the biggest thing in the file and the one most often summarized wrongly. It was an active-comparator-controlled, double-blind noninferiority trial in patients with type 2 diabetes and atherosclerotic cardiovascular disease, randomized 1:1 to weekly subcutaneous tirzepatide up to 15 mg or dulaglutide 1.5 mg — dulaglutide being chosen precisely because it had already been shown to reduce cardiovascular events.[3] 13,299 patients underwent randomization; 134 were later excluded, leaving 6,586 and 6,579 in the modified intention-to-treat groups.[3]

Primary composite end point as reported in the trial publication. The noninferiority margin was 1.05 for the upper confidence limit; superiority required an upper limit below 1.00.[3]

So the honest sentence is: in more than thirteen thousand patients over more than four years, tirzepatide was not worse than an established cardioprotective comparator on major adverse cardiovascular events, and the prespecified test of whether it was better did not succeed.[3] The confidence interval crosses 1.00 by a hair, which is what makes the difference easy to elide and important not to. The trial also reported that adverse events appeared similar between groups, with more gastrointestinal events on tirzepatide.[3]

The same discipline applies to the liver data, which travel further than their design supports. The MASH trial was phase 2, dose-finding, in 190 randomized participants with biopsy-confirmed steatohepatitis and stage F2 or F3 fibrosis; 157 had evaluable week-52 biopsies, with missing values imputed under the assumption that they would follow the placebo pattern.[6] Resolution of MASH without worsening of fibrosis was 10 percent on placebo against 44, 56 and 62 percent at 5, 10 and 15 mg.[6] Large effect, small phase 2, biopsy endpoint, one third of biopsies imputed. PubMed also records a published erratum attached to that paper in March 2026; pepmg did not obtain the text of the correction and so does not characterize what it changed.[7]

The label

What is approved, at what dose, and which warning is rodent data

Drugs@FDA records two Lilly applications. Mounjaro, NDA 215866, was approved on May 13, 2022; Zepbound, NDA 217806, on November 8, 2023.[13][14] The current labels are recent: the Zepbound prescribing information in FDA's label database carries an effective date of August 28, 2026, the Mounjaro label August 27, 2026.[11][12]

Zepbound is indicated, in combination with a reduced-calorie diet and increased physical activity, "to reduce excess body weight and maintain weight reduction long term in adults with obesity or adults with overweight in the presence of at least one weight-related comorbid condition" and "to treat moderate to severe obstructive sleep apnea (OSA) in adults with obesity."[11] Mounjaro is indicated as an adjunct to diet and exercise for glycemic control in adults and pediatric patients 10 years and older with type 2 diabetes, and "to reduce the risk of major adverse cardiovascular (CV) events (CV death, non-fatal myocardial infarction, or non-fatal stroke) in adults with type 2 diabetes mellitus who are at high risk for these events."[12]

Set that alongside the shelf without doing any arithmetic on a reader's behalf: the largest single tirzepatide item in the index generated for this note records a size of 1,500 mg, and several exceed 800 mg.[23] pepmg reports the label's maximum weekly dose and the listing's recorded pack size as two published numbers and does not translate one into the other, because turning a vial size into a schedule is exactly the step this site does not take.

The following paragraph is animal data. Both labels carry a boxed warning that reads, in part: "In both male and female rats, tirzepatide causes dose-dependent and treatment-duration-dependent thyroid C-cell tumors at clinically relevant exposures."[12] The same warning then states that "it is unknown whether MOUNJARO causes thyroid C-cell tumors, including medullary thyroid carcinoma (MTC), in humans as human relevance of tirzepatide-induced rodent thyroid C-cell tumors has not been determined."[12] The drug is contraindicated in patients with a personal or family history of MTC or with Multiple Endocrine Neoplasia syndrome type 2.[12] It is a rodent finding that FDA considered serious enough to box and to contraindicate against, while stating in the same paragraph that its human relevance is undetermined. Both halves are the label's own words.

The change

The shortage-based compounding discretion ended in 2025

FDA ended the temporary enforcement discretion tied to the tirzepatide shortage. That is not a blanket prohibition on every patient-specific compounded prescription; the statutory conditions still matter.

FDA determined the shortage resolved on December 19, 2024.[15] It subsequently extended the 503A transition beyond the original February 18 deadline until a court decision. After the court denied the preliminary injunction on March 5, 2025, FDA stated that 503A discretion had ended; the 503B window ended March 19, 2025.[16] FDA's April 2026 reminder still describes conditional 503A exemptions, including a prescriber-documented change producing a significant difference for an identified patient. These conditions do not make compounded products FDA-approved.[16]

Dates from FDA's shortage order and March 10, 2025 timeline update, read alongside its April 1, 2026 reminder of compounding conditions.[15][16]

This is the part of the file that changed while the trials kept reading the same way. Nothing in SURMOUNT-1 or SURPASS-CVOT is different because of a compounding deadline. What changed is the set of lawful ways a person could obtain a tirzepatide vial that was not Lilly's — and the research-chemical shelf is not one of the ways that reopened.[16][17]

The shelf

FDA's position on "research use only," in its own words, to sellers in this index

FDA's page on unapproved GLP-1 drugs, dated September 1, 2026, states that the agency "has warned companies that have illegally sold unapproved drugs containing semaglutide, tirzepatide, retatrutide, survodutide or mazdutide that are falsely labeled 'for research purposes' or 'not for human consumption.'"[17] All five of those compounds are in pepmg's index.[23]

The individual letters make the finding in near-identical language, and the reasoning is consistent across them: the disclaimer does not decide the question, the surrounding marketing does. One letter dated August 24, 2026 states that "despite statements on your product labeling marketing your products 'for research use only' and 'not for human or veterinary use,' evidence obtained from your website establishes that your products are intended to be drugs for human use," and names semaglutide, tirzepatide, retatrutide, elamipretide, tesamorelin and bremelanotide among the unapproved new drugs at issue.[18] A second letter of the same date uses the same construction against products "for laboratory, research, and analytical use," and goes further on the solvent: because the seller offered bacteriostatic water alongside the peptides, FDA concluded that "the sale of these products together demonstrates that you intend for your 'Bacteriostatic water for Peptides (BAC Water)' to be used in combination for injection. Therefore, your 'Bacteriostatic water for Peptides (BAC Water)' is a drug."[19]

pepmg matched four of the 108 vendor domains in its index, by exact hostname, to websites named in published FDA warning letters about unapproved GLP-1 products. Two of those letters are dated December 10, 2024 and name semaglutide and retatrutide but not tirzepatide; two are dated August 24, 2026 and do name a tirzepatide product.[18][19][20][21] That is a count of domain matches against letters FDA has published, checked on September 8, 2026; it is not a claim that the remaining 104 have not received one, and it is not a statement about what any vial contains. No vendor is named on this page.

Index generated September 7, 2026. Price statistics exclude blends and outliers under pepmg's standard rules.[23]

One more figure, and it cuts against pepmg's own numbers. The 36 is a floor, not a total. pepmg only records a compound identity a vendor has itself published, so a product sold under a coded name that the seller never decodes is left coded and drops out of the index entirely. At least one of the four matched sellers appears in the index with survodutide, mazdutide and a tesamorelin blend but with no tirzepatide row at all — while FDA's August 24, 2026 letter to that same website names a "GLP-2 Tirz Peptide" among the unapproved new drugs.[19][23] FDA has done the same elsewhere: a March 31, 2026 letter to a seller that is not in pepmg's index identifies that site's "GLP-2 peptide" as tirzepatide and its "GLP-1-R peptide" as retatrutide.[22] The conservative rule that keeps pepmg from guessing also keeps some products out of its counts — and it means a regulator has, in these cases, published an identification that this index deliberately withholds. Readers should treat 36 of 108 as an undercount of unknown size.

Four things this note is not saying

First, it is not saying tirzepatide does not work. It plainly does what the trials measured: −20.9 percent mean weight change at 72 weeks in a 2,539-person randomized phase 3, a 20-events-per-hour reduction in apnea-hypopnea index at 52 weeks, a hazard ratio of 0.62 for a heart-failure composite in 731 patients.[1][4][5] These are substantial results for the populations and endpoints studied.

Second, it is not saying the cardiovascular trial failed. SURPASS-CVOT met its prespecified primary objective, which was noninferiority to an active comparator that itself reduces events.[3] It is saying that noninferior and superior are different findings, that the paper reports the superiority test and its P value of 0.09, and that a summary which skips that step is reporting something the trial did not conclude.[3]

Third, it is not saying anything about what is in any vendor's vial. pepmg indexes published prices and reports published evidence; it does not test products, and an FDA warning letter is a finding about how a website marketed a product, not an assay of its contents.[18][23]

Fourth, it is not offering a route to obtaining anything. The approved product is a prescription drug with a boxed warning, a contraindication and a dose-escalation schedule, and every trial above ran under medical supervision.[11][12]

What it is saying is narrower, and it is the whole file in one sentence: tirzepatide has a large clinical development program, whose registry totals mix actual and planned enrollment. The pivotal trials reviewed here do not validate a research vendor's vial. Shortage-based compounding discretion ended in March 2025, with other statutory conditions still applying, and FDA has since told sellers — four of whose domains sit in this index — that a research-use-only sticker does not decide what a product is.[10][16][17][18]

Questions people are asking