Tirzepatide: The Trials Behind the Approvals

Tirzepatide was the first approved medicine to act on two gut hormone receptors at once, GIP and GLP-1. This article follows it from the laboratory to the regulators: how the molecule is built, what each major trial measured and found, what was approved and where, and what its label warns about.

Two hormones the gut already makes

After a meal, cells in the lining of the small intestine release two hormones: glucose-dependent insulinotropic polypeptide, known as GIP, and glucagon-like peptide-1, known as GLP-1. Both tell the pancreas to release more insulin while blood glucose is high. This is called the incretin effect, and research reviewed by the Danish physiologist Jens Juul Holst found the two hormones to be roughly equally important to it.

Neither lasts long in the body. An enzyme called dipeptidyl peptidase-4, or DPP-4, cuts both of them, and about 90 percent of GLP-1 is broken down before it even reaches the general circulation. Before that happens, GLP-1 acts on sensory nerves in the gut wall that signal to the brainstem and the hypothalamus, which is thought to explain part of its effect on appetite and on how the gut moves. GIP appears to influence how fat is stored.

Tirzepatide is a single molecule that activates the receptors for both hormones. Its US prescribing information describes three main actions: it increases insulin secretion, it lowers glucagon, and it does both only when glucose is raised. It also reduces food intake.

How the molecule is built

Tirzepatide is a chain of 39 amino acids based on the sequence of human GIP. Three changes to that starting point explain most of how it behaves in the body.

• Two non-standard amino acids. Positions 2 and 13 carry aminoisobutyric acid, usually shortened to Aib, an amino acid the body does not use to build its own proteins. The same substitution appears in semaglutide, whose designers describe the aim as full stability against metabolic breakdown.
• A fatty acid chain. The lysine at position 20 is attached, through a linker, to a fatty diacid with 20 carbon atoms. The label explains that this lets tirzepatide bind to albumin, the most abundant protein in blood, and that binding prolongs its half-life. About 99 percent of the tirzepatide in plasma is bound to albumin.
• A capped end. The last amino acid carries a C-terminal amide rather than a free acid group.

A half-life of about five days is what allows a medicine given once a week. The body eventually clears tirzepatide the way it clears other proteins and fats: enzymes cut the peptide backbone, the fatty diacid is broken down by beta-oxidation, and the amide is hydrolysed. Intact tirzepatide is not found in urine or faeces, only its breakdown products.

From first human studies to phase 3

The compound first appeared in the scientific literature under the code LY3298176. In 2018 Eli Lilly scientists published its laboratory work and first human studies together. In cell assays it activated both receptors. In mice it reduced body weight and food intake more than a selective GLP-1 receptor agonist did.

In the phase 1 studies, 142 people received at least one dose of the compound, the comparator dulaglutide, or placebo. The first studies were in healthy volunteers, followed by a four-week study in people with type 2 diabetes. The blood levels supported a once-weekly schedule. The most frequent side effects were gastrointestinal, namely vomiting, nausea, decreased appetite, diarrhoea and abdominal distension, and all of them increased with dose and were described as mild to moderate.

Two large phase 3 programmes followed. SURPASS studied people with type 2 diabetes and SURMOUNT studied people with obesity or overweight. The sections below walk through the main trials from both programmes.

SURPASS-2: head to head with semaglutide

Many diabetes trials compare a new medicine with placebo. SURPASS-2 compared tirzepatide with semaglutide, another weekly medicine in the same family that acts on the GLP-1 receptor alone. It enrolled 1,879 adults with type 2 diabetes and followed them for 40 weeks. Participants were randomly assigned in equal numbers to one of three tirzepatide dose levels or to semaglutide at the dose then used for diabetes, which is lower than the dose later approved for weight management.

The trial was open label, which means participants and investigators knew which medicine each person received. At the start, the average HbA1c, a blood test that reflects glucose levels over the previous two to three months, was 8.28 percent. Average age was 56.6 years and average weight 93.7 kg.

All three tirzepatide groups lowered HbA1c more than the semaglutide group did. Body weight also fell further: the difference against semaglutide was 1.9 kg, 3.6 kg and 5.5 kg across the three dose levels.

The side effects were similar in kind across both medicines. Nausea was reported by 17 to 22 percent of people on tirzepatide and 18 percent on semaglutide. Diarrhoea was reported by 13 to 16 percent against 12 percent, and vomiting by 6 to 10 percent against 8 percent. Serious adverse events were reported in 5 to 7 percent of the tirzepatide groups and 3 percent of the semaglutide group. Blood glucose below 54 mg/dL was uncommon: 0.6, 0.2 and 1.7 percent on tirzepatide and 0.4 percent on semaglutide.

SURMOUNT-1: the obesity trial

SURMOUNT-1 is the tirzepatide trial quoted most often. It enrolled 2,539 adults with a body mass index (BMI) of 30 or more, or 27 or more with at least one weight-related complication, and it excluded people with diabetes. Participants were randomised in equal numbers to one of three tirzepatide dose levels or to placebo for 72 weeks, including a 20-week period in which the dose was raised step by step.

At the start, the average participant weighed 104.8 kg and had a BMI of 38.0. Almost all of them, 94.5 percent, had a BMI of 30 or higher.

An average hides the spread around it, so the trial also reported how many people crossed fixed thresholds. A reduction of at least 5 percent was reached by 85, 89 and 91 percent of the three tirzepatide groups and by 35 percent of the placebo group. A reduction of 20 percent or more was reached by half of the middle dose group and 57 percent of the highest dose group, compared with 3 percent on placebo. Put the other way round, even at the highest dose level more than four in ten participants did not reach 20 percent.

Gastrointestinal events were the most common adverse events. Most were mild to moderate and occurred while the dose was being raised. Adverse events led people to stop treatment in 4.3, 7.1 and 6.2 percent of the three tirzepatide groups, compared with 2.6 percent on placebo.

Reading the headline number

Two reports about the same trial can quote different numbers because they answer slightly different questions, known as estimands. SURMOUNT-1 reported the treatment-regimen estimand, which counts every randomised participant, including those who stopped the medicine. It answers the question: what happens to a group of people who are prescribed it? A second kind of analysis estimates what would have happened if everyone had kept taking it, and it usually gives larger numbers. When two headlines disagree, check which question each one answers.

SURMOUNT-4: what happened when treatment stopped

Any medicine used for a long-term condition has to answer what happens after it is stopped. SURMOUNT-4 tested this directly with a randomised withdrawal design, run at 70 sites in four countries.

From week 36 to week 88, the group that continued tirzepatide lost a further 5.5 percent of body weight on average. The group switched to placebo regained 14.0 percent. The difference between the two groups was 19.4 percentage points, with a 95 percent confidence interval of 17.7 to 21.2. The authors concluded that withdrawing tirzepatide led to substantial regain of the weight that had been lost, while continued treatment maintained it and added to it.

This result shapes how the medicine is understood. In obesity it is studied and approved as a long-term therapy, which is why the approved use is called chronic weight management, rather than as a course with an end date.

SURMOUNT-OSA: sleep apnoea

Obstructive sleep apnoea is the repeated collapse of the upper airway during sleep, and excess body fat is one of its causes. It is measured with the apnoea-hypopnoea index, or AHI, which counts how many times an hour breathing stops or becomes shallow. SURMOUNT-OSA ran two 52-week trials in adults with obesity and moderate to severe sleep apnoea. One enrolled people not using positive airway pressure (PAP) therapy, and the other enrolled people who were. Together the trials enrolled 469 adults without type 2 diabetes.

In the first trial, the average AHI at the start was 51.5 events an hour. It fell by 25.3 events an hour with tirzepatide and by 5.3 with placebo, an estimated difference of 20.0 events an hour with a 95 percent confidence interval of 14.2 to 25.8. The trials also reported reductions in hypoxic burden, high-sensitivity C-reactive protein and systolic blood pressure, and better patient-reported sleep outcomes.

SURPASS-CVOT: heart attacks, strokes and deaths

Glucose and weight are what doctors call surrogate outcomes. What people with diabetes and heart disease most need to know is whether a medicine changes heart attacks, strokes and deaths. SURPASS-CVOT answered that against an active comparator rather than placebo: dulaglutide, a GLP-1 receptor agonist already shown to reduce cardiovascular events.

The trial randomised 13,299 adults with type 2 diabetes and established atherosclerotic cardiovascular disease in 30 countries. After 134 were excluded for not meeting the entry criteria, 6,586 received tirzepatide and 6,579 received dulaglutide, and they were followed for a median of about four years. The average participant was 64.1 years old with a BMI of 32.6 and an HbA1c of 8.4 percent, and 29.0 percent were women.

The trial was built to test non-inferiority first, with a margin of 1.05 for the upper limit of the confidence interval. The upper limit came in at 1.01, so tirzepatide met non-inferiority. Showing superiority would have needed that upper limit to sit below 1.00, and it did not. In plain terms, tirzepatide was at least as good as a medicine already known to protect the heart, and the 8 percent lower rate may be due to chance. Gastrointestinal adverse events were more frequent with tirzepatide, while overall adverse event rates appeared similar.

Where it is approved, and for what

A regulator does not approve a molecule in general. It approves a specific product, made by a specific manufacturer, for the specific uses written on its label. Eli Lilly sells tirzepatide under two brand names: Mounjaro, and in the United States also Zepbound for weight management and sleep apnoea.

The diabetes and weight management approvals all pair the medicine with diet and physical activity. When BPOM announced its decision, it said it had evaluated safety, efficacy and quality against international standards for laboratory, clinical and manufacturing practice, and that monitoring continues after approval through pharmacovigilance. The same statement notes that BPOM has also registered semaglutide and liraglutide for similar uses. BPOM's guidance on the approved medicine is that it is used according to a doctor's prescription and under a doctor's supervision.

Safety: what the label warns about

The US label for Mounjaro opens with a boxed warning, the strongest warning the FDA uses. In rats, tirzepatide caused thyroid C-cell tumours that depended on dose and duration, at exposures relevant to people. Whether it causes these tumours in humans, including medullary thyroid carcinoma, is not known. It must not be used by anyone with a personal or family history of medullary thyroid carcinoma, or with multiple endocrine neoplasia syndrome type 2.

The warnings and precautions section then covers:

• acute pancreatitis
• low blood glucose when it is combined with insulin or with medicines that increase insulin release
• serious allergic reactions
• acute kidney injury, which can follow dehydration from vomiting or diarrhoea
• severe gastrointestinal disease
• complications of diabetic retinopathy in people with a history of it
• acute gallbladder disease
• inhaling stomach contents during general anaesthesia or deep sedation

The most common side effects are gastrointestinal: nausea, diarrhoea, decreased appetite, vomiting, constipation, indigestion and abdominal pain. For Zepbound, the FDA also lists injection site reactions, fatigue, allergic reactions, burping, hair loss and acid reflux, and it includes suicidal behaviour and thinking among the warnings.

The approved medicine and everything else sold under its name

The results in this article belong to Mounjaro and Zepbound as manufactured by Eli Lilly and used under medical supervision. They do not transfer to other products that carry the word tirzepatide, and regulators have spent several years saying so.

In the United States, compounded medicines are prepared by pharmacies. The FDA states that it does not verify their safety, effectiveness or quality before they are marketed. Its page on unapproved GLP-1 medicines, updated on 1 September 2026, reports more than 730 adverse event reports linked to compounded tirzepatide and 990 linked to compounded semaglutide. It describes adverse events, some needing hospital care, that followed errors in measuring doses, and cases where people received more than the approved label allows.

The same page describes warning letters to companies selling semaglutide, tirzepatide, retatrutide and similar compounds labelled "for research purposes" or "not for human consumption" while giving instructions for personal use. It also confirms that counterfeit Ozempic has been found in the US supply chain, and it advises people to use only medicines prescribed by a licensed prescriber and dispensed by a licensed pharmacy.

What these trials do and do not tell you

• They are averages from selected groups. Every trial has entry criteria. SURMOUNT-1 excluded people with diabetes, and SURPASS-CVOT only enrolled people who already had cardiovascular disease.
• The comparator shapes the result. A result against placebo and a result against another active medicine answer different questions.
• The manufacturer funded them. Every trial in this article was sponsored by Eli Lilly, which is standard for trials submitted to regulators.
• Stopping matters. SURMOUNT-4 shows that the effect on weight depends on continuing treatment.
• Rare harms need large numbers and time. Uncommon adverse events only become visible in trials with thousands of participants followed for years, and in monitoring after approval.