Tirzepatide side effects by dose: the label table, read properly

Every tirzepatide side-effect list on the internet descends from one table: Table 1 in section 6.1 of the Zepbound label. Here it is in full, with what the label says around it, because the sentences around the table matter as much as the numbers in it.

Where the numbers come from

The table pools two placebo-controlled trials, Study 1 (SURMOUNT-1, adults with obesity or overweight without diabetes, PubMed 35658024) and Study 2 (SURMOUNT-2, adults with type 2 diabetes, PubMed 37385275). Together they cover 2,519 people treated with Zepbound for up to 72 weeks plus a 4-week follow-up. Mean age 47, 37% male, mean BMI 37.4. Arm sizes: placebo 958, 5 mg 630, 10 mg 948, 15 mg 941. The table lists reactions reported by at least 2% of Zepbound patients and more often than on placebo.

The table

Each figure is the percentage of people in that arm who reported the reaction at least once during the trial. It is not a per-week rate and it is not a severity score.

How to read a row

Take nausea at 15 mg: 28% versus 8% on placebo. The placebo figure is the background rate of people reporting nausea over 72 weeks of a diet-and-activity program with regular injections of nothing. The drug's contribution is the excess, 20 percentage points, which means about one person in five reports nausea because of tirzepatide. Turned round, 72% of people at 15 mg did not report nausea at all.

Constipation runs the other way: 17% at 5 mg, 14% at 10 mg, 11% at 15 mg. The label offers no explanation; it is simply what the pool showed.

The explorer above does this arithmetic for every row and lets you sort by excess over placebo, so the rows that matter most rise to the top.

The sentences around the table

Any gastrointestinal adverse reaction: 56% at 5 mg, 56% at 10 mg, 56% at 15 mg, versus 30% on placebo. Discontinuation because of a GI reaction: 1.9%, 3.3% and 4.3%, versus 0.5%. Then the line that answers most timing questions: "The majority of nausea, vomiting, and/or diarrhea events occurred during dose escalation and decreased over time."

Permanent discontinuation for any adverse reaction: 4.8%, 6.3% and 6.7% at 5, 10 and 15 mg, versus 3.4% on placebo, with most of those stopping "during the first few months of treatment due to gastrointestinal adverse reactions". The NEJM paper for SURMOUNT-1 alone reports 4.3%, 7.1%, 6.2% and 2.6% (PubMed 35658024). In Study 3, the intensive-lifestyle lead-in trial, 10% of Zepbound patients discontinued for adverse reactions versus 2% on placebo; in Study 4, 7% stopped before the week-36 randomisation.

Some specific items the label singles out: hypotension in 1.6% of Zepbound patients versus 0.1% on placebo, more often in people on blood-pressure medication (2.2% versus 1.2%) and in association with GI events and dehydration. Immediate hypersensitivity reactions (within a day of injecting) in 2.1% versus 0.4%, mostly skin reactions. Hair loss in 7.1% of women versus 0.5% of men on Zepbound, against 1.3% and 0% on placebo, associated with the weight reduction itself; nobody on Zepbound stopped because of it. Acute pancreatitis confirmed by adjudication in 0.2% of both Zepbound and placebo patients in this pool, 0.14 versus 0.15 per 100 patient-years.

What the table cannot tell you

It cannot tell you which week a reaction will start, how long it will last, or whether you will be in the 28% or the 72%. The arms were fixed doses: someone whose prescriber holds them at 7.5 mg has no row of their own. And the population was people with obesity in a trial, weighed and counselled every few weeks. The warnings page covers the rarer, serious items that do not appear in a frequency table. For how the same reactions look in the diabetes label, see Mounjaro vs Zepbound.