Tirzepatide Week by Week: What the SURMOUNT Trials Recorded

Tirzepatide climbs a longer ladder than most people are told. The label starts at 2.5 mg and raises the dose no faster than every four weeks, so reaching the top of the range takes months rather than weeks — and in SURMOUNT-4, the trial built to test what continuing is worth, participants spent 36 weeks in open-label treatment before the real experiment started, losing about 20.9% along the way.^[3] Here is what each SURMOUNT trial recorded, and what happened to the people who stopped.

The ladder comes first

Zepbound and Mounjaro both begin at 2.5 mg once weekly, and that opening dose is there to be tolerated rather than to work. Increases follow no sooner than every four weeks. Zepbound is made in four strengths and Mounjaro in six, which means a Zepbound titration steps straight from 5 mg to 10 mg where Mounjaro can stop at 7.5 mg on the way.

The practical consequence is that anyone measuring their progress in the first month or two is measuring the ladder, not the drug at its working dose. Our titration planner lays the schedule out against real dates from each label, and the Zepbound pen chart covers which strength is which.

What 72 weeks produced

SURMOUNT-1 is the trial the rest are measured against. It enrolled 2,539 adults living with obesity and without diabetes, across three doses — 5, 10 and 15 mg — against placebo, over 72 weeks. The top dose finished at −20.9% against −3.1% on placebo, and roughly 57% of that arm lost a fifth or more of their starting weight.^[1]

That last figure is the one worth sitting with. A majority of people on the top dose lost at least 20% — which also means a substantial minority did not, and some lost considerably less than the average. The mean is a description of a group, never a prediction for a person.

Diabetes changes the number

SURMOUNT-2 ran the same drug in people who had both excess weight and type 2 diabetes, and this is the population where GLP-1 drugs reliably underperform. Tirzepatide managed −15.7% at 15 mg and −13.4% at 10 mg against −3.3% on placebo, with A1C falling about 2.1 points.^[2]

So roughly two-thirds to three-quarters of the effect seen without diabetes survives. If you have type 2 diabetes and have been comparing yourself against the 20.9% headline, you have been comparing yourself against the wrong trial.

What stopping did

SURMOUNT-4 is the clearest evidence on this question that exists for any GLP-1. Everyone in it first took tirzepatide openly for 36 weeks and lost about 20.9%. Then they were randomized: half continued, half moved to placebo, and both were followed for another year.

The continuers lost a further 5.5%. Those switched to placebo regained 14.0% — roughly two-thirds of a very large weight loss, back within a year, in people who had already done everything the protocol asked of them.^[3]

Two-thirds of a 20.9% loss came back inside a year, in people who had done everything right.

Why this is not a race against semaglutide

It is tempting to set −20.9% beside semaglutide’s −14.9% and call it settled. That comparison is not available from these trials. SURMOUNT-1 and STEP 1 enrolled different people, ran for different lengths — 72 weeks against 68 — and each had its own placebo arm. Reading across two separate trials as though they were arms of one is precisely the move this register criticizes when a seller does it.

A genuine head-to-head does exist, and it is a separate piece of work. In the meantime, our semaglutide timeline sets out that drug’s own record on its own terms.

What this timeline cannot tell you

• Every figure is a group mean. Individuals scattered widely around each one, in both directions.
• Trial conditions are not ordinary conditions. Participants had free medication, structured support and regular contact with a study team.
• It describes the brand product at labeled doses. A compounded vial raises separate questions, taken up in is compounded tirzepatide safe.
• It says nothing about who should take it. That is a clinical judgment about your history, not an arithmetic one about a curve.

One thing the published record does not settle: how much of the 36-week run-in loss in SURMOUNT-4 came from the dose ladder versus the maintenance dose that followed it. The trial reports the total, not the split, so the shape of that first stretch is less well described for tirzepatide than for semaglutide.