SURMOUNT-1 produced the largest weight figures in obesity medicine, and those figures are quoted far more often than the trial that generated them is described. The dose-by-dose results are laid out in the tirzepatide weight article. What is below is the construction: four parallel arms rather than a ladder, an estimand that keeps the quitters, three substudies that complicate the headline, and a seventeen-week window at the end that almost nobody quotes.
Four arms, not four rungs
The trial assigned 2,539 adults with a body-mass index of 30 or more, or 27 or more with at least one weight-related complication, in a 1:1:1:1 ratio to once-weekly subcutaneous tirzepatide at 5 mg, 10 mg or 15 mg, or to placebo, for 72 weeks including a 20-week dose-escalation period. Diabetes was an exclusion.[1]
Four parallel arms is not the same shape as a titration schedule. A participant assigned to 5 mg stayed at 5 mg for the full 72 weeks; nobody climbed from one published figure to the next. The difference matters when a seller quotes the top number beside a plan that caps lower, and the escalation itself is covered in the titration article.
The coprimary endpoints were the percentage change in weight from baseline and a weight reduction of 5% or more. The treatment-regimen estimand assessed effects regardless of treatment discontinuation, in the intention-to-treat population.[1] As in the semaglutide program, that means the published percentages already carry the people who stopped — the same estimand logic described in the STEP 1 article.
Who was measured
At baseline the mean body weight was 104.8 kg, the mean body-mass index was 38.0, and 94.5% of participants had a body-mass index of 30 or higher.[1] Adverse events caused treatment discontinuation in 4.3%, 7.1% and 6.2% of the 5 mg, 10 mg and 15 mg groups against 2.6% on placebo, and the most common events were gastrointestinal, occurring primarily during dose escalation.[1]
A mean body-mass index of 38 sits in class II obesity. Someone entering a telehealth intake at a body-mass index of 31 is outside the center of this distribution, and the trial offers no separate estimate for them. The comparison a buyer usually wants — against the same drug in people with type 2 diabetes, whose mean baseline weight was 100.7 kg and mean glycated hemoglobin 8.02% — is a different trial with a different answer.[6]
The 1:1:1:1 allocation also sets how much any one figure can be trusted. Splitting 2,539 people four ways leaves roughly 635 in each arm, which is why the confidence intervals around the headline percentages are narrow enough to separate the doses from one another rather than merely from placebo. That number is worth remembering, because the substudies below are built on far smaller ones.
A subgroup where the doses changed places
A prespecified subpopulation analysis reported the 102 Japanese adults in the trial. Least-squares mean percent change in body weight at week 72 was −12.0% (SE 1.7) at 5 mg, −22.4% (1.7) at 10 mg and −22.1% (1.6) at 15 mg, against −0.3% (1.6) on placebo. A reduction of 5% or more was reached by 91.7%, 100% and 96.6% of those arms against 15.4% on placebo.[2]
The 10 mg arm finished ahead of the 15 mg arm. That is not evidence that the middle dose outperforms the top one; it is what roughly twenty-five people per arm look like, with standard errors of about 1.7 percentage points around each estimate. A reader who would reject a subgroup result pointing the other way should reject this one too, and the discipline of reading a stratified result is the same discipline the rest of this page applies.
Where the weight came from, in both arms
A substudy scanned 160 of the 2,539 participants by dual-energy X-ray absorptiometry at baseline and week 72 — 124 on pooled tirzepatide doses and 36 on placebo, 73% female, mean weight 102.5 kg and mean body-mass index 38.0. Changes were −21.3% in body weight, −33.9% in fat mass and −10.9% in lean mass with tirzepatide, against −5.3%, −8.2% and −2.6% with placebo (p < 0.001 for all).[3]
The comparison is the finding. Of the weight lost, approximately 75% was fat mass and 25% lean mass in both groups, and those proportions held across most subgroups by sex, age and size of weight loss.[3] Losing weight without the drug cost a similar fraction of lean tissue; what the drug changed was how much weight came off, not the composition of it. The pooled body-composition literature across the class is in the muscle article.
When the curve flattened
A post hoc analysis defined a weight plateau as a change of less than 5% over a twelve-week interval and every interval after it, in participants adherent to treatment who reached at least 5% loss by the primary endpoint. Among the 1,438 SURMOUNT-1 participants who qualified, the median time to plateau across body-mass index categories — overweight, class I, class II and class III — was 24.3, 26.0, 36.1 and 36.1 weeks. By week 72, 90.2%, 88.9%, 87.6% and 87.8% of those categories had plateaued.[4]
Higher doses, younger age and female sex were each associated with reaching the plateau later.[4] For a buyer, the practical reading is that the steep part of the curve is measured in months rather than years, and that a plan priced on the expectation of continuous visible progress will disappoint most people well before the end of the first year. What a plateau is and is not is covered in the plateau article.
Blood pressure, and what carried it
Stratified analyses found a rapid decline in systolic and diastolic blood pressure over the first 24 weeks, then stabilization, ending in a net reduction at 72 weeks of 6.8 mm Hg systolic and 4.2 mm Hg diastolic against placebo. Participants assigned to any tirzepatide group were more likely to have normal blood pressure at week 72, at 58.0% against 35.2%. A mediation analysis attributed 68% of the systolic and 71% of the diastolic reduction to weight loss.[5]
Two qualifications belong next to that. Roughly a third of the blood pressure effect was not explained by the weight change, so it is not simply a downstream consequence of a smaller body. And low blood pressure adverse events, while infrequent, were more common on tirzepatide than on placebo.[5] A benefit with a direction of harm attached is still a benefit; it is also a reason the decision belongs to a prescriber holding a medication list.
Three years on, and seventeen weeks off
Treatment continued to week 176 for the 1,032 participants who had prediabetes as well as obesity, followed by a 17-week off-treatment period. Over the 176 weeks, type 2 diabetes was diagnosed in 1.3% of the tirzepatide groups against 13.3% on placebo, a hazard ratio of 0.07 (95% CI, 0.0 to 0.1).[7]
Then the drug stopped. After the 17 weeks off treatment, 2.4% of those who had received tirzepatide and 13.7% of those on placebo had type 2 diabetes, a hazard ratio of 0.12 (95% CI, 0.1 to 0.2).[7] The placebo figure barely moved. The tirzepatide figure nearly doubled in four months.
The prevention result survives that, and the gap between the groups remains large. But the direction of travel in those seventeen weeks says the effect is a treatment effect rather than a durable change of state, which is the same conclusion the weight data reach in the stopping article.
What the trial does not describe
Every number above comes from the branded, FDA-approved product at labeled doses, escalated on a protocol, supplied without interruption, with lifestyle support in all four arms. Sellers on the tirzepatide board overwhelmingly dispense compounded preparations, which are not FDA-approved and are not reviewed by the FDA for safety, efficacy or quality before they are dispensed. The difference between the two is set out in the compounded-versus-brand article.
The trial also ran on a schedule nobody had to pay for. Escalation followed a protocol rather than a budget, doses did not lapse between shipments, and the assigned milligram did not change when a participant found it expensive. Each of those is a variable in a cash-pay plan and none of them is a variable in SURMOUNT-1, which is why a trial figure describes the best version of a treatment course rather than the average one.