Tirzepatide, on current evidence. In SURMOUNT-5, a randomized trial at weight-management doses, weight fell 20.2% on tirzepatide and 13.7% on semaglutide over 72 weeks [5]. Participants knew which drug they were on, and the sponsor was Eli Lilly, tirzepatide's manufacturer. Routine-care studies point the same way, by three to four percentage points.
The randomized comparisons, and their limits
SURMOUNT-5 randomized 751 adults with obesity but without type 2 diabetes [5]. Each took the maximum tolerated dose of one drug: tirzepatide 10 or 15 mg, or semaglutide 1.7 or 2.4 mg, weekly for 72 weeks. Weight fell 20.2% (95% CI −21.4 to −19.1) against 13.7% (−14.9 to −12.6). Waists shrank by 18.4 cm on tirzepatide and 13.0 cm on semaglutide.
Two limits sit beside that result. Participants knew which drug they were taking, and the sponsor makes one of them. Gastrointestinal events were the most common adverse events in both arms, mostly mild to moderate and mostly during dose escalation.
An earlier trial compared them in type 2 diabetes. SURPASS-2 randomized people with type 2 diabetes to tirzepatide at 5, 10 or 15 mg, or semaglutide at 1 mg [1]. HbA1c fell 2.01, 2.24 and 2.30 points against 1.86, and weight differed by 1.9 kg, 3.6 kg and 5.5 kg.
The comparator matters. Semaglutide 1 mg is the diabetes dose; weight management uses 2.4 mg. A reader setting the SURPASS-2 figures beside the STEP trials is comparing different doses — the mismatch this site keeps finding, set out in the tirzepatide network comparison.
What routine practice shows
| Study | Tirzepatide | Semaglutide | Design |
|---|---|---|---|
| Chart review, 511 patients, 12 months | 16.6% | 13.4% | Retrospective, multicenter |
| Matched cohort, 2 years | 14.7% | 10.8% | Propensity-matched records |
| Reached ≥15% (chart review) | 56.8% | 40.7% | — |
| Reached ≥15% in year 1 (cohort) | 42.6% | 21.6% | — |
The matched cohort also measured pace: 2.54% of body weight a month against 2.18% [4]. Among patients without diabetes in the chart review, the gap widened to 18.5% against 14.2% [2].
A third and much larger matched cohort points the same way. Among 18,386 propensity-matched adults in routine US care, tirzepatide was significantly more likely to reach every weight-loss threshold [3]. It also recorded something the smaller studies do not: follow-up ended in discontinuation for 55.9% of the tirzepatide group and 52.5% of the semaglutide group. More than half of each arm stopped.
The finding most likely to change a decision
Side effects: the direction is unusual, the evidence is weak
The chart review recorded gastrointestinal side effects in 28.5% of tirzepatide patients against 50.7% on semaglutide, which runs against the usual pattern where the stronger drug is harder to tolerate.
Read how that was collected. In a randomized trial, adverse events are solicited from everyone on a schedule; in a chart review they appear when a patient raised them and a clinician wrote them down. The cohort study reached a similar conclusion using AI curation of clinical notes, a method it reports no validation for here.
So the tolerability claim is directionally interesting and weakly evidenced. It sits against the network meta-analysis pattern in the drugs that lose most and are tolerated worst.
Who responded, and who did not
The two-year matched cohort reports a demographic pattern it does not explain. Women were over-represented among high responders for both drugs, as were White patients. Black and Hispanic patients were over-represented among the group losing less than 5%.
The paper offers no mechanism and this site will not invent one. It is a description of who did well in the records examined. It is also the kind of finding that ought to shape who future trials enroll — a representativeness problem this site set out in who actually gets into trials.
What none of this prices. Which of the two is cheaper depends entirely on the seller, and the stopping question applies to both — see what happens when you stop tirzepatide.