New analysis narrows the perceived edge of tirzepatide over semaglutide in weight loss for type 2 diabetes

A recent study challenges previous assumptions by showing that, when comparing doses directly, semaglutide’s effectiveness in weight reduction approaches that of tirzepatide, prompting a nuanced re-evaluation of their relative benefits.

Average results have recently made tirzepatide look like the more powerful of the two leading incretin injections, but a new analysis suggests the gap narrows sharply when doctors compare dose with dose rather than drug with drug. In Diabetes Therapy (link.springer.com), researchers reported that semaglutide 2.4mg a week had a 99.4% probability of producing weight loss equivalent to tirzepatide 10mg in adults with type 2 diabetes, while semaglutide 7.2mg and tirzepatide 15mg reached a 94.8% probability. Lower-dose semaglutide did not keep pace with higher-dose tirzepatide. The study was published online on 7 June 2026 and appeared in the journal’s July issue. (link.springer.com)

The paper drew together phase III trial data from the semaglutide SUSTAIN and STEP programmes and the tirzepatide SURPASS and SURMOUNT-2 programmes. Springer’s version of the paper says 23 trials contributed 48 treatment arms and 16,524 participants in total, split between 23 semaglutide arms involving 9,902 people and 25 tirzepatide arms involving 6,622. (link.springer.com) Treatment lasted from 26 to 104 weeks, mean baseline age was about 55, and baseline HbA1c and body mass index were broadly similar across the two drug groups. The pooled mean baseline body weight was 94.8kg in semaglutide arms and 90.6kg in tirzepatide arms, although the ranges overlapped. (link.springer.com)

What makes the study notable is not just the answer but the way it was reached. The authors used generalised additive models and a Bayesian hierarchical model to trace how weight loss changed across dose ranges, adding zero-dose “pseudo-observations” to stop the curves behaving implausibly near the bottom end. (link.springer.com) They defined equivalence as no more than a two-percentage-point difference in predicted weight change, a pragmatic threshold rather than a biological law. Because the work was built from published trial arms rather than individual patient records, the authors said fresh ethics approval and consent were not required for this secondary analysis. (link.springer.com)

The clinical message is more nuanced than a simple winner-loser verdict. According to the paper, both drugs show a curved dose-response pattern: larger gains come early, then taper at higher doses. Tirzepatide appeared to deliver bigger incremental effects at lower doses, whereas semaglutide retained benefit over a broader escalation range. (link.springer.com) In the authors’ simulated intensification scenarios, moving a patient from semaglutide 1mg to 2.4mg was more likely to produce a clinically meaningful extra benefit than switching that patient to tirzepatide 5mg, if semaglutide was still tolerated. At the top end, however, the model suggested that switching from tirzepatide 10mg to semaglutide 7.2mg could yield more extra weight loss than stepping up tirzepatide to 15mg. (link.springer.com)

That helps explain why previous comparisons have been hard to interpret. Endotext, the long-running NCBI clinical reference, notes that SURPASS-2 compared tirzepatide against semaglutide 1mg, not the 2.4mg dose used for weight management, so it could not settle whether higher-dose semaglutide might close the gap. (ncbi.nlm.nih.gov) The same reference points to SURMOUNT-2, in which adults with obesity and type 2 diabetes lost 9.6% of body weight on tirzepatide 10mg and 11.6% on 15mg over 72 weeks versus placebo. (ncbi.nlm.nih.gov) The new paper is, in effect, an attempt to connect those scattered findings into a single dose-based framework.

However, other evidence still points to tirzepatide doing better on average in routine practice. A June study in PNAS Nexus followed matched real-world cohorts of 10,339 patients on each drug and found mean maximum weight loss of 14.7% with tirzepatide against 10.8% with semaglutide over two years. (academic.oup.com) The proportion achieving at least 15% loss was 43% with tirzepatide and 22% with semaglutide, and high responders on tirzepatide lost weight faster in the first six months. (academic.oup.com) That study also found women and White patients were overrepresented among high responders, while Black and Hispanic patients were more common in the minimal-response groups, and it reported lower documented rates of nausea, vomiting, constipation, headache and fatigue with tirzepatide in some response categories. (academic.oup.com)

A separate real-world signal, this time publicised by Novo Nordisk and highlighted by Reuters in June, pulled in a different direction. In a retrospective US claims analysis of more than 64,000 adults already taking semaglutide 1mg, the company said those escalated to 2mg were as likely to reach HbA1c below 7% within a year as those switched to tirzepatide, and were more likely to achieve weight loss of at least 5%. (novonordisk.com) Michael Radin, an executive medical director at Novo Nordisk Inc, said the findings addressed “an important clinical consideration” for people already established on therapy. The comparison was not a randomised trial, and the tirzepatide group started on 2.5mg or 5mg with later titration allowed, but it echoes the new paper’s broader argument that dose optimisation within a drug can matter as much as switching molecules. (novonordisk.com)

The new modelling study also comes with caveats that matter. Its estimates are based on aggregated trial-arm data, not patient-level records, so the authors could not fully adjust for differences in background treatment, coexisting illness or adherence. The semaglutide 7.2mg estimate rests on a single trial, adverse events were extracted but not formally modelled, and the whole exercise was confined to weight loss rather than blood sugar control, cardiovascular outcomes or long-term safety. (link.springer.com) There are also conflicts to weigh: Springer’s disclosures state that two authors are Novo Nordisk employees and that the lead author has received consulting and speaking fees from several drugmakers. (link.springer.com) Even so, the paper offers something clinicians have largely lacked in this debate: a quantified way to think about what counts as a like-for-like switch when formularies, cost or supply push patients from one injectable to another.