Explore this article's sources with AI
Follow PeptideMethods on Google
Semaglutide vs orforglipron: who paid?
A new indirect comparison says oral semaglutide beats orforglipron on weight loss. Five of its eight authors work for Novo Nordisk. Here is what it shows.
Why we wrote this. The headline says one obesity drug beat another. The author list says Novo Nordisk, and the method is an indirect comparison, so we put the funding and the limits up front.
In this article (5 sections)
A paper published in Diabetes, Obesity and Metabolism in June 2026 reports that oral semaglutide 25 mg produced more weight loss than orforglipron 36 mg in adults with overweight or obesity, and fewer dropouts from side effects[1]. No trial has compared the two drugs head to head. This paper did not either. It is an anchored indirect treatment comparison that borrows the placebo arms of two separate trials, OASIS 4 and ATTAIN-1, to bridge between them, and five of its eight authors are employees and shareholders of Novo Nordisk, the company that sells semaglutide[1]. Both of those facts belong in the first paragraph, not a footnote at the bottom.
What the paper reported
After adjusting for baseline differences in sex, body weight and glycaemic status, the analysis put oral semaglutide 25 mg ahead of orforglipron 36 mg by 3.2 percentage points of body-weight change (95% confidence interval minus 5.9 to minus 0.4) on the treatment-regimen estimand, and by 3.0 points (95% CI minus 5.8 to minus 0.3) on the efficacy estimand[1]. An estimand is the precise question an analysis answers, and the two here differ in how they handle people who stopped treatment early. Tolerability pointed the same way. Stopping treatment for any adverse event was more common on orforglipron, at an odds ratio of 4.1 (95% CI 1.3 to 13.0), and stopping specifically for gastrointestinal adverse events carried an odds ratio of 13.9[1]. Look at the interval on that second figure before repeating the number: 2.0 to 96.0. An estimate that could plausibly be a doubling or a fiftyfold difference is not a precise finding.
The two trials were not built to be compared
OASIS 4 randomised 307 adults, 205 to once-daily oral semaglutide 25 mg and 102 to placebo, across 22 sites in four countries, with its coprimary endpoints measured at week 64. Mean body-weight change was minus 13.6% against minus 2.2% on placebo, a difference of 11.4 percentage points (95% CI minus 13.9 to minus 9.0)[2]. ATTAIN-1 was a different animal. It randomised 3127 adults with obesity across three orforglipron doses and placebo and read out at week 72, with the 36 mg group losing 11.2% (95% CI minus 12.0 to minus 10.4) against 2.1% on placebo[3]. Orforglipron is not a peptide at all. It is a small-molecule GLP-1 receptor agonist taken by mouth, which is why a pill without the peptide manufacturing chain has drawn so much attention[3].
So the two treatment arms in this comparison were measured eight weeks apart, in trials separated by roughly a factor of ten in size. Set the two published headline numbers side by side and the raw gap is 2.4 percentage points. The population-adjusted analysis reports 3.2. Adjustment is a normal statistical operation and the paper is explicit about the methods it used. An adjustment that widens the sponsor's lead still deserves more scrutiny than one that narrows it.
Who wrote it, and who paid for it
The disclosure is not hidden. Five of the eight authors are employees and shareholders of Novo Nordisk. Two more work at Petauri Evidence, an agency the paper states was paid by Novo Nordisk to support the study. The eighth author, based at Northwestern University, discloses consulting relationships with both Eli Lilly and Novo Nordisk among several other companies, and the copyright line on the article belongs to Novo Nordisk Inc[1]. Eli Lilly develops orforglipron[3].
None of that makes the result wrong. It does place it in a category with a measured track record. A Cochrane methodology review covering 75 papers found that industry-sponsored drug and device research more often produced favourable efficacy results (risk ratio 1.27, 95% CI 1.17 to 1.37) and more often reached favourable conclusions (RR 1.34, 95% CI 1.19 to 1.51) than research funded from other sources, and its authors reported that the pattern could not be explained by standard risk-of-bias assessment[4]. The fair reading is that this is sponsor-generated evidence of a kind that historically tends to flatter the sponsor, and it should carry the weight that implies.
What an anchored indirect comparison can and cannot do
The method is legitimate and health-technology agencies rely on it when no head-to-head trial exists. An anchored comparison uses each trial's placebo arm as a shared reference point, so what gets compared is each drug's effect against its own placebo rather than one raw result against another. On that axis the two trials cooperate: the placebo groups lost 2.2% and 2.1% respectively[2][3]. Population adjustment then reweights one trial's individual patient data to resemble the other trial's population on the variables that were recorded, here sex, body weight and normoglycaemic status[1].
The limitation sits in that last clause. Adjustment can only touch variables somebody measured. It cannot repair differences in how sites recruited participants, how each protocol titrated the dose, how adverse events were captured and coded, or how a 64-week endpoint lines up against a 72-week one. That is why regulators and guideline committees treat an indirect comparison as supporting evidence and not as proof of superiority. The thing that would settle this question is a trial that randomises the same people to both drugs, and nobody has run one.
What we don't yet know
Whether oral semaglutide genuinely outperforms orforglipron in clinical practice is unresolved. The two trials each tested their drug against placebo, not against each other, and the reported difference of 3.2 percentage points has a confidence interval reaching to within 0.4 points of no difference at all[1]. A head-to-head trial could land on either side of that. The tolerability comparison is looser still, because the gastrointestinal discontinuation odds ratio rests on a small number of events in both arms. ATTAIN-1 described orforglipron as under investigation for obesity when it published in late 2025[3], while semaglutide is a prescription-only medicine across every market we cover. Country-by-country detail is on our semaglutide regulation page. Decisions about starting, switching or stopping either medicine belong with a prescribing clinician who knows your history. This article is for educational and journalistic purposes only and does not constitute medical advice.
Frequently asked
Has oral semaglutide been compared with orforglipron in a head-to-head trial?
No. As of September 2026 there is no randomised trial that gives the same participants either oral semaglutide 25 mg or orforglipron 36 mg. The June 2026 comparison in Diabetes, Obesity and Metabolism is an anchored indirect treatment comparison, which stitches together the separate OASIS 4 and ATTAIN-1 trials using their placebo arms as a shared reference point. That is a recognised method, but it is not the same class of evidence as a direct comparison.
What did the indirect comparison actually find?
It reported that oral semaglutide 25 mg produced 3.2 percentage points more body-weight loss than orforglipron 36 mg on the treatment-regimen estimand (95% confidence interval minus 5.9 to minus 0.4) and 3.0 points more on the efficacy estimand. It also reported higher odds of stopping treatment on orforglipron, with an odds ratio of 4.1 for any adverse event and 13.9 for gastrointestinal adverse events. The interval on that last figure runs from 2.0 to 96.0, which is very imprecise.
Does the Novo Nordisk funding mean the results are wrong?
Not automatically. It does mean the result should be read with the sponsor's interest in view. Five of the eight authors are Novo Nordisk employees and shareholders, two more work for an agency the paper says Novo Nordisk paid, and Novo Nordisk holds the copyright. A Cochrane methodology review of 75 papers found industry-sponsored research more often produces favourable efficacy results and favourable conclusions than research funded elsewhere, by a margin that standard risk-of-bias tools do not explain.
How do the underlying trial numbers compare?
OASIS 4 randomised 307 adults and reported a mean body-weight change of minus 13.6% with oral semaglutide 25 mg at week 64, against minus 2.2% on placebo. ATTAIN-1 randomised 3127 adults and reported minus 11.2% with orforglipron 36 mg at week 72, against minus 2.1% on placebo. The endpoints sit eight weeks apart and the trials differ roughly tenfold in size, which is exactly the kind of mismatch an indirect comparison has to adjust for and cannot fully remove.
Sources
- [1]Michalak W, et al. Oral Semaglutide 25 mg Versus Orforglipron 36 mg in Obesity: A Population-Adjusted Indirect Treatment Comparison. Diabetes Obes Metab. 2026;28(8):7247-7256.Tier 1 · primary↩
- [2]Wharton S, et al. Oral Semaglutide at a Dose of 25 mg in Adults with Overweight or Obesity (OASIS 4). N Engl J Med. 2025;393(11):1077-1087.Tier 1 · primary↩
- [3]Wharton S, et al. Orforglipron, an Oral Small-Molecule GLP-1 Receptor Agonist for Obesity Treatment (ATTAIN-1). N Engl J Med. 2025;393(18):1796-1806.Tier 1 · primary↩
- [4]Lundh A, Lexchin J, Mintzes B, Schroll JB, Bero L. Industry sponsorship and research outcome. Cochrane Database Syst Rev. 2017;2(2):MR000033.Tier 1 · primary↩
No revisions yet. First published .