A tablet that works as well as an injection is the thing most readers actually want, and this is one of the trials that decides whether it exists. The answer is encouraging and comes with three conditions attached, all of which get dropped when the headline travels — the same fate as the trial figures collected in what the semaglutide trials are worth.
What was tested
Two hundred and one adults with overweight or obesity, with or without type 2 diabetes, were randomized 2:1 to oral semaglutide 50 mg daily or placebo [1]. Mean age was 49, 43.3% were female, a quarter had type 2 diabetes, and mean body weight was 91.9 kg.
What it found
Mean weight change was −14.3% on the drug against −1.3% on placebo, an estimated treatment difference of 13.07 percentage points (95% CI −15.61 to −10.52). More of the treated group reached a 5% reduction: 84.3% against 17.2%, an odds ratio of 23.00.
Gastrointestinal adverse events were reported by 63.4% of the semaglutide arm against 34.8% on placebo, and adverse events led 4.5% of the treated group to discontinue. That discontinuation figure is lower than the injectable trials generally report, though with 134 people on the drug it is not a precise estimate.
The part that matters on a storefront
That last condition is the one with commercial consequences. A seller offering “oral semaglutide” is almost certainly not offering 50 mg daily of the product this trial used, and in the compounded market the milligram is frequently not published at all — the silence counted in who publishes a dose ladder.
A reader who has seen this result and gone looking for a tablet may be sold something at an unstated fraction of the dose, with no way to tell from the page. What an oral formulation does and does not deliver is covered in what the pill actually does, and the chemically different small-molecule tablet now in trials is a separate matter again, covered in orforglipron, a pill without a needle.
What it establishes
That a peptide tablet at a high enough dose can produce double-digit weight reduction in a randomized trial. That is a real advance and it is worth stating plainly. What it does not establish is that any particular tablet a reader can buy will do it, which depends entirely on a dose most sellers decline to name.