A pill cut body weight by as much as 7.8 percent in people who had both obesity and type 2 diabetes. The pill is orforglipron, an oral version of the appetite-and-blood-sugar drugs that until now meant a weekly injection. And the 7.8 percent figure does not come from a single headline trial. It comes from the first attempt to pool everything published on the drug in this group of patients, which turns out to be three trials and 2,505 people.
That pooled analysis, published June 26 in Diabetology & Metabolic Syndrome ↗, is worth reading less for the size of the effect than for how thin the evidence underneath it still is.
What orforglipron is
Semaglutide and tirzepatide, the drugs sold as Ozempic, Wegovy, Mounjaro and Zepbound, are peptides. They lean on the GLP-1 receptor ↗, the switch that tells the body it is full and helps the pancreas manage glucose (tirzepatide pulls a second gut receptor, GIP ↗, at the same time), and they have to be injected, because a peptide swallowed as a pill gets digested before it can work. Orforglipron, Eli Lilly's oral candidate, is not a peptide. It is a small molecule built to flip the same GLP-1 receptor, which means it survives the stomach and can be taken as a daily tablet. That is the entire commercial pitch: same target, no needle, better odds that people actually keep taking it.
So the question the meta-analysis asked is narrow and practical. Across the trials run so far, how much does the pill actually move weight and blood sugar, and how sure can anyone be about the answer?
The dose curve
The effect climbs with dose, cleanly. At the lowest 3 mg dose, body weight fell 2.43 percent more than placebo. At 24 mg it fell 7.80 percent more. Waist circumference shrank at every dose of 6 mg and above. Blood sugar followed the same shape: HbA1c, the three-month average of blood glucose, dropped 0.80 percent at the low dose and 1.67 percent at the highest dose tested, and fasting glucose fell further at every step up. A consistent dose-response like that is the signature drug developers want, because it suggests the effect is real and not a fluke of one trial.
The safety read was the ordinary GLP-1 story. More people on orforglipron quit because of side effects, the nausea-and-gut complaints that come with hitting this receptor, and that was true across doses. Serious adverse events, the hospital-grade problems, were no more common than on placebo at any dose.
The part the effect size hides
Here is what a clean dose curve cannot tell you: three trials is not much. The pooled cohort is 2,505 people, and the authors themselves rated several of the key efficacy outcomes as low certainty under GRADE, the standard system for scoring how much weight a body of evidence can bear. Their own conclusion says the results should be read cautiously, pending larger confirmatory studies.
That is the honest shape of the orforglipron story right now. A dose-response from three trials is a slope with a hypothesis attached, not a settled drug. The number that gets quoted, 7.8 percent, is real, but it sits on an evidence base small enough that the next trial could move it in either direction.
For the injectable incumbents the bar is different. Semaglutide ↗ and tirzepatide ↗ have been through cardiovascular outcome trials with tens of thousands of patients and years of follow-up. Orforglipron's pitch is that an oral drug people will keep taking can close the real-world gap that adherence opens up. The pooled trial data says the molecule does what it should at the receptor. Whether it holds the line at scale is the trial that has not reported yet.