The Obesity Pill Arrived: What the Trials Measured

You have probably seen both headlines: there is now a weight-loss pill instead of a weekly injection, and a new combination injection that was meant to beat everything missed its target. The useful question is narrower than "does it work" — what was measured, in whom, for how long, and which numbers come from a paper rather than a company release.

This material is informational and does not replace a consultation with a physician. Every dose below is a dose used inside a trial, not an instruction.

A pill instead of a weekly shot — what was measured?

On 1 April 2026 the US FDA approved orforglipron tablets as Foundayo. The label is worth reading literally: together with a reduced-calorie diet and more physical activity, to reduce excess body weight and keep it off in adults with obesity, or overweight with at least one weight-related condition. The approval letter (NDA 220934) defers paediatric studies in ages 6 to 17.

The pivotal trial, ATTAIN-1, appeared in the New England Journal of Medicine on 16 September 2025: phase 3, double-blind, placebo-controlled, 3,127 adults randomised to once-daily oral orforglipron at 6, 12 or 36 mg or to placebo, all on a reduced-calorie diet and more activity. The primary endpoint was percent change in body weight to week 72, counted across everyone randomised whether or not they stayed on the drug.

• 36 mg: −11.2% (95% CI −12.0 to −10.4)

• 12 mg: −8.4%; 6 mg: −7.5%

• Placebo: −2.1% (95% CI −2.8 to −1.4); P<0.001 for each dose

In the 36 mg group, 54.6% lost at least 10% of body weight, 36.0% at least 15% and 18.4% at least 20%, against 12.9%, 5.9% and 2.8% on placebo. Between 5.3% and 10.3% stopped the drug for adverse events, against 2.7% on placebo; those events were mostly gastrointestinal, mostly mild or moderate. Waist, systolic blood pressure, triglycerides and non-HDL cholesterol moved more than on placebo — risk-factor numbers, not heart attacks or deaths.

Headlines said about 12% because two analyses answer two questions. The figure above counts everyone randomised, including those who stopped early. The other estimates what happens if everyone stays on treatment: the manufacturer's same-day release gave −12.4% at 36 mg versus −0.9% on placebo. Comparing two drugs means comparing the same kind of figure.

Is the pill as good as the injections?

ATTAIN-1 cannot answer that: it had no active comparator, only placebo. Setting its 11.2% beside a figure from another trial, another drug and another population is arithmetic, not comparison.

The one published head-to-head is ACHIEVE-3 (The Lancet, online 26 February 2026): open-label, 52 weeks, 1,698 adults with type 2 diabetes on metformin, mean baseline HbA1c 8.3%, BMI at least 25. The primary endpoint was HbA1c change at week 52. Orforglipron gave −1.91 points at 36 mg and −1.71 at 12 mg, against −1.47 for oral semaglutide 14 mg and −1.23 for 7 mg — a difference of −0.44 points between the top doses (95% CI −0.62 to −0.26). The cost side moved too: gastrointestinal events in 59% and 58% versus 37% and 45%, discontinuation for adverse events 9% and 10% versus 4% and 5%, and a pulse rise of 3.7 and 4.7 beats per minute versus 1.0 and 1.5.

Three limits matter more than that result. The comparator is oral semaglutide at diabetes doses, not the injectable obesity dose and not tirzepatide; the population has type 2 diabetes; and the endpoint is blood sugar, not weight — the abstract gives no verified weight change.

What are amylin drugs, and why is CagriSema everywhere?

Cagrilintide is an amylin analogue — a different hormone pathway from GLP-1. CagriSema is cagrilintide given with semaglutide; petrelintide is another amylin analogue given alone.

REDEFINE 1 (NEJM, 22 June 2025) was phase 3a over 68 weeks, double-blind, with placebo and active arms: 3,417 adults without diabetes (BMI at least 30, or at least 27 with a complication), randomised to the combination (2,108), semaglutide 2.4 mg alone (302), cagrilintide 2.4 mg alone (302) or placebo (705), all with lifestyle support. Coprimary endpoints: relative weight change, and a loss of at least 5% at week 68. The combination gave −20.4% against −3.0% on placebo, a difference of −17.3 percentage points (95% CI −18.1 to −16.6), P<0.001 — counted across everyone randomised, whether or not they stayed on the drug. The abstract does not print the single-drug arms, so how much each drug contributed on its own cannot be read off the paper. The sponsor's own summary of the paper (18 December 2025) adds gastrointestinal events in 79.6% versus 39.9%, nausea in 55% versus 12.6%, discontinuation for adverse events in 5.9% versus 3.5%, and −22.7% versus −2.3% on the "if everyone stayed on treatment" analysis.

REDEFINE 2 ran the same combination in 1,206 adults who also had type 2 diabetes (904 on drug, 302 on placebo, 68 weeks, no active arm): −13.7% versus −3.4%, a difference of −10.4 points (95% CI −11.2 to −9.5) on the same count-everyone-randomised analysis, with gastrointestinal events in 72.5% versus 34.4%. Same drug, different population, a markedly smaller change — which is why the population belongs in every sentence.

Status: an NDA was filed in December 2025, and the sponsor then stated the drug was not approved in the US or the EU; a release of 23 September 2026 still called it investigational, with regulatory decisions then expected in the fourth quarter of 2026. No approval notice appeared in our check, and the cardiovascular-outcomes trial REDEFINE 3 is still running.

Why did the comparison against tirzepatide miss?

REDEFINE 4 exists only as a company announcement dated 23 February 2026, not a peer-reviewed paper. Open-label, 84 weeks, CagriSema 2.4/2.4 mg against tirzepatide 15 mg: 23.0% versus 25.5% if everyone stayed on treatment, 20.2% versus 23.6% counting everyone randomised. The pre-specified primary endpoint, non-inferiority on weight loss, was not met, and the sample size was absent from the text we retrieved. An open-label trial reported by its own sponsor with a missed primary endpoint is a reason to wait for the paper, not a verdict on either drug.

Petrelintide: how early is early?

Phase 2. ZUPREME-1 (Lancet Diabetes & Endocrinology, 29 September 2026) was double-blind and placebo-controlled, completed 7 March 2026, in 485 adults without diabetes — mean age 47, mean BMI 36.7, mean weight 107.1 kg — on weekly doses of 1.0 to 9.0 mg across 42 weeks. The primary endpoint was percent weight change at week 28, and it was counted only for people who stayed on treatment — a different count from the ATTAIN-1 figure above: −9.8% at 5 mg (95% CI −10.9 to −8.6), −9.3% at 7 mg, −9.4% at 9 mg and −7.9% at both 1 and 2.5 mg, against −1.7% on placebo (95% CI −2.8 to −0.5). Above 5 mg, more drug added nothing by week 28.

Weight was still falling at week 42, up to −10.7% on drug, but the abstract does not pair that with a week-42 placebo figure. Nausea was reported by 79 of 404 people on drug (20%) against 6% on placebo; vomiting, diarrhoea and constipation were 3%, 7% and 7% against 6%, 7% and 4%. No deaths in the abstract. The sponsor said a global phase 3 monotherapy programme had started; petrelintide is not approved anywhere we found.

What none of this shows

Cross-trial arithmetic is not a comparison. Trials, populations, durations and analysis conventions do not line up.

Risk factors are not events. Changes in waist, blood pressure and lipids do not demonstrate that heart attacks or deaths become less frequent, and no cardiovascular-outcomes result for orforglipron was verified in our check.

Phase 2 is not a treatment option. Cagrilintide alone sits there: a 26-week dose-finding trial in 706 adults without diabetes, against liraglutide 3.0 mg, an older GLP-1 drug. Petrelintide's phase 3 has not reported.

A company release is not a paper. REDEFINE 4 and several of the "if everyone stayed on treatment" figures above come from sponsors and have not been through peer review.

Averages are averages. These were adults with obesity inside structured programmes with diet and activity support; one person's result is not the group mean. A trial dose label is also not a marketed tablet strength.

The horizons are short. Sixty-eight and seventy-two weeks cannot see rare or late harm, and none measured what happens after the drug stops.

What to do with these numbers

If you are preparing a conversation about drug treatment for obesity, these figures work better as questions than conclusions. Which analysis produced the number — everyone randomised, or only those who stayed on treatment? Which population, and does it resemble mine? And was the comparison I care about ever run?

Approval status you can check yourself, and it decides what is available: orforglipron is approved in the US for the adults described in its label; CagriSema was still investigational in its sponsor's September 2026 release; cagrilintide alone and petrelintide are research programmes.

Give the side-effect numbers the same weight as the weight numbers: across these trials they differed between drugs more sharply than the average weight changes. Dose, suitability, interactions and monitoring belong to a physician who sees your history.

Sources

• Wharton et al. Orforglipron, an oral small-molecule GLP-1 receptor agonist for obesity treatment (ATTAIN-1). NEJM. 2025. PMID 40960239 · DOI 10.1056/NEJMoa2511774 · NCT05869903

• Rosenstock et al. Orforglipron versus oral semaglutide in type 2 diabetes (ACHIEVE-3). Lancet. 2026. PMID 41765029 · DOI 10.1016/S0140-6736(26)00202-3 · NCT06045221

• Garvey et al. Cagrilintide plus semaglutide in overweight or obesity (REDEFINE 1). NEJM. 2025. PMID 40544433 · DOI 10.1056/NEJMoa2502081 · NCT05567796

• Davies et al. CagriSema in overweight or obesity with type 2 diabetes (REDEFINE 2). NEJM. 2025. PMID 40544432 · DOI 10.1056/NEJMoa2502082 · NCT05394519

• Lau et al. Once-weekly cagrilintide for weight management, phase 2 dose-finding. Lancet. 2021. PMID 34798060 · DOI 10.1016/S0140-6736(21)01751-7 · NCT03856047

• Petrelintide phase 2 (ZUPREME-1). Lancet Diabetes & Endocrinology. 2026. DOI 10.1016/S2213-8587(26)00213-5 · NCT06662539

• REDEFINE 3, cardiovascular outcomes with CagriSema — ongoing, no result. NCT05669755

• Manufacturer release issued on the day of the ATTAIN-1 publication — the figures for people who stayed on treatment (−12.4% at 36 mg versus −0.9% on placebo). Sponsor statement, not peer-reviewed; cited by date, as no stable link was verified in our check.

• Novo Nordisk release, 18 December 2025 — the sponsor’s own summary of REDEFINE 1 (gastrointestinal events, nausea, discontinuation, and the −22.7% versus −2.3% "if everyone stayed on treatment" figures) and the NDA filing. Sponsor statement, not peer-reviewed; cited by date.

• Novo Nordisk release, 23 September 2026 — CagriSema still described as investigational, regulatory decisions then expected in the fourth quarter of 2026. Sponsor statement; cited by date.

• Zealand Pharma release, 30 September 2026 — the start of a global phase 3 monotherapy programme for petrelintide. Sponsor statement; cited by date.

• Novo Nordisk company announcement No. 13/2026, 23 February 2026 — REDEFINE 4 headline results. Sponsor statement, not peer-reviewed; cited by document number and date, as no stable link was verified in our check.

• US FDA news release, 1 April 2026, and approval letter for NDA 220934 (reference ID 5773652) — orforglipron (Foundayo) tablets. Regulatory documents; cited by document and date, as no stable link was verified in our check.

Informational material. It does not replace a consultation with a physician and is not a treatment recommendation.