What does a weight-loss trial tell someone whose BMI is below 30 but who already has a weight-related health condition? A new analysis of ATTAIN-1 offers more specific evidence for that group. It also shows why a promising average should not become an individual prediction.
The paper, published on October 14, 2026, examines 260 adults with BMI from 27 to below 30, at least one adiposity-related complication, and no diabetes. Orforglipron produced greater weight reduction than placebo over 72 weeks. But this was a small, post-hoc subgroup analysis, with results that depend on how treatment interruptions and missing measurements were handled.
A closer look at a group inside the larger trial
The ATTAIN-1 registry describes a randomized, double-blind trial enrolling 3,127 adults. People with BMI of at least 27 could qualify if they had conditions such as hypertension, abnormal blood lipids, obstructive sleep apnea or cardiovascular disease. Diabetes was excluded. Participants received orforglipron or placebo alongside diet and physical-activity support.
The new subgroup paper narrows that population. Its four groups contained 56, 64 and 62 people assigned to the three active regimens, and 78 assigned to placebo. The researchers actually applied an upper BMI cutoff of 29.9, excluding 16 participants whose BMI fell above 29.9 but below 30. Most subgroup participants were Asian, which limits how confidently the estimates can be extended to other populations.
This is a new analysis of an existing study, not a newly completed trial or a new approval. The registry records primary completion in July 2025 and an ongoing extension. Our orforglipron profile gives the broader medicine background.
The 13.5% result answers a particular question
At week 72, the highest studied regimen had an estimated mean weight change of −13.5%, compared with −0.3% for placebo. The difference was 13.2 percentage points in favor of orforglipron; its 95% confidence interval ranged from 10.2 to 16.1 percentage points. These are model-based group estimates, not a promise that each participant lost that amount.
Crucially, the paper’s main subgroup analysis used an efficacy estimand: it asks about effects assuming participants continued their assigned treatment and did not start prohibited weight-management therapies. Measurements after those events were excluded from this analysis, and the statistical model handled missing observations under a missing-at-random assumption.
A treatment-regimen analysis, which addresses a different treatment scenario, appears as supportive evidence in the supplement. We did not inspect that supplement and do not quote its estimates. Readers should therefore avoid comparing this 13.5% figure directly with a treatment-regimen headline from the overall trial. A different population and a different analysis can both change the number.
Why the dose numbers look different
The subgroup report labels its regimens 5.5, 9 and 17.2 mg, whereas earlier ATTAIN-1 reporting used 6, 12 and 36 mg. This reflects different formulations, not a new randomized test proving that a smaller milligram amount works better.
A separate phase 1 bioequivalence study tested capsules and corresponding dose-adjusted tablets in 429 otherwise healthy adults with overweight or obesity. It found that the paired formulations met prespecified criteria for comparable drug exposure. The subgroup paper uses the equivalent tablet strengths. These numbers identify research regimens; they are not instructions for converting, selecting or changing a prescription.
Risk markers improved, but outcomes remain unproven
The subgroup report describes reductions in waist circumference, systolic blood pressure and several blood-lipid measures. Those findings add context beyond body weight. They do not establish that this subgroup experienced fewer heart attacks, strokes or deaths. A change in a risk marker and a demonstrated reduction in clinical events are different claims.
Nor did every measure clearly separate from placebo. Diastolic blood-pressure changes were comparable across groups, and the report found no between-group difference in high-sensitivity C-reactive protein. The analyses involved many comparisons without adjustment for multiple testing. Their nominal P values should be read as exploratory evidence, consistent with a study not powered to confirm subgroup hypotheses. Our trial-reading guide explains these distinctions.
Tolerability belongs beside the benefit
In the highest-regimen group, six of 62 participants stopped treatment because of adverse events, compared with five of 78 on placebo. Gastrointestinal events were common and predominantly mild to moderate. The small subgroup cannot reliably characterize rare harms or establish that people below BMI 30 tolerate treatment better than the wider trial population.
There is already a US weight-management indication for Foundayo, the orforglipron product. The FDA label covers adults with obesity or overweight plus at least one weight-related condition, alongside reduced-calorie eating and increased physical activity. Its contraindications and warnings still apply; a favorable subgroup average does not override them. This analysis does not establish approval or availability in another jurisdiction.
The useful contribution is narrower: people below BMI 30 with complications now have a more focused estimate from ATTAIN-1, with substantial uncertainty attached. A care discussion can ask which health outcomes matter, how tolerability will be monitored and how closely the study population matches the individual. The broader evidence guide puts this finding alongside other orforglipron research. Larger, more diverse analyses and longer follow-up are needed before treating these subgroup results as dependable personal forecasts.



