How Does Oral Orforglipron Compare With Tirzepatide for Obstructive Sleep Apnea?
Orforglipron cut sleep apnea events 12 to 13 an hour more than placebo in ATTAIN-OSA, against 20 to 24 on tirzepatide's label, largely tracking weight loss.
Idea in Brief
The News
Two phase 3 ATTAIN-OSA trials published Nov. 9 in JAMA found once-daily orforglipron lowered AHI by 12.0 to 13.2 events an hour more than placebo at 52 weeks.
The Comparison
Tirzepatide's label shows placebo-adjusted AHI reductions of 20.0 and 23.8 with about twice the weight loss; by our arithmetic the AHI gain per point of weight lost is similar.
The Caveats
Cross-trial comparison only, populations and analyses differ, and orforglipron has no sleep apnea indication.
Orforglipron, Eli Lilly's once-daily oral GLP-1 receptor agonist, lowered the apnea-hypopnea index by about 12 to 13 events an hour more than placebo after 52 weeks in two phase 3 trials published Nov. 9 in JAMA. On tirzepatide's label, the injectable already approved for obstructive sleep apnea shows placebo-adjusted reductions of about 20 to 24 events an hour.
The ATTAIN-OSA results, led by Atul Malhotra and presented at the American Heart Association's Scientific Sessions by Vaishnavi Kundel, are the first phase 3 sleep apnea data for an oral incretin. Orforglipron is already on the U.S. market as Foundayo. Tirzepatide (Zepbound) has carried an obstructive sleep apnea indication since the FDA approved it on Dec. 20, 2024. Patients who cannot or will not inject will ask whether the pill does the same job. The short answer from the published numbers is that it helps, by less.
What ATTAIN-OSA tested
The two trials ran in parallel at 66 sites in nine countries, enrolling between Nov. 11, 2024, and July 23, 2025. Adults with moderate to severe obstructive sleep apnea and a body mass index of 27 or more were randomized 1:1 to orforglipron or placebo, titrated from 0.8 mg to the highest tolerated dose of 9, 14.5 or 17.2 mg. Trial 1 enrolled 363 people not using positive airway pressure. Trial 2 enrolled 349 PAP users, who stopped PAP for seven days before each sleep study. People with diabetes were excluded.
Baseline disease was severe: mean AHI was 50.7 events an hour in trial 1 and 57.6 in trial 2. At 52 weeks, AHI fell by 17.8 events an hour on orforglipron and 4.5 on placebo in trial 1 (difference, -13.2; 95% CI, -17.7 to -8.7), and by 17.3 and 5.3 in trial 2 (difference, -12.0; 95% CI, -17.1 to -6.9). Body weight fell 10.4 and 10.7 percent on the drug. Six of seven key secondary endpoints were met; the sleep-related impairment score was not.
How does it compare with tirzepatide?
Lilly ran both programs, and both used two 52-week trials split by PAP use. We lined up the published results from the JAMA paper and section 14 of the Zepbound label.
| Orforglipron, no PAP | Orforglipron, on PAP | Tirzepatide, no PAP | Tirzepatide, on PAP | |
|---|---|---|---|---|
| Baseline AHI, events/h | 50.7 | 57.6 | 51.5 | 49.5 |
| AHI change, drug vs placebo | -17.8 vs -4.5 | -17.3 vs -5.3 | -25.3 vs -5.3 | -29.3 vs -5.5 |
| Placebo-adjusted AHI difference | -13.2 | -12.0 | -20.0 | -23.8 |
| At least 50% AHI reduction | 42.2% vs 16.0% | 30.8% vs 13.1% | 61.2% vs 19.0% | 72.4% vs 23.3% |
| Remission or mild, nonsymptomatic OSA | 28.4% vs 13.0% | 19.2% vs 9.0% | 42.2% vs 15.9% | 50.2% vs 14.3% |
| Placebo-adjusted weight change | -8.3% | -9.6% | -16.1% | -17.3% |
| AHI events/h per point of weight lost, our arithmetic | 1.6 | 1.25 | 1.2 | 1.4 |
The last row is the one that explains the gap. Divide each placebo-adjusted AHI reduction by the placebo-adjusted weight loss and the two drugs land in the same narrow range, 1.2 to 1.6 fewer events an hour per percentage point of weight lost. That is a crude ratio across trials, not a mechanistic finding, but it is consistent with weight loss driving most of the benefit, and with tirzepatide's larger effect coming from roughly twice the weight loss. The JAMA authors say further studies are needed to establish whether mechanisms beyond weight loss contribute.
A pill that lowers AHI by a dozen events an hour is useful. It is not a substitute for PAP in a patient who starts with 50.
Clinical Pearls
- ATTAIN-OSA randomized 363 patients not on PAP and 349 PAP users to orforglipron or placebo for 52 weeks.
- AHI fell 17.8 vs 4.5 events an hour in trial 1 and 17.3 vs 5.3 in trial 2.
- Weight fell 10.4 and 10.7 percent on orforglipron, about 8 to 10 points more than placebo.
- Tirzepatide's label reports AHI reductions of 25.3 and 29.3 vs about 5 on placebo, with 16 to 17 points of placebo-adjusted weight loss.
- Discontinuation for adverse events on orforglipron was 12.2 percent vs 3.8 percent in trial 1 and 6.8 vs 0.6 percent in trial 2.
- People with diabetes, a BMI under 27 or mild OSA were excluded.
Where the comparison breaks down
The populations differ. ATTAIN-OSA admitted a BMI of 27 or more, with at least one failed weight-loss attempt, while tirzepatide's trials required a BMI of 30 or more. Trial 2 of ATTAIN-OSA started at a higher mean AHI than either tirzepatide trial. The statistical approaches differ too: ATTAIN-OSA reports a treatment-regimen estimand, and the Zepbound label reports a modified intention-to-treat analysis with multiple imputation. None of these differences is large enough to close a 7 to 12 point gap in placebo-adjusted AHI, but they are reasons not to quote the numbers as a head-to-head result.
Tolerability
Gastrointestinal adverse events occurred in 68.0 percent of orforglipron patients and 37.4 percent on placebo in trial 1, and 70.6 and 36.6 percent in trial 2. Discontinuation for any adverse event was 12.2 percent vs 3.8 percent in trial 1 and 6.8 vs 0.6 percent in trial 2; serious adverse events were 8.8 vs 4.9 percent and 6.2 vs 5.2 percent. Each trial had one death, both on placebo. The authors list their own limitations: no data past 52 weeks, low racial and ethnic diversity, mostly severe disease, and exclusion of people with diabetes, a BMI under 27 or mild OSA.
Orforglipron and tirzepatide are both among the 12 GLP-1 receptor agonist products that the FDA on Nov. 9 said must carry a new optic nerve warning. The trials were presented at the same American Heart Association meeting where two lipoprotein(a) and inflammation drugs missed on outcomes.
What it means in clinic
Orforglipron does not have a sleep apnea indication, and the sources we reviewed did not say when Lilly will seek one. Until it does, prescribing it for OSA is off-label and coverage will depend on its approved indications and the patient's plan. For a patient with moderate to severe OSA who is choosing between agents, the published data favor tirzepatide on both AHI and weight. For a patient who will not inject, orforglipron roughly doubled the share of patients reaching remission or mild disease compared with placebo, and that is a reasonable discussion to have, alongside a repeat sleep study at a year and a plan to keep PAP in place until it shows the change.
Reported, fact-checked and edited by The Script Pad's editorial staff. See How we use AI.
This article is for professional education and does not replace clinical judgment. Treatment decisions should be based on the individual patient and current guidelines.
