• Science & evidence
• Clinical trials
• FDA
• Wellness

Lilly's GLP-1 pill eased sleep apnea in two trials. The 44% drop isn't the only number.

By Peptide Policy Watch staff · Published November 10, 2026 · 8 min read · 4 sources, 4 checked at the source

A CPAP mask can keep the airway open, if you can sleep with it on. On November 9, 2026, Eli Lilly said its GLP-1 pill, Foundayo (orforglipron), cut breathing pauses in two Phase 3 sleep-apnea trials, and the figure the company led with was a 43.8% drop among people not using a breathing machine. [V] The same day, the results appeared in JAMA and on the main stage of the American Heart Association's Scientific Sessions in Chicago. The journal's lead number is smaller, because it used a different way of counting. [V] Foundayo is already approved for weight management. It is not approved for sleep apnea. Lilly says it plans to file with FDA for that use by the end of 2026. [V]

Facts are as of November 10, 2026. Sources, with retrieval dates, are at the bottom.

What those sleep numbers mean

AHI, the apnea-hypopnea index, counts how often breathing stops or becomes shallow during an hour of sleep. Lilly, citing the American Academy of Sleep Medicine, treats 15 to 29 events per hour as moderate and 30 or more as severe. [V] People in both trials started well into the severe range: 50.7 events per hour in the group not using PAP, and 57.6 in the group that was. [V]

Lilly also reported a combined "remission or mild disease" figure. In its definition, remission is an AHI below 5, and mild disease is an AHI from 5 to under 15 with no heavy daytime sleepiness, using an Epworth score at or under 10. [V] In the efficacy analysis, 31.5% of people on Foundayo in the non-PAP trial met that combined mark, as did 19.6% in the PAP trial. [V] Those are Lilly's figures, not an FDA conclusion.

Hypoxic burden is a second sleep metric. Lilly describes it as a way of scoring how far oxygen dips, and how long those dips last, when breathing is interrupted. [V] AHI counts the events. Hypoxic burden tries to capture how much oxygen is lost because of them.

The two ways of counting

Lilly's November 9 release is explicit: every result in that document is the efficacy estimand. The company defines that as the effect if every randomized participant had stayed on the assigned treatment for 52 weeks, without starting PAP (in the non-PAP trial), having a prohibited sleep-apnea procedure, or starting a prohibited weight-management treatment. [V]

Under that analysis, people not using PAP saw AHI fall by 21.1 events per hour (43.8%) on Foundayo, compared with 4.3 events per hour (6.3%) on placebo. Almost half, 47.2%, cut their AHI by at least half. People using PAP saw AHI fall by 18.3 events per hour (27.5%) on Foundayo, compared with 5.5 (3.0%) on placebo, and 33.7% cut their AHI by at least half. [V]

JAMA reported the same two trials — 363 and 349 people, the same starting AHI figures — but the abstract leads with a treatment-regimen estimand, which counts people as they were actually treated, including those who stopped. [V] In that analysis, the non-PAP trial showed a mean AHI change of 17.8 events per hour down on orforglipron versus 4.5 down on placebo, a gap of 13.2 events per hour (interval, 8.7 to 17.7 fewer; P < .001). The PAP trial showed 17.3 versus 5.3, a gap of 12.0 (interval, 6.9 to 17.1 fewer; P < .001). [V]

• 43.8%AHI decline on Foundayo in the non-PAP group, Lilly's efficacy analysis, vs 6.3% on placebo [V]
• 13.2fewer AHI events per hour than placebo in JAMA's non-PAP treatment-regimen analysis [V]
• 0FDA approvals of Foundayo for sleep apnea so far [V]

Neither analysis is a typo. The efficacy figures ask what the drug did if people stayed on it. The treatment-regimen figures fold in the messier reality of dropouts. Lilly's 43.8% is a percent change from a severe starting AHI in the first of those two approaches, compared with 6.3% on placebo — not a 44-point gap, and not the number JAMA put up front. [U] That reading is a way of placing the two reports side by side, not a finding from either paper.

The American Heart Association's newsroom summary of the presentation rounded the year-one change to about 18 fewer breathing interruptions per hour on orforglipron and about 5 fewer on placebo, across both studies, which sits with the journal's treatment-regimen results. [V]

People in the PAP trial were using a breathing machine and, Lilly says, were meant to keep using it during the study. [V] The AHA release adds that those participants paused the machine for seven days before sleep testing, so the overnight measurements were taken without it. [V]

Weight, oxygen and blood pressure

Lilly's efficacy analysis also reported weight change: down 11.5% (28.0 pounds) on Foundayo in the non-PAP trial, versus 1.8% on placebo, and down 11.2% (28.2 pounds) in the PAP trial, versus 0.9%. [V] Sleep-apnea-specific hypoxic burden fell 55.9% versus 18.3% in the non-PAP trial and 45.8% versus 17.0% in the PAP trial. [V] High-sensitivity C-reactive protein, a blood marker of inflammation, fell 48.9% and 38.6% on Foundayo in the two trials, and systolic blood pressure fell 8.6 and 7.3 mm Hg, according to the same Lilly analysis. [V]

The presenting author, Vaishnavi Kundel of Mount Sinai, told AHA that participants' weight loss likely contributed to the sleep-apnea improvements, and that later research would need to test whether anything beyond weight loss is at work. [V] These trials were not built to separate those effects. [U]

Limits and unknowns

• Stomach side effects were common. The JAMA abstract says gastrointestinal events were the most frequent adverse events, reported in 68.0% of people randomized to orforglipron versus 37.4% on placebo in the non-PAP trial, and in 70.6% versus 36.6% in the PAP trial. [V] Lilly says those effects were generally mild to moderate, and that the most common were nausea, diarrhea, constipation, dyspepsia and vomiting. [V]
• More people stopped the drug. In Lilly's account, 12.2% of Foundayo patients in the non-PAP trial discontinued because of adverse events, versus 3.8% on placebo. In the PAP trial the figures were 6.8% versus 0.6%. [V]
• Who wasn't enrolled. AHA says both trials left out anyone with diabetes (type 1 or type 2) and anyone with only mild sleep apnea, followed everyone for a year, and underrepresented some racial and ethnic groups. [V]
• This is not an FDA decision. A positive Phase 3 program is evidence for a filing, not an approval. Lilly funded the studies. [V]

Where Foundayo stands with FDA

Foundayo is a GLP-1-receptor drug taken as a pill, built from a small molecule rather than a peptide. That distinction, and the April 1, 2026 weight-management approval, are covered in the Foundayo approval explainer. [V] Lilly says it plans to submit the ATTAIN-OSA results to FDA, by the end of 2026, seeking a moderate-to-severe obstructive sleep apnea use in adults who have obesity or overweight. [V] Until that review happens, sleep apnea is not a labeled use.

Lilly also notes that its other medicine, Zepbound (tirzepatide), is already FDA-approved for that sleep-apnea use in adults who have obesity. [V] Our approved-medicines comparison describes that labeled use, which comes from Zepbound's official label. A Foundayo sleep-apnea use would need its own decision.

On November 9, FDA also required a new eye-risk warning on every approved GLP-1 receptor agonist, including Foundayo. That class-wide label change is covered in our NAION warning article. [V]

The useful next date is Lilly's planned filing, and then FDA's decision. The trial record is in JAMA and on ClinicalTrials.gov as NCT06649045. For how to read a listing like that, see our guide to ClinicalTrials.gov.

Sources (retrieved November 10, 2026 unless noted)

1. Eli Lilly and Company, "Lilly's Foundayo (orforglipron) significantly reduced obstructive sleep apnea severity, reinforcing its potential to improve health beyond weight loss," PR Newswire, November 9, 2026, 09:30 ET [V] — https://www.prnewswire.com/news-releases/lillys-foundayo-orforglipron-significantly-reduced-obstructive-sleep-apnea-severity-reinforcing-its-potential-to-improve-health-beyond-weight-loss-302932877.html
2. Malhotra A, Gottlieb DJ, Woehrle H, et al., "Orforglipron for Obstructive Sleep Apnea: Results From 2 Randomized Clinical Trials (ATTAIN-OSA)," JAMA, published November 9, 2026, doi 10.1001/jama.2026.24080; PubMed abstract, PMID 43002161 [V] — https://pubmed.ncbi.nlm.nih.gov/43002161/ and https://doi.org/10.1001/jama.2026.24080
3. American Heart Association, "Daily, oral GLP-1 medication reduced episodes of sleep apnea in adults with excess weight," Scientific Sessions 2026 news release, November 9, 2026 [V] — https://newsroom.heart.org/news/daily-oral-glp-1-medication-reduced-episodes-of-sleep-apnea-in-adults-with-excess-weight
4. ClinicalTrials.gov, NCT06649045 (ATTAIN-OSA) [V] — https://clinicaltrials.gov/study/NCT06649045

Spot a mistake? See how we source and our corrections policy.

Image: AI-generated. Image credits