- Orforglipron, a once-daily oral GLP-1 pill, reduced sleep apnea severity in adults with obesity, whether or not they used a PAP machine.
- Breathing interruptions during sleep fell by about 18 per hour with the study drug, vs about 5 with placebo.
- Patients lost about 10% of their body weight, and systolic blood pressure dropped by about 7 mm Hg.
For decades, the main treatment for obstructive sleep apnea (OSA) has been a positive airway pressure (PAP) machine worn at night. It helps patients sleep better, but 4 major randomized trials have not shown that it lowers cardiovascular risk. Obesity is one of the biggest drivers of OSA, which makes weight loss a natural target.
In a late-breaking presentation at the American Heart Association 2026 Scientific Sessions, Dr. Vaishnavi Kundel of the Icahn School of Medicine at Mount Sinai presented ATTAIN-OSA (NCT06649045), simultaneously published in the Journal of the American Medical Association. The program, funded by Eli Lilly, tested orforglipron, an oral GLP-1 receptor agonist, in two phase 3 trials. Both enrolled adults with overweight or obesity (BMI ≥ 27 kg/m2) and moderate to severe OSA, without diabetes. The first trial (n=363) included people who were not using PAP. The second (n=349) included people who were, and who paused it for a week before each sleep study. Patients were randomized to orforglipron or placebo for 52 weeks.
The main measure was the apnea-hypopnea index (AHI), the number of times per hour that breathing stops or becomes shallow during sleep. Patients started with severe OSA, with an average AHI greater than 50.
At 52 weeks, orforglipron cut AHI by about 18 events per hour in both trials, compared with about 5 with placebo (P<0.001). That was a 35% reduction in patients not using PAP and a 26% reduction in PAP users. Hypoxic burden, which measures how much oxygen levels drop during these events, fell more than twice as much with the drug as with placebo.
The benefits went beyond sleep. Patients on orforglipron lost about 10% of their body weight, compared with 1% to 2% on placebo. At 48 weeks, systolic blood pressure dropped by about 7 mm Hg, compared with about 2 mm Hg on placebo, and C-reactive protein, a marker of inflammation, fell by about 40%. Kundel noted that PAP lowers blood pressure in OSA by only 2 to 3 mm Hg. Patients on the drug also reported feeling less impaired by poor sleep, but the difference from placebo was not statistically significant in the main analysis.
Stomach side effects were the most common, mainly nausea, which affected about a third of patients on orforglipron. Side effects led 7% to 12% of patients on the drug to stop treatment, compared with 1% to 4% on placebo. The safety findings matched earlier orforglipron trials.
The trial excluded people with diabetes and was not designed to determine whether orforglipron reduces cardiovascular events. The blood pressure results were not adjusted for multiple comparisons.
Kundel and colleagues concluded that the results provide “evidence for the potential use of orforglipron as an oral treatment option for moderate to severe OSA in adults with overweight or obesity.”