Sleep apnea now has a drug. New data show what it fixes and what it doesn't.

For as long as obstructive sleep apnea has had a standard treatment, that treatment has been a machine. Continuous positive airway pressure works when it is used. The difficulty has always been the second clause.

On December 20, 2024, the FDA approved tirzepatide, marketed as Zepbound, for moderate-to-severe obstructive sleep apnea in adults with obesity — the first prescription medication ever approved for the condition [s1]. Twenty months on, the trial behind that approval has been re-analyzed in ways that make the picture considerably more specific, and a second drug with an entirely different mechanism is sitting in front of the agency.

What the original trial showed

SURMOUNT-OSA was run as two parallel 52-week randomized, double-blind, placebo-controlled phase 3 studies. Study 1 enrolled 234 participants who were not using positive airway pressure therapy; Study 2 enrolled 235 who were and intended to continue [s2]. Participants received the maximum tolerated dose of tirzepatide, either 10 or 15 mg once weekly [s2]. These were people with severe disease: mean baseline apnea-hypopnea index was 51.5 ± 31.0 events per hour in Study 1 and 49.5 ± 26.7 in Study 2 [s2].

At 52 weeks, tirzepatide reduced the apnea-hypopnea index by 25 to 29 events per hour, against 5 to 6 events per hour on placebo [s1]. Participants lost 18% to 20% of body weight, roughly 45 to 50 pounds, compared with about 2% on placebo [s1]. Remission or a drop to mild disease was reached by 42% of the non-PAP group and 50% of the PAP group on tirzepatide, versus 16% and 14% on placebo [s1].

Those are large effects by the standards of sleep medicine. They are also, importantly, effects on a physiological index — not on the outcomes that make apnea dangerous.

The January 2026 analysis: what came from where

The more useful question is why the drug helped, and a secondary-outcomes paper published in Nature Medicine on January 15, 2026 tried to answer it [s2].

The cardiometabolic results were substantial. At week 48, systolic blood pressure fell by 7.9 mmHg more than placebo in Study 1 (95% CI −11.0 to −4.9; P<0.001) and by 4.3 mmHg more in Study 2 (95% CI −7.3 to −1.2; P=0.007) [s2]. Diastolic pressure fell by 3.2 mmHg more than placebo in Study 1 (95% CI −5.4 to −1.0; P=0.005) but showed no significant difference in Study 2 [s2]. At week 52, high-sensitivity C-reactive protein — a marker of systemic inflammation — was 28.9% lower than placebo in Study 1 (P=0.003) and 45.1% lower in Study 2 (P<0.001) [s2]. Triglycerides fell 32.2% and 31.5% relative to placebo (both P<0.001), and a post hoc insulin-resistance measure fell 48.0% and 54.5% (both P<0.001) [s2].

Then the authors ran a mediation analysis, which is the part that changes how the drug should be understood. Changes in the sleep apnea measures — apnea-hypopnea index and sleep apnea-specific hypoxic burden — independently mediated the effect on C-reactive protein, insulin resistance, and triglycerides [s2]. But the blood pressure benefit was mediated by weight loss, and by the combination of weight and apnea measures, not by the apnea measures on their own [s2].

The authors' own conclusion is the honest summary: "treating both sleep-disordered breathing and obesity is likely required to optimize the treatment effect on cardiometabolic benefits" [s2]. Fixing the breathing and shrinking the body are doing different jobs.

The limits, stated plainly

The paper is explicit about what it cannot support. The studies "did not have sufficient duration of follow-up or statistical power to examine the impact of tirzepatide on cardiometabolic outcomes such as myocardial infarction or stroke" [s2]. Nobody has shown that this drug prevents a heart attack in this population.

The trial also excluded people with mild obstructive sleep apnea, people with diabetes, and people with a body mass index below 30 [s2]. That is a large share of the apnea population, and the results do not transfer to them. The authors note that emerging data suggest tirzepatide and GLP-1 receptor agonists produce less weight reduction in patients with type 2 diabetes than in those without [s2], which would likely blunt the effect in a group the trial never tested.

On how people actually felt, a patient-reported-outcomes analysis of the same 469 participants, published in Sleep Medicine in August 2025, found improvements that were real but modest. PROMIS Sleep Disturbance scores improved by 2.0 points more than placebo in Study 1 (95% CI −4.0 to −0.1) and 3.9 points in Study 2 (95% CI −6.2 to −1.6) [s3]. Daytime sleepiness on the Epworth scale improved by 1.4 points more than placebo in Study 1 (95% CI −2.5 to −0.3) but did not reach significance in Study 2 (−0.9; 95% CI −2.1 to 0.2) [s3]. The authors flag that most of these endpoints were exploratory and may lack adequate statistical power, and that thresholds for clinically meaningful change on several of these measures are poorly established [s3].

What is coming behind it

Tirzepatide treats apnea by treating obesity. A second candidate is trying to treat the airway directly.

AD109, a once-daily oral combination of aroxybutynin and atomoxetine developed by Apnimed, targets the neuromuscular control of the upper airway rather than body weight. In the phase 3 SynAIRgy trial, 646 participants across 73 centers in the US and Canada were randomized 1:1 against placebo [s4][s6]. Apnimed reports a 55.6% mean reduction in apnea-hypopnea index from baseline at week 26, with 39.6% of treated participants achieving at least a 50% reduction and 22.3% reaching an index below 5 events per hour — the threshold for disease control [s4]. Hypoxic burden fell 60.5% [s4]. The most common adverse events were dry mouth, insomnia, and nausea, and the company reports no serious adverse events related to the drug [s4]. Results were published in the American Journal of Respiratory and Critical Care Medicine in May 2026 [s4].

Two cautions. These figures come from company communications about the trial, and reductions from baseline are not the same as reductions relative to placebo — placebo groups in apnea trials typically improve as well, so the between-group difference is the smaller and more meaningful number. The FDA accepted Apnimed's new drug application, announced on July 14, 2026, and assigned a target action date of February 28, 2027 [s7]. No approval decision has been made. Published figures for the drug's component doses have not been consistent across company and trade communications, so they are omitted here.

The access question

A drug that works is not the same as a drug people can get.

Medicare has been prohibited by law from covering medications used specifically for weight loss since the Part D benefit began [s5]. The apnea indication threads that needle: Part D plans can cover a GLP-1 when it is used for "a medically accepted FDA-approved indication other than obesity, like type 2 diabetes, cardiovascular disease risk reduction, or sleep apnea" [s5].

Separately, a Medicare GLP-1 Bridge demonstration began on July 1, 2026 and runs through December 31, 2027, offering a $50 monthly copay for all formulations of Foundayo and Wegovy and the KwikPen formulation of Zepbound, at a negotiated net price of $245 per month [s5]. Eligibility requires meeting BMI and clinical criteria with provider attestation through prior authorization, and the $50 copay does not count toward a beneficiary's Part D deductible or out-of-pocket maximum [s5]. The broader BALANCE Model launched for Medicaid on May 1, 2026 and runs to December 31, 2031; its Medicare Part D component has been indefinitely delayed [s5].

What to watch

Whether a cardiovascular outcomes trial in this population is ever run, since the existing data cannot answer whether the drug prevents events. Whether the FDA acts on AD109 by its February 28, 2027 target date, which would give the field a mechanism that does not depend on weight. And whether coverage for the apnea indication holds as the demonstration programs shift.

This article describes published trial results and regulatory actions. It is not medical advice, and decisions about apnea treatment belong with a clinician who has seen the patient's sleep study.

Sources

• [s1] Zepbound approved by FDA as first sleep apnea medication, American Academy of Sleep Medicine, 2024-12-21 (FDA approval 2024-12-20).
• [s2] Tirzepatide on obstructive sleep apnea-related cardiometabolic risk: secondary outcomes of the SURMOUNT-OSA randomized trial, Nature Medicine, online 2026-01-15; 32(2):653–659.
• [s3] Effect of tirzepatide treatment on patient-reported outcomes among SURMOUNT-OSA participants with obstructive sleep apnea and obesity, Sleep Medicine, online 2025-08-06.
• [s4] Apnimed Announces Publication of its Phase 3 SynAIRgy Trial of AD109 for Obstructive Sleep Apnea in the American Journal of Respiratory and Critical Care Medicine, Apnimed, May 2026.
• [s5] What to Know About the BALANCE Model for GLP-1s in Medicare and Medicaid and the Medicare GLP-1 Bridge, KFF, 2026-05-11.
• [s6] Apnimed Announces Positive Topline Results in the First Landmark Phase 3 Clinical Trial of AD109, Apnimed, 2025-05-19.
• [s7] FDA Accepts NDA for Apnimed Sleep Apnea Pill, RT: Respiratory Therapy, 2026-07-14.