SURMOUNT-OSA: How Much Does Zepbound Improve Sleep Apnea?

SURMOUNT-OSA (Malhotra 2024 NEJM^[1]) ran two parallel phase-3 trials of tirzepatide in adults with obesity and moderate-to-severe obstructive sleep apnea — one in patients who were not using CPAP, one in patients who were. Both hit their primary endpoint. Apnea-hypopnea index dropped by 25 events per hour on the CPAP-naive arm and 29 events per hour on the CPAP-treated arm by week 52, compared with placebo reductions of about 5 events per hour. In December 2024 the FDA used these data to approve Zepbound for moderate-to-severe OSA in adults with obesity, and within weeks the Centers for Medicare and Medicaid Services confirmed Medicare Part D coverage of tirzepatide on that indication — the first time a weight-loss-class molecule has been a covered Medicare benefit. This article goes deeper than the headline result: what the trial-1 vs trial-2 split actually shows, how AHI response distributes across subgroups, and the practical eligibility math.

The honest summary

• Two trials, same drug, different baseline. SURMOUNT-OSA trial 1 randomized 234 adults with moderate-to-severe OSA who were not using CPAP. Trial 2 randomized 235 adults who were already on CPAP and continued it. Both ran 52 weeks on tirzepatide titrated to the maximum tolerated dose (10 or 15 mg) vs placebo. Both met their primary endpoint of AHI change from baseline (Malhotra 2024^[1]).
• AHI improved by about half. Mean baseline AHI was roughly 50 events per hour in both trials. At week 52, tirzepatide arms dropped to about 25 events per hour in trial 1 and about 21 in trial 2 — absolute reductions of 25.3 and 29.3 events per hour. Placebo dropped by ~5.
• Body weight loss was -18 to -20%. Consistent with SURMOUNT-1 in adults with obesity but no OSA endpoint. Most of the AHI improvement tracks the weight change, but a smaller independent fat-redistribution effect on parapharyngeal soft tissue is plausible (Schwartz 2008^[5]).
• Tirzepatide does not reach the CPAP floor. CPAP titrated to residual AHI typically delivers an AHI of ~5 events per hour. Tirzepatide brings a baseline of 50 down to ~25. For severe OSA, that is meaningful but not equivalent to therapy — combination is the most aggressive option (Malhotra 2024^[1]).
• Medicare 2024: a new coverage pathway opens. The FDA OSA indication unlocks Medicare Part D coverage in a population for which the obesity indication alone is not a covered Part D benefit. The qualification math — BMI, documented AHI, polysomnography vs home test — is the new gatekeeper.

Trial 1 vs trial 2: the split that matters

The two SURMOUNT-OSA trials were designed in parallel to answer two different clinical questions about the same drug. Trial 1 addressed the patient who has been diagnosed with moderate-to-severe OSA but is not on CPAP — either intolerant, unwilling, or not yet treated. Trial 2 addressed the patient who is on CPAP and stays on it during the study. The randomization, dosing, and 52-week duration were identical; the only structural difference was the baseline OSA treatment status.

Trial 1 (CPAP-naive, n=234) reported a placebo-adjusted change in AHI of −20.0 events per hour at week 52 (tirzepatide −25.3, placebo −5.3). Trial 2 (on CPAP, n=235) reported a placebo-adjusted change of −23.8 events per hour (tirzepatide −29.3, placebo −5.5). The marginally larger effect in trial 2 is not a sign that CPAP and tirzepatide were synergistic in a pharmacologic sense — AHI is measured during the study sleep nights with the CPAP machine in use, so the residual AHI is what is being reduced. The cleaner reading is that both populations responded to the underlying weight-loss-driven anatomy change.

For the clinician, the implication is that the CPAP-vs-tirzepatide question is not either-or. Trial 2 is the evidence base for adding tirzepatide on top of CPAP and watching the residual events fall. Trial 1 is the evidence base for offering tirzepatide to a patient who has refused or failed CPAP — with the caveat that the post-treatment AHI is not equivalent to a titrated CPAP floor.

Subgroup heterogeneity: who responds most

The published trial reported pre-specified subgroup analyses of the primary endpoint (Malhotra 2024^[1]). The broad pattern: AHI reduction tracks the percentage of body weight lost, with secondary modifiers from sex, baseline BMI, and baseline AHI severity.

• BMI tertiles. Higher baseline BMI was associated with larger absolute AHI reduction, consistent with the known dose-response between weight loss and pharyngeal patency (Schwartz 2008^[5]). Patients in the lowest BMI tertile (BMI 30–34) still benefited but with a narrower margin over placebo.
• Sex. Both men and women showed clinically meaningful AHI reductions. Men had a higher baseline AHI on average and therefore a larger absolute change. The percentage reduction was similar across sex.
• Age. Younger patients (less than ~50) had slightly larger AHI reductions, plausibly because of more plastic upper-airway soft tissue and greater absolute weight loss. Patients over 65 still hit the primary endpoint, which matters for the Medicare population.
• Baseline AHI severity. Severe OSA at baseline (AHI ≥ 30) saw the largest absolute reductions. Moderate OSA at baseline (AHI 15–29) saw smaller absolute changes but a larger fraction of patients reaching AHI < 5, the threshold often used for clinical remission.
• Epworth Sleepiness Scale. Patient-reported sleepiness improved with tirzepatide on both trials. The effect was moderate — not the dramatic improvement most patients describe with first-night CPAP — but meaningful at week 52.

Magnitude: AHI events/hour at week 52

The pathophysiology in two sentences

Obstructive sleep apnea in adults with obesity is mechanically driven by fat deposition in the parapharyngeal soft tissue and at the base of the tongue, narrowing the upper airway and increasing collapsibility during sleep (Schwartz 2008^[5]). Weight loss reduces neck circumference by 5 to 10 cm at the magnitudes seen in SURMOUNT-OSA, which is the proximate cause of the AHI improvement — not a central respiratory-drive effect of the GLP-1 / GIP agonism itself.

FDA approval December 2024 and the Medicare pathway

On December 20, 2024, the FDA approved a supplemental indication for Zepbound (tirzepatide) covering moderate-to-severe obstructive sleep apnea in adults with obesity (BMI ≥ 30). The approval was based directly on the SURMOUNT-OSA trial-1 and trial-2 data (Malhotra 2024^[1]). The labeling specifies that tirzepatide is for use in combination with a reduced-calorie diet and increased physical activity, consistent with the obesity indication.

The Medicare implication is the load-bearing change. The 2003 Medicare Modernization Act explicitly excluded “drugs used for the treatment of obesity” from Part D coverage, which is why semaglutide for obesity (Wegovy) and tirzepatide for obesity (Zepbound) had been ineligible for Medicare Part D. The FDA OSA indication moves the same molecule onto a covered medical indication. CMS clarified in early 2025 that Part D plans may cover tirzepatide for beneficiaries with the FDA-approved OSA indication, subject to plan-level prior authorization. This is the first weight-loss-class molecule with a Medicare coverage pathway, and it materially changes the access math for patients aged 65+ with documented OSA.

The practical eligibility math

Three independent filters determine which patient qualifies under the OSA indication on Medicare Part D today.

1. BMI ≥ 30. About 42% of US adults meet this threshold in the most recent NHANES data. Documented via a recent clinic weight and height.
2. Moderate-to-severe OSA, AHI ≥ 15. Established via either in-laboratory polysomnography (the SURMOUNT-OSA inclusion standard, and the AASM guideline gold standard per Epstein 2009^[6]) or a validated home sleep apnea test (HSAT) when AHI > 15 on the home test — AASM 2017 diagnostic guideline (Kapur 2017^[7]). Lower-acuity HSAT results (AHI 5–14.9) generally require confirmatory PSG.
3. Obesity-driven, not lean OSA. The label and the trial population both require co-existing obesity. Lean OSA patients (BMI < 30) do not qualify under this indication; they remain primarily CPAP or oral-appliance candidates.

For patients who clear all three filters, the workflow is: recent sleep study report, recent BMI documentation, ICD-10 coding for OSA (G47.33) plus obesity (E66.x), and a plan-specific prior authorization form. Some plans will additionally require documentation of CPAP intolerance or suboptimal response, mirroring the trial-1 inclusion logic.

Comparison to bariatric surgery

The classical Buchwald 2004 JAMA meta-analysis^[9]of bariatric surgery reported OSA resolution or improvement in 85% of patients across procedures, with sleeve gastrectomy and gastric bypass producing similar absolute AHI reductions in the range of 25 events per hour at one year — comparable in magnitude to the SURMOUNT-OSA tirzepatide arm. The trade-off is well known: bariatric surgery carries operative and long-term nutritional risk, while tirzepatide carries the recurring cost and the GI-side-effect profile. For patients on Medicare who clear the OSA pathway, tirzepatide is now a non-surgical option with a similar magnitude of effect.

OSA, cardiovascular risk, and the longer-term question

The American Heart Association scientific statement on OSA and cardiovascular disease (Yeghiazarians 2021^[8]) summarized the evidence that untreated moderate-to-severe OSA is associated with hypertension, atrial fibrillation, and increased risk of major adverse cardiovascular events. CPAP adherence in real-world cohorts is poor — the Sawyer 2011 Sleep Medicine Reviews systematic review^[3]documented average adherence rates in the 30 to 60 percent range, depending on how adherence is defined. Tirzepatide offers a once-weekly injection that does not require nightly device wear, and the AHI reduction it produces is meaningful even when it does not reach the CPAP residual floor. Whether that translates into the same cardiovascular event reduction as well-titrated CPAP is the open question for the next decade of OSA outcomes research.

Provider routing in practice

Many telehealth GLP-1 clinics do not currently offer a sleep medicine workup; conversely, most sleep medicine clinics do not prescribe tirzepatide. The patients who clear the eligibility filters fastest are the ones with a recent in-laboratory PSG already on file. For patients without that record, the practical path is: primary care or sleep medicine referral for HSAT or PSG, ICD-10 documentation, then GLP-1 prescribing through either the same clinic or a partnered telehealth route. Combined sleep medicine and obesity medicine clinics are an emerging model that consolidates the workup and the prescription under one prior authorization bundle.

Related research and tools

• SURMOUNT-OSA tirzepatide sleep apnea overview — the higher-level explainer of the trial result and the FDA approval
• GLP-1 sleep apnea patient guide — the patient-facing workup, expectations, and insurance basics
• GLP-1 insurance coverage: Medicare, Medicaid, commercial — how the OSA indication fits into the broader coverage landscape
• Tirzepatide vs semaglutide head-to-head — why semaglutide is not (yet) on the OSA pathway
• Mounjaro vs Zepbound complete comparison — same molecule, different indications; which one your OSA prescription will reference

Important disclaimer. This article is educational and does not constitute medical advice. Eligibility decisions for the OSA indication require a clinician review of a documented sleep study, current weight and height, and plan-specific prior authorization requirements. Tirzepatide is not a substitute for CPAP in severe or symptomatic OSA without prescribing-clinician oversight; combination therapy is the strongest evidence base in the trial-2 cohort. Medicare and commercial plan coverage rules change — verify current plan policies before making a clinical or financial commitment. PMIDs were verified live against the PubMed E-utilities API on 2026-05-29.

Last verified: 2026-05-29. Next review: every 6 months, or sooner if new SURMOUNT-OSA secondary analyses, CMS coverage determinations, or FDA labeling updates are published.