GLP-1 Agonists and Sleep Apnea: SURMOUNT-OSA, 63% AHI Reduction, and What It Means vs. CPAP

What SURMOUNT-OSA Actually Found

Published in The New England Journal of Medicine in June 2024, the SURMOUNT-OSA trial is the definitive Phase 3 study of a GLP-1/GIP receptor agonist in obstructive sleep apnea. It enrolled 469 adults with moderate-to-severe OSA (apnea-hypopnea index ≥15 events per hour) and obesity (BMI ≥30) across two parallel cohorts: those who were CPAP-naïve or unwilling to use CPAP, and those currently on CPAP who paused it for the trial.

The primary endpoint was change in AHI from baseline to 52 weeks. Participants randomized to tirzepatide 10–15 mg weekly experienced a mean AHI reduction of approximately 27 events per hour — a 63% improvement — compared to just 4.8% in the placebo group (p<0.0001). In the CPAP-naïve cohort, 42.3% of tirzepatide patients achieved AHI remission, defined as fewer than 5 events per hour, compared to 15.7% on placebo.

Beyond the primary outcome, tirzepatide patients showed significant improvements in hypoxic burden (the cumulative time spent with oxygen saturation below 90%), patient-reported sleepiness on the Epworth Sleepiness Scale, and cardiovascular biomarkers. Mean body weight loss in the tirzepatide arms was 18–20% over 52 weeks. Crucially, the drug was simultaneously improving metabolic risk — HbA1c, blood pressure, lipids — that CPAP alone cannot address.

The FDA approved tirzepatide (Zepbound) specifically for moderate-to-severe OSA in adults with obesity in December 2024, making it the first drug ever approved for this indication. This approval represents a paradigm shift: treating sleep apnea pharmacologically via root-cause fat reduction rather than purely mechanical airway management.

The Mechanism: How Upper Airway Fat Drives OSA — and How GLP-1 Removes It

Obstructive sleep apnea is fundamentally an anatomical problem exacerbated by obesity. During wakefulness, upper airway dilator muscles — primarily the genioglossus (tongue muscle) — maintain patency despite soft-tissue pressure. During sleep, especially during REM, this neuromuscular tone drops. In lean individuals, the airway remains open. In people with obesity, fat deposited in specific anatomical compartments makes airway collapse far more likely.

The critical anatomical sites are the parapharyngeal fat pads, the tongue body, and the soft palate. MRI studies show that individuals with OSA have significantly greater tongue fat volume and expanded lateral pharyngeal walls compared to weight-matched controls without OSA. This fat mechanically reduces airway cross-sectional area and increases the collapsibility index (Pcrit — the critical closing pressure of the upper airway).

GLP-1 receptor agonists do not preferentially burn upper airway fat in isolation, but the distribution of weight loss matters enormously. Multiple imaging studies have now documented that semaglutide and tirzepatide drive preferential reduction in visceral and ectopic fat depots — the metabolically active, hormonally responsive fat stored in and around organs and soft tissues. Tongue fat and parapharyngeal deposits fall into this category. In the SURMOUNT-OSA imaging sub-study, participants showed disproportionate reductions in upper airway soft tissue compared to subcutaneous body fat.

Additionally, GLP-1 agonists reduce systemic inflammation (lowering CRP and IL-6), which may reduce upper airway mucosal edema. There is also emerging evidence that GLP-1 receptors in brainstem nuclei modulate respiratory drive and arousal thresholds, though the clinical magnitude of this central effect remains under investigation. The dominant mechanism is mechanical: less fat = more airway space = fewer obstructions.

Semaglutide and OSA: The Evidence Before SURMOUNT

While tirzepatide's SURMOUNT-OSA data is more comprehensive, semaglutide built the evidentiary foundation for GLP-1 drugs in sleep medicine. In June 2024 — the same month as SURMOUNT-OSA — the FDA approved semaglutide (Wegovy, 2.4 mg weekly) for moderate-to-severe OSA in adults with obesity, based on data from the STEP-HFpEF trial and dedicated OSA sub-analyses.

The STEP-HFpEF trial (heart failure with preserved ejection fraction) enrolled patients with both obesity and HFpEF, many of whom also had OSA. Semaglutide 2.4 mg produced an average 25% reduction in AHI at 52 weeks versus placebo, alongside substantial improvements in 6-minute walk distance, Kansas City Cardiomyopathy Questionnaire scores, and CRP. The OSA benefit appeared independent of CPAP use, suggesting a physiological effect beyond simple weight loss.

Sub-analyses of the pivotal STEP 1 trial (semaglutide vs. placebo in 1961 adults with obesity) showed that among participants with baseline OSA, those who achieved ≥10% weight loss experienced roughly a 30% reduction in self-reported apnea frequency. Polysomnography substudies of STEP 1 confirmed AHI improvements correlating tightly with total weight lost — approximately 2.5 events/hour per percent of body weight reduction.

Prior to GLP-1 data, the most robust weight-loss intervention for OSA was bariatric surgery. A landmark meta-analysis by Greenburg et al. (2009, Annals of Internal Medicine) found bariatric surgery reduced AHI by 38.2 events/hour — a result now surpassed by tirzepatide's pharmacological effect at a fraction of the procedural risk. This comparison underscores how transformative the SURMOUNT-OSA data truly is.

GLP-1 Agonists vs. CPAP: Complementary, Not Competing

The most important clinical question is not "which is better" but rather "what does each treatment actually do, and for whom?" CPAP (Continuous Positive Airway Pressure) and GLP-1 agonists solve different problems with different mechanisms, time courses, and risk profiles.

CPAP is mechanically superior at eliminating apnea events during the night it is used. A well-titrated CPAP machine reduces AHI to <5 events/hour in the large majority of users regardless of body weight. Its benefits on nocturnal oxygenation are immediate and, across large trials (SAVE, ISAACC, RICCADSA), have been shown to reduce cardiovascular events in patients who are highly adherent (>4 hours/night). CPAP is non-pharmacological, reversible, and has decades of safety data.

However, CPAP has critical limitations. Adherence is the core problem: despite best clinical effort, 30–50% of patients use CPAP fewer than 4 hours per night, and 20–30% abandon it within 12 months. CPAP treats the nightly airway obstruction but does nothing for the underlying adiposity driving it. It does not improve insulin resistance, blood pressure trajectory, or cardiovascular inflammation — the co-morbidities that make severe OSA a mortality risk.

GLP-1 agonists address root cause. By reducing parapharyngeal fat and tongue volume, they structurally improve the airway over months. The 63% AHI reduction in SURMOUNT-OSA represents a durable, metabolically meaningful change that persists even during the nights you don't "take" anything. Patients who achieve remission (<5 AHI) on tirzepatide may no longer need CPAP at all — though this should be confirmed by repeat polysomnography before discontinuing CPAP.

Current expert consensus (including American Academy of Sleep Medicine 2025 guidance) recommends a combinatorial approach: maintain CPAP while titrating GLP-1 therapy to ≥10% weight loss, then re-assess with polysomnography. Many patients will find CPAP pressure requirements drop substantially, improving comfort and adherence. A subset will achieve OSA remission and may discontinue CPAP under clinical supervision.

Weight-Dependent Effects and What to Expect at Different Levels of Loss

The relationship between weight loss and AHI improvement is well-characterized and non-linear — early weight loss yields proportionally larger OSA improvements than equivalent additional loss at lower body weights. This has implications for how patients on GLP-1 therapy should track their progress.

The 10% threshold is clinically meaningful. A 2013 meta-analysis by Araghi et al. in Sleep found that 10% weight loss produced approximately 26% AHI reduction across lifestyle and surgical interventions. This aligns with the dose-response observed in GLP-1 trials: semaglutide patients averaging 12–15% weight loss typically achieve 28–35% AHI reductions.

SURMOUNT-OSA participants averaged 18–20% weight loss, producing the 63% AHI reduction. Modeling data suggest that the 42% achieving OSA remission were concentrated among those losing more than 20% of body weight — consistent with the curvilinear dose-response. Patients who lose 25% or more of body weight, as some tirzepatide users do at maximum dose (15 mg), may see AHI improvements exceeding 70–75%.

Practically, patients should expect: 5% weight loss → minimal but measurable AHI change (10–15%), improved oxygen nadir; 10% weight loss → meaningful AHI reduction (25–30%), likely improved daytime sleepiness; 15–20% weight loss → major AHI improvement (40–65%), substantial chance of remission; >20% weight loss → highest remission rates, may enable CPAP discontinuation with clinical monitoring.

Notably, body composition at baseline predicts response. Patients with predominantly upper-body and visceral fat distribution respond better than those with more subcutaneous, lower-body fat — because GLP-1 agonists preferentially mobilize the visceral depot. Positional OSA (symptoms predominantly supine) also tends to respond better than pan-positional severe OSA, because the anatomical contribution of fat is proportionally larger in supine-predominant cases.

Evidence Summary: Key Studies in GLP-1 and Obstructive Sleep Apnea