CPAP adherence rates in real-world practice are around 50–65% at best, and that's among the patients who actually fill their prescription. For a condition affecting an estimated 1 billion people globally, that represents an enormous treatment gap — one that has motivated decades of research into alternatives, with limited success. Two developments in the past few years have genuinely changed this conversation. First, tirzepatide's SURMOUNT-OSA program showed that dramatic weight loss produces dramatic AHI reductions — enough for FDA approval specifically for OSA in 2024. Second, a growing body of evidence shows that a meaningful subset of OSA patients has significant non-anatomical drivers: reduced muscle responsiveness, low arousal threshold, or unstable ventilatory control. For those patients, a CPAP is treating the symptom, not the cause.
This article is for informational purposes only and does not constitute medical advice. Clinical trial eligibility and availability vary. Always consult a qualified healthcare professional before making any medical decisions or considering participation in a clinical trial.
Summary
Obstructive sleep apnea affects an estimated 1 billion people worldwide, but fewer than 30% of diagnosed patients consistently use CPAP. In June 2024, tirzepatide (Zepbound) became the first drug ever approved specifically for moderate-to-severe OSA in adults with obesity — based on SURMOUNT-OSA Phase 3 data showing 55–63% reductions in apnea-hypopnea index. In parallel, next-generation hypoglossal nerve stimulation devices are expanding surgical options, and the atomoxetine-oxybutynin combination is in Phase 3 targeting the non-anatomical endotypic drivers of OSA that CPAP and weight loss don't address.
ClinicalMetric Analysis
- Tirzepatide's OSA benefit tracks directly with weight loss magnitude — which means it's most effective in patients who achieve the greatest weight reduction, not uniformly across all users. The SURMOUNT-OSA AHI reductions of 55–63% from baseline were achieved in patients with significant obesity (mean BMI ~39) who reached average weight loss of ~18%. Patients with moderate obesity who achieve less dramatic weight reduction will see proportionally smaller AHI benefits. The OSA effect is not independent of weight — there's no mechanism by which GLP-1 agonism directly affects pharyngeal muscle tone outside of its weight-loss-mediated fat reduction in upper airway tissue. Patients with OSA and modest obesity should have realistic expectations about the magnitude of AHI benefit.
- The complete concentric palatal collapse exclusion for Inspire affects 25–30% of surgical candidates — and many patients told they "might qualify" haven't actually had the required DISE evaluation. Drug-induced sleep endoscopy (DISE) is the definitive pre-surgical evaluation, revealing upper airway collapse patterns under sedation that polysomnography can't assess. Patients told by a sleep medicine physician that they're "a candidate for the implant" based on CPAP failure and BMI criteria alone have not been fully evaluated. Nyxoah Genio's bilateral stimulation approach is being specifically developed for complete concentric collapse patients — a defined unmet need within the surgical OSA population.
- Atomoxetine-oxybutynin targets non-anatomical OSA endotypes — and endotype characterization, not polysomnography alone, determines which patients are most likely to respond. OSA pathophysiology involves four endotypic contributors: anatomical (pharyngeal soft tissue burden), arousal threshold (low threshold means patients wake before muscles can compensate), muscle responsiveness (genioglossal muscle response to airway narrowing), and loop gain (respiratory control instability). Ato-oxy primarily addresses low arousal threshold and poor muscle responsiveness. Patients whose OSA is primarily anatomical — high BMI, large tonsils, retrognathia — are unlikely to respond. Endotype characterization requires specialized sleep study analysis beyond standard AHI measurement, and trials that don't use this stratification will underestimate ato-oxy's efficacy in the responsive subgroup.
Tirzepatide for OSA: The SURMOUNT-OSA Results in Detail
Tirzepatide (Zepbound/Mounjaro, Eli Lilly) is a dual GIP/GLP-1 receptor agonist that produces average weight loss of 20–22% of body weight in obesity trials — substantially greater than any prior non-surgical intervention, including semaglutide's 15% average in STEP trials. Because pharyngeal fat accumulation is the primary modifiable anatomical contributor to OSA, the hypothesis that tirzepatide would dramatically reduce OSA severity was well-grounded. The SURMOUNT-OSA Phase 3 program tested it in two populations: OSA patients not using PAP therapy, and those using PAP therapy.
Both trials showed tirzepatide reduced the apnea-hypopnea index (AHI — breathing interruptions per hour of sleep) by approximately 27–30 events per hour from baseline. In percentage terms, these represent 55–63% reductions from baseline. Many patients shifted from severe (AHI >30) to mild (AHI <15) or normal (AHI <5) categories. Oxygen saturation nadirs improved. Patient-reported daytime sleepiness (Epworth Sleepiness Scale) and sleep quality scores improved significantly. Secondary cardiovascular endpoints also trended favorably. The FDA approved tirzepatide specifically for moderate-to-severe OSA in adults with obesity in June 2024 — the first drug approval for this indication in the drug's history. Current trials are evaluating tirzepatide in combination with CPAP, in central sleep apnea, and in patients with OSA without obesity.
Hypoglossal Nerve Stimulation: The Inspire Data and What Comes Next
Upper airway stimulation (UAS) via the Inspire device (Inspire Medical Systems) stimulates the hypoglossal nerve — which controls tongue protrusion and upper airway patency — in synchrony with respiratory effort detected by a chest pressure sensor. The STAR trial (NCT01490970) reported 68% reduction in median AHI at 12 months and durable efficacy sustained at five-year follow-up, with favorable quality-of-life outcomes. Inspire received FDA approval in 2014 and is now standard-of-care for CPAP-intolerant OSA patients with appropriate anatomy — specifically, without complete concentric palatal collapse on drug-induced sleep endoscopy (DISE), which currently excludes roughly 25–30% of candidates evaluated.
The next-generation device landscape in 2026 is active:
Nyxoah Genio System
A bilateral hypoglossal nerve stimulator implanted entirely at the chin — no chest-implanted pulse generator, no sleep remote required. The DREAM trial (NCT03678259) demonstrated AHI reductions comparable to Inspire in European cohorts. The device is approved in Europe and Australia. The BETTER SLEEP IDE trial is currently enrolling US patients for FDA approval. The bilateral stimulation approach may extend eligibility to patients with complete concentric collapse who currently cannot receive Inspire.
Palate Stiffening Combined with HNS
For patients excluded from HNS due to complete concentric palatal collapse, trials are testing palate stiffening procedures (Pillar implants, radiofrequency ablation of soft palate tissue) performed before or concurrently with HNS implantation to address the palatal component. Early data suggests this combination can extend HNS eligibility to the previously excluded anatomical phenotype.
Endotyping OSA: Treating the Cause, Not Just the Anatomy
The insight that changed how sleep medicine researchers think about drug therapy is this: OSA is not uniformly anatomical. Approximately 30–40% of patients have a significant non-anatomical component to their apnea — one or more "endotypic traits" that are measurable and pharmacologically targetable:
- Reduced upper airway muscle responsiveness: The genioglossus and other upper airway dilators normally stiffen during sleep to maintain patency. In a subset of patients, this muscle response is blunted — the airway collapses despite adequate anatomical dimensions.
- Low arousal threshold: Some patients wake at minimal respiratory stimulus — before the event is even classified as an apnea — disrupting sleep architecture without reducing AHI, and creating unstable ventilatory patterns by interrupting sleep repeatedly.
- High loop gain: An unstable ventilatory control system that over-responds to chemical stimuli (CO2, O2), creating oscillating breathing patterns independent of anatomical airway narrowing.
These traits explain why many non-obese patients have severe OSA, why weight loss eliminates OSA in some patients but not others, and why CPAP fails to adequately treat some patients from a cardiovascular-outcome perspective even when AHI is controlled.
AD109: Targeting Muscle Tone and Arousal Threshold
Atomoxetine plus oxybutynin (AD109, Apnimed) is the most advanced pharmacological program targeting non-anatomical OSA. Atomoxetine, a norepinephrine reuptake inhibitor used in ADHD, increases upper airway muscle tone during sleep via noradrenergic signaling to the hypoglossal motor nucleus. Oxybutynin, a muscarinic antagonist, raises the arousal threshold, allowing deeper sleep without the reflexive arousals that destabilize ventilatory control. The combination was tested in a landmark proof-of-concept study that showed ~50% AHI reduction in a selected population — specifically patients with measurable genioglossal muscle responsiveness deficits.
The Phase 2 MARIPOSA trial enrolled 100 patients and confirmed meaningful AHI reductions in the target population. The Phase 3 MARIPOSA-2 trial is actively recruiting for primary endpoint data in 2026. Critically, eligibility requires characterization of OSA endotype — a drug-induced sleep endoscopy or upper airway physiology testing to confirm that the patient has the muscle responsiveness and arousal threshold traits that the drug targets. This enrichment step is essential: in unselected OSA populations, the effect size is much smaller.
Key Takeaways
- Tirzepatide (Zepbound) is FDA-approved for moderate-to-severe OSA in adults with obesity — the first drug approval for this indication — with SURMOUNT-OSA Phase 3 data showing 55–63% AHI reductions and significant improvements in oxygen saturation and patient-reported sleep quality.
- Inspire UAS has 5-year durability data; Nyxoah Genio (bilateral, no pulse generator) is enrolling in the US BETTER SLEEP trial with European approval already in hand.
- Atomoxetine + oxybutynin (AD109) targets non-anatomical OSA — specifically patients with reduced upper airway muscle responsiveness and low arousal threshold. Phase 3 MARIPOSA-2 is recruiting; this drug only works in patients whose OSA is driven by these endotypic traits.
- OSA endotyping — DISE, upper airway physiology phenotyping, loop gain estimation — is becoming a practical tool to match patients to appropriate therapy rather than defaulting to CPAP for everyone.
- Semaglutide and tirzepatide in central sleep apnea are in active trials — the benefit in central vs. obstructive mechanisms is an open question and the subject of ongoing work.