SURMOUNT-OSA post-hoc: who benefits most
SURMOUNT-OSA post-hoc (Sleep, 2026): more-symptomatic patients saw larger subjective gains; objective AHI and weight changes were consistent across subgroups.
Why we wrote this. The primary SURMOUNT-OSA result answers whether tirzepatide works; this post-hoc answers who notices it most.
In this article (7 sections)
- What SURMOUNT-OSA measured in the first place
- How the post-hoc divided participants
- Subjective gains were larger in people with more symptoms at baseline
- Objective measures stayed consistent across subgroups
- Where this fits in the broader SURMOUNT-OSA evidence base
- What this does not tell us
- Medical disclaimer
Tirzepatide for obstructive sleep apnea (OSA) earned its first regulatory approval in December 2024, when the FDA cleared Zepbound for moderate-to-severe OSA in adults with obesity.[4] The approval rested on the SURMOUNT-OSA programme, two 52-week randomised, placebo-controlled trials reported in the New England Journal of Medicine that showed large reductions in the apnea-hypopnea index (AHI), the standard count of breathing interruptions per hour of sleep.[2]
What that primary result could not fully answer was whether tirzepatide's benefits were shared evenly across the range of patients enrolled, or whether people who started with more severe symptoms experienced a different pattern of improvement. A post-hoc analysis published on 7 July 2026 in the journal Sleep, led by Chisom Kanu and Terri Weaver, takes that question apart subgroup by subgroup.[1]
What SURMOUNT-OSA measured in the first place
The two SURMOUNT-OSA trials enrolled adults with obesity and moderate-to-severe OSA, defined as an AHI of at least 15 events per hour. Study 1 excluded people already on positive airway pressure (PAP) therapy; Study 2 enrolled people who were using PAP and wished to discontinue or already had. Both trials ran 52 weeks and tested tirzepatide at the maximum tolerated dose of 10 or 15 mg weekly.
In Study 1, tirzepatide cut AHI by 25.3 events per hour versus a 5.3-event reduction on placebo, a treatment difference of 20.0 events per hour (95% CI, 16.1 to 23.9; P<0.001). In Study 2, the figures were 29.3 versus 5.5, a difference of 23.8 (95% CI, 19.5 to 28.0; P<0.001).[2] Beyond AHI, the primary trials also collected patient-reported outcome measures (PROMs): questionnaires capturing how rested, alert, and functional participants felt. Improvement in those self-reported scales ran alongside the objective AHI change, though the relationship between objective and subjective improvement is not always linear in sleep medicine.
How the post-hoc divided participants
Kanu et al. stratified Study 1 participants at baseline into two subgroups on each of four symptom dimensions: fatigue (present or absent), daytime sleepiness (present or absent), snoring severity, and sleep quality (good or poor).[1] The goal was to determine whether the size of tirzepatide's PROM benefit tracked with baseline symptom load, and whether the objective changes in AHI and body weight were equally distributed regardless of starting severity.
Subjective gains were larger in people with more symptoms at baseline
Across every PROM examined in Study 1, participants who were more symptomatic at baseline showed bigger improvements over placebo than those who were less symptomatic.[1]
On the Functional Outcomes of Sleep Questionnaire (FOSQ) Activity Level subscale, the fatigued subgroup showed a treatment difference of 0.33 points above placebo; the non-fatigued subgroup showed 0.05 points. On the SF-36v2 Vitality domain, fatigued participants gained 8.5 points more than placebo versus 1.0 points in the non-fatigued group.
The sleepiness stratification showed a similar pattern. Participants who reported daytime sleepiness at baseline saw PROMIS Sleep-Related Impairment improve by 7.1 points more than placebo; those without sleepiness at baseline showed a 1.6-point difference. For sleep quality, participants with poor baseline quality showed a 5.8-point advantage over placebo on PROMIS Sleep Disturbance; those with good baseline quality showed a 1.2-point difference.
These numbers reflect differences from placebo, not absolute symptom scores, and the post-hoc nature of the analysis means the subgroups were not pre-specified in the trial protocol. The findings are hypothesis-generating, not confirmatory.
Objective measures stayed consistent across subgroups
The finding that separates this paper from the PROM results is its report on objective metrics. AHI reductions and body-weight reductions showed no meaningful variation between the more-symptomatic and less-symptomatic subgroups.[1] Tirzepatide cut AHI and weight to a similar degree regardless of whether participants entered the trial fatigued or not, sleepy or not, with good or poor subjective sleep quality.
This divergence, stable objective effects alongside symptom-severity-dependent PROM effects, has a straightforward interpretation. AHI measures airway obstruction events per hour, a physiological count. How tired or alert someone feels in daily life integrates many more variables, including sleep duration, anxiety, pain, and pre-existing fatigue unrelated to OSA. Someone already close to normal on those dimensions has less room to show measurable subjective improvement even if their AHI moves the same distance.
Where this fits in the broader SURMOUNT-OSA evidence base
A February 2026 secondary outcomes analysis of SURMOUNT-OSA in Nature Medicine documented that tirzepatide's effects on cardiometabolic risk markers, including high-sensitivity C-reactive protein (hsCRP), insulin resistance (HOMA-IR), and triglycerides, were partly mediated by improvements in sleep apnea metrics independently of weight loss.[3] That paper positioned OSA resolution as a mechanism contributing to broader metabolic benefit, not just a side-effect of weight change. The Kanu et al. post-hoc adds granularity from the patient experience side: it identifies the subset of patients for whom that metabolic benefit is also most likely to translate into felt improvement in daily functioning.
What this does not tell us
The post-hoc was restricted to Study 1 participants for the PROM analyses, so the PAP-user population in Study 2 is not covered in the same detail. The subgroup stratifications were not pre-specified, which limits the inferential weight each comparison can carry. No formal interaction tests are reported in the abstract, so the differences between fatigued and non-fatigued subgroups cannot yet be called statistically significant differences in treatment effect. And this analysis says nothing about what happens after 52 weeks, the question that most directly shapes long-term prescribing decisions.
Practical prescribing guidance belongs with a clinician who can weigh individual history, severity, and other treatment options including PAP. See the tirzepatide page for the full evidence summary.
Medical disclaimer
This article is for educational and journalistic purposes. PeptideMethods does not provide medical advice, diagnosis, or treatment recommendations. Obstructive sleep apnea is a medical condition that requires assessment and management by a qualified clinician. Do not start, stop, or change any treatment based on what you read here. Consult a healthcare professional who knows your history.
Frequently asked
What is the apnea-hypopnea index (AHI)?
AHI counts the number of complete breathing pauses (apneas) and partial breathing pauses (hypopneas) per hour of sleep. A score of 15 to 29 events per hour is classified as moderate OSA; 30 or above is severe. SURMOUNT-OSA enrolled adults with a baseline AHI of at least 15.
Does this post-hoc analysis change who should be offered tirzepatide for OSA?
No. The post-hoc is exploratory and not pre-specified. It cannot alter prescribing criteria. The FDA-approved indication is moderate-to-severe OSA in adults with obesity; eligibility for any individual is a clinical decision.
Why do objective AHI reductions and subjective symptom gains not always move together?
AHI tracks a single physiological measure, airway events per hour. Patient-reported fatigue, sleepiness, and sleep quality reflect many overlapping factors including sleep duration, mood, pain, and baseline fitness. Someone with mild fatigue at baseline may have less room to register a gain on a fatigue questionnaire even if their AHI improves by the same amount as a more-fatigued participant.
Is tirzepatide a replacement for CPAP in OSA?
Not necessarily. The SURMOUNT-OSA trials show large AHI reductions over 52 weeks, and the FDA approved Zepbound for OSA in December 2024. Whether tirzepatide is appropriate as a standalone treatment or alongside PAP therapy is a clinical question that depends on OSA severity, weight status, comorbidities, and patient preference. Speak with a sleep medicine clinician or your prescriber.
Sources
- [1]Kanu C et al. Changes in patient-reported outcomes and objective assessments based on baseline symptom severity in SURMOUNT-OSA: post-hoc analyses. Sleep. 2026 Jul 7. doi:10.1093/sleep/zsag184. PMID 42412744.Tier 1 · primary↩
- [2]Malhotra A et al. Tirzepatide for the treatment of obstructive sleep apnea and obesity. N Engl J Med. 2024;391(13):1193-1205. doi:10.1056/NEJMoa2404881. PMID 38912654.Tier 1 · primary↩
- [3]Malhotra A et al. Tirzepatide on obstructive sleep apnea-related cardiometabolic risk: secondary outcomes of the SURMOUNT-OSA randomized trial. Nat Med. 2026;32(2):653-659. doi:10.1038/s41591-025-04071-1. PMID 41540105.Tier 1 · primary↩
- [4]Hassan MG et al. Tirzepatide gains US Food and Drug Administration approval for the management of obstructive sleep apnea. J Am Dent Assoc. 2025;156(8):620-625. PMID 40569229.Tier 2 · expert↩
No revisions yet. First published .