SURMOUNT-OSA: Symptoms and Readouts

A 2026 post hoc analysis of SURMOUNT-OSA found that participants with worse fatigue, sleepiness or sleep quality at baseline generally reported larger improvements with tirzepatide than participants who began with milder symptoms. Objective changes, including the apnea-hypopnea index (AHI) and body weight, were generally similar across the baseline symptom groups^[1]. This does not show that baseline symptoms caused a different response. It is an exploratory subgroup analysis of two existing randomized trials, not a new randomized comparison.

What the parent trials established

SURMOUNT-OSA comprised two 52-week, phase 3 randomized trials in adults with obesity and moderate-to-severe obstructive sleep apnea. Study 1 enrolled people who were not using positive airway pressure (PAP) at baseline. Study 2 enrolled people who were using PAP. Within each trial, participants were assigned to the maximum tolerated dose of tirzepatide or placebo. The primary outcome was change in AHI, which counts apneas and hypopneas per hour of sleep^[2].

The randomized report found an estimated treatment difference in AHI of -20.0 events per hour in Study 1 and -23.8 events per hour in Study 2 at week 52. It also reported greater improvements with tirzepatide treatment in body weight and sleep-related patient-reported outcomes. Gastrointestinal events were the most frequently reported adverse events and were mostly mild to moderate^[2]. Those placebo comparisons come from the parent randomized trials. They should not be confused with the later paper's exploratory comparison of symptom-defined subgroups.

How the post hoc groups were formed

The later analysis sorted participants by how they described four baseline domains: fatigue, daytime sleepiness, snoring and sleep quality. Each domain was divided into two groups, such as fatigued versus non-fatigued or poor versus good sleep quality. Researchers then assessed week-52 changes in questionnaires and objective measures within those groups^[1]. The original treatment assignment remained randomized, but symptom-group membership did not. Readers considering the broader tirzepatide evidence should therefore treat these splits as descriptive rather than as prespecified treatment strata.

Patient-reported measures ask whether a person feels or functions differently. They do not duplicate an objective sleep measurement. The paper examined the Functional Outcomes of Sleep Questionnaire, SF-36v2 domains and PROMIS short forms for sleep-related impairment or disturbance^[1]. A person can have substantial room to improve on one questionnaire and little room on another. That baseline headroom matters when subgroup changes are interpreted alongside the clinical evidence summarized on our tirzepatide page.

What changed with worse baseline symptoms

In Study 1, the placebo-adjusted difference on the FOSQ Activity Level subscale was 0.33 points for the fatigued subgroup and 0.05 for the non-fatigued subgroup. On SF-36v2 Vitality, the corresponding differences were 8.5 and 1.0 points. The SF-36v2 General Health differences were 6.2 and 1.4 points^[1]. The paper calls these numerically greater improvements. That wording matters: the abstract does not turn the contrasts into a validated threshold for selecting tirzepatide.

The daytime-sleepiness split followed the same direction. The placebo-adjusted PROMIS Sleep-Related Impairment difference was -7.1 points in the sleepy subgroup and -1.6 in the non-sleepy subgroup. For baseline sleep quality, PROMIS Sleep Disturbance differences were -5.8 points in the poor-quality group and -1.2 in the good-quality group^[1]. Changes were similar between the two snoring subgroups. Study 2 patterns were generally consistent with Study 1, so the result was not limited to the non-PAP trial alone.

Objective results told a different story

The post hoc paper reports that AHI, sleep apnea-specific hypoxic burden and body weight generally improved more with tirzepatide than placebo across the symptom groups. OSA remission also followed that pattern. Unlike the questionnaires, these objective changes were generally similar across baseline symptom severity groups in both studies^[1]. Baseline symptom burden therefore appeared to modify how much change participants reported more than it modified the measured physiological response.

That pattern should not be read as proof that subjective improvement is less real or that AHI tells the whole story. It shows that the instruments answer different questions. AHI records breathing events, while questionnaires capture a participant's reported experience. Clinical decisions about tirzepatide for OSA also have to account for PAP use, comorbidities and tolerability. The post hoc groups did not test a decision rule for any of those factors.

Limits and sponsor context

The analysis was designed after the parent trial framework and used baseline symptom categories rather than new random assignment. Subgroup contrasts can reflect starting scores and measurement ceiling effects. They can also be unstable when several outcomes are examined. The PubMed-indexed record describes the results as post hoc and reports numerical patterns, not a causal symptom threshold^[1]. It does not establish that a person with mild baseline fatigue will receive less objective benefit from tirzepatide.

Sponsor involvement also belongs in the interpretation. The parent randomized trials were funded by Eli Lilly^[2]. PubMed affiliation metadata for the post hoc paper lists six of its seven authors with Eli Lilly, while the remaining author has university affiliations^[1]. Industry funding does not negate the results, but it strengthens the case for reading the exploratory claims narrowly and waiting for independent replication where it becomes available.

What we don't yet know

This analysis does not show whether baseline questionnaires should influence eligibility, dose decisions or the choice between medication and PAP. It does not establish a minimum symptom score needed for benefit. The randomized evidence supports average treatment effects over 52 weeks in the enrolled populations, while the subgroup paper describes how reported and objective changes were distributed within those trials. For current US regulatory context, see the US tirzepatide page.

The practical takeaway is limited but useful. Worse baseline fatigue, sleepiness or sleep quality left more room for participants to report improvement, while AHI and weight changes were broadly consistent across symptom strata. That is a finding about trial groups, not a promise for an individual. Review the wider tirzepatide safety evidence and discuss OSA treatment with a clinician who knows your sleep study, PAP history and medical record.

Medical disclaimer: This article is for educational and journalistic purposes only and does not constitute medical advice. Peptides discussed may be classified as prescription medicines or research chemicals depending on your jurisdiction. Always consult a qualified healthcare professional before using any peptide product. PeptideMethods.com does not sell, distribute, or facilitate the sale of any peptide product.