GLP-1 receptor agonists and obstructive sleep apnoea: what this study means for primary care

Evidence review for UK healthcare professionals. This article summarises an observational study and current NICE OSAHS guidance. It does not replace individual assessment, product information, local pathways or prescribing judgement.

The short version

A large US real-world study found that people with obstructive sleep apnoea (OSA) who were exposed to a glucagon-like peptide-1 receptor agonist (GLP-1RA) had lower recorded hazards of ischaemic stroke, intracranial haemorrhage, emergency-department attendance, inpatient hospitalisation and all-cause mortality than matched people with OSA who were not exposed.¹

That is an important signal, but it is not proof that GLP-1RAs prevent stroke in OSA, and it does not replace standard OSA management. The study was retrospective and observational. The authors describe the findings as hypothesis-generating and call for prospective studies.¹

What did the study do?

Rai and colleagues used the TriNetX US Collaborative Network to identify adults with OSA between 1 January 2016 and 31 December 2025. Exposure was initiation of a GLP-1RA between six months before and one month after the OSA diagnosis. The study compared people exposed to a GLP-1RA with people with OSA who were not exposed.

The investigators used 1:1 propensity-score matching, leaving 438,844 people in each cohort. They assessed ischaemic stroke, intracranial haemorrhage, emergency-department visits, inpatient hospitalisations and all-cause mortality at one, three and five years. They also reported CPAP-restricted and tirzepatide-specific subgroup analyses.¹

What did it find?

Across the follow-up periods, GLP-1RA exposure was associated with lower hazards for every outcome measured. The published hazard ratios at one, three and five years were:

At one year, the source study reported ischaemic stroke in 2.7% of the GLP-1RA cohort and 3.7% of the comparison cohort. Intracranial haemorrhage was reported in 0.4% and 1.0%, respectively. These figures are observed outcomes in matched cohorts; they are not estimates of a guaranteed treatment effect for an individual patient.¹

The direction of association was reported as consistent in CPAP-restricted and tirzepatide-specific subgroup analyses. That is reassuring as a consistency check, but it does not remove the limits of an observational design.¹

Why the result needs careful interpretation

Propensity-score matching can make recorded groups more comparable. It cannot randomise treatment or fully remove differences that are not measured or not reliably coded. People prescribed GLP-1RAs may differ from comparison patients in ways that affect outcomes, including the care they receive and the conditions that led to treatment.

The study used US electronic health-record data. Its prescribing, diagnostic coding, access to sleep services and follow-up context are not the same as UK primary care. It also records exposure rather than proving medication use or defining the mechanism behind any observed association.

The clinically appropriate conclusion is therefore: the study identifies a promising association worth testing, not a new indication for GLP-1RA treatment in OSA.

What should remain unchanged in UK primary care?

NICE guideline NG202 remains the framework for recognising, investigating and managing obstructive sleep apnoea/hypopnoea syndrome (OSAHS) in adults. NICE recommends lifestyle discussion for all severities, including weight management, smoking cessation, alcohol reduction and sleep hygiene. NICE recommends CPAP, alongside lifestyle advice, for people with moderate or severe OSAHS.²

This study does not show that GLP-1RAs replace sleep-study assessment, CPAP, mandibular advancement splints where appropriate, lifestyle support or referral pathways. It should not be used on its own as a reason to start a GLP-1RA solely to reduce OSA-related cerebrovascular risk.

Where a GLP-1RA is already being considered for an established, licensed indication, this paper provides useful context for a wider cardiometabolic discussion. It does not establish that the observed benefits are caused by direct treatment of OSA.

Bottom line

This very large real-world analysis found lower recorded cerebrovascular, healthcare-utilisation and mortality outcomes among people with OSA exposed to GLP-1RAs. The findings are clinically interesting, especially given the overlap between OSA, obesity, diabetes and vascular risk.

However, the evidence is observational. For now, standard OSAHS assessment and management remain central, while prospective trials determine whether incretin-based therapies can become a validated adjunct to usual OSA care.

References

1. Rai P, Bathla G, Praveen N, et al. GLP-1 receptor agonists in obstructive sleep apnea: A propensity score-matched real-world analysis. Respiratory Medicine. 2026;261:109001. https://doi.org/10.1016/j.rmed.2026.109001
2. National Institute for Health and Care Excellence. Obstructive sleep apnoea/hypopnoea syndrome and obesity hypoventilation syndrome in over 16s (NG202). https://www.nice.org.uk/guidance/ng202