GLP-1 receptor agonists were not linked to fewer new sleep apnoea diagnoses than DPP-4 inhibitors

Design

Population-based, active-comparator, new-user cohort study using UK primary care data

Population

47,315 GLP-1 receptor agonist and 159,066 DPP-4 inhibitor initiators with type 2 diabetes and BMI of 30 or more

Primary outcome

Incident clinical diagnosis of obstructive sleep apnoea

Effect

5.8 vs 5.4 per 1,000 person-years; HR 1.07 (95% CI 0.93 to 1.23)

This UK cohort study used primary care records (CPRD, 2007 to 2023) in an active-comparator, new-user design. It included 47,315 adults with type 2 diabetes and a BMI of 30 or more who started a GLP-1 receptor agonist and 159,066 who started a DPP-4 inhibitor, excluding anyone with prior sleep apnoea. Groups were balanced with propensity score weighting.

During follow-up of up to three years there were 612 incident sleep apnoea diagnoses among GLP-1 users and 1,197 among DPP-4 users, rates of 5.8 and 5.4 per 1,000 person-years. The hazard ratio was 1.07 (95% CI 0.93 to 1.23). Results were consistent across BMI strata, sex, drug type and sensitivity analyses.

The outcome was a recorded clinical diagnosis, which depends on who is tested, and sleep apnoea often goes undiagnosed. The result does not contradict trials showing that GLP-1 receptor agonists reduce the severity of established sleep apnoea. It says that, in routine care, starting one did not lower the rate of new diagnoses compared with a weight-neutral drug.

• Do not assume that starting a GLP-1 receptor agonist makes sleep apnoea screening unnecessary.
• Keep screening patients with type 2 diabetes and obesity for sleep apnoea whatever drug they take.
• Trials of treatment effect in established sleep apnoea are a separate question and are not contradicted here.
• A recorded diagnosis depends on testing; this study cannot tell who had undiagnosed disease.

Why it matters

It tests whether the sleep apnoea benefit seen in trials extends to prevention in routine care.

Don't overread it

A diagnosis-based outcome in an observational cohort cannot show whether GLP-1 receptor agonists reduce the underlying disease.

The statistics, in plain English

A hazard ratio of 1.07 with an interval of 0.93 to 1.23 means the data are compatible with a small reduction or a small increase in diagnoses, centred on no difference.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.