Is there good evidence of mortality benefits from positive airway pressure?

Authors’ reply

Obstructive sleep apnoea (OSA) is an important condition,¹ but its role as a reversible cardiovascular risk factor has been questioned. Randomised controlled trials (RCTs) of positive airway pressure (PAP) therapy have yielded inconclusive results, with limitations including inadequate sample sizes, short follow-up durations, low numbers of endpoint events—especially death—poor adherence to therapy, and enrolment of patients with minimal symptoms. Furthermore, only a minority of sleep clinic patients diagnosed with OSA meet the eligibility criteria of these RCTs, potentially limiting generalisability to clinical patients.² In contrast, non-randomised studies in OSA have larger sample sizes of diverse, symptomatic clinical patients with long-term follow-up and thousands of death events. However, these studies are susceptible to healthy user effects and residual confounding; issues that are extensively discussed in our meta-analysis.³

Several points raised in the letter by Craig L Phillips and colleagues are addressed in our limitations section, including differences between RCTs and confounder-adjusted non-randomised controlled studies (NRCSs), known and unknown confounders, the healthy user effect, and results being driven by NRCSs. Many publications on the effects of PAP also have issues with residual confounding and the healthy user effect, as well as multiple comparisons based on non-prespecified subgroup analyses.^4,5 We were transparent in our rationale for inclusion of both randomised and confounder-adjusted non-randomised studies, recognising that both have limitations, but are complementary. We are grateful for thorough peer reviewers whose tough critiques helped to yield a final Article that responsibly conveys these important findings.

Contrary to the letter’s assertions, we believe that our new findings help to restore equipoise regarding the long-term risks and benefits of PAP therapy for preventing cardiovascular disease in patients with OSA. We encourage ongoing rigorous OSA research using complementary research methodologies to draw meaningful conclusions and develop new therapeutic approaches. In the absence of conclusive RCTs, physicians should consider the best available evidence to optimise shared clinical decision making with their patients. We believe our findings contribute valuable evidence while actively encouraging continued rigorous investigation through well-designed clinical trials.