AHRQ Review Highlights Uncertainty Regarding Sleep Apnea Definitions and Outcomes

Aaron B. Holley, MD

January 10, 2023

The Agency for Healthcare Research and Quality (AHRQ) just released a new document on long-term health outcomes in patients with obstructive sleep apnea (OSA). AHRQ is a critically important organization that helps health systems, insurers, and providers develop policies for and approaches to care delivery at the individual and population levels. Its findings tend to influence reimbursement rates across the healthcare industry. AHRQ is free of commercial biases. In short, when AHRQ publishes a summary document, it’s worth paying attention to it.

The new document is monstrous in size, so I’ll highlight what I feel is an important point, using my own academic and intellectual biases as a guide. In the past, my colleagues and I have drawn attention to sleep medicine’s failure to properly define OSA. Remarkably, no one seems to agree on this. Given its widespread recognition, how can OSA remain undefined? How can you diagnose, study, and treat a condition that can’t be identified? AHRQ’s answer to this question is simple: You can’t.

Here’s the crux of the problem. OSA is currently diagnosed by scoring partial (hypopneas) and full (apneas) “obstructions” to breathing during sleep. The number of hypopneas and apneas per hour are added together to form what’s called the apnea and hypopnea index (AHI). AHI thresholds connote the presence of OSA. For the majority of patients, hypopneas constitute the majority of the AHI (ie, in all but the most severe forms of OSA, the AHI is driven by the number of hypopneas scored). But no one, including the American Academy of Sleep Medicine (AASM), knows what a hypopnea is.

We’ve found more than 13 separate definitions for hypopnea used in the literature. In a given patient, the AHI varies widely by the definition employed. Across studies, patients with OSA are quite different from one another. Even if readers paid attention to the criteria employed to score breathing events — and in my experience they don’t — the AHRQ authors found that most studies are short on detail. The majority of papers they examined did not “explicitly report full definitions or criteria” for breathing events during sleep. Although 26 of 32 studies reviewed reported that they followed AASM scoring criteria, there was “no discernable consistency in choice of a threshold and citation of a specific AASM version.”

So, what about the AASM scoring manual? Is this the problem? Sort of. The AHRQ document chronicles the scoring manual’s evolution since 1999. This manual is the “final word” on scoring breathing events during a sleep study, but there’s been nothing final about their hypopnea criteria. They started with one definition in 1999, went to three in 2007, and then consolidated to one (different from the 1999 version) again in 2012. Unfortunately, a series of minor revisions issued after 2012 vacillated between two recognized hypopnea types. Knowing this, it’s easy to see how 26 of 32 studies could claim to be following AASM scoring criteria while simultaneously studying very different patients with “OSA.”

Despite its importance, the AASM scoring manual isn’t an evidence-based document. As the AHRQ authors note, most of the scoring rules are based on the consensus of the panel members. That being the case, the factors driving changes to scoring rules often aren’t evidence-based. The mercurial updates published since 2012 aren’t a reflection of new and compelling data; rather, they’ve been driven, at least in part, by the need to reconcile the liberal definition endorsed in 2012 with reimbursement factors.

In fairness, the AASM has a somewhat thankless job here. Good luck extracting consensus from the data void that’s historically been filled by competing interests and perverse incentives. Even when the evidence is robust, creating guidelines isn’t easy. The current reality is unfortunate, though. OSA has dominated the field for close to 20 years, yet a shared, epistemologic definition for what it is remains elusive. Even worse, clinicians possess the apocryphal belief that every patient with OSA is the same. I’ve been forthright about what I believe should be done, but mine is the minority opinion. I’d urge those who treat patients with OSA to read the AHRQ document for themselves.

Aaron B. Holley, MD, is a professor of medicine at Uniformed Services University in Bethesda, Maryland, and a pulmonary/sleep and critical care medicine physician at MedStar Washington Hospital Center in Washington, DC. He covers a wide range of topics in pulmonary, critical care, and sleep medicine.