Diagnosing Obstructive Sleep Apnea: The Glass is Half Empty, Despite What You've Been Told

Aaron B. Holley, MD


February 27, 2019

Believe it or not, the sleep medicine field still can't figure out how to diagnose obstructive sleep apnea.

The apnea-hypopnea index (AHI) used to detect obstructive sleep apnea and grade its severity is calculated by tallying apneas and hypopneas during sleep. Apneas represent complete cessation of breathing and are easy to measure. Conversely, hypopneas are "partial" breathing cessations and no one really knows what they look like.

The first scoring manual published by the American Academy of Sleep Medicine (AASM) contained three separate hypopnea definitions,[1] resulting in 3 dramatically different AHI calculations for the same patient.[2,3,4] In 2012, the AASM updated their scoring manual and recommended using one, very liberal definition. They then backtracked and added a second, more conservative definition when certain payers balked at using the first one.[5]

A Victory for the 'Liberal' Definition of Hypopnea?

A recently published paper seeks to provide clarity to this hypopnea dilemma.[6]

The authors applied both hypopnea definitions retrospectively to 1400 Veterans Health Administration patients, with data from in-lab polysomnography (PSG). Using the more liberal hypopnea definition, initially recommended by the AASM 2012 scoring manual, that includes events ending in arousal without desaturation, they identified an additional 175 patients with obstructive sleep apnea. The overwhelming majority (93.1%) of these additional patients had mild obstructive sleep apnea, and only one had severe obstructive sleep apnea.

The authors also found that only severe disease was associated with cardiovascular outcomes when the liberal 2012 AASM scoring criteria for hypopneas was applied. There was no significant relationship between mild to moderate obstructive sleep apnea and cardiovascular events.

When they used the conservative definition, in which desaturation is required to score a hypopnea, all obstructive sleep apnea (not just severe disease) was associated with cardiovascular events.

True to form, the authors of the article and those of an accompanying editorial[7] claim victory for the liberal, arousal-based hypopnea definition.

The Perspective of an Obstructive Sleep Apnea Cynic

For sleep clinicians and researchers, the obstructive sleep apnea glass is always half full. Stretching the term "phenotype" to within an inch of its overused life, they argue they have identified a novel group that can only be found using the liberal, arousal-based hypopnea definition endorsed in 2012. Therefore, they believe it's important we keep this definition. Huh?

Allow me to be the obstructive sleep apnea cynic that I am, and let's evaluate this new phenotype they have identified. These newly identified patients almost exclusively had mild obstructive sleep apnea. The field has yet to prove that mild obstructive sleep apnea is associated with symptoms or adverse outcomes,[8] and randomized controlled trials show that patients with mild disease (diagnosed using more conservative hypopnea criteria) have minimal, if any, response to treatment with continuous positive airway pressure.[9,10] Furthermore, when the authors used the liberal definition, they lost most of the obstructive sleep apnea association with cardiovascular events. With the liberal definition, only severe disease matters, and the liberal definition only identified one new patient with severe obstructive sleep apnea.

The new phenotype they've identified is not associated with adverse outcomes and does not benefit from treatment. Their addition to the group diagnosed using oxygen-based hypopnea scoring weakened the relationship between obstructive sleep apnea and cardiovascular outcomes. Ergo, the new phenotype doesn't identify disease, and it doesn't identify a group who benefits from continuous positive airway pressure.

In my opinion, arousal-based hypopnea criteria are not helpful. So why did the authors of the article and the editorial come to a different conclusion then? Consider, without arousal-based hypopnea scoring, the sleep field risks losing in-lab PSG (level I), because oxygen-based hypopneas are detected using cheaper, less labor-intensive home-tests (level III). Losing reimbursement for in-lab PSG hurts the viability of sleep medicine, and in the case of this manuscript and its editorial, self-preservation bias may be clouding reasonable interpretation.

This glass isn't half full. In fact, it's probably empty.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.