Issue Navigator

Volume 15 No. 05
Earn CME
Accepted Papers


Why a Single Index to Measure Sleep Apnea Is Not Enough

Vishesh K. Kapur, MD, MPH, FAASM1; Lucas M. Donovan, MD, MS1,2
1University of Washington School of Medicine, Seattle, Washington; 2Veterans Affairs Puget Sound Health Care System, Seattle, Washington

The apnea-hypopnea index (AHI) measures the frequency of potentially significant breathing events and plays a central role in the management of obstructive sleep apnea (OSA). Classically the AHI is compared to cutoffs to determine whether OSA is present and, if present, its level of physiologic severity. Although this dual use is common, we will argue that a single metric is not adequate for both purposes. Causing further confusion, two popular definitions of AHI exist that incorporate different criteria to define partial breathing pauses (hypopneas). The definition recommended by the American Academy of Sleep Medicine (AASM: 3% ODA) requires that a hypopnea be associated with either an electroencephalogram arousal or ≥ 3% oxygen desaturation, whereas the definition accepted by the Centers for Medicare and Medicaid Services (4%OD) is more restrictive, requiring ≥ 4% oxygen desaturation.

The systematic review and meta-analysis presented in this issue of Journal of Clinical Sleep Medicine by Mansukhani et al. assesses the effect of different AHI definitions.1 Specifically the authors focus on OSA diagnosis, severity, and clinical implications. The authors identify 33 studies that compare AHI based on differing hypopnea criteria. Based on the pooled meta-analysis of 6 studies comparing the 3%ODA criteria to 4% OD, 17% of individuals meeting a cutoff using the first criteria would be missed by the second. Not surprisingly, more inclusive definitions of hypopnea lead to higher AHI values and estimates of severity. Considering this discordance, the common practice of using uniform cutoffs (< 5, 5–15, 15–30, > 30) across hypopnea definitions is problematic. For example, an individual with mild/no disease in the 4%OD measure could have severe OSA defined by the 3%ODA definition. These discrepancies also give rise to two natural questions: which measure should clinicians focus on, and when should they choose one measure over another?

A core issue is whether hypopnea events without oxygen desaturation have clinical significance. This has been evaluated by studies not covered by the systematic review. In a noteworthy study using the Wisconsin Sleep Cohort, breathing events with > 30% decrease in nasal pressure amplitude of breaths for > 10 seconds were classified as nondesaturation events if they were associated with oxygen saturation drops of less than 2%.2 The index of nondesaturation events was associated with objective sleepiness as measured by the Multiple Sleep Latency Test. This finding supports the common clinical observation that patients who have OSA with only nondesaturating hypopneas can experience reduced sleepiness with positive airway pressure therapy.3

In contrast, there is considerable evidence that respiratory events associated with transient hypoxemia are more important in predicting cardiovascular outcomes. In the same Wisconsin Sleep Cohort study, the 3% oxygen desaturation index (ODI3%) was associated with prevalent hypertension although events without oxygen desaturation or lower magnitude of desaturation were not.2 Further, in the Sleep Heart Health Study cohort, respiratory events associated with ≥ 4% oxygen desaturation were associated with mortality; measures of sleep fragmentation were not.4

Therefore, including potentially symptomatic nondesaturating events comes at the cost of weakening a measure's relationship to cardiovascular risk. The effect of this issue is highlighted by the current diagnostic criteria for OSA.5 The International Classification of Sleep Disorders, Third Edition relies on the respiratory distress index (RDI: apneas, hypopneas, and respiratory effort-related arousals per hour of sleep) for diagnosis. Individuals with an RDI ≥ 5 who also have signs and symptoms of OSA, or an associated medical disorder, qualify as having OSA. This criterion is sensitive in identifying symptomatic individuals, as events without desaturation are incorporated in the index. Asymptomatic individuals with RDI ≥ 15 also meet criteria for OSA diagnosis. The inclusion of the second group recognizes that a significant proportion of individuals with more severe obstructive disordered breathing may be asymptomatic but may still be at risk for the development of adverse cardiovascular consequences over time.6 Unfortunately, a high RDI is not equivalent to high exposure to intermittent hypoxemia, the exposure of concern. Therefore, this criterion identifies many asymptomatic individuals with low ODI who will not suffer long-term adverse consequences from sleep-disordered breathing.

As a solution, we propose that an index closely tied to the frequency of intermittent hypoxemia (such as the 3% ODI, 4% ODI, or 4%OD) be used in conjunction with an inclusive respiratory event index, such as the current preferred AASM AHI criteria (3%ODA) or the RDI. This combination allows for a sensitive assessment of potentially symptomatic respiratory events, while also providing a direct measure of intermittent hypoxemia, a better predictor of cardiovascular outcomes. Our suggestion by no means mitigates the other limitations of the AHI as a measure of sleep-disordered breathing that have been previously discussed,7 but is a simple and unburdensome improvement that in the case of the ODI can be automatically identified by many polysomnography software packages. This minor change advances the personalization of sleep apnea care: rather than conveying the same counsel to two patients with similar AHI but differing levels of intermittent hypoxemia, a provider incorporating the ODI into the discussion of severity would more accurately convey cardiovascular risk.


All authors have seen and approve of the manuscript. The authors report no conflicts of interest. The views expressed here are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs.


Kapur VK, Donovan LM. Why a single index to measure sleep apnea is not enough. J Clin Sleep Med. 2019;15(5):683–684.



Mansukhani MP, Kolla BP, Wang Z, Morgenthaler TI. Effect of varying definitions of hypopnea on the diagnosis and clinical outcomes of sleep-disordered breathing: a systematic review and meta-analysis. J Clin Sleep Med. 2019;15(5):687–696


Koch H, Schneider LD, Finn LA, et al. Breathing disturbances without hypoxia are associated with objective sleepiness in sleep apnea. Sleep. 2017;40(11)


Guilleminault C, Stoohs R, Duncan S. Snoring (I). Daytime sleepiness in regular heavy snorers. Chest. 1991;99(1):40–48. [PubMed]


Punjabi NM, Newman AB, Young TB, Resnick HE, Sanders MH. Sleep-disordered breathing and cardiovascular disease: an outcome-based definition of hypopneas. Am J Respir Crit Care Med. 2008;177(10):1150–1155. [PubMed Central][PubMed]


Adams R, Appleton S, Taylor A, McEvoy D, Wittert G. Are the ICSD-3 criteria for sleep apnoea syndrome too inclusive? Lancet Respir Med. 2016;4(5):e19–e20. [PubMed]


Kapur VK, Baldwin CM, Resnick HE, Gottlieb DJ, Nieto FJ. Sleepiness in patients with moderate to severe sleep-disordered breathing. Sleep. 2005;28(4):472–477. [PubMed]


Punjabi NM. Counterpoint: is the apnea-hypopnea index the best way to quantify the severity of sleep-disordered breathing? No. Chest. 2016;149(1):16–19. [PubMed]