0
Editorials: Point and Counterpoint |

COUNTERPOINT: Is the Apnea-Hypopnea Index the Best Way to Quantify the Severity of Sleep-Disordered Breathing? No FREE TO VIEW

Naresh M. Punjabi, MD, PhD, FCCP
Author and Funding Information

FINANCIAL/NONFINANCIAL DISCLOSURES: N. M. P. has received grant support from ResMed and Philips Respironics.

FUNDING/SUPPORT: This work was supported by National Institutes of Health [Grant HL07578].

CORRESPONDENCE TO: Naresh M. Punjabi, MD, PhD, FCCP, Division of Pulmonary and Critical Care Medicine, Johns Hopkins Asthma and Allergy Center, 5501 Hopkins Bayview Cir, Baltimore, MD 21224


Copyright 2016, American College of Chest Physicians. All Rights Reserved.


Chest. 2016;149(1):16-19. doi:10.1378/chest.14-2261
Text Size: A A A
Published online

Since the early clinical descriptions of OSA in the 1970s, our understanding of the pathogenesis and adverse consequences of this chronic disease has advanced substantially. Initially, it was recognized that apneic events fragmented sleep, induced cardiovascular instability, and led to excessive daytime sleepiness., Given the hemodynamic and sleep-related effects of obstructive apneas, it comes as no surprise that the apnea index, which tallies the number of apneas per hour of sleep, became the primary disease-defining metric for OSA. Over time, as the full spectrum of upper-airway collapse during sleep became more apparent, the simple concept of only quantifying apneas quickly evolved into something more complex. It is now obvious that obstructive apneas, the original sine qua non for the disease, are not the only events of interest because obstructive hypopneas can also have similar effects (eg, arousals, BP swings)., As the clinical impact of hypopneas became widely recognized, these events were incorporated into quantifying disease activity, and the original apnea index gave way to the now commonly used apnea-hypopnea index (AHI). Although defining hypopneas continues to be plagued with many challenges,,, the AHI has become a ubiquitous measure in sleep medicine. In fact, the AHI not only is used to diagnose OSA but also is central to assessing disease severity.

Since the initial formulation of the AHI, a large body of clinical and epidemiologic evidence has accumulated linking OSA severity to numerous clinical outcomes, including daytime sleepiness, impaired quality of life, motor vehicle accidents, incident hypertension, myocardial infarction, stroke, heart failure, diabetes, and all-cause mortality.,,, In most, if not all, available studies on the health significance of OSA to date, the AHI has been used as the primary exposure or independent variable and has been correlated with specific outcomes of interest. Moreover, interventional studies, which have examined the effects of positive airway pressure therapy, have shown that treatment of varying degrees of sleep apnea, as indexed by the AHI, is associated with favorable effects. Because of the consistency across studies in correlating the AHI with clinical sequelae, a strong foundation has formed supporting its use in characterizing OSA. Indeed, advocates of the AHI often cite the considerable evidence base of clinical outcomes to suggest that the AHI has criterion validity generalizable to several outcomes. Although the argument that AHI is a valid measure of disease activity because it predicts the presence or development of clinical outcomes is reasonable, it is not without limitations and should be, at best, considered a crude and imprecise metric of OSA.

Several lines of reasoning exist to refute the use of the AHI as the “holy grail” for assessing OSA. The first and perhaps simplest argument is that calculation of the AHI assumes that apneas and hypopneas are fundamentally equal in their biologic effects. Although evidence to date is limited on the relative contribution of apneas vs hypopneas in mediating clinical outcomes, it is easily argued that because apneas occur as a consequence of the complete upper-airway collapse, they may impose a greater pathophysiologic impact (ie, larger surges in sympathetic activity) than hypopneas which result from partial collapse of the upper airway. The lack of evidence examining the relative clinical impact of apneas and hypopneas does not necessarily obviate the concern that apneas and hypopneas are not pathophysiologically equal.

Second, it is now generally accepted that the clinical impact of OSA is partly attributable to the cyclical hypoxemia associated with apneas and hypopneas. Although the hypopnea part of the AHI certainly incorporates a component of oxygen desaturation, the threshold for the desaturation is arbitrary, and little evidence supports a specific cut point (ie, 3%, 4%)., The choice of a desaturation threshold notwithstanding, an inherent problem with any threshold chosen is that it assumes that all hypopnea events that exceed the chosen threshold are equivalent; that is, a hypopnea with a 4% desaturation is considered biologically equivalent to a hypopnea that has an 8% or a 10% oxygen desaturation. Certainly, greater degrees of hypoxemia are likely to have greater effects, and as such, the AHI disregards a crucial biologic mechanism through which OSA may arbitrate its clinical impact.

Third, computation of the AHI also neglects the temporal distribution of the events. Apneas and hypopneas can occur evenly across the night, or they can cluster during a particular segment of the sleep period. Events that temporally cluster may not have as large an influence as events that are distributed throughout the night. For example, consider two patients who have a total of 160 disordered breathing events over 8 h of sleep. Although the AHI is exactly the same (approximately 20/h), events in the first patient are clustered within the initial 2 h of sleep, whereas in the second patient, they are spread equally throughout the night such that the instantaneous rate at any time is 20/h. It could be easily argued that the second patient may be more prone to the clinical effects of OSA because disordered breathing events disrupt the entire sleep period. In contrast, the first patient may not be as prone to the health-related effects of OSA because the concentrated exposure to disordered breathing events, although disruptive for the first 2 h of the sleep period, still leaves the patient with a sufficient period of unperturbed sleep. Although variants of AHI, such as sleep stage-dependent AHI (non-rapid eye movement-AHI and rapid eye movement-AHI) and position-dependent AHI (supine vs nonsupine), are used to capture the heterogeneity of sleep physiology, they do not entirely remedy the problem of temporal distribution of events.

A fourth and relevant consideration in assessing the utility of the AHI relates to controversy that surrounds the definition of a hypopnea. Substantial disagreements remain regarding the degree of hypoventilation (or decrease in flow), severity of desaturation, and incorporation of arousals in defining a hypopnea.,, Thus, can we really consider the AHI to be the ultimate criterion for characterizing OSA if even the experts who use it cannot agree on an acceptable definition for a hypopnea? The lack of consensus is highly visible in the numerous hypopnea definitions that have been proposed and the rate at which these definitions have been changed just within the past few years. The final concern with the AHI as a metric for OSA is that it is fraught with errors of omission. The AHI, which tabulates the number of apneas and hypopneas, does not incorporate any information about the duration of these events. It is inconceivable that an event lasting 10 s is physiologically equivalent to an event lasting 2 or 3 min. Figure 1 is a 10-min sleep recording in a patient with OSA. The AHI in this patient would be approximately 18/h, which certainly does not fully reflect that in this patient, the events are long and associated with significant hypoxemia. Thus, omission of event duration information drills yet another hole in the armor of the AHI as a disease-defining metric for OSA. Another error of omission in the AHI is that it does not incorporate the increased work of breathing during sleep associated with the occurrence of disordered breathing events. Complete or partial collapse of the upper airway increases the work of breathing, which in turn results in acute cardiovascular effects (ie, increased afterload) that may be of clinical significance. Operationalizing any measure of disease presence or severity for OSA should capture the pathophysiologic diversity of the disease process. The AHI falls far from this goal and is disappointing as a disease metric because it only provides a rate at which events occur during sleep.

Figure Jump LinkFigure 1 Illustrative tracing from a patient with prolonged hypopneas. SpO2 = oxygen saturation as measured by pulse oximetry.Grahic Jump Location

Much of the foregoing discussion assumes that calculating a rate (apneas and hypopneas per hour of sleep) is an appropriate approach to characterize the severity of OSA. Shahar recently argued that the use of the AHI as an exposure variable is rarely justified because it does not measure what it purports to represent. An argument is made that the frequency of disordered breathing events should be used as the exposure and not the rate at which these events occur (ie, AHI). By calculating the AHI, bias is introduced that can be avoided by the use of alternative variables. Shahar provided an illustrative example in which two patients have the same AHI but one sleeps 7 h each night and the other sleeps 5 h. Given that the AHI is similar in these two patients, the conclusion is that the exposure level based on the AHI is comparable, which in fact is not true given that the cumulative number of events experienced by these patients is different because of differences in their habitual sleep time.

After decades of research, it is now accepted that OSA is associated with major cardiovascular and noncardiovascular end points. Given the wealth of accumulated evidence in the five decades after the initial clinical description of OSA, we should have a robust disease-defining measure. Unfortunately, that is not the case. Although the AHI has generally served the sleep medicine community well, it is not without several major limitations. It is surprising that despite the amount of data collected during an overnight sleep study, the field has settled on one number (a rate) to characterize disease severity. Without doubt, using the AHI is a gross oversimplification of a complex disease phenomenon. When other disease states are examined, such oversimplification is not the case. Take the example of hypertension. BP usually is clinically represented by not one, but several metrics (eg, systolic, diastolic, average). The field of sleep disorders medicine should not settle for taking a series of time-varying signals from the polysomnogram that often exceed 1 GB of digital data and just use one metric to summarize all of the key features. Certainly, a one-size-fits-all solution will never exist that can embody the physiologic diversity of apneas and hypopneas. It is time for the field of sleep medicine to accept that the AHI is a parameter that crudely integrates the pathophysiology of OSA. It is likely that several different parameters are needed to reflect disease complexity (eg, average desaturation amount, event length, frequency of events, AHI). It does not have to be just one measure. Five decades have lapsed since the original descriptions of OSA, and the understanding of the pathophysiology and clinical outcomes has improved greatly. However, our ability to characterize OSA severity has been stagnant and is in need of change.

References

Jordan A.S. .McSharry D.G. .Malhotra A. . Adult obstructive sleep apnoea. Lancet. 2014;383:736-747 [PubMed]journal. [CrossRef] [PubMed]
 
Lugaresi E. .Coccagna G. .Mantovani M. .Cirignotta F. .Ambrosetto G. .Baturic P. . Hypersomnia with periodic breathing: periodic apneas and alveolar hypoventilation during sleep. Bull Physiopathol Respir (Nancy). 1972;8:1103-1113 [PubMed]journal. [PubMed]
 
Guilleminault C. .Tilkian A. .Dement W.C. . The sleep apnea syndromes. Annu Rev Med. 1976;27:465-484 [PubMed]journal. [CrossRef] [PubMed]
 
Hudgel D.W. . “Apnea index”: need for improving the description of respiratory variability during sleep. Am Rev Respir Dis. 1986;133:708-709 [PubMed]journal. [PubMed]
 
Gould G.A. .Whyte K.F. .Rhind G.B. .et al The sleep hypopnea syndrome. Am Rev Respir Dis. 1988;137:895-898 [PubMed]journal. [CrossRef] [PubMed]
 
Moser N.J. .Phillips B.A. .Berry D.T. .Harbison L. . What is hypopnea, anyway? Chest. 1994;105:426-428 [PubMed]journal. [CrossRef] [PubMed]
 
Redline S. .Sanders M. . Hypopnea, a floating metric: implications for prevalence, morbidity estimates, and case finding. Sleep. 1997;20:1209-1217 [PubMed]journal. [PubMed]
 
Ruehland W.R. .Rochford P.D. .O’Donoghue F.J. .Pierce R.J. .Singh P. .Thornton A.T. . The new AASM criteria for scoring hypopneas: impact on the apnea hypopnea index. Sleep. 2009;32:150-157 [PubMed]journal. [PubMed]
 
Vijayan V.K. . Morbidities associated with obstructive sleep apnea. Expert Rev Respir Med. 2012;6:557-566 [PubMed]journal. [CrossRef] [PubMed]
 
Mannarino M.R. .Di Filippo F. .Pirro M. . Obstructive sleep apnea syndrome. Eur J Intern Med. 2012;23:586-593 [PubMed]journal. [CrossRef] [PubMed]
 
Usmani Z.A. .Chai-Coetzer C.L. .Antic N.A. .McEvoy R.D. . Obstructive sleep apnoea in adults. Postgrad Med J. 2013;89:148-156 [PubMed]journal. [CrossRef] [PubMed]
 
Heatley E.M. .Harris M. .Battersby M. .McEvoy R.D. .Chai-Coetzer C.L. .Antic N.A. . Obstructive sleep apnoea in adults: a common chronic condition in need of a comprehensive chronic condition management approach. Sleep Med Rev. 2013;17:349-355 [PubMed]journal. [CrossRef] [PubMed]
 
Giles T.L. .Lasserson T.J. .Smith B.J. .White J. .Wright J. .Cates C.J. . Continuous positive airways pressure for obstructive sleep apnoea in adults. Cochrane Database Syst Rev. 2006;:CD001106- [PubMed]journal
 
Punjabi N.M. .Newman A.B. .Young T.B. .Resnick H.E. .Sanders M.H. . Sleep-disordered breathing and cardiovascular disease: an outcome-based definition of hypopneas. Am J Respir Crit Care Med. 2008;177:1150-1155 [PubMed]journal. [CrossRef] [PubMed]
 
Stamatakis K. .Sanders M.H. .Caffo B. .et al Fasting glycemia in sleep disordered breathing: lowering the threshold on oxyhemoglobin desaturation. Sleep. 2008;31:1018-1024 [PubMed]journal. [PubMed]
 
Shahar E. . Apnea-hypopnea index: time to wake up. Nat Sci Sleep. 2014;6:51-56 [PubMed]journal. [PubMed]
 

Figures

Figure Jump LinkFigure 1 Illustrative tracing from a patient with prolonged hypopneas. SpO2 = oxygen saturation as measured by pulse oximetry.Grahic Jump Location

Tables

References

Jordan A.S. .McSharry D.G. .Malhotra A. . Adult obstructive sleep apnoea. Lancet. 2014;383:736-747 [PubMed]journal. [CrossRef] [PubMed]
 
Lugaresi E. .Coccagna G. .Mantovani M. .Cirignotta F. .Ambrosetto G. .Baturic P. . Hypersomnia with periodic breathing: periodic apneas and alveolar hypoventilation during sleep. Bull Physiopathol Respir (Nancy). 1972;8:1103-1113 [PubMed]journal. [PubMed]
 
Guilleminault C. .Tilkian A. .Dement W.C. . The sleep apnea syndromes. Annu Rev Med. 1976;27:465-484 [PubMed]journal. [CrossRef] [PubMed]
 
Hudgel D.W. . “Apnea index”: need for improving the description of respiratory variability during sleep. Am Rev Respir Dis. 1986;133:708-709 [PubMed]journal. [PubMed]
 
Gould G.A. .Whyte K.F. .Rhind G.B. .et al The sleep hypopnea syndrome. Am Rev Respir Dis. 1988;137:895-898 [PubMed]journal. [CrossRef] [PubMed]
 
Moser N.J. .Phillips B.A. .Berry D.T. .Harbison L. . What is hypopnea, anyway? Chest. 1994;105:426-428 [PubMed]journal. [CrossRef] [PubMed]
 
Redline S. .Sanders M. . Hypopnea, a floating metric: implications for prevalence, morbidity estimates, and case finding. Sleep. 1997;20:1209-1217 [PubMed]journal. [PubMed]
 
Ruehland W.R. .Rochford P.D. .O’Donoghue F.J. .Pierce R.J. .Singh P. .Thornton A.T. . The new AASM criteria for scoring hypopneas: impact on the apnea hypopnea index. Sleep. 2009;32:150-157 [PubMed]journal. [PubMed]
 
Vijayan V.K. . Morbidities associated with obstructive sleep apnea. Expert Rev Respir Med. 2012;6:557-566 [PubMed]journal. [CrossRef] [PubMed]
 
Mannarino M.R. .Di Filippo F. .Pirro M. . Obstructive sleep apnea syndrome. Eur J Intern Med. 2012;23:586-593 [PubMed]journal. [CrossRef] [PubMed]
 
Usmani Z.A. .Chai-Coetzer C.L. .Antic N.A. .McEvoy R.D. . Obstructive sleep apnoea in adults. Postgrad Med J. 2013;89:148-156 [PubMed]journal. [CrossRef] [PubMed]
 
Heatley E.M. .Harris M. .Battersby M. .McEvoy R.D. .Chai-Coetzer C.L. .Antic N.A. . Obstructive sleep apnoea in adults: a common chronic condition in need of a comprehensive chronic condition management approach. Sleep Med Rev. 2013;17:349-355 [PubMed]journal. [CrossRef] [PubMed]
 
Giles T.L. .Lasserson T.J. .Smith B.J. .White J. .Wright J. .Cates C.J. . Continuous positive airways pressure for obstructive sleep apnoea in adults. Cochrane Database Syst Rev. 2006;:CD001106- [PubMed]journal
 
Punjabi N.M. .Newman A.B. .Young T.B. .Resnick H.E. .Sanders M.H. . Sleep-disordered breathing and cardiovascular disease: an outcome-based definition of hypopneas. Am J Respir Crit Care Med. 2008;177:1150-1155 [PubMed]journal. [CrossRef] [PubMed]
 
Stamatakis K. .Sanders M.H. .Caffo B. .et al Fasting glycemia in sleep disordered breathing: lowering the threshold on oxyhemoglobin desaturation. Sleep. 2008;31:1018-1024 [PubMed]journal. [PubMed]
 
Shahar E. . Apnea-hypopnea index: time to wake up. Nat Sci Sleep. 2014;6:51-56 [PubMed]journal. [PubMed]
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
PubMed Articles
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543