NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Balk EM, Moorthy D, Obadan NO, et al. Diagnosis and Treatment of Obstructive Sleep Apnea in Adults [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2011 Jul. (Comparative Effectiveness Reviews, No. 32.)

Cover of Diagnosis and Treatment of Obstructive Sleep Apnea in Adults

Diagnosis and Treatment of Obstructive Sleep Apnea in Adults [Internet].

Show details

Future Research

General Recommendation

  • High dropout rates (as high as 40 percent in a matter of weeks) and relatively short followup are recurrent problems with the studies evaluated in the present review. In particular, the issue of high dropout rates bears further investigation. Is this a problem peculiar to this field? Are patients’ sleep apnea symptoms interfering with their desire to serve as research participants? Are patients with sleep apnea not sufficiently well-informed about the serious consequences of sleep apnea and therefore less motivated to find out which therapies or methods could effectively help relieve their symptoms? Are treatments so onerous that patients are refusing to continue with them? Although we recommend additional trials that delve more deeply into sleep apnea diagnoses and treatments (see below), they are likely to be of little value if these questions remain unaddressed and high dropout rates and short followup durations continue to be the standard.

Diagnostic Tests

  • The most clinically useful evaluation of prediction rules and questionnaires (to screen for or diagnose OSA) would be trials examining whether use of the tests result in improved clinical outcomes, as opposed to simple studies of test accuracy.
  • A meta-analysis of individual patient measurements with various portable monitors as compared with facility-based PSG may be the best opportunity to gain insights on the relative contribution of different neurophysiologic signals used by portable monitors. Such a meta-analysis would not provide direct insight on the effects of testing on patients; however, it would provide valuable information for medical device developers and sleep physicians.
  • Future studies of the accuracy or bias of diagnostic tests should focus more on head-to-head comparisons of portable monitors, questionnaires, and prediction rules to determine the optimal tool for use in a primary care setting to maximize initial evaluation of OSA and triage high-risk patients for prompt PSG. It is our conclusion that the field would be better served by comparing the existing array of diagnostic tools with each other to try to determine which is most useful for screening or diagnosing, rather than evaluating new devices, questionnaires, or models. It is unlikely that new tests will be sufficiently more accurate or less biased to warrant the expended resources or effort.
  • The concept of phased testing, with the goals of maximizing efficiency to OSA diagnosis while minimizing overtesting, is appealing, but has not been properly evaluated by any study. Randomized trials comparing potential phased testing strategies with direct PSG (or portable) testing are needed. Similarly, it would greatly inform decisionmaking to have studies that evaluate which tests would be most appropriate to use for which patients based on the type and severity of their symptoms.
  • In comparing different OSA diagnostic devices, it is important that the findings from all participants – regardless of whether the results were above or below the particular device’s diagnostic threshold – be verified against an accepted reference standard such as laboratory PSG. Diagnostic test studies that systematically fail to fully test all participants are generally impossible to meaningfully interpret and have little scientific or clinical value.
  • No trials have addressed the value of routine (or selected) preoperative screening for OSA. Well-conducted trials are needed.


  • Of the 172 studies of treatments covered in this report, only three studies reported clinical outcomes (not including adverse events): one on heart failure with CPAP126 and two on mortality (surgery versus CPAP).257,260 We need comparative studies focusing on clinical outcomes like mortality, cardiovascular disease, hypertension, and type 2 diabetes. To be of value, studies must have long-term followup durations (measured in years) and must be adequately powered to detect statistically significant differences.
  • Primary studies on the modifying effects of different patient characteristics, baseline disease severity, and other relevant parameters on various treatment outcomes should be undertaken so that treatment options can be optimized for or can be focused on patients with specific profiles, thus maximizing treatment benefits. Studies should be large enough, of sufficient duration, and bear minimal loss-to-followup rates to allow meaningful subgroup or regression analyses.
  • Fixed CPAP is clearly an effective treatment to minimize AHI and to improve sleepiness symptoms. With the exception of the evaluation of long-term clinical outcomes, no further trials are needed to compare fixed CPAP with no treatment or with sham CPAP. Given the large effects of CPAP on AHI across all trials, it does not appear to be necessary to determine which subgroups of patients may have relatively larger or smaller benefits from CPAP. Again, the major exception would be for major clinical outcomes. Trials are needed to assess which groups of patients would experience the most long-term clinical benefits from likely life-long CPAP treatment. Probably, the most important subgroups to examine are those based on severity of OSA (as measured by AHI) and those patients with comorbidities.
  • Given the effectiveness of fixed CPAP, all other interventions should either be
    • Directly compared with fixed CPAP, among patients naïve to CPAP, or
    • Compared with no treatment or alternative treatment, among patients who have failed to comply with CPAP treatment.
  • Mandibular advancement devices also improve AHI and sleepiness symptoms, though their effects on long-term clinical outcomes are unknown (from trial data). While future trials on new variations of these devices are inevitable, they are of secondary interest.
  • More trials are needed to determine if different degrees of mandibular advancement would offer corresponding degrees of sleep apnea symptom improvement.
  • Studies of interventions other than fixed CPAP should restrict their analyses to groups of patients who are either naïve to CPAP treatment or who have failed (or refused) CPAP treatment. Alternatively, trials should perform sufficiently powered subgroup analyses based on these groups of patients. It is reasonable to assume that patients who have not used CPAP will respond differently to treatments than patients who have tried but stopped using CPAP. Thus trials performed in one group might not be applicable to patients in the other group.
  • To understand better treatment options for patients who do not tolerate fixed CPAP, head-to-head comparisons of alternative treatments, including simple adjunctive treatment (like humidification or nasal spray), nonfixed CPAP, MAD, surgery, as well as others, are needed to identify those that are most likely to benefit from these alternative treatments.
  • Rigorously conducted head-to-head comparisons of surgical interventions and CPAP are necessary to determine the relative merits of these two forms of treatments as almost all the existing data are from cohort studies, often retrospective, with all their attendant limitations. The trials should clearly identify the criteria used to consider patients for surgery and then investigate surgery versus CPAP in those patients. Trials should not mix in patients who are not seriously considering surgical treatment.
  • Many interventions (including drugs, specific oropharyngeal exercises, etc.) have been evaluated by only one or two small studies. Further, more well-conducted studies are needed per intervention to accurately determine the effects of these therapies on sleep apnea symptoms and clinical outcomes.
    • The incremental benefit of weight loss programs in addition to accepted treatments for OSA (e.g., CPAP) should be examined.
    • Suction tongue-retaining devices may be associated with better outcomes than nonsuction tongue-retaining devices, although more studies are needed to confirm this finding.
  • Future trials are needed to find effective treatments to improve compliance with both CPAP and MAD. Some specific forms of intensive patient education, cognitive behavioral therapy, telemonitoring, and habit-promoting audio-based interventions hold promise but need further investigation. Ideally, these (and other, superior interventions) should be tested against each other to enable determination of which are the most effective and should be further pursued.

Predictors of Clinical Outcomes and Compliance

  • Regarding the question of whether OSA severity (as measured by AHI) is associated with long-term outcomes, it may be of interest to perform patient-level meta-analyses of the Sleep Heart Health Study, the Wisconsin Sleep Cohort Study, and other cohorts of individuals who had sleep testing. All long-term clinical outcomes except all-cause mortality, have yet to be adequately evaluated.
  • Studies are needed to assess whether and when patients should be treated with CPAP, MAD, or surgery. This additional research should involve an assessment of how well patients tolerate each of these three treatment modalities. Further high quality studies are also necessary to determine which factors predict compliance with CPAP and MAD, and which predict successful outcomes following surgery.
PubReader format: click here to try


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...