Purpose: To systematically review the evidence on screening and treating asymptomatic adults or those with unrecognized symptoms for obstructive sleep apnea (OSA).
Data Sources: PubMed/MEDLINE, the Cochrane Library, EMBASE, and trial registries through October 2015; reference lists of retrieved articles; outside experts; and reviewers, with surveillance of the literature through October 5, 2016.
Study Selection: Two investigators independently selected English-language studies using a priori criteria. Eligible studies included randomized, controlled trials (RCTs) of screening for or treatment of OSA, studies evaluating accuracy of screening questionnaires or clinical prediction tools in asymptomatic adults or persons with unrecognized symptoms of OSA, systematic reviews (and studies published after eligible systematic reviews) evaluating diagnostic accuracy or reliability of portable monitors (PMs), and prospective cohort studies (≥1 year) evaluating the association between apnea-hypopnea index (AHI) and health outcomes among community-based participants that adjusted for potential confounding through multivariable analyses.
Data Extraction: One investigator extracted data and a second checked accuracy. Two reviewers independently rated quality for all included studies using predefined criteria.
Data Synthesis: We included 110 studies. No RCTs compared screening with no screening. The only screening approach for which we found two eligible studies reporting accuracy was the Multivariable Apnea Prediction (MVAP) score followed by home PM testing; for detecting severe OSA syndrome (OSAS) (AHI ≥30 and Epworth Sleepiness Scale [ESS] score >10), areas under the curve were 0.799 (95% confidence interval [CI], 0.777 to 0.822) and 0.833 (95% CI, 0.765 to 0.902). However, both studies oversampled high-risk participants and those with OSA and OSAS. Studies reporting accuracy of PMs for diagnostic testing of persons with suspected OSA found wide ranges for sensitivity and specificity (Type II monitors: 85% to 94% and 77% to 95%; Type III monitors: 49% to 92% and 79% to 95%; Type IV monitors: 7% to 100% and 15% to 100%, respectively, for polysomnography AHI ≥15). Data were limited by imprecision and inconsistency for Type IV monitors. We found sparse data on reliability of PMs.
Our meta-analyses of RCTs found that continuous positive airway pressure (CPAP) effectively reduced AHI to normal or near-normal levels (weighted mean difference [WMD], -33.8 [95% CI, -42.0 to -25.6]; 13 trials; 543 participants), reduced excessive sleepiness as measured by the ESS (WMD, -2.0 [95% CI, -2.6 to -1.4]; 22 trials; 2,721 participants), reduced diurnal systolic blood pressure (WMD, -2.4 [95% CI, -3.9 to -0.9]; 15 trials; 1,190 participants), and reduced diurnal diastolic blood pressure (WMD, -1.3 [95% CI, -2.2 to -0.4]; 15 trials; 1,190 participants) compared with sham. Trial evidence for most health outcomes was too limited to make conclusions (e.g., mortality, cardiovascular events, motor vehicle accidents). However, our meta-analysis for sleep-related quality of life found a significant benefit for CPAP, albeit with a small effect size (Cohen's d, 0.28 [95% CI, 0.14 to 0.42]; 13 trials; 2,325 participants). The effect size was slightly greater among those with excessive sleepiness at baseline but still small (0.33 [95% CI, 0.17 to 0.50]). Mandibular advancement devices (MADs) and weight loss programs also reduced AHI and excessive sleepiness; effect sizes were generally smaller than those for CPAP. Reporting of harms was suboptimal. Common adverse effects of CPAP included oral or nasal dryness, eye or skin irritation, rash, epistaxis, and pain; common adverse effects of MADs included oral dryness, excess salivation, mucosal erosions, or pain (mucosal, dental, or jaw).
Consistent evidence from prospective cohort studies supports the association between AHI and all-cause mortality; persons with severe OSA die at about twice the rate of controls (pooled hazard ratio [HR], 2.07 [95% CI, 1.48 to 2.91]; 5 studies; 11,003 participants). Risk of cardiovascular mortality was also increased (HRs from 2.9 [95% CI, 1.1 to 7.3] to 5.9 [95% CI, 2.6 to 13.3]).
Limitations: Data on screening accuracy for the MVAP followed by home PM testing were limited by risk of spectrum bias, which may substantially overestimate the accuracy that would be achieved in the general population of asymptomatic adults (or those with unrecognized symptoms). We found no studies that prospectively evaluated screening questionnaires or clinical prediction tools to report calibration or clinical utility for improving health outcomes. Treatment studies did not focus on screen-detected, asymptomatic patients (or those with unrecognized symptoms). Reporting on harms was scant; no studies evaluated overdiagnosis, overtreatment, or psychosocial harms (e.g., anxiety, labeling).
Conclusions: There is uncertainty about the clinical utility of all potential screening tools. Although screening with MVAP followed by home PM testing may have promise for distinguishing persons in the general population who are more or less likely to have OSA, current evidence is limited. Multiple treatments for OSA reduce AHI, ESS, and blood pressure. Although good evidence has established that persons with severe OSA die at twice the rate of controls, trials of CPAP and other treatments have not established whether treatment reduces mortality or improves most other health outcomes, barring evidence of some possible benefit for sleep-related quality of life.