The meta-analysis included 518 unique patients undergoing 518 unique procedures. CPAP indicates continuous positive airway pressure; MMA, maxillomandibular advancement; and PSG, polysomnography.
Differences in the apnea-hypopnea index (ΔAHI) and respiratory disturbance index (ΔRDI) outcomes are shown as means (data markers) with 95% CIs (error bars) to include 2 SEs from the mean. The references are ranked in descending order of sample size. Results demonstrate a symmetric inverted funnel shape and reflect a data set for which publication bias has been minimized.
Four hundred fifty-five patients had AHI data. Mean differences are displayed for each preoperative AHI cohort. A direct linear correlation between preoperative AHI and ΔAHI is seen (R2 = 0.84; P < .001) (ΔAHI = 3.76 − [0.90 × preoperative AHI]). Patients with more severe preoperative AHI values experienced the greatest magnitude of reduction in the postoperative AHI. The mean ΔAHI of the preoperative AHI cohort with fewer than 30 events/h was −14.1 (11.6) events/h compared with a mean ΔAHI of −94.5 (23.5) events/h for the preoperative AHI cohort with 60 to fewer than 90 events/h.
eAppendix. Individual Study Quality Questionnaire for MMA Meta-analysis
eTable 1. Results of Individual Study Quality Questionnaire
eTable 2. Predictors of Post-MMA OSA Surgical Cure by AHI
eTable 3. Predictors of Post-MMA OSA Surgical Success by AHI
Customize your JAMA Network experience by selecting one or more topics from the list below.
Zaghi S, Holty JC, Certal V, et al. Maxillomandibular Advancement for Treatment of Obstructive Sleep Apnea: A Meta-analysis. JAMA Otolaryngol Head Neck Surg. 2016;142(1):58–66. doi:10.1001/jamaoto.2015.2678
Maxillomandibular advancement (MMA) is an invasive yet effective surgical option for obstructive sleep apnea (OSA) that achieves enlargement of the upper airway by physically expanding the facial skeletal framework.
To identify criteria associated with surgical outcomes of MMA using aggregated individual patient data from multiple studies.
The Cochrane Library, Scopus, Web of Science, and MEDLINE from June 1, 2014, to March 16, 2015, using the Medical Subject Heading keywords maxillomandibular advancement, orthognathic surgery, maxillary osteotomy, mandibular advancement, sleep apnea, surgical, surgery, sleep apnea syndrome, and obstructive sleep apnea.
Inclusion criteria consisted of studies in all languages of (1) adult patients who underwent MMA as treatment for OSA; (2) report of preoperative and postoperative quantitative outcomes for the apnea-hypopnea index (AHI) and/or respiratory disturbance index (RDI); and (3) report of individual patient data. Studies of patients who underwent adjunctive procedures at the time of MMA (including tonsillectomy, uvulopalatopharyngoplasty, and partial glossectomy) were excluded.
Three coauthors systematically reviewed the articles and updated the review through March 16, 2015. The PRISMA statement was followed. Data were pooled using a random-effects model and analyzed from July 1, 2014, to September 23, 2015.
Main Outcomes and Measures
The primary outcomes were changes in the AHI and RDI after MMA for each patient. Secondary outcomes included surgical success, defined as the percentage of patients with more than 50% reduction of the AHI to fewer than 20 events/h, and OSA cure, defined as a post-MMA AHI of fewer than 5 events/h.
Forty-five studies with individual data from 518 unique patients/interventions were included. Among patients for whom data were available, 197 of 268 (73.5%) had undergone prior surgery for OSA. Mean (SD) postoperative changes in the AHI and RDI after MMA were −47.8 (25.0) and −44.4 (33.0), respectively; mean (SE) reductions of AHI and RDI outcomes were 80.1% (1.8%) and 64.6% (4.0%), respectively; and 512 of 518 patients (98.8%) showed improvement. Significant improvements were also seen in the mean (SD) postoperative oxygen saturation nadir (70.1% [15.6%] to 87.0% [5.2%]; P < .001) and Epworth Sleepiness Scale score (13.5 [5.2] to 3.2 [3.2]; P < .001). Rates of surgical success and cure were 389 (85.5%) and 175 (38.5%), respectively, among 455 patients with AHI data and 44 (64.7%) and 13 (19.1%), respectively, among 68 patients with RDI data. Preoperative AHI of fewer than 60 events/h was the factor most strongly associated with the highest incidence of surgical cure. Nevertheless, patients with a preoperative AHI of more than 60 events/h experienced large and substantial net improvements despite modest surgical cure rates.
Conclusions and Relevance
Maxillomandibular advancement is an effective treatment for OSA. Most patients with high residual AHI and RDI after other unsuccessful surgical procedures for OSA are likely to benefit from MMA.
Maxillomandibular advancement (MMA) is an invasive yet potentially effective surgical option in the treatment of obstructive sleep apnea (OSA) for patients who have difficulty tolerating continuous positive airway pressure and whose OSA has been refractory to other surgical modalities.1 Maxillomandibular advancement achieves enlargement of the nasopharyngeal, retropalatal, and hypopharyngeal airway by physically expanding the facial skeletal framework via Le Fort I maxillary and sagittal split mandibular osteotomies. Advancements of the maxilla and mandible increase tension on the pharyngeal soft tissue, thereby enlarging the medial-lateral and anteroposterior dimensions of the upper airway.2 A previous meta-analysis3 demonstrated a mean decrease in the apnea-hypopnea index (AHI) from 63.9 to 9.5 events/h with a pooled surgical success rate of 86.0% and OSA cure rate of 43.2% using study-level data. Despite a large number of studies reporting excellent outcomes on the cohort level, baseline individual variables that might be associated with a highly effective outcome remain to be elucidated. Indeed, assessment of whether any preoperative factors could be consistently associated with postoperative outcomes could help to shape patient selection criteria and to counsel patients regarding their chances to achieve a significant improvement with MMA.
We performed a systematic review of the literature and meta-analysis of studies reporting individual patient data among adults who underwent MMA for the treatment of OSA. The purpose of our meta-analysis was to use aggregated individual patient-level data from a large number of studies to assess whether any baseline preoperative factors might be predictive of postoperative AHI and respiratory disturbance index (RDI) outcomes, surgical success, and/or OSA cure. Our specific aim was to elucidate factors associated with outcome effect size and the likelihood of surgical success and cure.
Three of us (V.C., J.A., and M.C.) independently performed a literature search to identify potentially relevant studies via search of the Cochrane Library, Scopus, Web of Science, and MEDLINE. These same three of us came to a consensus as to which studies met the inclusion criteria and submitted these to another one of us (S.Z.), who independently reviewed each article to ensure that they met the inclusion and exclusion criteria.
The 4 databases were searched from June 1, 2014, through March 16, 2015. The Medical Subject Heading keywords and phrases searched included maxillomandibular advancement, orthognathic surgery, maxillary osteotomy, mandibular advancement, sleep apnea, surgical, surgery, sleep apnea syndrome, and obstructive sleep apnea.
We included studies in all languages of (1) adult patients (aged >18 years) who underwent MMA (with or without genial tubercle advancement) as a treatment for OSA; (2) reports of preoperative and postoperative quantitative outcomes for the AHI and/or RDI; and (3) reports of individual patient data. We excluded studies of patients who underwent adjunctive procedures at the time of MMA (including tonsillectomy, uvulopalatopharyngoplasty, and partial glossectomy).
We screened 1280 MMA studies for potential relevance, and 117 nonduplicated articles were downloaded for detailed evaluation (Figure 1). An effort was made to include all available studies in all languages, including library requests and direct contact with the authors. Five more articles were added based on a review of references. After reviewing the full-text versions of 122 articles, a total of 45 studies were included. One of these was a randomized clinical trial.4 The other 77 articles were excluded because preoperative and/or postoperative polysomnographic data were missing (n = 21), individual patient data were not available (n = 43), MMA was combined with other surgeries (n = 4), or the articles were reviews or editorials (n = 9). The 45 included studies were written in English (n = 40), French (n = 2), German (n = 1), Dutch (n = 1), and Chinese (n = 1). A quality control questionnaire was developed to evaluate the methodologic quality of each study (eAppendix in the Supplement). Results of the questionnaire are given in eTable 1 in the Supplement.
Individual patient data from each article were abstracted into a spreadsheet (Excel 2013; Microsoft Corporation). Abstracted data included age, sex, prior OSA surgery, body mass index (BMI; calculated as weight in kilograms divided by height in meters squared), AHI, nadir of the pulse oximeter oxygen saturation level (Spo2), RDI, Epworth Sleepiness Scale score, posterior airway space, length of maxilla advancement, length of mandible advancement, and sella-nasion points A and B angles. All articles were reviewed at least 3 times to ensure accurate transposition of the data. Email correspondence was used to contact authors of included studies to acquire individual patient data for pertinent missing variables (ie, age, sex, and BMI).
The main outcome measure was the change in AHI (ΔAHI, calculated as preoperative AHI – postoperative AHI) or the change in RDI (ΔRDI, calculated as preoperative RDI – postoperative RDI) after surgery for each patient. The secondary outcome measures were rates of surgical success and OSA cure. Surgical success was defined as the percentage of patients with greater than 50% reduction of the AHI to fewer than 20 events/h after MMA; surgical cure was defined as a post-MMA AHI of fewer than 5 events/h. We selected variables agreed on by consensus5 to grade RDI severity on the same scale as AHI severity. On this scale, 0 to fewer than 5 events/h indicates normal; 5 to fewer than 15, mild sleep apnea; 15 to fewer than 30, moderate sleep apnea; and 30 or more, severe sleep apnea. If the RDI was reported, the same criteria were used for surgical success and OSA cure, respectively. Seventeen studies6-22 reported RDI data (without AHI), and, among these, 8 studies6,8-12,14,20 provided an explicit definition or reference for RDI that is consistent with AHI according to present guidelines5; these data points were corrected as AHI data.
Data were analyzed from July 1, 2014, to September 23, 2015. Meta-analysis was performed to assess for heterogeneity of the studies included and to assess the overall effect size of the MMA intervention. Heterogeneity was assessed by the following 3 methods: (1) graphic inspection of forest plots; (2) review of the I2 statistic with cutoffs of 25% (low), 50% (moderate), and 75% (high)23; and (3) review of the Cochran Q statistic23 with a heterogeneity cutoff of P ≤ .10. Data were pooled using a random-effects model.
Univariate and multivariate analyses of the individual patient data were performed to assess for preoperative factors that could predict differences in patient outcomes. Multivariate analysis was performed with a standard least squares–effect leverage model using backward elimination to select the variables for the model. The Hosmer-Lemeshow test for goodness of fit was performed to assess for adequacy of the models analyzed. Results are reported according to the PRISMA statement guidelines.24 Unless otherwise indicated, data are expressed as mean (SD).
Individual data from 518 unique patients undergoing 518 unique procedures were extracted from 45 studies, including 9 studies7,13,15-19,21,22 with RDI data alone (63 patients), 34 studies4,6,8-12,14,20,25-51 with AHI data alone (450 patients), and 2 studies52,53 with RDI and AHI data (5 patients). Of 339 patients with sex data available, 282 (83.2%) were male; the mean patient age was 45.3 (10.0) years with a mean preoperative BMI of 33.8 (9.7). The median minimum follow-up time reported by the studies was 6 months, with a range of 2 to 6 months. Table 1 summarizes pre-MMA and post-MMA characteristics.
Mean post-MMA ΔAHI was −47.8 (25.0); mean post-MMA ΔRDI, −44.4 (33.0). A negative ΔAHI or ΔRDI value represents a net decrease in the postoperative AHI or RDI outcome and characterizes improvement of OSA after surgery. Forest plots for AHI and RDI outcomes (Figure 2) show symmetric inverted funnel shapes and are consistent with minimal publication bias. We found statistically significant heterogeneity and a low to moderate level of inconsistency among the studies included in the meta-analysis for the ΔAHI (Cochran Q35 = 90.42; P < .001; I2 = 61.3% [moderate]) and the ΔRDI (Cochran Q10 = 17.04; P < .001; I2 = 41.3% [low]). A random-effects model was assumed to control for heterogeneity, and appropriateness was confirmed by analysis of variance. Graphs of study reference vs each baseline factor were inspected visually, and we found no significant outliers in the data series.
In this series, 512 of 518 patients (98.8%) experienced an improvement with respect to the primary outcome measures of ΔAHI and ΔRDI. Among patients with AHI data (n = 455), 90% experienced improvements in the AHI of at least −19 events/h; 75%, at least −30 events/h; 50%, at least −46 events/h; 25%, at least −62 events/h; and 10%, exceeding −78 events/h. Three hundred eighty-nine of the 455 patients with AHI data (85.5%) had greater than 50% reduction and AHIs of fewer than 20 events/h after MMA (surgical success); 366 of 455 (80.4%), fewer than 15 events/h; 290 of 455 (63.7%), fewer than 10 events/h; and 175 of 455 (38.5%), fewer than 5 events/h (OSA surgical cure). Similarly, among the 68 patients with RDI data, 90% experienced improvements in the RDI of at least −10 events/h; 75%, at least −18 events/h; 50%, at least −39 events/h; 25%, at least −71 events/h; and 10%, exceeding −86 events/h. Forty-four of 68 patients with RDI data (64.7%) had a greater than 50% reduction and an RDI of fewer than 20 events/h after MMA (surgical success); 41 of 68 (60.3%), fewer than 15 events/h; 27 of 68 (39.7%), fewer than 10 events/h; and 13 of 68 (19.1%), fewer than 5 events/h (OSA surgical cure).
On multivariate analysis using a standard least squares–effect leverage model with backward elimination, the following baseline preoperative factors were found to be statistically significantly associated with OSA surgical cure by AHI: age (P = .03), preoperative AHI (P < .001), and preoperative Spo2 nadir (P = .04). Patients with surgical cure were characterized as younger with a lower preoperative AHI and higher Spo2 nadir compared with patients without a surgical cure (eTable 2 in the Supplement). The factor associated with OSA surgical success by AHI was preoperative AHI (P = .02); patients who achieved surgical success were characterized by lower preoperative AHI compared with patients who did not achieve surgical success (eTable 3 in the Supplement). Preoperative AHI was the single factor consistently associated with outcome and correlated with other measures of OSA disease severity (RDI, Spo2 nadir, Epworth Sleepiness Scale score, and BMI). Other preoperative factors (including female sex, higher preoperative BMI and Epworth Sleepiness Scale score, and lower preoperative Spo2 nadir) were shown to have an association with outcome effect size only on univariate analyses.
The individual patient data were divided into the following 4 cohorts with respect to preoperative AHI: fewer than 30 events/h, 30 to fewer than 60 events/h, 60 to fewer than 90 events/h, and 90 or more events/h. Pearson χ2 analysis showed a greater likelihood of surgical success (P = .009) and surgical cure (P < .001) for patients in the lower preoperative AHI cohorts (Table 2). Patients with a higher preoperative AHI are less likely to achieve the constructs of surgical success and cure. Preoperative mean BMI was significantly higher among patients in the higher preoperative AHI cohorts at 27.2 (1.5) for fewer than 30 events/h, 29.9 (7.0) for 30 to fewer than 60 events/h; 32.8 (7.1) for 60 to fewer than 90 events/h; and 38.8 (8.2) for 90 or more events/h (P < .001, Pearson χ2 test). Preoperative Spo2 nadir was lower among patients in the higher preoperative AHI cohorts (P < .001, Pearson χ2 test). Cohorts with a higher preoperative AHI experienced a greater degree of improvement to the Spo2 nadir outcome (P = .005, Pearson χ2 test). We otherwise found no significant differences in any of the other factors among the preoperative AHI cohorts.
We found a direct linear correlation between preoperative AHI and ΔAHI (R2 = 0.84; P < .001). Patients with more severe preoperative AHI values experienced the greatest magnitude of reduction in the postoperative AHI values compared with patients with lower preoperative AHI values (Figure 3). Similar results were obtained for RDI data (R2 = 0.60; P < .001).
Genial tubercle advancement was performed in 174 of 518 patients (33.6%) at the time of MMA surgery; on multivariate analysis using the standard least squares–effect leverage model, the percentage of change of the sella-nasion point A angle was the only factor that had a statistically significant difference between the groups (mean [SD] MMA + genial tubercle advancement, 7.1% [0.4%]; MMA, 4.1% [0.5%]; P = .02). An increase in the sella-nasion point A angle is known to be a direct surgical consequence of genial tubercle advancement; no other apparent differences in outcomes existed between the MMA–genial tubercle advancement vs MMA groups.
Maxillomandibular advancement is a highly effective OSA surgical treatment that is associated with substantial improvements to AHI and RDI. Among 518 patients, 512 (98.8%) experienced improvement in outcomes (2 patients had no reported change from preoperative to postoperative AHI and RDI, and 4 patients had worse postoperative polysomnographic outcomes).6,7,25-27 The mean AHI improved from a mean (SD) preoperative value of 57.2 (25.4) to a postoperative value of 9.5 (10.4). Similarly, the mean (SD) RDI improved from a preoperative value of 65.8 (31.9) to a postoperative value of 21.4 (21.7). The mean (SE) reduction for AHI and RDI outcomes was 80.1% (1.8%) and 64.6% (4.0%), respectively. We also found significant improvements in the postoperative Spo2 nadir and Epworth Sleepiness Scale score outcomes.
The overall surgical success and cure rates for MMA as a treatment for OSA were 85.5% and 38.5%, respectively, for AHI data and 64.7% and 19.1%, respectively, for RDI data. Patients with higher preoperative OSA severity were less likely to achieve the defined constructs of surgical success and OSA cure. For example, the cure rate was only 20% among patients with a preoperative AHI of 90 or more events/h but was as high as 56% for patients with preoperative AHI of fewer than 30 events/h. However, patients with higher preoperative OSA severity were most likely to experience the greatest magnitude of improvement. The mean ΔAHI of the cohort with a preoperative AHI of fewer than 30 events/h was −14.1 (11.6) events/h compared with a mean ΔAHI of −94.5 (23.5) events/h for the cohort with a preoperative AHI of greater than 90 events/h.
Our results show that patients with a high residual RDI and AHI after failure of other surgical procedures for sleep apnea are highly likely to benefit from MMA. However, although these patients with the most severe and refractory OSA conditions experience a substantial improvement with respect to postoperative AHI and RDI, the current definitions of surgical success and OSA cure54,55 may not sufficiently represent the benefit achieved from surgery. Indeed, some authors56 have argued that surgical intervention for patients with sleep apnea is reserved only for those who cannot or will not accept continuous positive airway pressure therapy; as such, the goal of surgery is not to cure a condition that is obviously incurable but rather to offer a treatment that will help abate symptoms and minimize ongoing multisystem damage.
Surgical success is a vague and controversial entity that is not based on objective systematic data collection.57 Neuropsychological OSA symptoms58 and associated cardiovascular sequela59 may still be present with an AHI of at least 5 events/h despite a 50% reduction in AHI with treatment. Patients may complain of persistent daytime fatigue, tiredness, difficulty concentrating, and memory deficits if the OSA is not completely treated.58 Patients with residual OSA may be at risk for cardiovascular disease, heart failure, hypertension, and reduced insulin sensitivity.57 The matrix of an AHI of 5 events/h is old but was suggested based on polysomnographic recording of 200 healthy individuals in 1976.60 Many authors5,54,56,57 agree that polysomnographic measures alone are insufficient to assess the severity of OSA and response to treatment and that better matrices based on objective validated testing should be developed. The area of reporting outcomes for OSA clearly requires further investigation as demonstrated by the inadequacies of these constructs in this article.
Holty and Guilleminault3 performed a prior meta-analysis of 22 studies reporting AHI outcomes describing 627 adults undergoing MMA to treat OSA. They report a significant reduction in the mean (SD) AHI (63.9 [26.7] vs 9.5 [10.7] events/h; P < .001) at a mean follow-up of 5.3 months after MMA. The percentages of participants with a reduction greater than 50% and an AHI of fewer than 20, 15, 10, and 5 events/h after MMA were 86.0%, 77.6%, 63.4%, and 43.2%, respectively. Clinical factors associated with surgical success in the analysis by Holty and Guilleminault3 included age, preoperative BMI, and preoperative AHI. The present meta-analysis differs in that it excludes studies that do not provide individual patient data and includes studies published since 2010. Our results are similar, with a few notable differences. In the prior study, patients with higher preoperative RDI and AHI were found to have a lower chance for surgical success and cure (which we corroborate herein) but with the implication that patients with high severity of disease are worse candidates for MMA surgery. The results of the present meta-analysis qualify the prior finding by demonstrating that patients with higher preoperative RDI and AHI experienced the greatest magnitude of improvement, although they did in fact have the lowest chance of achieving the end points of surgical success and cure.3,61,62
One important limitation of this meta-analysis is that we only included studies that reported individual patient data; as such, we may have introduced selection bias by systematically excluding studies with very large sample sizes. The following are some notable studies that did not meet the inclusion criteria owing to a lack of reported individual patient data: Riley et al61 (n = 306), Prinsell63 (n = 50), Li et al8 (n = 175), and Bettega et al64 (n = 51). Results of these studies (from centers with larger sample sizes and experience) reported higher levels of surgical success than reported in this meta-analysis. For example, Li et al8 reported a 95% overall success rate with a mean change in RDI from 69.6 (27.9) to 7.7 (5.3) events/h. Prinsell63 reported a 100% success rate with a mean change in AHI from 59.2 (28.4) to 4.7 (5.9) events/h. We show a moderate level of heterogeneity for studies included in the present meta-analysis, which reflects a broad range of experiences among different surgeons and populations with respect to use of the surgical technique. However, we were also limited by the heterogeneity in that the terms of the variables reported were inconsistent between the studies.
Maxillomandibular advancement is a highly invasive surgical procedure with risks that include pain, swelling, malocclusion, poor cosmetic result, facial numbness, tingling, jaw stiffness, and postsurgical relapse of advancement. Minor hemorrhage, local infection, and extrusion of hardware have also been reported.1,3,28,62,65 Facial paresthesia due to stretching or injury to the inferior alveolar nerve is universally common (100% of patients) but has been reported to resolve in 85% to 90% of patients by 6 to 12 postoperative months.8,66 Patient perception of facial aesthetics has been generally positive after MMA; modified MMA techniques, such as using counterclockwise rotation and presurgical or postsurgical orthodontics, have been developed to prevent maxillary protrusion and to improve facial aesthetics.9 The mean (SD) duration of surgery (from tracheostomy and intermaxillary fixation to the final imaging in the operating room) according to 1 study was 6.0 (1.0) hours.4 After undergoing MMA, patients require a mean of 3.5 days of hospitalization. Most patients are able to return to their regular functional status within 2 to 10 weeks after surgery.63 Major complications are rare (approximately 1%) and are associated with being older and having a preoperative medical comorbidity.3 Because many patients with OSA undergoing MMA are obese (mean BMI, 30.2) and have compromised airways, careful postoperative care is warranted, including postoperative evaluation by nasopharyngolaryngoscopy.67 No deaths attributable to or related to MMA were identified in the literature search for this meta-analysis; however, the US surgical community is aware of 2 or 3 deaths that occurred during or immediately after MMA in the past 5 years (C.G., email communication, January 5, 2015).
Additional issues that are not addressed in this meta-analysis are the effect of ethnicity on the surgical approaches27 and the amount of minimum advancement (particularly at the level of the maxilla) that is needed to achieve long-term improvement. Surgery performed on Far Eastern Asian patients (especially Asian women) requires aesthetic concerns that are different than those for white patients, and specific modifications of MMA have been developed.29,30 Additional experience with the use of advancement measurements predefined by standardized imaging and virtual surgical planning may help to address some of these issues.52 In addition, longer follow-up is needed because recurrences of OSA have been noted at 10 to 15 years after MMA surgery based on the experiences of one of us (C.G.). One of the limitations of MMA (observed clinically in many of the recurrences) is good long-term gain in anteroposterior direction but limited gain in the lateral dimension of the pharyngeal airway. Last, additional studies reporting outcomes in morbidly obese patients are necessary because currently only 33 morbidly obese patients have been identified in this and a recent meta-analysis68 on the topic.
Maxillomandibular advancement is a highly effective treatment for OSA. Preoperative severity of OSA is the most reliable predictor of outcome effect size and the likelihood of surgical success and cure. Those patients with the most severe measures of OSA tend to benefit to the greatest degree. Patients with less severe measures of OSA experience a smaller magnitude of change in AHI or RDI postoperatively, but they have the highest chance of achieving surgical success and cure. Patients with high residual RDI and AHI scores (despite prior treatments by means of uvulopalatopharyngoplasty, partial glossectomy, and/or nasal surgery) are highly likely to benefit from management of OSA by means of MMA. Future studies will provide additional insights to help optimize patient selection for this treatment option.
Corresponding Author: Soroush Zaghi, MD, Department of Head and Neck Surgery, David Geffen School of Medicine at UCLA, 10833 LeConte Ave, Room 62-132, Center for Health Sciences, Los Angeles, CA 90095 (email@example.com).
Submitted for Publication: June 6, 2015; final revision received August 21, 2015; accepted September 23, 2015.
Published Online: November 25, 2015. doi:10.1001/jamaoto.2015.2678.
Author Contributions: Dr Zaghi had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Zaghi, Holty, Certal, Powell, Riley, Camacho.
Acquisition, analysis, or interpretation of data: Zaghi, Holty, Abdullatif, Guilleminault, Camacho.
Drafting of the manuscript: Zaghi, Abdullatif, Guilleminault, Camacho.
Critical revision of the manuscript for important intellectual content: Holty, Certal, Abdullatif, Guilleminault, Powell, Riley, Camacho.
Statistical analysis: Zaghi, Holty, Camacho.
Administrative, technical, or material support: Abdullatif, Powell.
Study supervision: Holty, Certal, Guilleminault, Powell, Riley, Camacho.
Conflict of Interest Disclosures: None reported.
Disclaimer: The views herein are the private views of the authors and do not reflect the official views of the US Department of the Army or the US Department of Defense.
Previous Presentation: This article was presented as a poster at the Sixth International Surgery, Sleep and Breathing Symposium of the International Sleep Surgery Society; October 24-25, 2014; Detroit, Michigan.