Alt JA, DeConde AS, Mace JC, Steele TO, Orlandi RR, Smith TL. Quality of Life in Patients With Chronic Rhinosinusitis and Sleep Dysfunction Undergoing Endoscopic Sinus SurgeryA Pilot Investigation of Comorbid Obstructive Sleep Apnea. JAMA Otolaryngol Head Neck Surg. 2015;141(10):873-881. doi:10.1001/jamaoto.2015.1673
Patients with chronic rhinosinusitis (CRS) have reduced sleep quality linked to their overall well-being and disease-specific quality of life (QOL). Other primary sleep disorders also affect QOL.
To determine the impact of comorbid obstructive sleep apnea (OSA) on CRS disease-specific QOL and sleep dysfunction in patients with CRS following functional endoscopic sinus surgery.
Design, Setting, and Participants
Prospective multisite cohort study conducted between October 2011 and November 2014 at academic, tertiary referral centers with a population-based sample of 405 adults.
Functional endoscopic sinus surgery for medically refractory symptoms of CRS.
Main Outcomes and Measures
Primary outcome measures consisted of preoperative and postoperative scores operationalized by the Rhinosinusitis Disability Index (RSDI) survey, the 22-item Sinonasal Outcome Test (SNOT-22), and the Pittsburgh Sleep Quality Index (PSQI). Obstructive sleep apnea was the primary, independent risk factor.
Of 405 participants, 60 (15%) had comorbid OSA. A total of 285 (70%) participants provided preoperative and postoperative survey responses, with a mean (SD) of 13.7 (5.3) months of follow-up. Significant postoperative improvement (P < .05) was reported across all mean disease-specific QOL measures for both participants with and without comorbid OSA. Participants without OSA reported significant greater improvement in unadjusted mean (SD) RSDI global scores (−25.0 [23.3] vs −16.5 [22.1]; P = .03), RSDI physical (−10.7 [9.2] vs −7.3 [9.1]; P = .03) and functional (−8.4 [8.7] vs −5.1 [7.5]; P = .03) subdomain scores, and SNOT-22 rhinologic symptom domain scores (−9.1 [7.7] vs −5.7 [6.9]; P = .008). Participants without OSA also reported greater improvements on mean (SD) PSQI global (−1.9 [4.0] vs −0.5 [3.7]; P = .03), sleep quality (−0.4 [0.8] vs −0.03 [0.7]; P = .02), and sleep disturbance (−0.4 [0.7] vs −0.1 [0.7]; P = .03) scores. The majority of these associations were found to be durable after adjustment for alternate independent cofactors using stepwise linear regression modeling.
Conclusions and Relevance
Patients with CRS and comorbid OSA have poor QOL with substantial disease-specific QOL improvements following surgery. Patients who present with CRS should be assessed for primary sleep disorders and, if identified, should be treated concurrently for both CRS and OSA to improve sleep dysfunction to optimize surgical outcomes.
clinicaltrials.gov Identifier: NCT01332136
Sleep disabilities impair the physical, psychological, and social aspects of a patient’s well-being and quality of life (QOL). A growing body of literature has highlighted the important links between QOL, sleep, and chronic rhinosinusitis (CRS), such that disease severity has been correlated with worse QOL and patients with worse QOL have poor sleep. It has been demonstrated that more than 75% of patients with CRS report abnormal sleep quality, with worse sleep seen in patients with higher levels of CRS-specific disease severity.1 Higher levels of sleep dysfunction are also associated with patients electing surgical therapy over continued medical management for CRS.2 Although the precise pathophysiologic mechanism of this phenomenon is unknown, it is plausible that CRS propagates sleep dysfunction through many cofactors including nightly arousals, nasal obstruction,3 depression, sex of the patient, pain, and somnogenic mediators.4
Patients with CRS and comorbid obstructive sleep apnea (OSA) are known to have an overall reduced QOL.5,6 Furthermore, both diseases are known to have substantial effects on sleep dysfunction.1 It is unknown whether patients with comorbid OSA have more severe disease, worse QOL, and sleep dysfunction. It is accepted, however, that lasting improvements in CRS-specific QOL can be achieved after functional endoscopic sinus surgery (FESS) in patients with medically refractory CRS.5- 7 It remains unknown whether similar improvements are seen in patients with CRS with comorbid OSA after FESS, and there has yet to be an investigation that has examined the role that OSA plays in CRS disease-specific QOL and outcomes in this patient population. The objectives of this investigation were to determine whether the comorbidity of OSA in patients with CRS negatively affects their reported baseline QOL and to subsequently examine the impact of FESS on QOL and sleep dysfunction.
Adult patients (>18 years) were recruited into a continuing, multi-institutional, prospective study using rhinology practices located within the University of Utah (Salt Lake City, Utah), Oregon Health and Science University (OHSU) (Portland, Oregon), the Medical University of South Carolina (Charleston, South Carolina), Stanford University (Palo Alto, California), and the University of Calgary (Calgary, Alberta, Canada). Preliminary results from this cohort study have been previously reported.1,7,8 All patients received a diagnosis of recalcitrant CRS according to diagnostic criteria endorsed by the American Academy of Otolaryngology9 and elected to pursue FESS as a treatment option following unsuccessful therapeutic regimens including at least 1 course (>14 days) of broad-spectrum or culture-directed antibiotics and a trial of topical corticosteroid application (>21 days) or a 5-day course of systemic corticosteroid therapy.
Study participants provided written, informed consent in English and agreed to complete all preoperative evaluations. Participants were asked to provide personal baseline characteristics, as well as a complete social and medical history. To determine whether self-reported diagnosis of OSA was accurate, we performed additional retrospective medical record review of all participants to identify physician-defined OSA as classified by the International Classification of Diseases, Ninth Revision, Coding Manual.10(p89) This approach found that self-reported comorbid OSA in study participants with CRS was 100% sensitive with 95% accuracy. These frequencies are consistent with prior published investigations showing that self-reported diagnosis is commonly found to be as accurate as the information obtained by physicians.11,12 The institutional review board at each enrollment location provided study approval and annual review of all study protocols (see Trial Protocol in the Supplement). Coordination services and central data review were conducted within OHSU. Study participants were observed for up to 18 months after FESS and completed survey evaluations postoperatively at 6-month intervals, either during physician-directed clinical appointments or via follow-up mailings using postal services and self-addressed return envelopes.
Participants were excluded from final analyses if they presented with exacerbations of recurrent acute rhinosinusitis, ciliary dyskinesia–associated sinusitis, or cystic fibrosis–associated sinusitis because of the heterogeneity of disease. Patients with oral corticosteroid dependency were excluded because of known influence of oral steroid use on sleep function. Because of the continuous nature of this cohort investigation, participants who enrolled within the prior 6 months were also excluded because they had not yet entered the initial follow-up period. Any participants who failed to complete any study evaluations within the preceding 18 months were categorized as lost to follow-up.
Standard measures of sinonasal disease severity collected during preoperative appointments were used simultaneously for research study purposes. Patients were evaluated using 30° rigid endoscopes (SCB Xenon 175, Karl Storz) by each enrolling physician. Endoscopic examinations were staged by each enrolling physician using the Lund-Kennedy scoring system (score range, 0-20).13
Disease severity was also evaluated using high-resolution computed tomography using 1.0-mm contiguous images. Images were scored by each enrolling physician in accordance with the Lund-Mackay scoring system (score range, 0-24).14 Higher scores on both staging systems represent worse disease severity.
All participants were expected to complete 2 disease-specific QOL surveys and 1 general sleep quality survey. The Rhinosinusitis Disability Index (RSDI) is a 30-item survey instrument composed of 3 subdomains to evaluate direct impacts of CRS on a patient’s physical (score range, 0-44), functional (score range, 0-36), and emotional (score range, 0-40) status.15 Higher subdomain and total RSDI scores (range, 0-120) represent worse QOL and greater impacts of sinonasal disease on daily function.
The 22-item Sinonasal Outcome Test (SNOT-22) is a validated survey instrument developed to evaluate symptom severity of CRS (©2006, Washington University, St Louis, Missouri).16 Exploratory factor analysis of SNOT-22 scores in this cohort identified 5 distinct subdomains of the SNOT-22, which have been previously described.2 Subdomains include rhinologic symptoms (score range, 0-30), extranasal rhinologic symptoms (score range, 0-15), ear and/or facial symptoms (score range, 0-25), psychological dysfunction (score range, 0-35), and sleep dysfunction (score range, 0-25). Higher subdomain and total SNOT-22 scores (range, 0-110) represent worse QOL and symptom severity.
The Pittsburgh Sleep Quality Index (PSQI) is a validated 18-item measure of self-reported sleep quality and duration that evaluates the 4-week period prior to survey completion. Total PSQI scores (score range, 0-21) consist of 7 subdomain component scores including those for sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbance, sleep medication use, and daytime dysfunction (scores range, 0-3). Higher PSQI scores represent worse sleep quality and duration; a PSQI total score less than 5 is considered the threshold for “good” sleep quality, whereas a PSQI total score of greater than 5 is characterized as “poor” sleep quality.17 Each enrolling physician was blinded to all QOL data for the study duration.
The extent of FESS was directed by the intraoperative discretion of each enrolling physician and reflected disease progression on an individual patient basis. Study participants underwent either primary or revision surgery. Endoscopic sinus surgery consisted of either unilateral or bilateral maxillary antrostomy, partial or total ethmoidectomy, sphenoidotomy, or frontal sinusotomy procedures (Draf I, IIa, IIb, or III), with septoplasty and inferior turbinate reductions as adjunctive procedures as needed. Image guidance was used judiciously when deemed necessary. All surgical procedures were followed by postoperative therapeutic regimens including daily nasal saline rinses and subsequent medical therapy if necessary.
Deidentified data were transferred from each enrollment center to a coordinating center (OHSU) for entry and storage into a central database (Microsoft Access; Microsoft). Statistical analyses were completed using commercial software (SPSS, version 22; IBM). Participant characteristics, clinical measures of disease severity, and QOL survey measures were evaluated descriptively and data distribution normality was verified for all continuous measures. Final cohort data were dichotomized between participants with and without comorbid OSA. Significant differences between participants with and without OSA were evaluated using 2-tailed sample t tests, Mann-Whitney U tests, and χ2 testing. Matched-pairing t tests or Wilcoxon signed rank tests were used to evaluate significant differences over time. The primary outcomes of interest include the postoperative improvement in mean endoscopy and QOL survey scores, operationalized by subtracting preoperative scores from last available postoperative scores. Improvement in the percentage of poor sleep was determined using the McNemar paired-sample χ2 test.
Simple stepwise linear regression analysis was used to identify significant independent risk factors associated with significantly different mean postoperative improvement. Preliminary models included a binary measure of OSA at the main exposure variable of interest and all independent factors screened for univariate significance (P < .25) while adjusting for potential enrollment site differences. Twenty-four variables were additionally screened for univariate significance including baseline characteristics and descriptors of surgical extent. Without adjustment for preoperative scores final models used a manual forward selection (P < .10) and backward elimination (P < .05) process and multicollinearity was evaluated using variance inflation factors. Unadjusted and adjusted regression coefficients (β), standard errors, 95% confidence intervals, and estimates of type I error (P values) are reported. The percentage of final model variance was evaluated using coefficients of multiple determination (R2).
After all inclusion and exclusion criteria were applied, a total of 405 participants completed all eligibility requirements and were included for final study analysis (Figure). A total of 285 (70%) participants meeting inclusion criteria had at least 6-month follow-up evaluations. Sixty study participants (15%) had a diagnosis of comorbid OSA. In regards to baseline characteristics, all participants with OSA were found to have a significantly lower prevalence of both female sex and nasal polyposis compared with those without OSA (Table 1).
The frequency of each discrete surgical procedure was compared between participants with and without OSA (Table 2). Participants with comorbid OSA were found to have a significantly lower prevalence of sphenoidotomy, Draf I frontal sinusotomy, and Draf IIa frontal sinusotomy when comparing the total number of sides undergoing surgical intervention. Participants with OSA were found to also have a significantly higher prevalence of inferior turbinate reduction when comparing total number of sides compared with counterparts without comorbid OSA.
Mean postoperative scores significantly improved over time for all disease-specific QOL outcome measures in both CRS and CRS with OSA subgroups (Table 3). When comparing the magnitude of mean improvement between subgroups, participants without OSA were found to have a significantly greater improvement in RSDI total scores, the physical and functional domain RSDI scores, and SNOT-22 rhinologic symptom scores. Mean improvement was also greater for SNOT-22 extranasal rhinologic symptom domain scores; however, that difference was not significant at the .05 level of determination. Similarly, mean (SD) endoscopy scores significantly improved over time for study participants with OSA (n = 33; −2.6 [3.8]; P < .001) and without OSA (n = 167; −3.4 [3.9]; P < .001) to a statistically similar magnitude (P = .26).
Participants without comorbid OSA reported significant sleep quality score improvement from preoperative scores for both PSQI total and subdomain scores, with the exception of sleep medication use (Table 3). Participants with comorbid OSA did not report significant improvement following FESS for either PSQI total or subdomain scores. The magnitude of improvement was significantly greater for participants without OSA for PSQI total scores, sleep quality, and sleep disturbance scores.
Postoperative PSQI total scores were dichotomized into those with “good” and “poor” sleep quality. The preoperative prevalence of poor sleep between participants with and without OSA is reported in Table 1. Between all participants with follow-up, patients with OSA were found to have similar prevalence of poor sleep compared with participants without OSA following sinus surgery. Interestingly, among those without comorbid OSA there was a significant reduction in the percentage of patients with poor sleep (P < .001) whereas the percentage of participants with comorbid OSA reporting poor sleep scores did not significantly change (P = .18) (Table 3).
Comorbid OSA was found to be significantly associated with less postoperative improvement in mean scores for a number of outcome measures (Table 3) without controlling for any additional covariates. Regression modeling was used to identify and adjust for the influence of additional covariates on significant bivariate relationships between OSA and outcome measures (Table 4). Comorbid OSA was associated with significantly less improvement on reported RSDI functional domain, SNOT-22 rhinologic symptom, and PSQI sleep quality scores after covariate elimination criteria were applied, compared with participants without a history of OSA. After controlling for all independent covariates, presence of OSA was found to be associated with a significantly smaller mean (SE) improvement of 3.0 (1.4) units on RSDI functional domain scores (P = .04), a difference of approximately 8.0% of the domain score range. No evidence of multicollinearity was discovered, with all variance inflation factors less than 2.0. Depression was not found to be a significant independent risk factor for any final model. Overall final models were able to explain only between approximately 5% and approximately 13% of total model variance.
The prevalence of OSA in women is estimated to range from 5% to 6%, compared with a prevalence in men of 13% to 14%.18 The prevalence of OSA in our cohort was found to be 15%, resembling the prevalence in the general US population. Similarly in our cohort, patients with CRS and OSA were significantly less likely to be female and have nasal polyps. There was no difference found between those with and without OSA in regards to disease severity or CRS disease-specific QOL, poor sleep (PSQI total score, >5), or mean PSQI total and subdomain scores prior to surgery (Table 1). Following FESS, substantial gains in QOL and disease severity were observed for patients with CRS with and without OSA, and these gains were statistically significant. Without controlling for other covariates, patients with CRS without OSA reported better mean improvement in RSDI total scores, the physical and functional domain scores of the RSDI, and the rhinologic symptom scores of the SNOT-22 compared with those with a comorbidity of OSA. Interestingly, patients without OSA reported significant sleep quality score improvement from preoperative scores for the PSQI, whereas those with comorbid OSA did not report significant postoperative improvements as measured by any aspect of the PSQI survey instrument.
Two bodies of literature exist that suggest that QOL and sleep are reduced in patients with OSA19 and in patients with CRS.1,20 Furthermore, poor sleep quality in patients with CRS significantly correlates with worse CRS disease-specific QOL as measured by the SNOT-22 and RSDI. Therefore, we hypothesized that patients with CRS with a comorbidity of OSA would report overall worse QOL and sleep. Unexpectedly, baseline QOL and reported sleep dysfunction did not differ between OSA subgroups. Following FESS, both those participants with and without OSA had significant improvements in QOL (Table 3). Because the magnitude of OSA severity was unknown, it is possible that participants in our cohort represent mild forms of OSA and this may be contributing to the lower QOL impairments seen in those with both CRS and OSA. Although the severity of OSA was not assessed in this investigation, current literature does suggest that the severity of the QOL impairment is not always proportional to the severity of OSA.21 This also holds true in patients with CRS because disease severity does not correlate well to PSQI scores.1 The mechanisms underlying these final common pathways leading to QOL and sleep dysfunction in OSA and CRS have not yet been elucidated.
The impact of CRS and the nasal airway on sleep quality and OSA may occur via some combination of 2 putative mechanisms: increased nasal airway resistance and the pro-inflammatory milieu.4 Increased airway resistance may be playing a role in CRS-associated sleep dysfunction, although the interplay between inflammation such as edema and nasal polyposis and increased airway resistance is complex. For instance, nasal polyposis has been associated with sleep-disordered breathing19 and reduced QOL.22,23 Nasal polyps have been associated with both nasal obstruction and sleep disturbance,24 with a reduced daytime sleepiness following surgery without associated changes in the apnea-hypopnea index (AHI).25 In contrast, surgery for nasal obstruction in patients with OSA has not consistently been proven to provide improvements in sleep disorders.26 This is likely the result of the multifactorial, multilevel obstruction that contributes to OSA. Functional endoscopic sinus surgery has been shown to reduce AHI in patients with moderate to severe OSA, but these improvements were not clinically significant.27 Likewise, we demonstrate that patients with CRS and OSA do not have significant improvements in sleep quality using the PSQI following FESS. Unfortunately, we did not objectively evaluate nasal obstruction and multilevel airway obstruction in this investigation and are unable to decipher the precise role that FESS plays in improving nasal obstruction as it relates to sleep in our cohort. Future studies, which intend to explore the mechanisms underlying sleep improvement in CRS, should explore both reduction in nasal obstruction and objective measures of sleep function.
Sinus surgery may lead to better improvements in QOL and sleep in patients with CRS through improved control of chronic inflammation rather than improved nasal obstruction. The current literature supports this, as Alt et al7 previously demonstrated sleep dysfunction in patients with CRS and that nasal polyposis was not an independent risk factor for sleep dysfunction.1 Furthermore, other investigators have shown that patients with CRS and OSA undergoing polypectomy do not reduce their AHI, whereas patients without nasal polyposis did demonstrate significant improvement in AHI scores.27 Likewise, herein we demonstrate that the difference in polyp status between those with and without OSA did not appear to contribute to the observed difference in sleep dysfunction improvement.
Obstructive sleep apnea has been associated with many health consequences that are thought to result from sleep disruption, including manifestations such as depression, which is also highly prevalent in CRS.28,29 Obstructive sleep apnea has been linked to increased morbidity from the increased risk of automobile and industrial accidents secondary to sleepiness, pulmonary and systemic hypertension, congestive heart failure, arrhythmias, myocardial infarction, stroke, and even death.30- 33 Sleep disruption has also been linked to increased burden of disease in other chronic diseases.34 It is unknown whether patients with CRS and comorbid OSA have these associated health consequences, although we believe that further investigations to look into these associations are warranted. Given the prevalence of OSA and CRS and their detrimental effects on physical and neuropsychiatric function, understanding the contribution that OSA makes in CRS has important implications.
This study demonstrates that sleep quality is diminished in patients with CRS with and without OSA, but the underlying etiology and pathophysiologic mechanism of this sleep dysfunction is still unknown. We have previously posited that sleep function in CRS is regulated in the central nervous system by highly interconnected neuronal groups that are characterized by altered input-output signaling.35 This is controlled via local signaling by growth factors and cytokines, which can influence neurons to adjust and even change the input-output properties of the neuronal groups both in sleep centers and in the cortex of the central nervous system. Interestingly, CRS is a chronic inflammatory disease that is associated with changes in cytokines, cytokine receptors, and downstream products. Obstructive sleep apnea also has strong links to systemic inflammation and specifically to prosomnogenic mediators. It is plausible and exciting to postulate that OSA has its own specific somnogenic-inflammatory-airway phenotype36 that is independent of CRS inflammation, and this may explain the lack of response to treatment seen in patients with CRS and OSA following FESS.
There are several caveats to consider when interpreting the findings of this study. First, there is potential follow-up bias associated with the fact that 29.6% of participants opted out of postoperative follow-up evaluations. Comparisons between all baseline measures found that participants with follow-up were significantly more likely to be older (mean [SD] age, 52.4 [15.1] vs 45.8 [13.8] years; P < .001) and have slightly worse baseline mean (SD) endoscopy (6.2 [3.7] vs 5.3 [3.6]; P = .02) and computed tomography (12.3 [6.2] vs 10.7 [5.8]; P = .02) scores. In addition, patients with and without OSA did not undergo identical surgical procedures. Patients with CRS and comorbid OSA had less complete surgery overall compared with those without OSA (Table 2). Because of the fact that inferior turbinate surgery could potentially improve upper airway obstruction, regression models evaluated measures of surgical extent; however, these did not identify significant associations between inferior turbinate reduction procedures and any measure of improved sleep quality. Although the literature is not clear in regards to nasal obstruction contributing to both QOL and sleep dysfunction in CRS, our findings are consistent with those of Yalamanchali et al,27 which suggest that concurrent FESS with nasal airway surgery does not play a major role in improving sleep dysfunction in patients with CRS with comorbid OSA. In addition, inferior turbinate reduction alone has no significant effect on AHI compared with combined septoplasty with inferior turbinate reduction in patients with OSA.37 Furthermore, this study did not evaluate differences in the frequency of patients undergoing simultaneous therapeutic management for comorbid OSA using continuous positive airway pressure, nor adherence to use of continuous positive airway pressure during the study. This study also did not include a control cohort, so we cannot fully exclude the influence of either placebo effects, the natural history of disease, or regression to the mean of reported outcomes scores.
Patients with CRS have a high prevalence of sleep dysfunction that significantly improves following FESS as measured by the PSQI. Patients with OSA and CRS demonstrate substantial improvements in QOL but not in sleep quality as measured by the PSQI. Patients with OSA should be treated concurrently for both CRS and OSA to optimize sleep dysfunction and QOL improvement. Future investigations are needed to further elucidate the discordance and underlying mechanisms of sleep improvement between those patients with and without OSA with objective sleep measures.
Submitted for Publication: May 4, 2015; final revision received May 27, 2015; accepted July 1, 2015.
Corresponding Author: Timothy L. Smith, MD, MPH, Oregon Health and Science University, Division of Rhinology and Sinus/Skull Base Surgery, Department of Otolaryngology–Head and Neck Surgery, Oregon Sinus Center, 3181 SW Sam Jackson Park Rd, PV-01, Portland, OR 97239 (firstname.lastname@example.org).
Published Online: September 10, 2015. doi:10.1001/jamaoto.2015.1673.
Author Contributions: Dr Alt and Mr Mace had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: DeConde, Smith.
Acquisition, analysis, or interpretation of data: Alt, Mace, Steele, Orlandi, Smith.
Drafting of the manuscript: Alt, DeConde, Mace, Steele, Smith.
Critical revision of the manuscript for important intellectual content: DeConde, Mace, Steele, Orlandi, Smith.
Statistical analysis: Mace.
Obtained funding: Smith.
Administrative, technical, or material support: DeConde, Mace, Smith.
Study supervision: DeConde, Smith.
Conflict of Interest Disclosures: Drs DeConde and Smith are consultants for Intersect ENT (Menlo Park, California), which is not affiliated with this investigation. Dr Orlandi is a consultant for Medtronic ENT (Jacksonville, Florida), which is not affiliated with this research. No other disclosures are reported.
Funding/Support: This study was supported by a grant from the National Institute on Deafness and Other Communication Disorders (NIDCD), one of the National Institutes of Health (R01 DC005805; principal investigator: T.L.S.). Dr Alt, Mr Mace, and Dr Smith receive partial support from the NIDCD.
Role of the Funder/Sponsor: The funding organization had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Previous Presentation: The abstract associated with this manuscript was accepted for oral presentation to the American Rhinologic Society at the annual Combined Otolaryngology Spring Meetings; April 22-26, 2015; Boston, Massachusetts. This manuscript will not be published as a result of this meeting presentation.