2017-18 HSC Section 3 Green Book

Research Original Investigation

Complications After Obstructive Sleep Apnea Upper Airway Surgery

patient surgical procedures. Similarly, amore recent and com- prehensive study 17 examining the effect of OSA on the rate of complications of bariatric surgery discovered no association between the severity of OSA as assessed preoperatively and perioperative complications. Increasing evidence shows that the incidence of operative complications is greater in unrecognized OSA than in known OSA. 1,25 Recently, a matched cohort study 26 found that, al- though the presence of OSA increased the rate of complica- tions by 2-fold, it is the patients with unrecognized OSA who were twice as likely to experience complications than those whose conditionswerediagnosedandmanagedaccordinglybe- fore surgery. Alarmingly, the prevalence of OSAwas estimated to increasemore than4-fold from1993until recently. 27,28 Inour study population, SDB is reported to affect 46 of 317 children (14.5%) aged 7 to 15 years and 549 of 1611 adults (34.1%) older than 30 years. 29,30 It was estimated that more than 80% of pa- tientshaveundiagnosedOSA, 27 andunrecognizedOSA ispreva- lent among adult surgical patients. 31,32 This findingmay lead to a significantly increased risk of perioperative complications. 32 We found that, rather than the AHI, Lsat (odds ratio, 1.03; 95%CI, 0.96-1.45; P = .04) and longest apneaduration (odds ra- tio, 1.03; 95% CI, 0.99-1.08; P = .02) were the polysomno- graphic variables that predict the perioperative complications. Many past studies 2,3,8,9 on upper airway surgery for OSA re- ported that it is the preoperative Lsat or longest apnea dura- tion that consistently linkedwithoperativecomplications rather than the AHI severity. These studies 2,3,8,9 might have indi- cated that, in the patients with polysomnography-confirmed OSAundergoing surgery, thesevariables aremore important and significant than AHI in determining the rate of complication. The current series includes 95patientswhounderwent sur- gery for the treatment of OSA. Therewere no deaths in our se- ries, but 48 of 95 patients developed at least 1 operative com- plication. This figure is high compared with many previous studies. 8-10,33 We think this occurred partly because we have lowered the definitionof complications to include even a slight deterrent fromthe normal variables (eg, temporary oxygende- saturation, CPAP or BiPAP intervention, and atelectasis are sometimes not included as respiratory complications in stud- ies). Furthermore, comparedwithprevious studies, 2,8,9,26 most of our patients had severe OSA (54 of 95 [57%]) (Table 1). Simi- lar tomany other OSA surgery series, 2,8,9,26 pulmonary compli- cationswere themost commonand consistently reported com- plications of our cohort. The risk of respiratory complications after upper airway surgery is enhanced inpatientswithOSAbe- cause of the underlying anatomical andphysiologic abnormal- ity associatedwith thedisease. Anesthetic andanalgesic agents used during the perioperative period can decrease upper air- way muscle tone and depress ventilatory responses, particu- larly during the early postoperative period. 34 Inparticular, opi- oid anesthesia adversely affects SDB and leads to suggestions that it shouldbeavoided inOSAsurgery. 5,34,35 Theseheightened concerns have prompted physicians to recommend routine postoperative intensive care unit stays in patientswith comor- bid factors and those undergoingmultiple procedures. 36-38 In our study, however, we didnot find any significant association between the types of analgesics used and the rates of operative

Table 4. Logistic Regression Analysis of Postoperative Complications Given Any Complication as the Dependent Variable a

Any Complication OR (95% CI) 1.18 (0.94-1.43) 1.11 (1.07-1.32)

Variable

P Value

Age (per 1 y)

.04 .03

BMI (per 5 units) b

Smoking No

1 [Reference]

.51

Yes

1.16 (0.54-3.33)

Surgery type Single

1 [Reference]

.68

Combination

1.04 (0.99-4.75)

OSA category c Mild

1 [Reference]

Moderate

0.99 (0.36-2.59) 1.05 (0.26-5.46) 1.00 (0.99-1.15)

.40

Severe

AHI (per 5-unit increase)

.20

Polysomnography

Lowest desaturation (per 5% increase) Longest apnea duration (per 5-s increase)

1.03 (0.96-1.45)

.04

1.03 (0.99-1.08)

.02

tive complications for airway and nonairway operations have published contrasting results. We hypothesized that, if the di- agnosis of OSA is an independent risk for operative complica- tions, there should be a significant association between the se- verity of OSA and the rate of complications. As expected in our patients, age, BMI, and smoking increase the risk of compli- cations in upper airway surgery. Nonetheless, univariate and multivariable analyses revealed no association between pre- operative AHI severity and complication rates. Because sleep laboratories differ in their criteria for detecting episodes of ap- nea and hypopnea, many experts believe that the sleep labo- ratory’s assessment (none, mild, moderate, or severe) should take precedence over the actual AHI (the number of episodes of SDB per hour) in measuring the severity of OSA for re- search purposes. 5,6 Even with the use of mild, moderate, and severe category grading, no significant associationwas found betweenOSA severity and the complication rates found in our data. Thus, either OSA is not an independent risk factor for op- erative complications or the preoperative recognition, and hence the subsequent preoperativemedicalmanagement,miti- gates this risk. Studies 3,22,23 have found that the rate of surgi- cal complications was reduced in patients withOSA preopera- tively treated with CPAP. Large studies 17,24 also found that, regardless of OSA severity, a planned elective surgery in pa- tientswithpreoperativelydiagnosedandmanagedOSAdidnot result in increased risk of perioperative complications. For ex- ample, in a study published in 2003, Sabers and colleagues 24 found that preoperative diagnosis of OSA was not a risk fac- tor for surgical complications among patients undergoing out- a Surgery types, OSA category, and smoking were modeled as categorical variables, and all other characteristics were modeled as continuous variables. b Calculated as weight in kilograms divided by height in meters squared. c The OSA categories are defined by AHI as mild (AHI, 5-15), moderate (AHI, 16-30), and severe (AHI, >30). Abbreviations: AHI, apnea-hypopnea index; BMI, body mass index; OR, odds ratio; OSA, obstructive sleep apnea.

JAMA Otolaryngology–Head & Neck Surgery March 2017 Volume 143, Number 3 (Reprinted)

jamaotolaryngology.com

Copyright 2017 American Medical Association. All rights reserved.

140