Postoperative Complications in Patients with Obstructive Sleep Apnea

Abstract & Commentary

By Barbara A. Phillips, MD, MSPH, Professor of Medicine, University of Kentucky; Director, Sleep Disorders Center, Samaritan Hospital, Lexington. Dr. Phillips serves on the speakers bureaus for Cephalon, Resmed, and Respironics.

Synopsis: More than half of patients undergoing non-cardiac surgery and polysomnography had obstructive sleep apnea, which was associated with an increased risk of perioperative complications, including hypoxemia, ICU transfer, and prolonged length of stay.

Source: Kaw R, et al. Postoperative complications in patients with obstructive sleep apnea. Chest 2012;141:436-441.

The cohort for this analysis was assembled from nearly 40,000 patients who underwent preoperative assessment over a 5-year period. The investigators cross-referenced the electronic health record with the sleep laboratory database to identify patients who had noncardiac surgery within 3 years of having a sleep study. Demographic, clinical, diagnostic, and postoperative data were collected from outpatient electronic records, inpatient hospital admission records, and surgical procedure dictation records. Obstructive sleep apnea (OSA) was defined as an apnea plus hypopnea index (AHI) of more than 5 events per hour of sleep. In patients with more than one procedure, morbidity data and postoperative outcomes were collected for the surgical procedure done closest to the date of the overnight sleep study.

The endpoints for this study were significant postoperative complications, including postoperative hypoxemia, respiratory failure, congestive heart failure, myocardial infarction (MI), atrial fibrillation, delirium, death within 30 days, and hospital length of stay (LOS). Postoperative hypoxemia was defined as oxygen desaturations with a < 90% or > 4% reduction from last recorded value, or by arterial blood gas postoperatively. Postoperative respiratory failure was defined as the need for mechanical ventilation longer than 24 hours, endotracheal reintubation, or tracheostomy. Postoperative congestive heart failure was defined as new pulmonary edema, jugular venous pressure > 10 mmHg, use of diuretic or afterload/preload reducing agents, or physician documentation. Postoperative MI was defined as the appearance of new Q waves > 0.04 sec wide and 1 mV in depth accompanied by elevated levels of troponin T (0.03 ng/mL) and creatine kinase-MB (> 100 IU/L).

A propensity score for the likelihood of sleep apnea for each patient was calculated using logistic regression. The propensity model used in this study included age, sex, race, BMI, use of general anesthesia, American Society of Anesthesiology class, and several comorbidities and their interactions as covariates. The model was statistically quite strong as a predictor of sleep apnea.

The final cohort for the analysis was 471 patients. Of these, 282 (59.8%) had OSA. Patients with OSA were older, (55.9 vs 46.3 years), predominantly male (44.7% vs 21.7%), and heavier (BMI 38.3 vs 33 kg/m2). They had a higher American Society of Anesthesiologists risk class, and many more medical comorbidities (i.e., COPD, hypertension, diabetes mellitus, coronary artery disease) than patients without OSA. Most of the surgeries were intermediate risk, with abdomino-pelvic and orthopedic procedures dominating. No differences existed between the types of anesthesia used in the two groups, and general anesthesia was more commonly used overall by far (> 80%). After adjustment for the propensity score, the presence of OSA was associated with a higher incidence of overall complications (odds ratio [OR], 6.9; P = 0.003), postoperative hypoxemia (OR, 7.9; P = 0.009), ICU transfer (OR, 4.43; P = 0.069), and longer LOS (OR, 1.65; P = 0.049). Severity of OSA measured by the AHI was not associated with postoperative complications. The median LOS was 2 days in the OSA group and 1 day in the control group.


The two take-home messages from this study are that sleep apnea is very common in perioperative patients and that it is associated with an increased risk of complications, notably increased LOS, ICU transfer, and hypoxemia.

This study is important because it is the largest study to date to determine the prevalence of polysomnographically determined OSA in the general surgical population. Although several previous reports have reported that OSA is a risk factor for increased postoperative morbidity and mortality,1-4 most have based the diagnosis of OSA on screening questionnaires. A confounder in previous reports has likely been the presence of undiagnosed and unrecognized sleep apnea in the "control group." Because all of the patients in this analysis had sleep studies, those patients who were characterized as not having sleep apnea in this report probably did not. (On the other hand, this cohort was likely "enriched" by sleep apneics, because sleep studies were ordered on the basis of clinical suspicion). A further strength of the current study is that the authors used a propensity score that controlled for factors associated with OSA, including BMI.

Because the prevalence of obesity is increasing in this country, the prevalence of OSA is increasing as well. Between 1990 and 1998, there was a 12-fold increase in the diagnosis of OSA in surgical outpatients.5 The American Society of Anesthesiologists has published clinical guidelines for the perioperative management of OSA,2 but these are mostly based on consensus and are not widely used. Further, they are focused on the patients' care while in the facility.

Where to go from here? Awareness is an important first step. Patients who are obese, sleepy, hypertensive, have witnessed apneas, or big necks are at high risk for sleep apnea. The American Society of Anesthesiologists recommends longer postoperative monitoring in OSA patients after ambulatory surgery and 7 hours of monitoring after the last episode of airway obstruction or hypoxemia while breathing room air in an unstimulated environment prior to discharge from the facility.2 And our colleagues in anesthesiology will likely make sure that this happens. But what happens after discharge? In a recent study, Liao et al reported a higher AHI and oxygen desaturation index among OSA patients on the third postoperative night compared with preoperatively or on the first postoperative night.6 This is very likely due to the use of respiratory depressants, especially opioids7 and/or Rapid-Eye Movement sleep rebound.6 Options here are to delay discharge, increase at-home monitoring, or reduce respiratory depressant use. None of these choices is particularly attractive, but neither are postoperative complications.


1. Sharma VK, et al. Ann Intern Med 2003;139:707-708.

2. Gross JB, et al. Anesthesiology 2006;104:1081-1093.

3. Hwang D, et al. Chest 2008; 133:1128-1134.

4. Memtsoudis S, et al. Anesth Analg 2011;112:113-121.

5. Namen AM, et al. Chest 2002;121:1741-1747.

6. Liao P, et al. Sleep 2009;A223:0684.

7. Rosenberg J, et al. Anaesthesia 1999;54:323-328.