Sleep-disordered breathing (SDB) encompasses a spectrum of breathing disorder events during sleep ranging from the mildest form, known as respiratory effort–related arousals (RERA) to more severe events such as hypopneas and apneas. The mildest form of SDB is upper airway resistance syndrome, which consists mostly of RERAs. The most severe and most prevalent form of SDB is obstructive sleep apnea, which consists mostly of obstructive apneas and hypopneas. The current estimated prevalence of moderate-to-severe SDB (defined by an apnea-hypopnea index [AHI] > 15 events/h) is 17% among 50- to 70-year-old men and 9% among 50- to 70-year-old women.1 The prevalence of sleep apnea in surgical patients may vary with different surgical populations. Published data from a single-center large database reported a prevalence of polysomnographically proven SDB of 7% in preoperative patients.2 Analyses of large and nationally representative administrative databases using ICD-9-CM codes have yielded a similar SDB prevalence of 8% in surgical patients.2,3 Because many SDB patients remain clinically undiagnosed,4 and because, in morbidly obese patients, 7 of 10 were found to have SDB,5 the point estimates from these studies may be an underestimation.
SDB carries significant clinical implications for anesthesiologists. Both adults and children with SDB are more sensitive to general anesthetics and narcotics.6–9 SDB patients have an increase in upper airway collapsibility10–13 and are particularly vulnerable to upper airway obstruction during the postoperative period. Accumulating evidence supports the notion that SDB patients could have an increased incidence of perioperative adverse events. There are case reports of death.3,14–20 A recent study of the Premier Perspective database found that patients with SDB were more likely to receive ventilatory support, use more intensive care, step-down, and telemetry services, consume more economic resources, and have a longer hospital stay.3 To decrease the perioperative risk of SDB, the American Society of Anesthesiologists has recently updated its guidelines;21 perioperative functional algorithms for the management of SDB have been proposed22 and a recent editorial has called for action.23
Recently published data show that postoperative AHI was significantly increased in some patients without preoperative sleep apnea (non–sleep apnea).24,25 These patients may carry an increased risk for SDB-related perioperative adverse events. The objective of this study was to investigate the incidence of postoperative moderate-to-severe SDB (AHI > 15 events/h) in patients in whom sleep apnea was absent based on preoperative polysomnography (AHI ≤ 5 events/h). We were also interested in exploring predictive factors associated with an increase in the AHI in this subset of patients. We hypothesized that in non–sleep apnea patients, clinical characteristics and parameters from preoperative polysomnography may be predictive of the postoperative occurrence of moderate-to-severe SDB.
Study Design and Study Subjects
Our study had a prospective observational design without any interventions. Approvals from Institutional Review Boards were obtained from University Health Network and Mount Sinai Hospital, both at Toronto (UHN: 06-0135-AE and 07-0515-AE; MSH: 06-0143-E and 07-0183-E). All patients 18 years or older, who were ASA physical status I to IV and scheduled for elective surgical procedures, were approached by the study coordinators for written informed consent. Patients who were unwilling or unable to give informed consent or patients who were expected to have abnormal electroencephalographic findings (e.g., brain tumor, epilepsy surgery, patients with deep brain stimulators) were excluded. Methods have been previously reported, and some of the data from sleep studies were reported previously.24,25
Patient Recruitment and Follow-Up
Eligible patients visiting preoperative clinics for a scheduled procedure were approached for informed consent. Patients who consented underwent sleep studies with a 10-channel portable polysomnography device (Embletta X100, Embla, Broomfield, CO). A full-night polysomnography with Embletta X100 was performed as previously described,26 preoperatively at home and on postoperative night 1 and 3 in the hospital or at home. Embletta X100 is a level 2 diagnostic tool for SDB27 and has been validated against laboratory polysomnography.26 The polysomnographic recording montage consisted of 2 electroencephalographic channels (C3 and C4), left or right electroculogram, chin muscle electromyogram, nasal cannula (pressure transducer), thoracic and abdominal respiratory effort bands, body-position sensor, and pulse oximetry. At bedtime, a sleep technician connected the portable polysomnography device at the patient’s home or in the hospital. The overnight recording itself was unattended. The patients were taught how to disconnect the device, which was retrieved by the same sleep technician the following morning. The patients were asked to keep a sleep diary, and the sleep technician picking up the device ensured that the diary was completed.
The recordings from the portable polysomnography were scored by a certified polysomnographic technologist and reviewed by a sleep physician. Somnologia Studio 5.0 (Embla, Broomfield, CO) was the platform used for scoring polysomnography recordings. The polysomnography recordings were manually scored epoch-by-epoch by the sleep technologist, according to the manual published by the American Academy of Sleep Medicine in 2007.28 Apnea index was defined as the average number of apneas/h. Hypopnea index was the average number of hypopneas/h. Hypopnea was defined as at least 50% reduction in air flow that lasted at least 10 seconds and was associated with at least 3% decrease in arterial oxyhemoglobin saturation or with an arousal. AHI was defined as the average number of apneas and hypopneas/h of sleep. Rapid eye movement (REM) sleep AHI was defined as the sum of apneas and hypopneas during REM sleep divided by the total amount of REM sleep, and non-REM sleep AHI was defined as the sum of apneas and hypopneas during non-REM sleep divided by the total amount of non-REM sleep. Respiratory disturbance index (RDI) was defined as the sum of apneas, hypopneas, and RERAs divided by total sleep time in hours. RERA was defined as an event characterized by obstructive upper airway airflow reduction (which does not meet the criteria of apnea or hypopnea), associated with increased respiratory effort that resolves with the appearance of arousals. Respiratory arousal index was calculated as the number of arousals due to respiratory events/h of sleep. Oxygen desaturation index was calculated as the total episodes of desaturation ≥ 4% lasting ≥10 seconds divided by total sleep time. CT90 was defined as the cumulative sleep time percentage with pulse oxygen saturation (SpO2) <90%.
In this study, all polysomnography recordings were scored after the patients were discharged from the hospital. The anesthesiologists and the surgeons caring for the study patients were blinded to the results of the polysomnography recordings. During the study period, the health care team provided routine care to their patients. The decision to provide supplemental oxygen therapy or continuous positive airway pressure (CPAP) therapy was determined by the perioperative care team.
The patients with an AHI ≤ 5 events/h on the preoperative home polysomnography were defined as non–sleep apnea patients. According to the AHI on the postoperative polysomnography, the preoperative non-SDB patients were classified into 2 groups. The patients with AHI > 15 events/h on either postoperative night 1 or night 3 were classified into postoperative SDB group. The patients with AHI ≤ 15 events/h on both postoperative night 1 and 3 were defined as postoperative non-SDB group. In this study, only preoperative and postoperative polysomnography data of the preoperative non-SDB patients were examined.
Anesthesia and Postoperative Pain Control
A balanced anesthetic technique was used in all patients. In general anesthesia, patients received an induction dose of propofol, a narcotic, an inhaled drug, and a neuromuscular blocking drug. The neuromuscular blocking drug may or may not have been reversed. In regional anesthesia, patients received epimorph and a propofol infusion. Both groups received a narcotic in the postoperative period.
All patients were reviewed twice daily by the Acute Pain Service team, as per our institutional standard of care. Pain was evaluated by a score of 0 to 10 with 0 as no pain and 10 as the most excruciating pain. IV morphine through patient-controlled analgesia was initiated when the verbal pain score was ≥4. The Acute Pain Service team increased the immediate-release oxycodone dose and/or added controlled-release oxycodone to achieve a verbal pain score of 4 or less for pain.
The clinical decision of oxygen therapy or CPAP therapy was made by the perioperative care team. According to the institution protocol, if SpO2 on oximeter monitoring is less than 94%, oxygen therapy would be provided.
DATA ANALYSIS AND STATISTICS
Sample Size Estimation
The hypothesis was that the clinical characteristics and parameters from preoperative polysomnography may be predictive of occurrence of moderate-to-severe postoperative SDB in non-SDB patients. To confirm or reject this hypothesis, we performed univariate logistic regression analysis for each candidate variable and multivariate logistic regression analysis to explore the adjusted predictive value of selected parameters. The simulation studies demonstrated that logistic regression models require 10 or more cases with events per predictor to produce stable estimates.29,30 If we include 3 predictors with a prevalence of postoperative moderate-to-severe SDB of 26%, the estimated sample size would be 10 × (3/0.26) = 115 or more.
Data were entered into a specifically designed Microsoft Access database and checked for possible errors. SAS 9.3 for Windows (SAS Institute, Cary, NC) was used for data analysis. All the statistical tests are 2-tailed tests, and P < 0.05 or adjusted P < 0.05 was accepted as statistically significant.
The clinical data and data from preoperative polysomnography of non–sleep apnea patients were summarized and compared between the postoperative SDB and postoperative non-SDB groups. Categorical data were presented as frequency with percentage and the statistical significance was checked by χ2 test or Fisher exact test. The mean ± SD was used for continuous data with normal distribution and the statistical significance was checked with Student independent 2-sample t test. The median (25th, 75th percentile) was used for continuous data with skewed distribution and Mann-Whitney U test was used to check the statistical significance for continuous data with skewed distribution. The Holm-Bonferroni method was used to adjust P value for multiple comparisons among the correlated variables.
If only 1 postoperative polysomnography was done either on postoperative night 1 or 3, the data from this polysomnography were used as postoperative polysomnographic data. If polysomnography was done on both postoperative night 1 and 3, the data from polysomnography with higher AHI was chosen as postoperative polysomnographic data. The perioperative change was the difference in values between postoperative polysomnography and preoperative baseline. Potential risk factors were first evaluated with univariate logistic regression, and the correlation between the significant factors was checked. If the correlation coefficient between predictive variables was equal or greater than 0.5, only one of these correlated variables was chosen for multivariate analysis. The predictive performance of the selected significant factor was also evaluated. The optimal cutoff, sensitivity privilege cutoff, and specificity privilege cutoff value were also explored. The determination of cutoffs was performed as previously described.31 The preoperative RDI value corresponding to the maximal Youden index (J = sensitivity − false-positive rate) was selected for each of 1000 bootstrapped populations, resulting in a set of 1000 “optimal” values. The mean of these values was defined as the optimal cutoff value, the lower limit of its 95% confidence interval was defined as the sensitivity privilege cutoff value, and the upper limit of its 95% confidence interval was defined as the specificity privilege cutoff value. The range between sensitivity privilege cutoff value and the specificity privilege cutoff value was marked as the gray zone (inconclusive zone). The patients with a value in the privilege sensitivity side of gray zone were classified as low risk for postoperative moderate-to-severe SDB. The patients with a value in the gray zone or inconclusive area were classified as intermediate risk, and the patients with a value in the privilege specificity side of gray zone were classified as high risk. The 95% confidence interval of sensitivity, specificity, positive predictive value and negative predictive value for the cutoffs were calculated with the exact binomial confidence interval method.
Patient recruitment and study implementation is shown in Figure 1. Six hundred fifty patients underwent preoperative polysomnography. Sixty-seven patients who wore CPAP during polysomnography and 406 patients who had an AHI > 5 events/h were excluded. One hundred seventy-seven patients with AHI ≤5 events/h were classified as non–sleep apnea patients. Because of withdrawal and failed postoperative polysomnography, 120 non–sleep apnea patients completed polysomnography on postoperative night 1 and/or night 3. Thirty-one (25.8 % [95% confidence interval (95% CI): 18.3%–34.6%]) patients had AHI > 15 events/h on postoperative night 1 and/or 3 (postoperative SDB group), and 89 (74%) patients with AHI ≤15 events/h on both postoperative night 1 and 3 (postoperative non-SDB group). Of these 120 patients, 13 (10.8%) patients had ambulatory surgery and were discharged on the same day, 3 (9.7%) in postoperative SDB group and 10 (11.2%) in nonpostoperative SDB group, P = 0.81.
The clinical data of 120 non–sleep apnea patients are summarized in Table 1. Compared with the patients in postoperative non-SDB group, the patients in postoperative SDB group were older (60 ± 13 years vs 53 ± 12 years, P = 0.008) with more smokers (32.3% vs 15.7%, P = 0.048). Among the inpatients, postoperative SDB group had longer hospital stay than the postoperative non-SDB group, 4.5 (3.0, 6.5) (median [25th, 75th percentile]) days (n = 28) vs 3.0 (2.0, 4.0) days (n = 79), P = 0.009.
Oxygen Saturation and Oxygen Therapy
On postoperative night 1, more patients in the postoperative SDB group were given oxygen therapy than the postoperative non-SDB group (89.3% vs 46.5%, P < 0.001). The average oxygen saturation did not change significantly from preoperative baseline, and there was also no difference in average oxygen saturation between the 2 groups (Fig. 2). On postoperative night 3, there was no difference in the percentage of patients receiving oxygen therapy between the 2 groups (Table 2). Compared with preoperative baseline, the average oxygen saturation was significantly lower in postoperative SDB group (Fig. 3, adjusted P = 0.0012); the average oxygen saturation was also significantly lower in the postoperative SDB group than postoperative non-SDB, 92.2% (88.6, 94.2) vs 95% (92.3, 96.2), adjusted P = 0.0102.
Preoperative Baseline Sleep Study Parameters
Preoperative baseline sleep parameters are summarized in Table 3. Compared with postoperative non-SDB group, the patients in postoperative SDB group had a statistically significant higher preoperative AHI (adjusted P = 0.015, Fig. 3) and higher AHI during both REM sleep (adjusted P = 0.030) and non-REM sleep (adjusted P = 0.030). They also had higher supine AHI, 3.8 (interquartile range, 1.8–10.1) vs 1.9 (interquartile range, 0.9–5.8), adjusted P = 0.030. Similarly, total arousal index (adjusted P = 0.012), respiratory arousal index (adjusted P = 0.004), and RDI (adjusted P = 0.015) were higher in the postoperative SDB group versus postoperative non-SDB group. RERA index also tended to be higher in postoperative SDB group (P = 0.043 and adjusted P = 0.086). The total combined cumulative apnea and hypopnea duration as a percentage of sleep time (TAHD %) were also significantly higher in the postoperative SDB group than in the postoperative non-SDB group (adjusted P = 0.021).
Changes in Sleep Study Parameters
The change (postoperative value minus preoperative value) in sleep parameters is summarized in Table 4 and shown in Figure 3. Compared with the postoperative non-SDB group, the postoperative SDB group had a greater increase in AHI (adjusted P = 0.0004, Fig. 3), AHI during either REM sleep (adjusted P = 0.0004) or non-REM sleep (adjusted P = 0.0008), obstructive apnea index (adjusted P = 0.0003), central apnea index (adjusted P = 0.0012), hypopnea index (adjusted P = 0.0004), and RDI (adjusted P = 0.0008). Postoperative SDB group also had a larger increase in respiratory arousal index (adjusted P = 0.0003) and a decrease in RERA index (adjusted P = 0.0087). Oxygen desaturation index, percentage of total apnea hypopnea duration (TAHD %), and the combined index (TAHD% × average oxygen desaturation) also had a greater increase in the postoperative SDB group than the postoperative non-SDB group (Table 4). There was no difference between the 2 groups in postoperative increase in supine sleep percentage, but the increase of AHI during supine sleep was significantly higher in postoperative SDB group versus postoperative non-SDB group (adjusted P = 0.0014).
Factors Associated with the Occurrence of Postoperative Moderate-to-Severe SDB
To explore the factors that may be associated with the occurrence of postoperative SDB, the variables with significant difference between the 2 groups from demographic characteristics and preoperative polysomnography were first evaluated by univariate logistic regression. The response variable was the occurrence of postoperative moderate-to-severe SDB. Age, preoperative AHI, preoperative RDI, total arousal index, respiratory arousal index, RERA index, awake average SpO2, and sleep average SpO2 were found to be significantly associated with the occurrence of postoperative moderate-to-severe SDB (data not shown). Since the correlation coefficient between preoperative RDI and AHI, total arousal index, respiratory arousal index, and total combined cumulative apnea and hypopnea duration as a percentage of sleep time (TAHD %) was greater than 0.5 (data not shown), preoperative RDI was chosen to represent the above variables in multivariate logistic analysis. Based on the same reasoning, the preoperative sleep average SpO2 and age were also selected for multivariate logistic regression analysis. The multivariate logistic regression analysis showed that age and preoperative RDI were significantly associated with the occurrence of postoperative moderate-to-severe SDB, P = 0.018 and P = 0.006, respectively (Table 5). An increase in age of 5 and 10 years elevated the risk of postoperative moderate-to-severe SDB by 26% and 60%, respectively. An increase in preoperative RDI of 5 and 10 events/h magnified the risk of postoperative moderate-to-severe SDB by 67% and almost 3-fold, respectively.
Predicting Postoperative Moderate-to-Severe SDB
The area under the receiver operating characteristic (ROC) curve was 0.728 (95% CI: 0.618–0.839) for multivariate logistic regression analysis including age and preoperative RDI, 0.677 (95% CI: 0.570–0.784) for univariate analysis with preoperative RDI only, and 0.662 (95% CI: 0.545–0.778) for univariate analysis with age only. Since there was no significant difference in the area under ROC between the 3 models, to keep it parsimonious, we explored the predictive performance of preoperative RDI for the occurrence of postoperative moderate-to-severe SDB. Based on the analysis of preoperative RDI values corresponding to the maximal Youden index for 1000 bootstrapped populations, the optimal cutoff for preoperative RDI to predict the occurrence of postoperative SDB was ≥6.8 events/h (Fig. 4), which combines moderate sensitivity (0.710 [95% CI: 0.520–0.858]) and specificity (0.572[95% CI: 0.464–0677]) (Table 6). The sensitivity privilege RDI cutoff (≥4.9 events/h) demonstrated a combination of high sensitivity (0.871 [95% CI: 0.702–0.964]) and low specificity (0.382 [95% CI: 0.281–0.491]), which is clinically useful to identify patients at low risk of developing postoperative SDB. On the other side, specificity privilege RDI cutoff (≥12.4 events/h) had a combination of low sensitivity (0.323 [95% CI:0.167–0.514]) and high specificity (0.865 [95% CI: 0.776–0.928]), which is useful for identifying patients at increased risk of developing postoperative SDB.
In 120 non–sleep apnea patients, 25.8% (95% CI:18.3%–34.6%) developed moderate-to-severe SDB during the postoperative period. The increase in postoperative AHI was mainly driven by obstructive apneas and hypopneas and, to a lesser degree, by central apneas. Higher preoperative RDI and age were significantly associated with the occurrence of postoperative moderate-to-severe SDB. The area under ROC curve was 0.677 (95% CI: 0.570–0.784) for preoperative RDI to predict the occurrence of postoperative moderate-to-severe SDB. The sensitivity privilege cutoff at RDI ≥ 4.9 events/h was able to identify 87.1% (95% CI: 70.2%–96.4%) of patients who developed postoperative moderate-to-severe SDB.
It is surprising that as many as 34.6% of non–sleep apnea patients may develop moderate-to-severe SDB after surgery. These patients were identified by polysomnography with AHI < 5 events/h. It is important to point out that AHI, similar to blood pressure, is not entirely static. That is to say, AHI can vary from night to night in the same patient. Therefore, it is reasonable to ask a clinically relevant question: How comfortable should we feel as clinicians when the preoperative AHI is <5 events/h (i.e., there is no clinically significant SDB)? One night of polysomnography in the preoperative period may not necessarily reflect what will happen postoperatively, particularly because postoperative patients may more likely lie in a supine position. Fluid shifts or medications (i.e., opioids or sedatives) may increase the AHI. Therefore, it is critical to raise clinical awareness of the possibility that non–sleep apnea patients may develop SDB postoperatively.
At present, there is a lack of evidence on the association between postoperatively developed moderate-to-severe SDB and the incidence of postoperative adverse events. More research is needed to determine the clinical implication of postoperative moderate-to-severe SDB. Clinically, some patients may develop respiratory arrest with postoperative opioids. Moderate-to-severe SBD and oxygen desaturation may be part of the clinical picture of respiratory depression. These areas need further exploration to determine the factors predisposing some patients to develop respiratory depression with opioids. Postoperatively developed moderate-to-severe SDB may have a safety implication for surgical patients and a potential practical application for anesthesiologists. Since these patients were identified by polysomnography to be free of sleep apnea, there were most likely no planned perioperative risk precautions and postoperative monitoring to prevent SDB-related adverse events. Conceivably, this group of patients may face a higher risk of adverse events similar to patients diagnosed with sleep apnea. Since the sample size was not large enough to provide a meaningful estimate for the occurrence of postoperative adverse events, we did not analyze the association between postoperative moderate-to-severe SDB and clinical outcomes. However, the postoperative SDB group had a significantly higher percentage of patients receiving oxygen therapy on postoperative night 1 and lower average oxygen saturation on postoperative night 3. The postoperative SDB group also had length of stay that was longer by 1.5 days.
Compared with the postoperative non-SDB group who did not develop postoperative moderate-to-severe SDB, the postoperative SDB patients had a significantly larger increase in obstructive apnea index (adjusted P = 0.0003), central apnea index (adjusted P = 0.0012), and hypopnea index (adjusted P = 0.0004), with the greatest increase in hypopnea index (13.2 events/h). At the same time, postoperative SDB patients had a trend of higher RERA index on preoperative polysomnography than postoperative non-SDB group (5.8 vs 4.1 events/h, adjusted P = 0.086). They also had a significantly larger decrease in postoperative RERA index than the postoperative non-SDB group (−3.3 vs −1.4 events/h, adjusted P = 0.0087). This suggests that some patients developing postoperative moderate-to-severe SDB may suffer from mild forms of upper airway resistance syndrome during the preoperative period. It is possible that during the immediate postoperative period, RERAs indicated a predisposition to more serious events (i.e., hypopneas and apneas) during postoperative sleep, resulting in an increase in the AHI.
Upper airway resistance syndrome is part of the spectrum of SDB characterized by increased airway resistance to breathing during sleep. The primary symptoms include daytime sleepiness and/or fatigue. These patients have respiratory events, which do not meet the criteria for apneas or hypopneas. On the sleep study, they have normal AHI but abnormal RDI due to increased number of RERAs.32 Upper airway resistance syndrome may occur in the absence of clinically significant or disruptive snoring and may be an occult cause of excessive daytime sleepiness.33 Postoperative medications, especially opioids, may increase the arousal threshold and increase upper airway collapsibility, resulting in a conversion of RERAs to apneas and hypopneas.
Opioids are associated with an increase in postoperative AHI.34 The dose of morphine is predictive of central apneas for both sleep apnea and non–sleep apnea patients.34 They can induce central respiratory depression through μ- and κ-opioid receptors.35 Opioids also inhibit central tonic outflow to the primary upper airway dilator, genioglossus muscle.35,36 Compared with the postoperative non-SDB group, the postoperative SDB patients did not receive significantly higher dose of opioids within the first 72 hours postoperatively. However, it is possible that patients with upper airway resistance syndrome are more vulnerable to opioid effects on upper airway function. Also, genetic variations may make some patients especially sensitive to opioids, resulting in more central apneas and hypopneas.37–40
In patients developing moderate-to-severe SDB postoperatively, increase of hypopnea index was most substantial (postoperative SDB group versus postoperative non-SDB group: 13.2 vs 1.1 events/h, adjusted P = 0.0004). To reliably differentiate obstructive hypopneas from central hypopneas in a noninvasive fashion (i.e., without esophageal manometry) remains a clinical challenge in the field of sleep medicine. Further research is needed to establish whether distinguishing central from obstructive hypopneas has clinical relevance as it relates to therapeutic interventions.41,42
Although not statistically significant (P = 0.056), there was a trend for patients developing postoperative moderate-to-severe SDB to have a lower body mass index, which is counterintuitive to the notion that the patients with sleep apnea are usually obese. In leaner patients, it is plausible that the main etiology for the upper airway resistance and postoperative SDB development was cephalometric differences leading to smaller maxillo-mandibular enclosure. Patients with small maxillo-mandibular enclosure are particularly vulnerable to the postoperative decrease in lung volume and perioperative fluid therapy.13,43
There was no significant difference between the postoperative SDB group and the postoperative non-SDB group in the percentage of supine sleep on preoperative polysomnography. However, the patients with postoperative SDB had higher preoperative supine AHI (adjusted P = 0.030) and significantly larger postoperative increase in supine AHI than the postoperative non-SDB group (21.1 vs 1.6 events/h, adjusted P = 0.0014). This suggests that postoperative SBD group was more prone to upper airway collapse in the supine position.
There were more smokers in the postoperative SDB group. Smoking may cause exacerbation of upper airway collapsibility at the level of the uvula.44 Smoking also affects the sleep architecture. Compared with nonsmokers, current smokers had a longer sleep latency, less total sleep time, more stage 1 sleep, and less slow wave sleep.45 Both nicotine in cigarette smoke and withdrawal may contribute to these changes and lead to the subjective experience of nonrestorative sleep.46
We found age to be associated with the occurrence of postoperative SDB. With aging, there is a preferential deposition of fat around the upper airway, suggesting that changes in fat distribution may compromise airway mechanics, independent of overall body fat.47 Opioid doses need to be curtailed in elderly patients because we found that an increase in age of 5 and 10 years increased the odds of postoperative, moderate-to-severe SDB by 26% and 60%, respectively. A recent experimental study reported that infusion of 22 mL/kg of normal saline over 30 minutes during sleep while wearing compression stockings (approximately 1.5 L in a 70 kg individual) caused significantly more increase in the AHI in elderly men compared with younger men (32.2 ± 22.1 vs 2.2 ± 7.1, P = 0.002).48 Therefore, age seems to be a significant factor affecting fluid accumulation in the neck in response to a saline bolus infusion. The significant increase in the AHI experienced in elderly men is in line with our findings. Therefore, perioperative fluid balance may be a significant risk factor in postoperative SDB in addition to age, opioids, and severity of surgery.24,49
Since patients with postoperative SDB may face extra perioperative risk, it is important to identify these patients preoperatively and prepare them for perioperative risk minimization and postoperative monitoring. As discussed above, older patients with upper airway resistance syndrome may account for a significant portion of the patients developing postoperative SDB because the postoperative SDB group were older, had higher RERA index on preoperative polysomnography, and significant postoperative decrease in RERAs. Analysis also found that RDI, which is equivalent to AHI + RERA, was associated with the occurrence of postoperative SDB. We explored the predictive value of RDI for the occurrence of postoperative SDB and found that the optimal cutoff was ≥6.8 events/h (Fig. 4). However, the sensitivity (71% [95% CI: 52%–85.8%]) was not high enough to detect most of the patients who would face increased risk of postoperative SDB. On the other side, the privilege sensitivity cutoff at ≥4.9 events/h identified 87.1% (95% CI: 70.2%–96.4%) patients developing postoperative SDB. Although it had a low positive predictive value (32.9% [95% CI: 22.9%–44.2%]), the privilege sensitivity cutoff of RDI would be a safer choice. This finding needs additional validation in other cohorts. Because RDI includes RERAs as well as hypopneas and apneas and can detect patients with upper airway resistance syndrome who are at increased risk of developing postoperative SDB, RDI may be a better preoperative predictor of postoperative SDB than AHI.
There are several limitations with our study. First, only one preoperative polysomnography was performed. Some patients with mild sleep apnea might have been missed due to night-to-night variability of the AHI. Second, because of the small sample size, we could not examine the effect of postoperative moderate-to-severe SDB on the incidence of perioperative adverse events. Third, the analysis on prediction of postoperative moderate-to-severe SDB was preliminary. The results need to be validated in other cohorts.
In conclusion, 26% of non–sleep apnea patients developed moderate-to-severe SDB after surgery. Some of these patients may have had undetected upper airway resistance syndrome. In these patients, RERAs may have been converted to apneas and hypopneas postoperatively due to either increased upper airway collapsibility and/or respiratory depression induced by opioids and sedatives. Preoperative RDI and age were significantly associated with the occurrence of postoperative moderate-to-severe SDB.
Name: Frances Chung, MBBS, FRCPC.
Contribution: This author designed and conducted the study, obtained funding, and wrote the manuscript.
Attestation: Frances Chung has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
Name: Pu Liao, MD.
Contribution: This author helped design the study, conduct the study, analyzed the data, and wrote the manuscript.
Attestation: Pu Liao approved the final manuscript.
Name: Yiliang Yang, MD.
Contribution: This author helped analyzing the data, and wrote the manuscript.
Attestation: Yiliang Yang approved the final manuscript.
Name: Maged Andrawes, MD.
Contribution: This author was involved in conducting the study.
Attestation: Maged Andrawes approved the final manuscript.
Name: Weimin Kang, MD.
Contribution: This author was involved in conducting the study.
Attestation: Weimin Kang approved the final manuscript.
Name: Babak Mokhlesi, MD, MSc.
Contribution: This author was involved in guiding data analysis and writing the manuscript.
Attestation: Babak Mokhlesi approved the final manuscript.
Name: Colin M. Shapiro, MD.
Contribution: This author supervised the sleep study scoring and helped write the manuscript.
Attestation: Colin M. Shapiro approved the final manuscript.
This manuscript was handled by: Peter S. Glass, MB, ChB, FFA (SA).
We acknowledge the help of Islam Sazzadual, MSc, Babak Amirshahi, MD, Hisham Elsaid, MD, and Hoda Fazel, MD, for their help in the conduct of the study.
1. Peppard PE, Young T, Barnet JH, Palta M, Hagen EW, Hla KM. Increased prevalence of sleep-disordered breathing in adults. Am J Epidemiol. 2013;177:1006–14
2. Ramachandran SK, Kheterpal S, Consens F, Shanks A, Doherty TM, Morris M, Tremper KK. Derivation and validation of a simple perioperative sleep apnea prediction score. Anesth Analg. 2010;110:1007–15
3. Memtsoudis SG, Stundner O, Rasul R, Chiu YL, Sun X, Ramachandran SK, Kaw R, Fleischut P, Mazumdar M. The impact of sleep apnea on postoperative utilization of resources and adverse outcomes. Anesth Analg. 2014;118:407–18
4. Singh M, Liao P, Kobah S, Wijeysundera DN, Shapiro C, Chung F. Proportion of surgical patients with undiagnosed obstructive sleep apnea. Br J Anaesth. 2013;110:629–36
5. Frey WC, Pilcher J. Obstructive sleep-related breathing disorders in patients evaluated for bariatric surgery. Obes Surg. 2003;13:676–83
6. Brown KA, Laferrière A, Lakheeram I, Moss IR. Recurrent hypoxemia in children is associated with increased analgesic sensitivity to opiates. Anesthesiology. 2006;105:665–9
7. Waters KA, McBrien F, Stewart P, Hinder M, Wharton S. Effects of OSA, inhalational anesthesia, and fentanyl on the airway and ventilation of children. J Appl Physiol (1985). 2002;92:1987–94
8. Bernards CM, Knowlton SL, Schmidt DF, DePaso WJ, Lee MK, McDonald SB, Bains OS. Respiratory and sleep effects of remifentanil in volunteers with moderate obstructive sleep apnea. Anesthesiology. 2009;110:41–9
9. Doufas AG, Tian L, Padrez KA, Suwanprathes P, Cardell JA, Maecker HT, Panousis P. Experimental pain and opioid analgesia in volunteers at high risk for obstructive sleep apnea. PLoS One. 2013;8:e54807
10. Eastwood PR, Szollosi I, Platt PR, Hillman DR. Collapsibility of the upper airway during anesthesia with isoflurane. Anesthesiology. 2002;97:786–93
11. Hillman DR, Walsh JH, Maddison KJ, Platt PR, Kirkness JP, Noffsinger WJ, Eastwood PR. Evolution of changes in upper airway collapsibility during slow induction of anesthesia with propofol. Anesthesiology. 2009;111:63–71
12. Dhonneur G, Combes X, Leroux B, Duvaldestin P. Postoperative obstructive apnea. Anesth Analg. 1999;89:762–7
13. Isono S. Obesity and obstructive sleep apnoea: mechanisms for increased collapsibility of the passive pharyngeal airway. Respirology. 2012;17:32–42
14. Lofsky A. Sleep apnea and narcotic postoperative pain medication: a morbidity and mortality risk. Anesthesia Patient Safety Foundation Newsletter. 2002;17:24–5
15. Gupta RM, Parvizi J, Hanssen AD, Gay PC. Postoperative complications in patients with obstructive sleep apnea syndrome undergoing hip or knee replacement: a case-control study. Mayo Clin Proc. 2001;76:897–905
16. Chung F, Yegneswaran B, Liao P, Chung SA, Vairavanathan S, Islam S, Khajehdehi A, Shapiro CM. Validation of the Berlin questionnaire and American Society of Anesthesiologists checklist as screening tools for obstructive sleep apnea in surgical patients. Anesthesiology. 2008;108:822–30
17. Liao P, Yegneswaran B, Vairavanathan S, Zilberman P, Chung F. Postoperative complications in patients with obstructive sleep apnea: a retrospective matched cohort study. Can J Anaesth. 2009;56:819–28
18. Memtsoudis S, Liu SS, Ma Y, Chiu YL, Walz JM, Gaber-Baylis LK, Mazumdar M. Perioperative pulmonary outcomes in patients with sleep apnea after noncardiac surgery. Anesth Analg. 2011;112:113–21
19. Kaw R, Pasupuleti V, Walker E, Ramaswamy A, Foldvary-Schafer N. Postoperative complications in patients with obstructive sleep apnea. Chest. 2012;141:436–41
20. Mokhlesi B, Hovda MD, Vekhter B, Arora VM, Chung F, Meltzer DO. Sleep-disordered breathing and postoperative outcomes after bariatric surgery: analysis of the nationwide inpatient sample. Obes Surg. 2013;23:1842–51
21. Gross JB, Apfelbaum JL, Caplan RA, Connis RT, Coté CJ, Nickinovich DG, Ward DS, Ydens L. Practice guidelines for the perioperative management of patients with obstructive sleep apnea: an updated report by the American Society of Anesthesiologists Task Force on Perioperative Management of patients with obstructive sleep apnea. Anesthesiology. 2014;120:268–86
22. Seet E, Chung F. Management of sleep apnea in adults - functional algorithms for the perioperative period: Continuing Professional Development. Can J Anaesth. 2010;57:849–64
23. Chung F, Mokhlesi B. Postoperative complications associated with obstructive sleep apnea: time to wake up! Anesth Analg. 2014;118:251–3
24. Chung F, Liao P, Elsaid H, Shapiro CM, Kang W. Factors associated with postoperative exacerbation of sleep-disordered breathing. Anesthesiology. 2014;120:299–311
25. Chung F, Liao P, Yegneswaran B, Shapiro CM, Kang W. Postoperative changes in sleep-disordered breathing and sleep architecture in patients with obstructive sleep apnea. Anesthesiology. 2014;120:287–98
26. Chung F, Liao P, Sun Y, Amirshahi B, Fazel H, Shapiro CM, Elsaid H. Perioperative practical experiences in using a level 2 portable polysomnography. Sleep Breath. 2011;15:367–75
27. Practice parameters for the use of portable recording in the assessment of obstructive sleep apnea. . Standards of Practice Committee of the American Sleep Disorders Association. Sleep. 1994;17:372–7
28. Iber C, Ancoli-Israel S, Chesson A Jr, Quan S The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specification. 20071st ed. Westchester, IL: American Academy of Sleep Medicine:17–49
29. Babyak MA. What you see may not be what you get: a brief, nontechnical introduction to overfitting in regression-type models. Psychosom Med. 2004;66:411–21
30. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol. 1996;49:1373–9
31. Cannesson M, Le Manach Y, Hofer CK, Goarin JP, Lehot JJ, Vallet B, Tavernier B. Assessing the diagnostic accuracy of pulse pressure variations for the prediction of fluid responsiveness: a “gray zone” approach. Anesthesiology. 2011;115:231–41
32. Pépin JL, Guillot M, Tamisier R, Lévy P. The upper airway resistance syndrome. Respiration. 2012;83:559–66
33. Kristo DA, Lettieri CJ, Andrada T, Taylor Y, Eliasson AH. Silent upper airway resistance syndrome: prevalence in a mixed military population. Chest. 2005;127:1654–7
34. Blake DW, Chia PH, Donnan G, Williams DL. Preoperative assessment for obstructive sleep apnoea and the prediction of postoperative respiratory obstruction and hypoxaemia. Anaesth Intensive Care. 2008;36:379–84
35. Macintyre PE, Loadsman JA, Scott DA. Opioids, ventilation and acute pain management. Anaesth Intensive Care. 2011;39:545–58
36. Hajiha M, DuBord MA, Liu H, Horner RL. Opioid receptor mechanisms at the hypoglossal motor pool and effects on tongue muscle activity in vivo. J Physiol. 2009;587:2677–92
37. Sadhasivam S, Chidambaran V. Pharmacogenomics of opioids and perioperative pain management. Pharmacogenomics. 2012;13:1719–40
38. Coulbault L, Beaussier M, Verstuyft C, Weickmans H, Dubert L, Trégouet D, Descot C, Parc Y, Lienhart A, Jaillon P, Becquemont L. Environmental and genetic factors associated with morphine response in the postoperative period. Clin Pharmacol Ther. 2006;79:316–24
39. Landau R, Kern C, Columb MO, Smiley RM, Blouin JL. Genetic variability of the mu-opioid receptor influences intrathecal fentanyl analgesia requirements in laboring women. Pain. 2008;139:5–14
40. Zhang W, Chang YZ, Kan QC, Zhang LR, Lu H, Chu QJ, Wang ZY, Li ZS, Zhang J. Association of human micro-opioid receptor gene polymorphism A118G with fentanyl analgesia consumption in Chinese gynaecological patients. Anaesthesia. 2010;65:130–5
41. Iber C. Are we ready to define central hypopneas? Sleep. 2013;36:305–6
42. Randerath WJ, Treml M, Priegnitz C, Stieglitz S, Hagmeyer L, Morgenstern C. Evaluation of a noninvasive algorithm for differentiation of obstructive and central hypopneas. Sleep. 2013;36:363–8
43. Malhotra A, Huang Y, Fogel RB, Pillar G, Edwards JK, Kikinis R, Loring SH, White DP. The male predisposition to pharyngeal collapse: importance of airway length. Am J Respir Crit Care Med. 2002;166:1388–95
44. Kim KS, Kim JH, Park SY, Won HR, Lee HJ, Yang HS, Kim HJ. Smoking induces oropharyngeal narrowing and increases the severity of obstructive sleep apnea syndrome. J Clin Sleep Med. 2012;8:367–74
45. Zhang L, Samet J, Caffo B, Punjabi NM. Cigarette smoking and nocturnal sleep architecture. Am J Epidemiol. 2006;164:529–37
46. Zhang L, Samet J, Caffo B, Bankman I, Punjabi NM. Power spectral analysis of EEG activity during sleep in cigarette smokers. Chest. 2008;133:427–32
47. Malhotra A, Huang Y, Fogel R, Lazic S, Pillar G, Jakab M, Kikinis R, White DP. Aging influences on pharyngeal anatomy and physiology: the predisposition to pharyngeal collapse. Am J Med. 2006;119:72.e9–14
48. Yadollahi A, Gabriel JM, White LH, Taranto Montemurro L, Kasai T, Bradley TD. A randomized, double crossover study to investigate the influence of saline infusion on sleep apnea severity in men. Sleep. 2014;37:1699–705
49. Ramachandran SK, Nafiu OO, Ghaferi A, Tremper KK, Shanks A, Kheterpal S. Independent predictors and outcomes of unanticipated early postoperative tracheal intubation after nonemergent, noncardiac surgery. Anesthesiology. 2011;115:44–53