Preoperative assessment for obstructive sleep apnoea and the prediction of postoperative respiratory obstruction and hypoxaemia.
Subject: Anesthesiologists (Practice)
General anesthesia (Health aspects)
Sleep apnea syndromes (Risk factors)
Sleep apnea syndromes (Diagnosis)
Sleep apnea syndromes (Care and treatment)
Surgery (Complications)
Surgery (Risk factors)
Surgery (Care and treatment)
Authors: Blake, D.W.
Chia, P.H.
Donnan, G.
Williams, D.L.
Pub Date: 05/01/2008
Publication: Name: Anaesthesia and Intensive Care Publisher: Australian Society of Anaesthetists Audience: Academic Format: Magazine/Journal Subject: Health Copyright: COPYRIGHT 2008 Australian Society of Anaesthetists ISSN: 0310-057X
Issue: Date: May, 2008 Source Volume: 36 Source Issue: 3
Topic: Event Code: 200 Management dynamics
Geographic: Geographic Scope: Australia Geographic Code: 8AUST Australia
Accession Number: 188796892
Full Text: SUMMARY

Patients scheduled for elective surgery requiring general anaesthesia and hospital admission were assessed for risk of obstructive sleep apnoea (OSA) using history, body mass index and upper airway examination to determine any relation between OSA risk and the rate of respiratory events after surgery. Anaesthesia and postoperative analgesia were at the discretion of the treating anaesthetist, who was made aware of any suspicion of OSA. Respiratory monitoring for apnoeas (central or obstructive), hypopnoeas and oxygen desaturations was continuous for a 12-hour period on the first postoperative night. We used automated analysis and visual scanning of respiratory recordings, but sleep stages were not assessed. Patients classified as OSA risk had more respiratory obstructive events per hour than controls (38 [+ or -] 22 vs. 14 [+ or -] 10) and an increased proportion of the 12-hour monitored period with oxygen saturation < 90% (7 [+ or -] 12% vs. 2 [+ or -] 5% of the 12-hour period). Perioperative morphine dose was predictive of central apnoeas for both OSA risk and control patients (P=0.002). This study suggests that preoperative suspicion of OSA should lead to increased postoperative monitoring and efforts to minimise sedation and opioid doses It also supports the routine use of supplemental oxygen with patient-controlled opioid analgesia.

Key Words: obstructive sleep apnoea, postoperative respiratory obstruction, hypoxaemia

**********

Obstructive sleep apnoea (OSA) is increasingly recognised in patients prior to surgery and its prevalence is strongly related to obesity, increasing age and cardiovascular disease. Although OSA symptoms are often routinely sought in preoperative assessment in accordance with American Society of Anesthesiologists practice guidelines (1), there are limited data on which to base guidelines for perioperative management or the assessment of perioperative risk (1-3). Often a definite diagnosis of OSA is not made because of difficulty in obtaining a polysomnography study and delaying surgery until a study is performed is generally unreasonable. There is limited direct evidence linking OSA to postoperative morbidity. However, in a retrospective study, serious complications after joint replacement surgery increased from 9% to 24% with a definite diagnosis of OSA (4). Obstructive sleep apnoea is also associated with other conditions likely to increase perioperative risk such as hypertension, heart failure or cardiac arrhythmia, asthma, chronic obstructive airways disease and gastro-oesophageal reflux (5,6). In the postoperative period the incidence of respiratory obstruction associated with hypoxaemia increases, particularly with the use of opioids and with upper abdominal surgery (7-9).

In this study we aimed to investigate the relationship between preoperative assessment of OSA risk and the incidence and severity of respiratory obstructions and oxygen desaturation in the early postoperative period.

METHODS

The study was approved by the Human Research and Ethics Committee of the Royal Melbourne Hospital and carried out according to the National Statement on Ethical Conduct in Research (NHMRC 1999). We screened patients at the preoperative clinic for OSA risk according to symptoms and airway features previously reported to be predictive of OSA 9 (1,10). Patients in OSA risk and control groups were continuously monitored for respiratory events during the first postoperative night.

Eighty-seven patients were recruited from the surgical preadmission clinic, 46 assessed as having OSA risk and 41 controls. Questions and physical examination related to OSA risk included: 1) history of snoring, witnessed apnoeas or excessive daytime somnolence (Epworth sleepiness scale >9) (1,11); 2) difficulty with endotracheal intubation or airway management with previous anaesthesia (3); 3) hypertension or heart failure; 4) body mass index (BMI) >27 kg/[m.sup.2]; 5) thyromental distance <6.5 cm (10); 6) pharyngeal grades III to IV (reduced space at base of tongue on oral examination (10)); 7) absence of overbite (ability to place lower incisors in front of upper if teeth present (10)); and 8) cricomental space > 1.5 cm (perpendicular to line between cricoid and mandible, used to exclude OSA risk (10)).

If the BMI was >27 kg/[m.sup.2] with any two symptoms or with one symptom plus a physical feature, patients were included in the OSA risk group. All OSA related symptoms and physical features were negative in the control group patients.

Patients were included if scheduled for elective surgery using general anaesthesia and requiring hospital admission for >24 hours. Patients were excluded if: 1) they had a previous diagnosis of OSA based on polysomnography or were already treated with continuous positive airway pressure; 2) they were scheduled for cardiothoracic or major upper abdominal surgery; or 3) they were expected to require elective postoperative ventilation or tracheostomy.

A portable respiratory monitor (Somte, Compumedics Ltd, Abbotsford, Victoria, Australia) was connected to patients in the recovery room immediately after surgery and recording continued to 0800 hours on the following day. Data recorded included thoracic and abdominal respiratory movement measured via inductive bands, airflow (nasal pressure), finger pulse oximetry, movement and position sensors and ECG. An oxygen mask was applied over the nasal prongs and postoperative supplemental oxygen was routinely prescribed. The monitor could be disconnected if required for patient care, but generally remained undisturbed overnight. The studies were unattended except for nursing intervention to adjust the monitors.

Anaesthesia and analgesia were at the discretion of the treating anaesthetist, but the anaesthetist was warned of the possible OSA diagnosis if the patient was in the OSA risk group. About 50% of patients received postoperative morphine via a patient-controlled analgesia system. Additional data obtained from the acute pain service charts included hourly pain and sedation scores, respiratory rate and analgesic drug dose. Fentanyl doses given in the perioperative period were converted to morphine equivalent (100 [micro]g=10 mg).

Apnoea was defined as an absence of airflow for >10 s. Apnoeas were considered "obstructive" if they were accompanied by respiratory movements as detected by the thoracic and abdominal induction bands. Apnoeas were considered "central" if respiratory movements were absent. Hypopnoea was defined as a reduction in amplitude of the primary airflow trace of at least 50% of the preceding amplitude and accompanied in <30 s by oxygen desaturation of 4% (12). Automated analysis was first applied to a 12-hour recording period to 0800 hours for each patient and then this was checked by visual inspection of the tracings. A sample of each recording (30 to 60 minutes) was visually inspected and if corrections to automatically identified respiratory events were required (>50%), the complete recording was inspected. In the automated analysis, desaturations due to artefact were defined as a >50% decrease or decrease >10% /s. Adjustment was frequently applied to the number of hypopnoeas, due to a tendency of the automatic analysis to overestimate these. Evidence of significant obstructed breathing was defined as >15 obstructive events per hour (obstructive and mixed apnoeas and hypopnoeas). Desaturations of >4% were recorded and percentage of recording time with saturation <90% and <85% was calculated. Relative time spent in the prone, supine or lateral position was also obtained.

Unpaired t-tests and Fisher's exact test were used to compare continuous and binary variables respectively between the OSA risk and control groups. Multiple linear regression was used to identify risk factors for respiratory events. Ability to fit a linear equation for each dependent variable was described by analysis of variance (Systat 7.0). Logistic regression was used for the binary dependent variable of significant respiratory obstruction (> 15 obstructive events per hour).

RESULTS

A total of 106 patients were approached to participate in this study. Of these, 87 were recruited and 73 subsequently underwent surgery when the respiratory monitor was available. After excluding patients with inadequate recording and one who was unfit for discharge from the operating theatre recovery ward, we included data from 33 patients in the OSA risk group and 30 in the control group.

The risk and control groups were comparable for age, gender and surgery type (Table 1). As expected, the OSA risk group had higher BMI and higher incidence of snoring history and hypertension. The average Epworth Sleepiness Scale score in the OSA risk group (6.4 [+ or -] 3.0) was less than the score >9 suggestive of OSA. However, there was a significant difference between the groups (P=0.021). Although the treating anaesthetist was aware of patients' OSA risk status prior to the surgery, there was no significant difference in the prescribing of patient-controlled morphine or the 12-hour morphine dose between the OSA risk and control groups.

The OSA risk group showed significantly more episodes per hour of obstructive apnoeas and hypopnoeas (Table 2). There was no significant difference in the frequency of mixed and central apnoeas between the groups (Figure 1). The OSA risk group had more events with >4% desaturation per hour (P=0.048) and spent a higher percentage of time below 90% oxygen saturation (7.4%, P=0.033, Figure 2). The duration of oxygen desaturation was related to the rate of both obstructive events and central apnoeas (P <0.001). The amount of time spent supine and the cumulative postoperative morphine dose were similar between the groups (Table 2).

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

Linear regression analysis was performed to estimate factors predictive of the total number of respiratory obstructive events (Table 3). Inclusion in the OSA risk versus control group, evidence of airway abnormality and lying supine for >70% of the monitored period all predicted increased respiratory obstructive events (Table 3). In the OSA risk group BMI was always >27 kg/[m.sup.2] and combined with symptoms or airway abnormality. Coefficients of the linear regression (Table 3) indicate that BMI and airway abnormality were the significant independent predictors of respiratory obstruction. Age above 50 years, daytime somnolence and history of snoring were not predictive in this study, but the incidence of daytime somnolence was low (average Epworth sleepiness scale <6). BMI was predictive of the rate of obstructive apnoea and hypopnoea (P <0.001), but not of central apnoeas (P=0.9). The postoperative cumulative morphine dose was however predictive of central apnoea (P=0.002), but not of obstructive apnoea (P=0.3).

Thirty-three of the total 63 patients monitored had a rate of respiratory obstructions >15 per hour. If this frequency of respiratory obstructive events is considered clinically significant, analogous to diagnosis of OSA, logistic regression analysis (Table 4) showed that being in the OSA risk group and age >50 years were predictive of this rate of obstruction.

DISCUSSION

A simple preoperative OSA risk assessment using BMI >27 kg/[m.sup.2], history of witnessed apnoeas or somnolence and airway abnormality was predictive of significant airway obstruction and oxygen desaturations on the first postoperative night. The additional finding that our control patients spent on average more than 15 minutes of the monitored period with oxygen saturation <90% and had 14 obstructive events per hour was surprising. The rate of hypopnoeas may have been overestimated by our partial reliance on automated analysis of the respiratory recordings, although >50% were scanned manually for the full 12-hour period. The number of hypopnoeas can also be varied according to the desaturation criteria set. We used a change of 4% rather than an absolute value for oxygen saturation in the criteria for hypopnoea. Supplemental oxygen therapy reduces the severity of postoperative nocturnal hypoxaemia but does not prevent episodes of desaturation after major surgery (13). Although supplemental oxygen was prescribed for all patients in the present study this was rarely maintained over the 12-hour monitored period. We assume that after removal of the oxygen mask by the patient, this was rarely corrected by ward staff.

A survey of Canadian anaesthetists suggested that the majority treated between one and five OSA patients per month (2). A quarter reported personal experience of complications in such patients such as postoperative decrease in oxygen saturation, postoperative apnoea and difficult intubation. Although the majority reported asking patients about symptoms or signs of OSA, there was no clinical consensus for the postoperative monitoring of these patients. Our study confirms the high rate of respiratory events in patients considered as high OSA risk and the need for evidence-based practice guidelines for perioperative management of patients with OSA risk.

In the present study, linear regression analysis was used to assess factors predicting respiratory obstructive events in all the patients monitored. Inclusion in the OSA risk group as a result of history and examination obtained at the preadmission clinic was a predictor of respiratory obstructions (coefficient of 17) and observation or past record of any upper airway abnormality was also a predictor (coefficient of 11). Sample size for the study was determined by the likely increase in respiratory events resulting from inclusion in the OSA risk group. It was therefore not powered to detect effects of individual factors such as the Epworth sleepiness scale or pharyngeal grading. OSA is diagnosed and graded according to the overnight rate of respiratory events. We selected a rate of > 15 obstructive events per hour as a clinically significant rate for the early postoperative period and used logistic regression to assess predictors of the severity of obstruction. In contrast to the previous linear regression analysis, this suggested age >50 years as a predictor in addition to our criteria for OSA risk.

The American Society of Anesthesiologists guidelines suggest that a presumptive diagnosis of OSA can be made on patient characteristics such as BMI and abnormal cephalometric measurements (1). We have confirmed a high positive predictive value for postoperative respiratory obstructions when physical examination criteria (10) are associated with high BMI and a history suggestive of apnoeas or daytime somnolence. We suggest that this screening should be routine at the preoperative visit. Although a higher rate of snoring was recorded in the OSA risk group over the monitored period, a preoperative history of snoring was not predictive of obstructive events. This may be due to a high prevalence of snoring in the community and failure to differentiate light snorers from loud habitual snorers, when the latter may be at higher risk of OSA. The high number of postoperative snoring events recorded in the controls may be partly attributable to artefact caused by speech.

In all patients, central depressant drugs diminish the action of the pharyngeal dilator muscles (14). Thus, sedatives or opioid analgesics and the residual effects of anaesthetic agents may worsen OSA by decreasing pharyngeal tone and increasing upper airway resistance (15). This may be a particular problem in obese patients who have a fat-laden pharynx (14). They may also attenuate the ventilatory and arousal responses to hypoxia, hypercarbia and obstruction, worsening the underlying sleep apnoea (16). In the OSA risk group studied, in the postoperative period predominantly obstructive respiratory events caused significantly greater duration and degree of oxygen desaturation compared to the control group. The prediction of central apnoeas by postoperative cumulative morphine dose and its additive effect to obstructive events is an additional reason to minimise opioid analgesia in patients likely to suffer from OSA. Prospective studies comparing the use of opioid versus non-opioid analgesia in high OSA risk patients are warranted.

Accepted for publication on February 8, 2008

REFERENCES

(1.) Gross JB, Bachenberg KL, Benumof JL, Caplan RA, Connis RT, Cote CJ et al. Practice guidelines for the perioperative management of patients with obstructive sleep apnea: a report by the American Society of Anesthesiologists Task Force on Perioperative Management of patients with obstructive sleep apnea. Anesthesiology 2006; 104:1081-1093.

(2.) Turner K, VanDenkerkhof E, Lam M, Mackillop W Perioperative care of patients with obstructive sleep apnea--a survey of Canadian anesthesiologists. Can J Anaesth 2006; 53:299-304.

(3.) Hillman DR, Loadsman JA, Platt PR, Eastwood PR. Obstructive sleep apnoea and anaesthesia. Sleep Med Rev 2004;8:459-471.

(4.) 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.

(5.) Bell RL, Rosenbaum SH. Postoperative considerations for patients with obesity and sleep apnea. Anesthesiol Clin North America 2005; 23:493-500.

(6.) Yaggi HK, Concato J, Kernan WN, Lichtman JH, Brass LM, Mohsenin V Obstructive sleep apnea as a risk factor for stroke and death. N Engl J Med 2005; 353:2034-2041.

(7.) Catley DM, Thornton C, Jordan C, Lehane JR, Royston D, Jones JG. Pronounced, episodic oxygen desaturation in the postoperative period: its association with ventilatory pattern and analgesic regimen. Anesthesiology 1985; 63:20-28.

(8.) Drummond GB. The abdominal muscles in anaesthesia and after surgery. Br J Anaesth 2003; 91:73-80.

(9.) Wu A, Drummond GB. Respiratory muscle activity and respiratory obstruction after abdominal surgery. Br J Anaesth 2006; 96:510-515.

(10.) Tsai WH, Remmers JE, Brant R, Flemons WW, Davies J, Macarthur C. A decision rule for diagnostic testing in obstructive sleep apnea. Am J Resp Crit Care Med 2003; 167:14271432.

(11.) Miletin MS, Handy PJ. Measurement properties of the Epworth sleepiness scale. Sleep Med 2003; 4:195-199.

(12.) Teichtahl H, Cunnington D, Cherry G, Wang D. Scoring polysomnography respiratory events: the utility of nasal pressure and oro-nasal thermal sensor recordings. Sleep Med 2003; 4:419-425.

(13.) Rosenberg J, Wildschiodtz G, Pedersen MH, von Jessen F, Kehlet H. Late postoperative nocturnal episodic hypoxaemia and associated sleep pattern. Br J Anaesth 1994; 72:145-150.

(14.) Benumof JL. Obstructive sleep apnea in the adult obese patient: implications for airway management. J Clin Anesth 2001; 13:144-156.

(15.) Boushra NN. Anaesthetic management of patients with sleep apnoea syndrome. Can J Anaesth 1996; 43:599-616.

(16.) Loadsman JA, Hillman DR. Anaesthesia and sleep apnoea. Br J Anaesth 2001; 86:254-266.

D. W BLAKE *, E H. CHIA ([dagger]), G. DONNAN ([double dagger], D. L. WILLIAMS ([section]) Department of Anaesthesia and Pain Management, Royal Melbourne Hospital, Melbourne Victoria, Australia

* Ph.D., F.A.N.Z.C.A, Associate Professor, Department of Pharmacology, University of Melbourne and Staff Anaesthetist.

([dagger]) Medical Student, University of Melbourne.

([double dagger]) M.B., B.S., F.A.N.Z.C.A., Staff Anaesthetist.

([section]) M.B., B.S., F.A.N.Z.C.A., Associate Professor and Director of Anaesthesia.

Address for reprints: A/Prof D. W. Blake, Department of Anaesthesia and Pain Management, Royal Melbourne Hospital, Parkville, Vic. 3050.
TABLE 1

Patient characteristic      OSA risk             Control (n=30)
                            (n=33)

Age (range) y               61 (23-82)           56 (18-79)

BMI (kg /[m.sup.2])         34.1[+ or -]5.4      23.5[+ or -]2.5

Epworth Sleepiness          6.4[+ or -]3.0       4.4[+ or -]3.1
Scale

Gender, M : F               19:14                15:15

Peripheral : Central        16:17                20:10
Surgery (b)

History of snoring          32:1                 16:14
Yes: No

History of                  24:9                 9:21
hypertension
Yes: No

Patient characteristic      Significance (a)

Age (range) y               P=0.283

BMI (kg /[m.sup.2])         P <0.001

Epworth Sleepiness          P=0.021
Scale

Gender, M : F               P=0.617

Peripheral : Central        P=0.203
Surgery (b)

History of snoring          P <0.001
Yes: No

History of                  P=0.001
hypertension
Yes: No

(a) t test or Fisher's Exact test.

(b) Central surgery-chest wall or abdominal surgery.

TABLE 2
Respiratory events, desaturations, position change and opioid
use for 12-hour monitored postoperative period

Characteristic                             Risk (n=33)

Respiratory events summary

Obstructive apnoeas per hour               14[+ or -]15
Central apnoeas per hour                   2[+ or -]5
Mixed apnoeas per hour                     2[+ or -]5
Hypopnoeas per hour                        20[+ or -]10
Obstructive events per hour (1)            38[+ or -]22
Snoring events per hour                    229[+ or -]125

Desaturation statistics, saturation
levels and Sp[O.sub.2] summary

Number of desaturations                    25[+ or -]24
[greater than or equal to]3% per hour

Number of desaturations                    16[+ or -]20
[greater than or equal to]4% per hour

% Time spent with Sp[O.sub.2] <90% (2)     7.4[+ or -]12.4

% Time spent with Sp[O.sub.2] <85% (2)     1.6[+ or -]4.6

Lowest Sp[O.sub.2] associated with         87.3[+ or -]5.6
a respiratory event (3) (%)

Position change summary

Supine >70% (4) (Yes: No)                  25:8

Use of opioids and sedatives

PCA (Yes: No)                              10:23

Postoperative morphine dose (mg) (5)       12.4[+ or -]14.6

Characteristic                           Control (n=30)   Significance

Respiratory events summary

Obstructive apnoeas per hour             3[+ or -]4       P <0.001
Central apnoeas per hour                 4[+ or -]8       P=0.126
Mixed apnoeas per hour                   1[+ or -]2       P=0.160
Hypopnoeas per hour                      10[+ or -]9      P <0.001
Obstructive events per hour (1)          14[+ or -]10     P <0.001
Snoring events per hour                  154[+ or -]95    P=0.011

Desaturation statistics, saturation
levels and Sp[O.sub.2] summary

Number of desaturations                  13[+ or -]18     P=0.030
[greater than or equal to]3% per hour

Number of desaturations                  7[+ or -]13      P=0.048
[greater than or equal to]4% per hour

% Time spent with Sp[O.sub.2] <90% (2)   2.1[+ or -]4.9   P=0.033

% Time spent with Sp[O.sub.2] <85% (2)   0.3[+ or -]1.2   P=0.122

Lowest Sp[O.sub.2] associated with       88.3[+ or -]     P=0.473
a respiratory event (3) (%)              4.9

Position change summary

Supine >70% (4) (Yes: No)                23:7             P=0.999

Use of opioids and sedatives

PCA (Yes: No)                            11:19            P=0.606

Postoperative morphine dose (mg) (5)     15.4[+ or -]     P=0.516
                                         21.6

Data expressed as mean [+ or -] SD or ratio.

(1) Number of obstructive apnoea, mixed apnoea and hypopnoea
events per hour.

(2) Based on the 12-hour monitoring period.

(3) Any of the four types of respiratory events: obstructive
apnoea, central apnoea, mixed apnoea and hypopnoea.

(4) >70% of time spent in supine position over the 12-hour
monitoring period.

(5) Cumulative morphine dose from the time of patient arriving at
theatre recovery ward until the end of the 12-hour monitoring
period.

Sp[O.sub.2]=peripheral oxygen saturation, PCA=patient-controlled
analgesia.

TABLE 3
Prediction of obstructive events

Factors              Coefficient (a)   Standard error   Significance

OSA risk vs.         16.828            5.316            P=0.002
Control Group

Airway               10.997            5.068            P=0.034
abnormality (b)

Supine >70% (c)      14.261            4.702            P=0.004

History of snoring   -0.881            5.430            P=0.872

Age >50              7.696             4.406            P=0.086

(a) Coefficient from multiple linear regression with analysis of
variance, F-ratio 11.04 (5,57 degrees freedom), P <0.001.

(b) Presence of any signs or symptoms of airway obstruction,
previous difficult intubation, thyro-mental distance <6.5 cm,
pharyngeal grades III-IV, absence of overbite.

(c) Referring to >70% of time spent in supine position over the
12-hour monitoring period.

Obstructive events=number of obstructive apnoeas, mixed apnoeas
and hypopnoeas per hour, CI =confidence interval.

TABLE 4
Prediction of obstructive events (a) > 15 /hour (n= 63)
--logistic regression analysis

Factors     Estimate    Standard    Significance    Odds ratio
                        error                       (95% CI)

OSA Risk    1.980       0.933       P=0.034         7.24
Group                                               (1.16- 45.11)

Age >50     2.141       0.834       P=0.010         8.51
                                                    (1.66- 43.62)

(a) Number of obstructive apnoea, mixed apnoea and hypopnoea
events per hour.

CI=confidence interval.
Gale Copyright: Copyright 2008 Gale, Cengage Learning. All rights reserved.