Three Sudden Postoperative Respiratory Arrests Associated with Epidural Opioids in Patients with Sleep Apnea

Current prominent textbooks on anesthesiology mention sleep apnea syndrome (SAS) only briefly. However, three patients in Illinois (two at one hospital) had sudden postoperative arrests associated with epidural opioids and sleep apnea. We therefore reevaluated the literature, which suggests that epidural opioids are the therapy of choice for patients with sleep apnea [1]. Of the 15 patients with SAS reported in the literature, 10 had severe respiratory problems due to postoperative analgesia. Only one report was associated with epidural analgesia, to which we now add three. Based on the experiences we now describe (Case Reports 1-3) and review of the literature (Case Reports 4-18), we believe that SAS patients are at particularly high risk of postoperative respiratory depression from any mode of analgesic therapy. Our theory is based on a review of a series of cases and not on mathematical estimates of relative risk. Because of our conclusion, we suggest guidelines for the anesthetic management of such patients. Case Report 1 A 41-yr-old woman was admitted to BM Hospital on October 27, 1993 for revision of a right hip arthroplasty. Her medical history consisted of juvenile rheumatoid arthritis, morbid obesity, and sleep apnea, with a tracheotomy in 1985 and a tracheostomy closure in 1988. She also had a cholecystectomy in 1987. According to a physical examination and preoperative clinical evaluation performed 1 wk earlier and prior to surgery, the patient was 160 cm tall and weighed 94.5 kg (body mass index [BMI, kg/m2] of 36.9). The patient underwent revision of a right hip arthroplasty facilitated by general anesthesia and controlled ventilation, followed by epidural administration of 0.5% bupivacaine and fentanyl (10 micro g/mL) at the rate of 7 mL/h. Postoperative Night 1 was uneventful except for vomiting, which was eliminated by administration of prochlorperazine. Postoperative Day/Night 2 was relatively unremarkable; nausea with movement necessitated one administration of 10 mg of prochlorperazine. The patient continued to receive epidural analgesia consisting of 0.06% bupivacaine (7 mL/h) plus fentanyl (10 micro g/mL). At 8 AM on Postoperative Day 3, the patient was alert, oriented, and behaving appropriately. Her pain score was 0 (highest possible 10). At 11 AM, the patient was asleep but easily arousable (pain score 0). She complained of a head cold but denied shortness of breath or chest pressure. At 12 PM (>48 h after surgery), the patient was found to be unresponsive, with slow, deep respirations (Kussmaul breathing). Arterial blood pressure was 84/40 mm Hg, heart rate was 70 bpm, and respirations were 18 breaths/min. Administration of 0.4 mg of naloxone produced no response in arterial blood pressure, level of consciousness, or respiration. The respiratory pattern consisted of 15-s periods of rapid breathing (20-30 breaths/min) combined with 20- to 40-s periods of apnea. The patient then had cardiac and respiratory arrest, necessitating ventilation by mask and endotracheal intubation. Resuscitation included repeated bolus doses of naloxone, controlled ventilation, and defibrillation. Defibrillation converted ventricular fibrillation to normal sinus rhythm. Cardiac arrest occurred three times later in the day, and the patient died. Autopsy did not reveal the cause of death. Case Report 2 A 66-yr-old man (height 170 cm, weight 105 kg, BMI, 36.3) underwent bilateral knee arthroplasty (also at BM Hospital) on May 2, 1994. His medical history and preoperative clinical evaluation revealed a 25-yr history of hypertension that was treated with diltiazem and enalapril, a deviated nasal septum, and sleep apnea. Preoperative blood gas values were 61 mm Hg for alveolar oxygen partial pressure (PaO2), 49 mm Hg for alveolar carbon dioxide partial pressure (PaCO2), and 7.42 for blood pH while breathing room air. Surgery was uneventful, involving spontaneous ventilation, epidural anesthesia with 2% lidocaine, and intraoperative sedation with 2 mg of midazolam. In the recovery room, an epidural infusion of 0.06% bupivacaine (7 mL/h) and fentanyl (10 micro g/mL) was begun. Pain scores ranged from 1 to 7 (highest possible 10). At 8:50 PM the rate of epidural infusion was increased to 8 mL/h, and at 10 PM, it was increased to 9 mL/h. Pain and vital signs were stable for several hours. At 11 PM and subsequently during the night, urinary output was 10-20 mL/h, for which the patient was given 500-mL infusions of lactated Ringer's solution; his intravenous rate increased to 150 mL/h after the infusions. At 4:30 AM on Postoperative Day 2, the patient complained of itching and was given 50 mg of diphenhydramine. At 8:20 AM, the patient was seen by the pain service attending physician. Respiratory rate was 20 breaths/min, arterial blood pressure was 110/90 mm Hg, and heart rate was 96 bpm. The pain score was 0 (highest possible 10), the sedation score was 0 (highest possible 3), and the patient was alert and oriented. At 9 AM, the orthopedic service recorded similar mental state and vital signs. The administration of oxygen by nasal prongs that had been initiated in the operating room was discontinued. At 1:45 PM, oxygen saturation (measured by pulse oximeter) was 86%. Two liters of oxygen was given by nasal prongs, and oxygen saturation increased to 98%. At 3 PM, in response to low urinary output (72 mL for the previous 8 h) and increased bleeding from the drainage tubes, the patient was given an infusion of 500 mL of lactated Ringer's solution and 20 mg of furosemide. Although the patient was troubled at night by an upset stomach, he did not complain of pain. Titration to reduce the rate of epidural infusion was not instituted on the judgment of the attending physicians and the lack of evident respiratory depression. At 6:30 AM on Postoperative Day 3, the patient was seen by the orthopedic service. At 7:00 AM (>40 h after surgery), the patient was seen by nurses and orthopedic physicians, who stated that he was sleeping; neither service disturbed him. However, at 7 AM, arterial blood pressure was 105/50 mm Hg, heart rate was 120 bpm, and respiratory rate was 14 breaths/min. Urinary output had been 300 mL for the last 8 h. At 7:25 AM, the patient was found unresponsive and could not be resuscitated. No details about the resuscitation attempt are available. Requests for autopsy were denied by the family and county medical examiner. Case Report 3 A 47-yr-old woman was admitted to LP Hospital on May 27, 1994 for ventral hernia repair. Her medical history was unremarkable, and no history of snoring or sleep apnea was reported. On physical examination, the patient was 160 cm tall and weighed 70.5 kg (BMI 27.5). Physical examination was otherwise unremarkable. No mention of a difficult airway, right ventricular heaves, or other respiratory abnormalities was made. The patient underwent an uneventful anesthetic and surgical procedure. The anesthetic consisted of epidural administration of 15 mL of a solution containing 0.1% bupivacaine and fentanyl (10 micro g/mL) for the surgical procedure, followed by continuous epidural infusion of a solution containing 0.1% bupivacaine and fentanyl (10 g/mL) at the rate of 8 mL/h. The patient continued to do well throughout the afternoon and evening. The patient was last seen regularly by the nursing staff at 6 AM on Postoperative Day 2, at which time vital signs were: arterial blood pressure 110/50 mm Hg, respiratory rate 14 breaths/min, and heart rate 78 bpm. The patient did not complain of pain. Slight and unimportant changes occurred in the vital signs between 12 midnight and 6 AM. The patient was receiving a continuous rate infusion and had not activated a dose by the patient-controlled analgesia PCA mode. At 7 AM on Postoperative Day 2 (>29 h after initiation of the epidural), the patient was found breathless, with cool skin, and cyanotic. She was asystolic, and advanced cardiac life support was given. Blood gas values 15 min after the start of the arrest were: pH 7.1; PaCO2 48 mm Hg, and PaO2 480 mm Hg while receiving 100% oxygen. Electrolyte levels and complete blood count were normal. The patient was resuscitated but died later. No details about the resuscitation are available. In retrospect, the patient's husband said that she had been evaluated 2 yr earlier for difficulty in sleeping and excessive daytime sleepiness. That evaluation resulted in a diagnosis of sleep apnea. He also stated that the patient had a history of snoring. Autopsy did not reveal the cause of her arrest. Detailed checking of the infusion pump revealed no abnormalities in function or in the infusion of bupivacaine or fentanyl. Case Reports in the Literature (Table 1) summarizes 18 case reports of sleep apnea syndrome (SAS): the 3 that we now report (Cases 1-3) plus 15 from the literature (Cases 4-18). Of the 18, 13 are reported complications (Cases 1-13) and are listed first; 5 (Cases 14-18) made no report of complications and are listed last. Table 2 and Table 4 show details of the 15 case reports of SAS that exist in the literature. The cases with reports of complications are listed first (Cases 4-13), followed by the cases with no report of complications (Cases 14-18). Table 3 lists the characteristics of the sleep apnea patients described in our three case reports.Table 1: Summary of 18 Case Reports on Management of Pain with Narcotics in Surgical Patients with Sleep Apnea Syndrome and Respiratory ComplicationsTable 2: Case Reports 4-18: Anesthesia, Pain Management, and Respiratory Complications in 15 Surgical Patients with Sleep Apnea Syndrome (SAS)Table 4: Table 2 ContinuedTable 3: Characteristics of Sleep Apnea PatientsDiscussion SAS is defined as the occurrence of at least 30 periods of apnea (cessation of airflow for at least 10 seconds) over a seven-hour period of nocturnal sleep; these periods of apnea must occur in both rapid eye movement (REM) sleep and non-REM sleep [13,14]. REM sleep, however, is associated with a greater number of apneic episodes than is deep non-REM sleep [15]. Sleep apnea is divided into three categories. These subtypes of apneic episodes rarely occur in isolation [1,13,15] and represent a continuum of destabilized ventilatory control [16]. Obstructive (peripheral) sleep apnea and mixed sleep apnea are characterized by obstruction of the upper airway and the presence of continuous respiratory effort. In contrast, central sleep apnea is rare (<10% of sleep apneas) and is characterized by loss of the central nervous system drive to respiration [1,13,14,17] without respiratory muscle movement [14,17] or respiratory muscle activity on electromyography. Mixed apnea occurs when a central respiratory pause is followed by obstructed ventilatory efforts. Approximately 4% of middle-aged adults have sleep apnea, the male/female ratio being 2:1 [18]. A 1976 study reported that 90% of patients who had SAS were Caucasians and 10% were African American [13]. However, when a 1995 investigation studied the elderly, the relative risk of severe sleep-disordered breathing was two times higher for African Americans than for Caucasians [19]. There is also strong evidence for a familial basis for SAS [20] and other conditions involving decreased chemosensitivity and respiratory failure [21]. The lumen of the pharynx is smaller in SAS patients than in nonapneic, nonsnoring controls [14,22]. The condition occurs because of increased pharyngeal compliance-the primary pathophysiologic characteristic of sleep apnea [23]-caused by relative hypotonia in the genioglossus and geniohyoid muscles and suggested by electromyographic findings. These muscles are important in keeping the upper airway patent. Their relaxation, as in central sleep apnea, causes obstruction of the upper airway because of rostral movement of the tongue into the pharyngeal lumen or sucking of the tongue into the airway by subatmospheric inspiratory pressure. In several cases of upper airway occlusion, nasopharyngeal intubation allowed return to a regular respiratory pattern. Despite the fact that the "floppy," easily collapsible pharynx is the most likely site of obstruction, any region from the nasopharynx to the larynx can be involved [24]. In addition, supine positioning reduces both the cross-sectional area of the pharynx and functional residual capacity, which emphasizes the importance of positioning. Increasing functional residual capacity has been shown to reverse obstructive sleep apnea [25]. During periods of obstructive sleep apnea, hypoxemia is the major stimulus for arousal, as carbon dioxide retention does not rise to a significant level while arterial oxygen partial pressure falls rapidly [13]. After arousal, with the subsequent return of breathing and relief of hypoxemia, the patient usually falls asleep quickly and has no awareness of the event. These multiple episodes of apnea reduce arterial oxygen saturation to less than 80%, a condition that is presumably responsible for the accompanying cardiac arrhythmias [1]. Patients with severe cardiac arrhythmia during periods of apnea often have unremarkable electrocardiograms while awake [26]. More than 50% of SAS patients have other hemodynamic changes, such as cyclic elevations in pulmonary artery pressure and systemic hypertension [11,13,27]. One report showed that for some patients, effective treatment of sleep apnea eliminated systemic hypertension [28]. Patients with sleep apnea present with intensive snoring, gasping, and choking while asleep; morning headache; histories of impaired daytime performance; and intellectual deterioration or personality changes. These symptoms are accompanied by cardiopulmonary dysfunction [1,13,15,27]. None of these features is present invariably [13]. Even small children can have SAS [6], especially if dysmorphologic conditions exist. Approximately 20%-40% of elderly people seem to have subclinical SAS. The classic presentation of obstructive sleep apnea, the pickwickian syndrome (obesity, hypersomnolence, cor pulmonale, and periodic breathing with hypoventilation), was found in only 5% of the sleep apnea patients described by Guilleminault et al. [13]. Of the 18 patients described in our table, 16 were male and 2 were female (Patients 1 and 3); 16 were middleaged, and at least 15 were obese. SAS is the most likely diagnosis in patients presenting with excessive daytime sleepiness. Because of the poor relationship between episodes of apnea and their signs and symptoms, symptoms such as excessive daytime sleepiness, fatigue, and lethargy can be sought for specifically with sleep characteristics obtained from the bed partner [29-32]. The history often reveals that the patient takes several nonrefreshing daytime naps a day [13]. Central sleep apnea can be associated with myopathic conditions [33] and neurologic disorders such as diabetic neuropathy or, less commonly, brainstem disorders. The differential diagnosis of excessive daytime sleepiness also includes narcolepsy, insufficient sleep, effects of medications, and central alveolar hypoventilation syndrome (Ondine's curse) [13]. Treatment consists of medical, surgical, and behavioral measures. For obese patients with obstructive SAS, weight loss to reduce fat deposits of the upper airway and tonsillar pillars may be important therapy [15,24]. Septal deviation or nasal polyps can be treated surgically [1]. A commonly performed procedure, uvulopalatopharyngoplasty [24], results in a great variability of improvement. Although tracheostomy has been effective in reversing cardiac arrhythmias [34], surgical treatment may reveal an unexpected central component [5,7,11]. Applying continuous positive airway pressure (CPAP, 2.5-20 cm H2 O) through the nares to prevent subatmospheric pharyngeal pressure has become the treatment of choice for a pneumatic splint to the nasopharyngeal airway [9,35]. For snoring adults, a negative airway pressure of -2 to -10 cm H2 O is sufficient for upper airway closure, whereas nonsnoring subjects may require a negative pressure of -25 cm H2 O [36]. The use of tongue-retaining and orthodontic devices has had limited success and a low compliance rate. Also, simple techniques such as the modification of sleep behavior to minimize sleeping on the back [29] (e.g., sewing a tennis ball in nightwear) can be included. Medical treatment of central apnea with use of respiratory stimulants such as nicotine or protriptyline or electrophrenic pacing is generally disappointing [31,37]. The use of respiratory depressants such as alcohol, narcotics, and sedative drugs is proscribed for all conditions related to sleep apnea [14,15,24,38,39] because benzodiazepines and ethanol decrease genioglossal activity selectively [40]. Preoperative Evaluations and Procedures For all patients with known or suspected SAS, it is particularly important to review previous medical records, to watch for related systemic disorders with emphasis on cardiopulmonary function, and to take a proper sleep history, including questioning the bed partner [30]. For Case Reports 3-6, the diagnosis of SAS was made only after the occurrence of respiratory depression. In Cases 3 and 5, it might have been possible to diagnose SAS by taking a more detailed history. Sedative Premedications The importance of sedative in SAS patients is evident in the case the SAS patient is at risk of respiratory depression. For all the cases that reported no complications (Cases 14-18) except one (Case 18, in which nasal was given and no sedative was For all the cases with except two (Cases 6 and sedative was (Cases and to to respiratory depression. For Cases given epidural to a and support and respiratory may be as as respiratory depressants or their are present [1]. consists of pulse or, of the patient to an intensive In retrospect, might have been for Patients while were receiving epidural nasal and it after for has been suggested based on the with 14 patients given and postoperative without complications occur only after the patient is and the upper airway However, Case Reports and 11 show that airway obstruction and ventilatory depression can occur when the patient's has been and the patient is alert and breathing Administration of can be for treatment of airway be with as the patient may be on the respiratory oxygen does not prevent apnea but with only limited and the of apnea Postoperative Respiratory and Respiratory depression with apnea occurs in SAS patients in the postoperative and apnea may be more likely with the use of case reports (Cases involved epidural The ventilatory depression that occurs in SAS patients given epidural consists of an respiratory related to levels after from the epidural and a respiratory depression between 4 and 12 after to rostral of the In Case Reports respiratory depression occurred after a period of to 48 after the start of epidural administration of after earlier of stable vital signs. respiratory depression has also been reported after intravenous administration of The use of partial drugs such as and mixed drugs to severe pain may have the of the risk of severe respiratory but we could no controlled of use at doses in SAS patients. reduce the postoperative for One can use drugs as a component of analgesia, as by after major There is some the use of epidural opioids in SAS patients. et al. (Case Report epidural opioids for SAS patients, whereas their use is reported by et al. (Case Report In addition, when respiratory depression associated with PCA at his he that repeated epidural of were for sleep apnea patients than patient-controlled administration of PCA or epidural opioids be for patients with SAS only when or anesthesia with is not or and only when patients are in an to epidural anesthesia for pain therapy have been described The three patients described in the case reports to a dose because were and pain. with the of pain by after the have continued with a smaller On the other the patients showed of the signs of respiratory that decreased ventilatory which to severe respiratory depression after epidural The of the signs is one between patients with sleep apnea and patients who analgesia with epidural to the infusion rate of epidural that a patient's pain