1 Hypoventilation Syndromes

The ASA Closed Claims Project found that 4% of 1175 closed claims resulted from critical respiratory events in the PACU. The highest proportion was attributed to inadequate ventilation, and many of these patients died or suffered brain damage.³¹

Many clinical conditions may give rise to postoperative ventilatory failure. A multicenter, prospective survey in France that looked at almost 200,000 general anesthetics administered between 1978 and 1982 found that postoperative respiratory depression accounted for 27 of 85 respiratory complications that were life-threatening or had serious sequelae. These complications were responsible for seven deaths and five cases of hypoxic encephalopathy.³² A respiratory rate of less than 8 breaths/min was observed by PACU nurses among 0.2% of 24,000 patients after general anesthesia.²²

Hypoventilation may be mediated centrally at the level of the upper motor neuron, anterior horn cell, lower motor neuron, neuromuscular junction, or respiratory muscles. Clinical correlates include central sleep apnea, carotid endarterectomy,³³ medullary injuries, demyelinating disorders, direct injury to peripheral nerves, poliomyelitis, Guillain-Barré syndrome, motor neuron disease, myasthenia gravis, and botulism. Hypoventilation may result from the loss of lung or pleural elasticity, diaphragmatic splinting caused by abdominal pain or distention, thoracic deformities such as kyphoscoliosis, or multifactorial entities such as morbid obesity and severe chronic obstructive pulmonary disease. Rarely, hypercapnia results from excess carbon dioxide production or a marked increase in physiologic dead space.

The residual effects of anesthetic drugs contribute to inadequate postoperative ventilation.^34–36 It may be aggravated by incomplete reversal of neuromuscular blockers,³⁷ hypocalcemia or hypermagnesemia, or the administration of other drugs, including antibiotics, local anesthetics, diuretics, and calcium channel blockers, which may potentiate neuromuscular blockade.

2 Hypoxemic Respiratory Failure

A review of the many causes of postoperative hypoxemia is beyond the scope of this chapter, but it can occur because of hypoventilation, a low inspired oxygen concentration, ventilation-perfusion mismatch, right-to-left shunting, increased oxygen consumption, diminished oxygen transport, or rarely, an impairment of oxygen diffusion. These events are more likely in some clinical situations because of preexisting medical conditions or anesthetic and surgical interventions. If sufficiently severe, there may be a requirement for continuous positive airway pressure (CPAP) or reintubation and mechanical ventilation.

3 Inability to Protect the Airway

ICU or postoperative patients may be unable to protect the airway because of preexisting obtundation, neurologic injury, or the effects of residual anesthesia. In the latter case, it may be possible to temporize by turning patients on their sides, altering their positions (head up or head down) or by reversing residual medications with antagonists. These measures may not restore airway competency, and lack of resolution may necessitate reintubation.

4 Failure of Pulmonary Toilet

Inadequate clearance of pulmonary secretions may result from a depressed level of consciousness with impaired airway reflexes, overproduction of secretions, alteration of sputum consistency leading to inspissation and plugging, impaired mucociliary clearance, or inadequate neuromuscular reserve. These problems may result in atelectasis or pneumonia with attendant hypoxemic respiratory failure. Alterations in pulmonary mechanics may also lead to hypercapnia, necessitating reintubation.

5 Inadvertent Extubations

Inadvertent extubations may result from movement of or by the patient with an inadequately secured tracheal tube. Intraoperatively, this may occur in the prone position, when the airway is shared with the surgeon, when the head and neck are extended, when draping obscures the view, or when drapes adhere to the endotracheal tube (ETT) or circuit and are carelessly removed. In the ICU, extubation may occur when the patient is repositioned for a radiograph or routine nursing care. Patients insufficiently sedated are at greater risk for deliberate self-extubation,³⁸ but those more heavily sedated are more likely to require reintubation.³⁹ Fastidious attention to securing and supporting the ETT and breathing circuit is essential. Self-extubation may occur during emergence from anesthesia, when the patient is confused, agitated, and distressed, which promotes premature extubation. In the ICU, it may not be possible to know whether a self-extubation is accidental or deliberate, but many of these patients require reintubation,⁴⁰ are more likely to exhibit post-extubation stridor, and may need multiple intubation attempts, increasing the likelihood of esophageal intubation and death.^41–43

6 Entrapment

The tracheal tube may become entrapped due to an inability to deflate the cuff,^44,45 or there may be difficulties with the pilot tube.^46,47 Difficulties include a crimped pilot tube, a defective pilot valve, and fixation of the tracheal tube by Kirschner wires,⁴⁸ screws,⁴⁹ ligatures,⁵⁰ or entanglement with other devices.^51,52 Entrapment can also occur during a percutaneous tracheostomy.⁵³ Mechanical obstruction of an entrapped tube is a life-threatening complication. Partial transection of the tracheal tube by an osteotome during a maxillary osteotomy has resulted in the partially cut tube forming a barb that caught on the posterior aspect of the hard palate.⁵⁴ One report of tube entrapment with fatal consequences involved a Carlens tube that was inadvertently sutured to the pulmonary artery.⁵⁵ Lang and colleagues recommended routine intraoperative testing for tracheal tube movement when fixation devices are used in proximity to the airway.⁴⁹ Uncertainty about tube movement should prompt fiberoptic examination before emergence from general anesthesia.

7 Hypertension and Tachycardia

Transient hemodynamic disturbances accompany extubation in most adults. These responses may be prevented by deep extubation,⁵⁶ insertion of a laryngeal mask airway (LMA) before emergence, or attenuated by concurrent medication.^57–59 Most healthy patients not on antihypertensive agents or other cardioactive drugs exhibit increases in heart rate and systolic blood pressure of more than 20% in association with extubation.⁶⁰ After coronary artery bypass surgery, these changes tend to be transient, lasting 5 to 10 minutes, and they usually are not associated with electrocardiographic evidence of myocardial ischemia.⁶¹ Coronary sinus lactate extraction measurements, however, indicate that among patients with poor cardiac function, extubation may be associated with myocardial ischemia.⁶² Patients with inadequately controlled hypertension, carcinoid, pheochromocytoma, hypertension associated with pregnancy, or hyperthyroidism may be expected to display even greater increases in blood pressure in response to tracheal extubation. The specific strategies needed to attenuate these usually transient changes are dictated by the clinical context. The strategies, which are not universally effective, include the use of intracuff,^63–65 intratracheal,⁶⁶ or intravenous lidocaine^61,67,68; beta-blockers^60,69–71; dexmedetomidine^72,73; and nitrates.

8 Intracranial Hypertension

Tracheal intubation and suctioning are associated with a rise in intracranial pressure. Extubation probably is associated with comparable or even greater increases in intracranial pressure. There is evidence, albeit contradictory, that intravenous and endotracheal lidocaine attenuate this effect.^74,75

9 Intraocular Pressure

Madan and colleagues compared the intraocular pressure changes of tracheal intubation and extubation in children with and without glaucoma.⁷⁶ They observed significantly greater increases 30 seconds and 2 minutes after deep extubation compared with the corresponding times after uncomplicated intubations. These differences were seen in both groups of children. It is likely that significant increases in intraocular pressure observed after deep extubation would have been even higher had extubation occurred after recovery of consciousness. Lamb and coworkers observed similar effects of extubation on intraocular pressure in adults and commented that this increase could be prevented by using an LMA rather than a tracheal tube.⁷⁷

10 Coughing

Coughing on emergence from general anesthesia is virtually ubiquitous, particularly when an ETT is used.⁷⁸ No difference between smokers and nonsmokers is observed. Although coughing is a protective reflex, it can be particularly troublesome in the setting of ophthalmologic, neurologic, oropharyngeal, and neck surgery.

Several strategies have been proposed to minimize coughing, including deep endotracheal extubation, primary use of or conversion to an LMA,^57,79,80 use of the sedative dexmedetomidine,⁷³ intravenous or topical application of a local anesthetic to the vocal folds, and use of intracuff lidocaine.^63,65,67,81 However, coughing on emergence is common and relatively benign for most patients.

The emergence of severe acute respiratory syndrome (SARS) and the high prevalence of drug-resistant droplet or airborne diseases in some locales make coughing potentially hazardous to the airway manager. In 2003, Toronto was the North American epicenter for SARS, and coughing assumed life-threatening proportions for medical personnel.^82–87 Patients with cough, fever, and pulmonary infiltrates were dangerous, and coughing on emergence was not protective, but rather posed a threat by dispersing infectious respiratory droplets on those in the patient’s vicinity. The strategies adopted at that time have been largely relaxed, although they may again become necessary when new risks threaten patients and care providers.

11 Laryngeal Edema

Several of the complications of endotracheal intubation do not become apparent until after extubation occurs.⁸⁸ Glottic or tracheal injury may occur despite a good laryngeal view or during awake fiberoptic intubation.^89–91 Anatomic or functional laryngeal problems are more likely to develop as a consequence of a difficult or prolonged intubation attempts.²² Possible airway injuries include laryngeal edema, laceration, hematoma, granuloma formation, vocal fold immobility, and dislocation of the arytenoid cartilages.⁹²

Glottic edema has been classified as supraglottic, retroarytenoidal, and subglottic.⁹³ Supraglottic edema may result in posterior displacement of the epiglottis, reducing the laryngeal inlet and causing inspiratory obstruction. Retroarytenoidal edema restricts movement of the arytenoid cartilages, limiting vocal cord abduction on inspiration. Subglottic edema, a particular problem in neonates and infants, results in swelling of the loose submucosal connective tissue that is confined by the non-expandable cricoid cartilage. In neonates and small children, this is the narrowest part of the upper airway, and small reductions in diameter result in a significant increase in airway resistance. In children, laryngeal edema is promoted by a tight-fitting ETT, traumatic intubation, intubation longer than 1 hour, coughing on the tracheal tube, and intraoperative alterations of head position.⁹⁴ Koka and coworkers found an incidence of 1% among children younger than 17 years. Laryngeal edema should be suspected when inspiratory stridor develops within 6 hours of extubation. Management of laryngeal edema depends on its severity. Treatment options include head-up positioning, supplemental humidified oxygen, racemic epinephrine, helium-oxygen administration, reintubation, and tracheostomy.

Clinical studies of children and adults that evaluated the role of prophylactic corticosteroids in the prevention of post-extubation stridor have yielded contradictory findings.^95–98 In the setting of adult ICUs, a large, multicenter, prospective, randomized, double-blind trial of methylprednisolone (20 mg administered 12 hours before and every 4 hours until extubation) versus placebo found that steroids significantly reduced post-extubation laryngeal edema (11 of 355 versus 76 of 343, P < 0.0001) and required reintubation due to laryngeal edema (8% versus 54%, P = 0.005). Laryngeal edema was defined by any two of stridor, inspiratory prolongation, and use of accessory muscles. It was classified as severe if reintubation was required within 36 hours. Laryngoscopy was not performed unless reintubation was required. Factors associated with a greater risk of laryngeal edema included female sex, shorter height, larger-diameter ETT, admission for trauma, and shorter duration of intubation. The absence of pretreatment with methylprednisolone was associated with a hazard ratio of more than 8.⁹⁹ It is possible that the benefits of steroids are restricted to high-risk populations and require administration of multiple doses.^{100, 101} Contradictory findings may relate to risk factors in the study populations and variability of dosage regimens.

An alternative classification has been proposed for laryngotracheal injury after prolonged intubation.^102–104 Immediate post-extubation airway obstruction results from glottic and subglottic granulation tissue, which may swell on removal of the ETT. Posterior glottic and subglottic stenosis due to contracting scar tissue results in increasing obstruction weeks or months after extubation. Benjamin found that fiberoptic evaluation or laryngoscopy with the tube in situ was of limited value.¹⁰² An ETT obscures the view of the posterior glottis and subglottis. These lesions were best identified using rigid telescopes with image magnification during general anesthesia. This approach permitted anticipation of problems and development of a management strategy.

12 Laryngospasm

Laryngospasm is believed to be a common cause of post-extubation airway obstruction, particularly in children.¹⁰⁵ Even in adults, Rose and colleagues found that it accounted for 23.3% of critical postoperative respiratory events, although the diagnosis was presumptive.¹⁰⁶ Olsson and Hallen observed an increased incidence among patients presenting for emergency surgery, those requiring nasogastric tubes, and patients undergoing tonsillectomy, cervical dilation, hypospadias correction, oral endoscopy, or excision of skin lesions.¹⁰⁵ A variety of triggers are recognized, including vagal, trigeminal, auditory, phrenic, sciatic, and splanchnic nerve stimulation; cervical flexion or extension with an indwelling ETT; or vocal cord irritation from blood, vomitus, or oral secretions.¹⁰⁷ A risk assessment questionnaire was used in a study to prospectively evaluate almost 10,000 children undergoing general anesthesia. A positive history of nocturnal dry cough, exertional wheezing, or more than three wheezing episodes in the prior 12 months was associated with a fourfold increase in the risk of laryngospasm in the PACU and a 2.7-fold increased risk of airway obstruction during surgery or in the PACU.¹⁰⁸ Twice as many children were managed with an LMA than an ETT, and an equal number had their devices removed awake and asleep. The depth of anesthesia at the time of device removal did not influence the incidence of laryngospasm.

Laryngospasm involves bilateral adduction of the true vocal folds, vestibular folds, and aryepiglottic folds that outlasts the duration of the stimulus. This is protective to the extent that it prevents aspiration of solids and liquids. It becomes maladaptive when it restricts ventilation and oxygenation. The intrinsic laryngeal muscles are the main mediators of laryngospasm, and they include the cricothyroids, lateral cricoarytenoids, and thyroarytenoid muscles. The cricothyroid muscles are the vocal cord tensors, an action mediated by the SLN. Management of laryngospasm consists of prevention by either extubating at a sufficiently deep plane of anesthesia or awaiting recovery of consciousness.⁵⁶ Potential airway irritants should be removed and painful stimulation should be discontinued. If laryngospasm occurs, oxygen by sustained positive pressure may be helpful, although this may push the aryepiglottic folds more tightly together.¹⁰⁹ Larson described a technique of applying firm digital pressure anteriorly directed to the “laryngospasm notch” between the ascending mandibular ramus and the mastoid process and observed that this technique is rapid and highly effective.¹¹⁰ Very small doses of a short-acting neuromuscular blocker with or without reintubation may be necessary.^111,112

13 Macroglossia

Massive tongue swelling may complicate prolonged posterior fossa surgery performed with the patient in the sitting, prone, or park bench position.^113–116 It is also seen with very steep or prolonged Trendelenburg positioning, hypothyroidism, acromegaly, lymphangioma, idiopathic hyperplasia, metabolic disorders, amyloidosis, cystic hygroma, neurofibromatosis, rhabdomyosarcoma, sublingual or submandibular infections, and chromosomal abnormalities such as the Beckwith-Wiedemann syndrome.¹¹⁷

The most common and dramatic presentation of macroglossia results from angioedema.¹¹⁸ It can be congenital or acquired. Hereditary angioedema results from a deficiency of C1 esterase inhibitor; acquired C1 esterase deficiency is associated with histamine release, a physical stimulus, or most commonly, a reaction to angiotensin-converting enzyme inhibitors or angiotensin receptor blockers.^118,119 Although involvement of the tongue is the most obvious manifestation, the uvula, soft tissues, and larynx may also be involved.

In the ICU setting, macroglossia may be seen as a complication of extreme volume overloading or tongue trauma, particularly when it is further complicated by a coagulopathic state. If this occurs or progresses after extubation, it can lead to partial or complete airway obstruction, making reintubation necessary but difficult or impossible.¹¹⁴ Lam and Vavilala postulate that positioning in most cases results in venous compression leading to arterial insufficiency and subsequent reperfusion injury.¹¹⁵ Alternatively, local compression may cause venous or lymphatic obstruction with resultant immediate and typically milder tongue swelling. The latter form is less severe but more apparent, and extubation is likely to be postponed.

14 Laryngeal or Tracheal Injury

Airway injuries, such as lacerations, edema, arytenoid dislocation, and vocal fold paralysis, may occur from the lips to the distal trachea. The lip or tongue may become entrapped between the laryngoscope blade and the mandibular teeth, resulting in swelling or bleeding, although this is unlikely to be severe enough to complicate extubation. The glottis may be injured as a result of insertion of a round tube through a triangular opening. The trachea can be lacerated or penetrated by the ETT or its introducer or by ischemic compression by the cuff on the tracheal mucosa. Palatopharyngeal injuries have been described as a consequence of blind insertion of an ETT during video laryngoscopy.^120–127 Although these injuries are not often apparent at the time of laryngoscopy, they have been managed conservatively and should not complicate extubation. The epiglottis can be downfolded during intubation, but the consequences of downfolding, even if prolonged, are unknown.¹²⁸

Laryngeal injuries accounted for 33% of all airway injury claims and 6% of all claims in the ASA Closed Claims Project database.¹²⁹ They range from transient hoarseness to vocal fold paralysis. Even when direct laryngoscopy provides a satisfactory glottic view or intubation is facilitated by fiberoptic instrumentation,^89,91 airway injury can occur and go unsuspected until after the ETT is removed. Airway injuries are presumed to be less likely if intubation is easy, but analysis of the ASA Closed Claims Project revealed that 58% of airway trauma and 80% of laryngeal injuries were associated with intubations that were not difficult.^129,130 Judging from the findings of the Closed Claims Project, difficult intubations were more likely to result in pharyngeal and esophageal than tracheal injuries.

Vocal fold immobility may result from injury to the recurrent laryngeal nerve or the arytenoid cartilages.^92,131–139 Arytenoid immobility has resulted from seemingly uneventful Macintosh and McCoy direct laryngoscopy,^137,140 double-lumen tube insertion, and lightwand intubation.¹³⁴ The mechanism of this injury is uncertain. It may be a consequence of a subluxation or a hemarthrosis with subsequent resolution or fixation. Prolonged or stressful contact between the ETT and the posteromedial aspects of the vocal cords, arytenoids, or posterior commissure may result in ulceration of the perichondrium, which heals with fibrous adhesions that produce an immobile glottis. This complication may be more common than the literature indicates.^92,141 Persistent post-extubation hoarseness, a breathy voice, and an ineffective cough should prompt assessment by an otolaryngologist. The diagnosis is confirmed by endoscopic visualization of an immobile vocal cord associated with a rotated arytenoid cartilage.⁸⁸ If the diagnosis is made early, before the onset of ankylosis, it may be possible to manipulate the arytenoid back into position.

Vocal fold paralysis results from injury to the vagus or one of its branches (i.e., recurrent laryngeal nerve [RLN] or external division of the superior laryngeal nerve [ex-SLN]) and may resemble arytenoid dislocation or ankylosis. Differentiation may require palpation of the cricoarytenoid joints under anesthesia or laryngeal electromyography.⁸⁸ When vocal fold paralysis occurs as a surgical complication, it is usually associated with neck, thyroid, or thoracic surgery. The left RLN can also be compressed by thoracic tumors, aortic aneurysmal dilatation, left atrial enlargement, or during closure of a patent ductus arteriosus. Occasionally, a surgical cause cannot be implicated. Cavo and coworkers postulated that an overinflated ETT cuff might result in injury to the anterior divisions of the RLN.¹⁴²

The RLN supplies all of the intrinsic laryngeal muscles except the cricothyroid, the true vocal cord tensor, which is innervated by the ex-SLNs. Unilateral ex-SLN injury results in a shortened, adducted vocal fold with a shift of the epiglottis and the anterior larynx toward the affected side. This produces a breathy voice but no obstruction and usually resolves within days to months. Bilateral ex-SLN injury causes the epiglottis to overhang, making the vocal folds difficult to visualize. If seen, they are bowed. This produces hoarseness with reduction in volume and range but no obstruction. Unilateral RLN injury causes the vocal fold to assume a fixed paramedian position and produces a hoarse voice. There may be a marginal airway with a weak cough. Bilateral RLN injury results in both vocal folds being fixed in the paramedian position and inspiratory stridor, often necessitating a surgical airway.¹⁴³

Pharyngeal, nasopharyngeal, and esophageal injuries include perforation, lacerations, contusions, and infections. These injuries may be associated with difficult laryngoscopy or intubation, but they may also result from the blind passage of a gum elastic bougie,¹⁴⁴ nasogastric tube,¹⁴⁵ nasotracheal tube,¹⁴⁶ suction catheter, esophageal stethoscope transesophageal echo probe, or temperature probe.¹⁴⁷ Penetrating injuries can communicate with the esophagus, resulting in a tracheoesophageal fistula, or with the mediastinum, which may go unrecognized and result in mediastinitis, retropharyngeal abscess, and death.¹³¹ After a brief intubation, soft tissue injuries resulting in airway obstruction are more likely to result from edema or hematoma than infection. Most of the described injuries do not significantly complicate extubation. Laryngeal and tracheal stenoses are serious complications, but they are rarely evident at the time of extubation.

15 Airway Injury

Burn patients can have intrinsic and extrinsic airway injuries. Circumferential neck involvement is an example of an extrinsic injury. Smoke inhalation or thermal injuries are examples of intrinsic injuries. Burn patients are at particular risk of requiring reintubation. They can have bronchorrhea, impaired mucociliary clearance and local defenses, laryngeal and supraglottic edema, increased carbon dioxide production, and progressive acute respiratory distress syndrome. Carbon monoxide may also diminish their level of consciousness and the ability to protect their airway. It may be difficult to secure the tracheal tube because of involvement of the adjacent skin, and burn victims may be agitated or uncooperative, increasing the risk of unintended extubation.¹⁴⁸ Kemper and coworkers reported their management of 13 burn patients younger than 15 years, 7 of whom exhibited post-extubation stridor. Patients treated with helium-oxygen mixtures had lower stridor scores than patients treated with an air-oxygen mixture.¹⁴⁹ They found that 11 of 30 extubated burn victims required treatment for stridor after extubation, consisting of racemic epinephrine, helium-oxygen, reintubation (n = 5), or tracheostomy (n = 1). The absence of a cuff leak was considered to be the best predictor of failure, with a sensitivity of 100% and a positive predictive value of 79%.¹⁴⁹

A variety of conditions may lead to airway edema severe enough to result in post-extubation stridor. It occurred in 2% to 37% of ICU patients after “prolonged intubation.”¹⁵⁰ A test was sought to predict patients with sufficient airway swelling to compromise safe extubation. The cuff-leak test was initially proposed for children with croup.¹⁵¹ The concept is that marked airway swelling is likely if air does not escape around the deflated ETT cuff when the ETT is occluded as the patient exhales. The cuff-leak test was the best predictor of successful extubation in a pediatric burn and trauma unit, and although sensitive, it was not specific for predicting stridor, necessitating reintubation in 62 adults.^152,153

The test has been refined by evaluating cuff-leak volume as the difference between inspiratory and expiratory tidal volumes during assist-control ventilation after cuff deflation.^41,154 Two studies found the cuff-leak volume could predict post-extubation stridor. In one study, 8 of 45 patients exhibited stridor, 4 of whom required reintubation.¹⁵⁴ In the other study involving 88 adult medical ICU patients, 6 patients exhibited stridor, 3 of whom required reintubation.⁴¹ They observed a significantly smaller cuff leak in patients who subsequently developed stridor, concluding that this measurement was the best predictor of the presence or absence of stridor. However, a study of 561 consecutive cardiothoracic patients who were extubated within 24 hours (median, 12 hours) defined a the cuff leak as the difference between the inspired and expired tidal volumes during assist-controlled ventilation. None of the (20) patients with leaks <110 mL developed stridor whereas one patient with a leak of 350 mL did and required reintubation. This led the investigators to conclude that the quantitative cuff-leak test was not reliable in this patient population.¹⁵⁵

In another study involving 110 trauma victims, cuff leaks of less than 10% had a 96% specificity for predicting post-extubation stridor or the need for reintubation.¹⁵⁶ De Bast and colleagues studied 76 adults in a combined medical-surgical ICU.¹⁵⁷ An equal number of patients had been intubated for more or less than 48 hours. Assist-control ventilation was reinstituted after cuff deflation, and the percentage of cuff leak was determined. Receiver operating characteristic (ROC) curves yielded a cutoff value of 15.5% cuff leak to equalize the false positives and false negatives for reintubation due to laryngeal edema, but this study was relatively small. Among patients intubated for more than 48 hours, only 2 of 22 with large-volume leaks and 6 of 16 with small-volume leaks required reintubation for laryngeal edema.

Kriner and colleagues evaluated the cuff-leak test for its ability to predict post-extubation stridor among 462 adult patients intubated for longer than 24 hours.¹⁵⁰ They evaluated the two thresholds previously described; a positive test was defined as a cuff-leak volume of 110 mL or less or of 15.5% or less of the exhaled tidal volume. Ten of 82 patients with leak volumes of 110 mL or less developed post-extubation stridor, and 10 of 380 with larger leaks developed post-extubation stridor, giving a positive predictive value for stridor of 0.12 and negative predictive value of 0.97. The sensitivity and specificity of the test under these conditions were 0.50 and 0.84, respectively.¹⁵⁰ With leak volumes of 15.5% or less, 7 of 48 patients developed post-extubation stridor, and 13 of 414 patients with volumes greater than 15.5% developed stridor, yielding positive and negative predictive values of 0.15 and 0.97, respectively. The sensitivity and specificity of this test for post-extubation stridor were 0.35 and 0.91, respectively. The prevalence of post-extubation stridor was 20 of 462. Seven of these patients required reintubation; 15 were managed with racemic epinephrine or helium-oxygen mixtures, or both, and 2 of the 15 failed, requiring intubation. The investigators concluded that neither threshold adequately predicted post-extubation stridor or could justify delaying extubation.

The reported variability of post-extubation stridor (2% to 37%) results in part from inconsistent diagnostic criteria,¹⁵⁰ but undoubtedly there are many other factors at play. Factors increasing the risk of post-extubation stridor include longer duration of intubation,¹⁵⁰ trauma and burns,^{14,148,154,158} pediatric age group,¹⁵⁸ female gender,^95,97,150 traumatic or emergent intubations, periods of hypotension, agitation, persistent attempts at phonation, inadequate ETT fixation, aggressive tracheal suctioning, highly positive fluid balance, low plasma oncotic pressure, gastroesophageal reflux, increased ETT diameter, and the presence of a nasogastric tube.^95,97,150

In summary, it appears that despite the intuitive appeal of the cuff-leak test, neither the qualitative nor quantitative test adequately predicts adult patients who will develop stridor after extubation.^150,155,159 It did not predict where patients were likely to require reintubation. Deem argued that absence of a cuff leak might result in an unnecessary delay of extubation, whereas a large cuff leak might produce false reassurance that there will be no difficulties.¹⁵⁹ At least in adults, the value of this test is uncertain.

16 Postobstructive Pulmonary Edema

Severe airway obstruction from any cause may complicate extubation and lead to postobstructive pulmonary edema, also called negative-pressure pulmonary edema.¹⁶⁰ This edema occurs when a forceful inspiratory effort is made against an obstructed airway (i.e., Mueller maneuver), often a closed glottis, generating large negative intrapleural pressures that promote venous return. The increase in venous pressure is aggravated by a lowered alveolar pressure, resulting in transudation of fluids into the pulmonary interstitium and alveoli. It may also result in a rightward shift of the interatrial and interventricular septa, raising left atrial and ventricular pressures. Some instances may be complicated by a permeability defect with exudative fluid and inflammatory cells.^161–166

Postobstructive pulmonary edema usually occurs in adult patients with upper airway tumors, severe laryngospasm, or rarely, bilateral vocal cord palsy,¹⁶⁷ whereas in children, it occurs most commonly as a complication of croup or epiglottitis.¹⁶⁸ The onset may be within minutes of the development of airway obstruction. It typically resolves with relief of the obstruction and supportive treatment for pulmonary edema.¹⁶⁶

1 Laryngoscopic Surgery

Mathew and colleagues looked at 13,593 consecutive PACU admissions from 1986 through 1989.²¹ Twenty-six (0.19%) of these patients required reintubation while in the PACU; seven of them had undergone ear, nose, and throat procedures. Of the seven patients, three had laryngeal edema, one was obstructed from a large thyroid, two bled at the operative site, and one developed postobstructive pulmonary edema after a tonsillectomy.

Patients undergoing laryngoscopy and panendoscopy (i.e., laryngoscopy, bronchoscopy, and esophagoscopy) are at an increased postoperative risk for airway obstruction and are approximately 20 times as likely to require reintubation as patients undergoing a wide variety of other surgical procedures.²⁰ Reviewing the records of 324 diagnostic laryngoscopies and 302 panendoscopies, Hill and colleagues found that patients who had undergone laryngeal biopsy were at the greatest postoperative airway risk. Thirteen (5%) of 252 patients required reintubation, most within 1 hour of extubation. Twelve of 13 had undergone laryngeal biopsy. Most of these patients had chronic obstructive pulmonary disease, and their need for reintubation was attributed largely to this.

Robinson prospectively studied 183 patients who had 204 endoscopic laryngeal procedures.¹⁷¹ Seven patients had tracheostomies before or after their surgery because of high-risk airways. Two of the remaining patients developed postoperative stridor; one required reintubation, and the other required a delayed tracheostomy. Indirect laryngoscopy carried out 4 to 6 hours after surgery revealed mucosal hemorrhage or laryngopharyngeal swelling in 32% of cases. Because the patients undergoing tracheostomy were not described, it is possible that the low incidence of reintubation resulted from an aggressive approach to preemptive tracheostomy.

2 Thyroid Surgery

A variety of airway-related injuries can be associated with thyroidectomies, including SLN and RLN injuries, wound hematoma, and tracheomalacia. Lacoste and colleagues retrospectively reviewed the records of 3008 patients who underwent thyroidectomies between 1968 and 1988.²⁷ The RLN had been identified intraoperatively in 2427 of these patients. Indirect laryngoscopy was performed on the third or fifth postoperative day. The RLN was damaged in 0.5% of patients with benign goiters and 10.6% of patients with thyroid cancer. RLN injury produces hoarseness, persistent coughing with phonation, and risk of aspiration. Unilateral RLN palsy was observed in 1.1% of patients. Three patients had bilateral RLN palsy and required tracheostomy. Six of a total of 16 deaths during the first 30 postoperative days were attributed to respiratory complications. One death occurred after failed intubation due to a deviated, constricted trachea. A second death was attributed to difficulties performing a tracheostomy. Two deaths resulted from aspiration or pneumonia, possibly related to RLN dysfunction.

SLN injury is more challenging to diagnose. It produces dysphonia and vocal fatigue, particularly in the higher registers. In a 5-year, multicenter study involving 42 centers and almost 15,000 thyroid operations, the diagnosis was suspected in 3.7% and confirmed in 0.4% of patients.¹⁷²

Local hemorrhage or hematoma occurs postoperatively in 0.1% to 1.6% of patients undergoing thyroid surgery and in 0.36% of the patients cared for by Lacoste and colleagues.^27,172–174 These complications occurred 5 minutes to 3 days postoperatively. Re-exploration within the first day was required only twice. Airway obstruction may result from significant laryngeal and pharyngeal edema, and wound evacuation may be of limited value in the relief of airway obstruction.^173,175,176 It may result from or be aggravated by ligature slippage, coughing, vomiting,¹⁷⁴ coagulopathies, and reoperation.¹⁷³ The prophylactic placement of surgical drains likely reduces the incidence of this complication. Wound evacuation may result in significant improvement; however this is not always the case. If time permits or intubation fails, wound evacuation should be considered.¹⁷⁶ Laryngeal edema may persist after the wound has been evacuated, necessitating postoperative intubation.

Tracheomalacia is rarely diagnosed after thyroidectomies, even in patients with significant retrosternal tracheal compression, although it may exist subclinically.^177–181 Although symptoms, computed tomography (CT), and pulmonary function test results make it easy to recognize airway compression preoperatively, tracheomalacia may be difficult to predict or even detect from the surgical field, and it does not become apparent until after the ETT is removed and spontaneous ventilation has resumed.¹⁸²

3 Carotid Artery Surgery

Neck swelling or hematoma formation after carotid endarterectomy may be relatively common. The New York Carotid Artery Surgery (NYCAS) study analyzed 9308 procedures performed between 1998 and 1999 at 167 hospitals.¹⁸³ A hematoma was identified in 5% of patients, substantially increasing the risk of death (OR = 4.30; 95% CI, 2.72 to 5.00) and stroke (OR = 3.89; 95% CI, 2.82 to 5.38). Hematoma occurrence reported in the literature ranges from 1.2% to 12%, depending on the definition used.¹⁸³ The overall rate of wound hematomas in the North American Symptomatic Carotid Endarterectomy Trial (NASCET), involving 1415 patients was 7.1%, 3.9% of which cases were considered to be mild (i.e., no delay in discharge), 3.0% were moderate (i.e., delay in discharge), and 0.3% were severe (i.e., permanent functional disability or death). The moderate and severe cases required re-exploration or wound evacuation (3.3%). Hematoma contributed to the death of four patients.¹⁸⁴ When wound hematomas are identified by a comparison of preoperative and postoperative CT scans, it occurs far more frequently (26%).¹⁸⁵ The postoperative reintubation or exploration rate is 1% to 3.3%.^184,186

Kunkel and colleagues described 15 patients who developed wound hematomas after carotid endarterectomy.¹⁸⁷ Eight of these were evacuated under local anesthesia. In six of seven cases in which general anesthesia was induced before opening the wound, difficulties arose with airway management, resulting in two deaths and one patient with severe neurologic impairment. O’Sullivan and coworkers reported a similar experience for six patients with airway obstruction after carotid endarterectomy.¹⁸⁸ Stridor was not relieved by wound evacuation. Administration of muscle relaxants made manual mask ventilation and tracheal intubation virtually impossible due to marked glottic or supraglottic edema. Cyanosis and extreme bradydysrhythmias or asystole occurred in four patients. The providers endorsed Kunkle’s recommendation for wound evacuation but thought that much of the airway compromise was caused by edema from venous or lymphatic congestion. They emphasized that the outward appearance may lead to an underestimation of the situation’s gravity. Voice changes are early signs of danger and may be relatively subtle. Rapid clinical deterioration can occur after stridor develops.¹⁸⁹

Studies by Carmichael and colleagues provided additional evidence for the role of swelling and bleeding in a small but elegant study. They compared the CT scans of 19 patients before and after carotid endarterectomy surgery.¹⁸⁵ Clinically, 1 patient had severe swelling, 4 had moderate swelling, and 3 had mild swelling, but 10 were deemed normal. However, postoperative CT scans demonstrated significant swelling of the retropharyngeal space and a reduction of the anteroposterior and transverse airway diameter, particularly at the level of the hyoid. Compared with preoperative CT scans, the calculated volume reduction averaged 32% ± 7% for extubated patients. The scans revealed a wound hematoma estimated to be greater than 10 mL (range, 44 to 94 mL) in 5 of 19 patients. Patients who remained intubated postoperatively showed a significantly greater volume reduction of 62% ± 9% (P < 0.025). Those remaining intubated as a result of swelling had contralateral extension of the swelling. These observations may help to explain why opening the wound frequently fails to provide benefit in many patients. Nonetheless, it may be difficult to clinically differentiate bleeding from swelling. After radiologic demonstration of swelling, the same group evaluated the benefits of prophylactic dexamethasone but failed to demonstrate any clinical benefit.¹⁹⁰

A 10-year, retrospective review of 3224 carotid endarterectomies performed at The Mayo Clinic revealed that 44 (1.4%) patients required wound exploration within 72 hours of surgery, despite the nonreversal of heparin.¹⁹¹ In two patients, re-exploration occurred before the initial extubation. The decision to re-explore was made in the PACU for 7 patients; the remaining 35 were identified in the ICU or ward. Only one patient required a surgical airway when direct laryngoscopy failed in the ICU. Several techniques were initially employed: awake bronchoscopic intubation (15 of 20 of which were successful), direct laryngoscopy after induction (13 of 15 were successful), and awake direct laryngoscopy (5 of 7 were successful). When awake bronchoscopy failed, direct laryngoscopy was successful whether the patient was awake (3 of 3) or asleep (2 of 2). When direct laryngoscopy initially failed after induction in two patients, it succeeded after opening the incision. When awake direct laryngoscopy failed, one patient required a surgical airway; in the other, direct laryngoscopy succeeded after opening the incision. Despite the size of this series, it is not possible to draw conclusions about which techniques are most successful. Success likely depends on the skill and judgment of the airway manager. It is also possible that this study differed from the other studies in that the decision to re-explore was made earlier and patients had less airway distortion.

Several nerve injuries can result from carotid artery surgery or the anesthetic technique. The range reported in the literature is 3% to 23%, although most of these cases resolve within 4 months of surgery.¹⁹² In the NYCAS study, cranial nerve palsies occurred in 514 (5.5%) of 9308 patients and involved, in descending order, the hypoglossal nerve (170), producing tongue deviation to the operative side; a branch of the facial nerve (126), resulting in lip or facial droop; the glossopharyngeal nerve (41); a branch of the vagus, which may involve the RLN and produce vocal cord paresis (31); the trigeminal nerve (19); a branch of the cervical plexus (10); or more than one nerve group (117).¹⁸³

Bilateral vocal cord and bilateral hypoglossal nerve palsies have been described after staged, bilateral carotid endarterectomies.^24,25 In the latter case, the first procedure, performed under regional anesthesia, had been complicated by a wound hematoma, resulting in numbness over the anterior neck and diminished sensation in the C2 and C3 distribution. The subsequent endarterectomy, done 4 weeks later under deep cervical plexus block with subcutaneous infiltration, caused intraoperative airway obstruction and asystole. The airway was secured, but repeated attempts at extubation resulted in persistent obstruction due to bilateral hypoglossal nerve palsy. In another case, performed under cervical plexus block, the patient developed bilateral vocal cord paralysis that required intubation and subsequent tracheotomy. It is suspected that she had a previously unrecognized contralateral vocal cord palsy from a prior thyroidectomy.¹⁹³ In this case, the vocal cord dysfunction was co-incident with retraction of the carotid sheath—although it can also be induced by a cervical plexus block—but it raises the importance of the preoperative assessment of patients who have had prior head and neck surgery.

4 Cervical Surgery

Cervical spine procedures may be followed by airway-related complications, including vocal cord paralysis and airway obstruction. Vocal cord dysfunction was seen in 5% of 411 patients undergoing anterior cervical discectomy and fusion.¹⁹⁴ Stridor was observed in one patient with bilateral vocal cord paralysis who required a tracheostomy. Fifteen of 17 patients had recovered by 12 months. One additional patient had recovered by 15 months, and the remaining patient was lost to follow-up.

Emery and colleagues studied the records of 133 patients who underwent cervical corpectomies with arthrodesis between 1974 and 1989.²⁶ The patients had undergone an anterior approach to achieve a three-level vertebral body and disc resection with bone grafting. This surgical approach requires tracheal and esophageal retraction toward the opposite side to permit exposure. Drains were placed, and all patients were immobilized by a halo vest or a rigid head-cervical-thoracic orthosis. They identified seven patients (5.3%) who required postoperative reintubation, and although they did not compare these patients with those not requiring reintubation, they attempted to identify common features that increased the risk of postoperative airway compromise. Three patients were immediately reintubated in the operating room, and four were reintubated 12 to 91 hours postoperatively. Severe hypopharyngeal swelling was observed at reintubation in four of seven patients and possibly in a fifth. Five of the seven reintubations had no serious sequelae; these patients were extubated within 2 to 8 days. One patient required a cricothyroidotomy, but delay resulted in hypoxic encephalopathy and death. Another patient was reintubated but developed and succumbed to severe adult respiratory distress syndrome. The investigators think that preexisting pulmonary disease, moderate or severe preoperative myelopathy, extensive multilevel decompression with prolonged surgery, and tissue retraction were risk factors for postoperative airway obstruction, but there were no controls.²⁶ They recommended 1 to 3 days of elective intubation postoperatively, a cuff-leak test, and direct laryngoscopy at extubation.

Venna and Rowbottom reviewed the records of 180 patients who had undergone a variety of cervical surgical procedures.²⁸ Based on the Emery study, they had made the decision to keep high-risk patients intubated until they met specified criteria, including a demonstrable cuff leak and the absence of significant airway edema on laryngoscopy. The average time to extubation was 33.5 hours. Despite the delay and the aforementioned criteria, 12 patients (6.6%) demonstrated post-extubation stridor and breathing difficulties, and 5 (2.7%) required reintubation. Two patients required tracheostomy, and two deaths were attributed to airway obstruction and unsuccessful reintubation.

Sagi and coworkers conducted a retrospective chart review of 311 anterior cervical procedures in an effort to identify the factors associated with airway complications.¹⁹⁵ In this series, 19 (6.1%) of patients had airway complications, but only 6 (1.9%) required reintubation. Most of these complications were attributed to pharyngeal edema. Risk factors included increased intraoperative bleeding, prolonged surgery (>5 hours), and exposure of more than three vertebral bodies, particularly when they included C2, C3, or C4. Reviewing the literature, these investigators identified an airway complication rate of 2.4% (from 1615 cases), 35 of whom required reintubation or tracheostomy. On average, those requiring reintubation did so at 24 hours.

Epstein and coworkers developed a collaborative protocol involving the neurosurgeon and anesthesiologist. Their objective was avoidance of reintubations.¹⁹⁶ Although their study enrolled only 58 patients, they required high-risk, lengthy procedures involving several cervical levels and significant blood loss. All patients remained electively intubated overnight and underwent fiberoptic airway examination before considering extubation. Most patients were extubated the day after surgery, but three remained intubated until day 7. Only one patient required reintubation. This reintubation rate was essentially the same as that observed by Emery and colleagues,²⁶ but Epstein’s cohort appeared to undergo higher-risk surgery.

In an effort to better understand the mechanism of postoperative airway obstruction, Andrew and Sidhu compared the soft tissue changes on the preoperative and postoperative cervical spine radiographs of 32 consecutive patients after a one- or two-level anterior cervical discectomy and fusion.¹⁹⁷ They found that the swelling was maximal at the C3-C4 level, corresponding to the area where Emery observed pharyngeal edema. None of their patients experienced dyspnea despite mean differences of 9.4 mm (95% CI, 7.41 to 11.09 mm) to 10.7 mm (95% CI, 7. 8.82 to 12.58 mm) between the preoperative and postoperative radiographs at C3 and C4, respectively. It appears that this area is most vulnerable, regardless of the level operated on. Compared with Sagi’s study, these patients underwent much shorter and more limited procedures.

Extubation criteria after anterior cervical spine surgery were discussed at length in the Society for Airway Management online Forum in 2011. The groups who perform this surgery frequently tend to extubate most of their patients at the conclusion of surgery, except when the blood loss is high (e.g., >500 mL), the surgery is prolonged (e.g., >5 hours), the operation involves more than three levels, the airway was difficult or expected to be difficult, or comorbidities exist, such as severe cervical myelopathy, obesity, and obstructive sleep apnea (OSA). Patients with multiple risk factors are cared for in an ICU and are extubated over an airway exchange catheter (AEC) (personal communication from C.A. Hagberg, SAM Forum, February 2011).

Patients undergoing posterior cervical surgery may face the risk of macroglossia and significant retropharyngeal and hypopharyngeal swelling, which may be aggravated by fixation of the cervical spine and make intubation more difficult.¹⁹⁸ There is a low probability (1.1% to 1.7%) that reintubation will be required,^196,198 but accomplishing this may be very difficult.

5 Maxillofacial Surgery and Trauma

Maxillary and mandibular surgery produces conspicuous and often worrisome swelling. Anxiety regarding postoperative care may be heightened by limited airway access, fear that airway intervention may disrupt the surgical repair, and anecdotal reports of near misses or actual fatalities. Because many of these patients are young and otherwise healthy and are undergoing elective surgery for functional or cosmetic improvement, there may be concerns about litigation. It is speculative about whether this results in more or less aggressive care.

Although these concerns demand special attention, deaths rarely occur. In a review of 461 perioperative deaths reported to the Ontario, Canada, coroner between 1986 and 1995, the investigators found only one death associated with orthognathic surgery, although they were unable to determine how many such cases had been performed (see “In Vivo Studies”).¹⁹⁹ They were unable to identify nonlethal complications. Meisami and others performed magnetic resonance imaging (MRI) approximately 24 hours after maxillary or mandibular surgery in 40 patients.²⁰⁰ Despite the significant facial swelling seen in almost all the patients, none exhibited soft tissue swelling from the base of the tongue to the glottis.

Complete airway obstruction after elective orthognathic surgery has been reported. Dark and colleagues described a case involving a young woman who underwent seemingly uneventful mandibular and maxillary osteotomies with submental liposuction.²⁰¹ Immediately after extubation, she developed airway obstruction requiring reintubation. Repeated fiberoptic examination and CT showed severe and extensive edema from the tongue to the trachea, which was maximal at the level of the hyoid. By the fourth postoperative day, a cuff leak was detected, and the patient was successfully extubated over a tube exchanger. Hogan and Argalious described a patient in whom maxillomandibular advancement was performed for OSA.²⁰² The procedure lasted 9 hours, during which he received 7200 mL of crystalloid and 500 mL hetastarch. He remained intubated overnight, and after demonstrating adequate spontaneous ventilation and a cuff leak, he was extubated (over a 19-F AEC). Extubation was immediately followed by clinical evidence of airway obstruction, and he was reintubated. The obstruction was attributed to fractured hardware and a hematoma in the piriform fossa that caused extrinsic compression. This could easily have resulted from periglottic edema. The investigators concluded that patients undergoing this type of surgery face a high risk of airway complications and recommended nasopharyngolaryngoscopy before extubation.

Clinical assessment of airway edema is unreliable,²⁰⁰ and studies indicate that the cuff-leak test is neither sufficiently sensitive nor specific to determine when to extubate these patients. Endoscopic assessment may help to identify patients who harbor occult clots (i.e., “coroner’s clot”) behind the soft palate or adjacent to the glottis, but it may miss or itself give rise to troublesome bleeding.

Maxillofacial injuries often result from unrestrained occupants of motor vehicles encountering an unyielding dashboard, windshield, or steering wheel. Gunshot wounds or physical altercations also cause maxillofacial injury. Airway obstruction is a primary cause of morbidity and mortality in these patients, and many die before they reach the hospital.²⁰³ Those with less life-threatening injuries are likely to present with a full stomach, and many have associated head and neck injuries, lacerations, loose or avulsed teeth, intraoral fractures, and fractures extending into the paranasal sinuses, into the orbit, or through the cribriform plate. They may also have an unstable cervical spine or damage to the neural axis. Injuries to the lower face raise the possibility of a laryngeal fracture. Intermaxillary fixation may be part of the surgical plan, necessitating a nasal intubation or a surgical airway. Timing of tracheal extubation is complex and must take into consideration factors such as the patient’s level of consciousness, ability to maintain satisfactory gas exchange, coagulation status, and integrity of protective airway reflexes. Attention must be paid to the difficulties originally encountered in securing the airway and an evaluation of whether reintubation would be easier or more difficult after surgery and resuscitation. Most of the trauma literature about airway management addresses intubation and offers little help with extubation, making cooperation between the anesthesiologist,^204,205 surgeon, and critical care physician essential. Intermaxillary fixation requires wire cutters to be immediately available and personnel to know which wires to cut. A flexible bronchoscope, provisions for an emergency surgical airway, and the required expertise should be immediately available at the time of extubation. Alternatives include prophylactic tracheotomy, submental intubation,^206–208 nasal intubation, and bronchoscopic airway evaluation performed before extubation,²⁰⁹ although assessment may be limited to supraglottic structures and exclusion of tube entrapment. Ideally, extubation should be accomplished in a reversible manner, permitting supplemental oxygenation, ventilation, and reintubation if needed (see “Extubation Strategies”).

6 Deep Neck Infections

Infections involving the submandibular, sublingual, submental, prevertebral, parapharyngeal, and retropharyngeal spaces are significant airway management challenges, whether intubation is achieved for surgical drainage or for protection during medical management. In expert hands, bronchoscopy-assisted intubation can often be achieved.²¹⁰ When this is unsuccessful or constitutes a significant risk of rupturing the abscess, a surgical airway before incision and drainage may be called for.²¹¹ Potter and colleagues retrospectively compared the outcomes of 34 patients in whom a tracheotomy was performed with 51 patients who remained intubated after surgical drainage.²¹² All patients had undergone surgical drainage for impending airway compromise and required airway support postoperatively. It was not always evident to the investigators why a particular strategy was chosen, and these groups were not likely identical. Airway loss occurred more commonly in the intubated patients, but this characteristic was not statistically significant. Two deaths occurred, one resulting from an unintended extubation and the other from post-extubation laryngeal edema and an inability to reestablish the airway. The latter patient had a cuff leak before extubation, and signs of obstruction developed 30 minutes after the ETT was removed. Surgical drainage rarely results in immediate airway improvement, and reintubation or emergent placement of a surgical airway, if required, may be complicated by edema, tissue distortion, and urgency.