1 Basic Features

Adverse respiratory events constitute the single largest source of injury in the Closed Claims Project (Table 55-1). A detailed analysis of these events was initiated when the database reached a total of 1541 claims.¹⁴ The contrast between adverse respiratory events and other claims was particularly unfavorable. Respiratory event–related claims were (and still are) characterized by a high frequency of devastating outcomes and costly payments (Table 55-2).

Just three mechanisms of injury accounted for almost two thirds of all claims for adverse respiratory events (Table 55-3). These mechanisms were inadequate ventilation (24% of cases), esophageal intubation (12%), and difficult intubation (DI; 24%). In the 1990s, after the adoption of pulse oximetry and end-tidal CO₂ (EtCO₂) as monitoring standards, DI (28%) and inadequate ventilation (18%) remained the most common adverse respiratory events, but esophageal intubation (5%) had decreased greatly compared with earlier decades (see Table 55-3). Evaluation of claims from the early 2000s suggests that this profile of adverse respiratory events is staying the same.

Aspiration was the third most common adverse respiratory event in the 1990s and 2000s. The remaining adverse respiratory events were produced by a variety of low-frequency mechanisms including airway obstruction, bronchospasm, premature and unintentional extubation, endobronchial intubation, inadequate inspired oxygen delivery, and equipment failure. Special features of low-frequency events are discussed later in this chapter.

A detailed display of outcome and payment data for the three most common types of adverse respiratory events since 1990 (inadequate ventilation, DI, and aspiration) is shown in Table 55-4. Death and permanent brain damage were more frequent in claims for respiratory events compared with nonrespiratory events (P < 0.05). Claims for inadequate ventilation exhibited the highest proportion of death and brain damage (89%, see Table 55-4). Overall, payment for respiratory-related claims ranged from $1260 to $11 million (in 2008$). Most claims (61%) resulted in payment. Claims for adverse respiratory events typically involved healthy adults undergoing nonemergency surgery with general anesthesia (GA) (Table 55-5).

TABLE 55-5 Basic Clinical Features of Cases Involving Adverse Respiratory Events Occurring in 1990 or Later*

ASAPS, American Society of Anesthesiologists Physical Status; SD, standard deviation.

^* P < 0.05 for respiratory compared with nonrespiratory events in all categories (chi square test for proportions, t-test for age). Total anesthesia claims = 8954. Claims for chronic pain management and all claims prior to 1990 are excluded.

^† Includes combined regional and general anesthesia, standby, and unknown.

Data from ASA Closed Claims Database, 2008.

The reviewers judged that better monitoring would have prevented the adverse outcome in 20% of the 867 claims for adverse respiratory events since 1990. In contrast, only 7% of nonrespiratory claims were judged preventable with better monitoring (P < 0.01). Half (54%) of all claims for inadequate ventilation were considered preventable with better monitoring, as opposed to 5% of claims for DI and only 1% of claims for aspiration.

For the claims considered preventable with better monitoring, the reviewers pointed to pulse oximetry, capnometry, or both in most cases. Data on the role of better monitoring in the prevention of adverse outcomes must be interpreted with particular care, because the reviewers were not asked to consider confounding factors such as equipment malfunction, diversion of attention, or the impact of false-positive and false-negative results. Therefore, the reviewers’ judgments should be regarded as a near-maximum (and probably unattainable) estimate of the efficacy of better monitoring. It should also be noted that this analysis is based on claims for events that occurred after the adoption of pulse oximetry and EtCO₂ as ASA standards.

2 Inadequate Ventilation

The largest class of adverse respiratory events was inadequate ventilation. The distinguishing feature in this group of claims was the reviewer’s inability to identify a specific mechanism of injury. In part, the inability to assign a mechanism of injury may reflect uncertainty on the part of the original health care providers. Because most adverse events occurred before the widespread use of pulse oximetry and capnometry, the uncertainty may be due to the limitations of traditional clinical signs, such as chest excursion, reservoir bag motion, and breath sounds. With increasing use of quantitative measures of ventilation, fewer cases have been assigned to the category of inadequate ventilation. These events have declined in occurrence from 37% of all respiratory events in the 1970s to 25% in the 2000s (see Table 55-3). It is also possible that a delayed rather than contemporaneous approach to the investigation of adverse outcomes is not powerful enough to provide an understanding of many events.

3 Esophageal Intubation

Prompt detection of esophageal intubation is a key concern in anesthesia practice. At present, anesthesiologists rely primarily on capnometry to confirm endotracheal intubation. The ASA Closed Claims Project database provides an important lesson about the pitfalls of less quantitative methods.

In 1990, we performed an in-depth analysis of 94 closed claims for esophageal intubation.¹⁴ Almost all (92%) of these claims occurred during the period 1975–1985, before routine use of intraoperative capnometry. The single most striking finding was that detection of esophageal intubation required at least 5 minutes in most cases (97%). Our immediate reaction was simply, “What took so long?” We wondered whether such delays in detection were caused by incompetence or negligence (e.g., intubation performed by a legally blind physician, minimal attention to the patient during the procedure), but we found only eight claims (9%) that could be explained in such a way.

A closer look at the esophageal intubation claims suggested that reliance on indirect tests of ventilation may have been an important factor contributing to delay. For example, auscultation is a test that is traditionally used during the first few minutes after intubation. In our set of claims, auscultation of breath sounds was documented in 62 of the 94 claims for esophageal intubation (63%). In 3 (5%) of these cases, breath sound auscultation led to a correct diagnosis of esophageal intubation. In 30 cases (48%), auscultation led to the erroneous conclusion that the endotracheal tube (ETT) was located in the trachea when it was actually in the esophagus. This result was termed a misdiagnosis of endotracheal intubation.

The diagnostic error in such cases was recognized in a variety of ways, including later reexamination with direct laryngoscopy, absence of any object in the trachea at the time of an emergency tracheostomy (despite ongoing “ventilation” through an ETT), resolution of cyanosis after reintubation (often by a second participant), and discovery of esophageal intubation at autopsy. In 29 (47%) of the 62 claims in which auscultation was documented, the records did not contain sufficient information to determine how the auscultatory findings were interpreted. In another 32 cases (34%), there was no information about the use of auscultation.

Using this information, we constructed a best-case scenario by assuming that auscultation led to a correct diagnosis in the 3 cases in which it actually did so, as well as in the 61 cases in which there was no information about its role or the information was unclear. Using this approach of constructing a hypothetical situation to demonstrate the greatest possible benefit, auscultation is still associated with a misdiagnosis rate of 32% (30 of 94 cases). Although the limitations of auscultation are well known,¹⁵ we think these claims emphasize the importance of confirming intubation with the quantitative and ongoing information provided by capnometry.

Another indirect test of ventilation is cyanosis. Almost all of the esophageal intubation claims that we studied took place before pulse oximetry became part of the ASA Standards for Basic Anesthetic Monitoring. The human eye is relatively insensitive to the changes in skin color that occur during arterial desaturation,^16,17 so it is not surprising that cyanosis preceded the recognition of esophageal intubation in only 34% of cases. One might also expect cardiovascular clues to accompany hypoxemia or hypercarbia and that these clues might alert the anesthesiologist to the possibility of esophageal intubation. Indeed, at least one major hemodynamic derangement was recorded in 79 (84%) of the 94 claims for esophageal intubation, and this knowledge preceded the recognition of esophageal intubation in 60 claims (65%). In order of frequency, the abnormalities included bradycardia, asystole, hypotension, unspecified dysrhythmia, tachycardia, and ventricular fibrillation (Table 55-6). Such changes certainly have the potential to serve as cues, but when they are extremely severe (e.g., ventricular fibrillation, asystole) they draw attention away from the underlying problem and leave very little time for effective remedies. These features point to the importance of confirmatory tests that provide early, direct, and ongoing confirmation of ventilation through the ETT.

TABLE 55-6 Major Hemodynamic Derangements Accompanying Esophageal Intubation Claims*

^* Percentages sum to more than 100 because of multiple derangements. Total anesthesia claims = 1541.

Data from ASA Closed Claims Database, 1990. Table adapted from Caplan RA, Posner KL, Cheney FW, et al: Adverse respiratory events in anesthesia: A closed claims analysis. Anesthesiology 72:828, 1990.

Is injury from undetected esophageal intubation no longer a concern? As shown in Table 55-3, claims for esophageal intubation decreased considerably in the 1990s and later compared with the prior two decades. A few claims for esophageal intubation still enter the database, but they usually involve cases in which capnometry was unavailable or not used or the procedure took place in a remote location.¹⁸

4 Difficult Intubation

Claims for DI in the ASA Closed Claims Project were distinguished by a relatively small percentage of cases in which care was considered less than appropriate (46%, compared with more than 80% for inadequate ventilation and esophageal intubation) and by a similarly small percentage of cases in which better monitoring would have prevented the complication (19%, again compared with more than 80% for inadequate ventilation and esophageal intubation). Although these findings seem outwardly favorable, the comparisons are not so attractive from the perspective of risk reduction. If most cases of DI cannot be linked to obvious inadequacies in care or deficiencies in monitoring, it is unlikely that claims analysis alone can point to effective or broad-based remedies. Simulators, algorithms, and drill routines have generated considerable interest in recent years. These newer educational tools and management strategies may provide an important opportunity for clinicians to gain concentrated exposure to relatively infrequent events.

An evidence-based guideline for management of the difficult airway (DA) has been developed by the ASA.¹⁹ Key features of this guideline are discussed in Chapter 10. The impact of this set of guidelines on claims involving DA management was studied in the ASA Closed Claims Project.²⁰ DA problems were encountered during all phases of anesthesia management, including preinduction, induction, maintenance, emergence, postanesthesia care unit (PACU) care, and intensive care unit settings. Death and brain damage were more common in DA claims arising from procedures performed outside the operating room (OR) or the PACU and in the settings of difficult mask ventilation (DMV); “cannot intubate, cannot ventilate” (CICV) emergencies; or persistent intubation attempts.²⁰

5 Aspiration

An in-depth analysis of claims for pulmonary aspiration was conducted in 1991, when such claims accounted for 3% of the database (56 cases).²¹ Almost all cases (95%) occurred in patients who received GA. The aspirated material was gastric contents in 88% of these cases; other cases involved aspiration of blood, pus, or teeth. Approximately one third (34%) of aspirations took place during anesthetic induction just before endotracheal intubation. In 6 of these cases, aspiration occurred during a rapid-sequence induction; in another 6, the aspiration occurred under circumstances in which the reviewer believed that rapid-sequence induction was indicated but in which it was not used. Another one third of aspiration cases (36%) took place during the maintenance phase of mask GA. Only 2 cases (4%) occurred during the maintenance phase of endotracheal GA. Aspiration occurred in one of these cases when the ETT was removed to facilitate the passage of a nasogastric tube. In the other instance, aspiration occurred while an ETT with a leaking cuff was being replaced with a new tube. The remaining cases (18%) took place during emergence from anesthesia.

Two clinical factors—pregnancy and emergency surgical status—were particularly prevalent in claims for aspiration. Obstetric patients accounted for 12% of the overall database but represented 29% of all aspiration claims (P < 0.05) in the 1991 study.²¹ This trend changed dramatically in subsequent years, with aspiration occurring in fewer than 1% of obstetric claims from 1990 or later.²² Emergency surgery patients accounted for 19% of the database in 1991 but represented 45% of all aspiration claims (P < 0.01). It is also noteworthy that 23% of aspiration claims involved a problem with airway management, such as DI (9 cases) or esophageal intubation (4 cases). This relationship was previously reported by Olsson and colleagues.²³

Overall, aspiration accounted for only 3% of claims in the database. This is consistent with the observation by Warner and colleagues that the incidence of aspiration in more than 200,000 patients who underwent elective or emergency surgery during 1985–1991 was very low (1 in 3216).²⁴ These observations suggest that current strategies used to prevent aspiration in the United States are generally successful, particularly in obstetric patients.

1 Basic Features

The foregoing discussion focused on the most common mechanisms of respiratory injury in the closed claims database. A formal study of low-frequency adverse respiratory events was conducted in 1991, when the ASA Closed Claims Project database had reached 2046 cases.²¹ Although these events were much less common (each category representing no more than 5% of the overall database), sufficient claims were collected to permit the identification of recurrent themes that might contribute to liability. Five categories of events have been studied in depth, each category containing at least 40 claims and together encompassing 300 claims, or about 15% of the overall database. These categories are airway trauma, pneumothorax, airway obstruction, aspiration, and bronchospasm. Death or brain damage occurred in almost half (47%) of these cases, and the median payment was $60,000 (Table 55-7). Airway trauma was reanalyzed in detail in 1999.²⁵ Aspiration now represents one of the most common adverse respiratory events.

2 Airway Trauma

Airway trauma was the most common type of low-frequency airway event, accounting for 266 of 4460 claims, or 6% of the overall database, in the 1999 study.²⁵ DI was associated with 103 (39%) of these claims. The most frequent sites of injury were the larynx, pharynx, and esophagus (Table 55-8), which together accounted for 70% of injuries associated with airway trauma claims. Esophageal and tracheal injuries were more likely than other airway trauma claims to be associated with DI. In contrast, laryngeal and temporomandibular joint (TMJ) injuries were rarely associated with DI. Pharyngeal and esophageal injuries most commonly consisted of lacerations or perforations leading to mediastinitis or mediastinal abscess. The most common laryngeal injuries were vocal cord paralysis (30 cases), granuloma (15 cases), and arytenoid dislocation (7 cases). None of the 27 TMJ injuries was associated with DI.

TABLE 55-8 Distribution of Airway Trauma Sites and Presence of Concomitant Difficult Intubation

^* P < 0.05 compared with other sites combined. Total anesthesia claims = 4460.

Data from ASA Closed Claims Database, 1999. Table adapted from Domino KB, Posner KL, Caplan RA, et al: Airway injury during anesthesia: A closed claims analysis. Anesthesiology 91:1703, 1999.

These claims provide several useful insights. Circumstances surrounding DI clearly put the tissues of the pharynx and esophagus at risk. The clinical implication is that patients in whom endotracheal intubation has been difficult should be observed for, or told to watch for, the development of signs and symptoms of pharyngeal abscess or mediastinitis. Because soft tissue infections may develop slowly over a period of days, an apparent lack of complications in the first few hours after surgery should not be regarded as a definitive outcome in patients who have experienced DI. This is especially important to remember in the ambulatory surgery setting.

Although it is easy to understand how DI may lead to trauma of the larynx, it is less apparent why laryngeal injuries appeared to be so infrequently associated with DI. The reasons for vocal cord paralysis, granuloma, and arytenoid dislocation with routine intubation were not apparent from the data available in the claim file. Similarly, it is curious that TMJ injury was present only with routine intubation. One might expect that TMJ injury would be more commonly associated with DI, in which forces applied to the jaw during airway manipulation and laryngoscopy might be more intense or prolonged than those encountered during routine intubation. These observations suggest that many injuries to the larynx and TMJ may be related to predisposing factors or underlying characteristics of patients that we do not yet understand. A similar phenomenon was observed in the review of closed claims for peripheral nerve injuries.^26,27

3 Pneumothorax

Pneumothorax was the second most common type of low-frequency airway event.²¹ Clinical activities that were not directly or clearly related to airway management were associated with 43 (64%) of the 67 claims for pneumothorax (Table 55-9). In particular, five types of nerve blocks (supraclavicular, intercostal, stellate ganglion, interscalene, and suprascapular) were responsible for 40% of pneumothorax claims. Airway instrumentation was associated with pneumothorax in 19% of cases. The actual mechanism of pneumothorax was not anatomically proved in most cases but was usually attributed to laryngoscopy, ETT placement, or bronchoscopy on the basis of clinical events and reviewer judgments. Barotrauma was the cause of pneumothorax in 11 claims (16%), mostly arising from obstruction of the expiratory limb of a mechanical ventilator or the use of excessive tidal volumes (7 cases).

TABLE 55-9 Clinical Factors Associated with Pneumothorax Claims*

^* Total anesthesia claims = 2046.

Data from ASA Closed Claims Database, 1991. Table adapted from Cheney FW, Posner KL, Caplan RA, et al: Adverse respiratory events infrequently leading to malpractice suit: A closed claims analysis. Anesthesiology 75:932, 1991.

A notable feature of pneumothorax claims was the marked disparity in outcome between events associated with airway instrumentation and events arising from nerve blocks or central-line placement. In the subset of 24 claims involving airway instrumentation, the outcome in 16 cases (67%) was death or permanent brain damage. In contrast, there were no instances of death or brain damage in the 16 cases of pneumothorax that arose after nerve blocks or central-line placement. We speculate that this difference may be due at least in part to the more rapid compromise of respiratory and circulatory function that occurs under conditions of mechanical ventilation and positive-pressure gas delivery. Not surprisingly, the median payment for pneumothorax associated with nerve block or central-line placement was only $6000, whereas that for cases associated with airway instrumentation was $75,000.

4 Airway Obstruction

Airway obstruction accounted for 56 claims, or approximately 3% of the database.²¹ Most cases (89%) occurred during GA. Obstruction was attributed to an upper airway site in 39 claims (70%), although an exact cause or site was identifiable in only half of these claims. Laryngospasm was the most common cause of upper airway obstruction, accounting for 11 (28%) of 39 cases. Other causes of upper airway obstruction included foreign body (4 cases), laryngeal polyps (2 cases), laryngeal edema (1 case), and pharyngeal hematoma (1 case). In 10 cases of upper airway obstruction, emergency tracheostomy was performed.

Causes of lower airway obstruction (21% of claims) included blood clots or mucus plugs in the tracheal lumen and external compression related to mediastinal tumor masses or blood. ETT obstruction accounted for 9% of cases and was attributed to blood clots in the lumen of the ETT or to kinking of the tube itself. Other factors associated with claims for airway obstruction included concurrent DI (17 cases, 30%), operation on the airway (13 cases, 23%), and pediatric age group (10 cases, 18%). The outcome in almost all claims for airway obstruction (87%) was death or brain damage.

5 Bronchospasm

Adverse outcomes arising from bronchospasm accounted for 40 claims, or almost 2% of the database.²¹ Most of these claims (80%) occurred during the administration of GA as the primary anesthetic technique. Almost half (48%) of the patients had a medical history that included asthma, chronic obstructive pulmonary disease, smoking, or some combination of these. In cases involving the administration of GA, the first occurrence of bronchospasm was more often at the time of intubation (69%) than during maintenance (25%) or emergence (6%).

Twenty percent of claims for bronchospasm were associated with the conduct of regional anesthesia. In most of these cases, bronchospasm occurred during cesarean section when endotracheal intubation was required for management of a failed block or a high block in a patient with a history of asthma. These cases illustrate the concept that regional anesthesia does not, in itself, obviate the risk associated with intraoperative management of reactive airway disease. In particular, the risk of bronchospasm may be especially pronounced in this setting, because relatively modest doses of intravenous agents are often used in an effort to minimize anesthetic effects on the fetus.

Bronchospasm claims were also notable in cases that involved a difficult differential diagnosis. The claims files in the 1991 study indicated that clinicians had difficulty distinguishing between bronchospasm and the presence of esophageal intubation (6 cases) or pneumothorax (4 cases). EtCO₂ monitoring was not used in any of the 6 cases in which the failure to make a correct and timely differential diagnosis between esophageal intubation and bronchospasm led to an adverse outcome. EtCO₂ is now an ASA standard for verification and monitoring of ETT placement, and bronchospasm appears to be on a steadily declining trend (see Table 55-3). However, it is important to recognize that failure to differentiate between bronchospasm and esophageal intubation may still occur in cases in which bronchospasm is so severe that ventilation is impossible and CO₂ cannot reach the detector in clinically useful amounts. In this circumstance, fiberoptic bronchoscopy can be helpful.