The cardiac arrest occurred at the time of block placement in 1 patient (4%), between block placement and surgical incision in 5 patients (19%), during the surgical procedure in 16 patients (62%), and after surgical closure in 4 patients (15%). The median time to arrest from the last local anesthetic administration (initial intrathecal injection or last epidural/caudal injection before arrest) was 50 min (range, 0–210 min). A sensory level at T6 or above (corresponding to a sympathetic level of T4 or above) was noted in 11 patients. In 12 (46%) patients, the cardiac arrest was associated with a specific surgical event, such as cementing of joint components, spermatic cord manipulation, intramedullary rod placement, rupture of amniotic membranes, hyponatremia secondary to prostatic resection, or reaming of the femur. In two patients, a preexisting cardiac condition was the etiology of the arrest. A vagally mediated response to block placement or postoperative nausea resulted in cardiac arrest in two patients. Sedation leading to respiratory depression was the mechanism of arrest in the remaining three patients (Tables 3–5).
Twelve of the 26 patients had a documented change in mental status or made specific complaints to the caregiver before the arrest, including nausea (n = 4), shortness of breath or restlessness (n = 4), light-headedness (n = 1), and tingling fingers (n = 1). One patient became less responsive but remained arousable, whereas another patient stated, “I’m going out.” The presenting cardiac rhythm was asystole in 15 (58%) patients, VF in 5 (19%) patients, and severe bradycardia, pulseless electrical activity, or ventricular tachycardia in 6 (23%) patients. Resuscitative efforts included chest compression in 18 (69%) patients and defibrillation in 8 (31%) patients. Atropine and epinephrine were administered to 21 (81%) and 18 (69%) patients, respectively. The median duration of resuscitation was 5 min (range, 0.5–85 min). Patients who survived were resuscitated more quickly than nonsurvivors: 9 ± 20 min versus 34 ± 12 min, respectively (P < 0.001). Eighteen (69%) patients who arrested during neuraxial anesthesia survived for at least an hour after the arrest. One patient died 25 days later of severe anoxic encephalopathy as a result of his cardiac arrest. The remaining 17 patients were discharged from the hospital without sequelae.
Among the 26 patients who arrested during neuraxial block, the percentage who survived to hospital discharge was not found to differ significantly between those who arrested during spinal versus epidural anesthesia (14 of 20 vs 3 of 6; P = 0.293) or according to type of surgery (4 of 8 vs 4 of 8 vs 9 of 10 for hip surgery, TURP, and other orthopedic or general surgical procedures, respectively; P = 0.115). However, hospital survival among these 26 patients was significantly associated with presenting cardiac rhythm: patients who experienced asystole had the most frequent hospital survival (13 of 15 vs 1 of 5 vs 3 of 6 for asystole, ventricular fibrillation, and “other” rhythm, respectively; P = 0.012).
During the same study period, 29 cardiac arrests occurred during general anesthesia in patients undergoing similar surgical procedures (Table 2). The mean patient age was 63 ± 20 yr (range, 7–94 yr). Eight patients were listed as ASA physical status I or II, 12 patients were ASA physical status III, and 9 patients were ASA physical status IV or V. Seventeen patients arrested because of a cardiac event (including myocardial infarction, high-degree cardiac conduction block, or dysrhythmia), five arrested because of hypoxia, six had a documented thromboembolism, and one experienced significant bleeding. Although there was no significant difference in the frequency of preoperative comorbidities, patients who arrested during general anesthesia had higher ASA classification than those who arrested during a neuraxial block (P = 0.031) and were more likely to have experienced hypotension before arrest (P = 0.045). The presenting cardiac rhythm at the time of arrest and the resuscitation efforts for patients who experienced a cardiac arrest during general versus neuraxial anesthesia were not statistically different. From univariate analysis, patients who arrested during neuraxial anesthesia had significantly better survival than those who arrested during general anesthesia (immediate survival: 69% vs 38%; P = 0.023; odds ratio [OR], 3.7; 95% CI, 1.2–11.2; hospital survival: 65% vs 31%; P = 0.013; OR, 4.2; 95% CI, 1.4–13.0).
To assess the influence of other potential patient or procedural characteristics and to determine whether neuraxial anesthesia was independently associated with improved survival after cardiac arrest after adjusting for these characteristics, a series of bivariate logistic regression analyses were performed. From these analyses, higher ASA classification (P = 0.043) and emergency surgery (P = 0.035) were found to be associated with worse survival, as were intensive resuscitation efforts (P = 0.022, P = 0.002, and P = 0.009 for use of chest compressions, defibrillation, and epinephrine, respectively). The association of neuraxial anesthesia with improved survival was found to remain statistically significant after adjusting for each patient/procedural characteristic, with the exception of ASA classification and emergency procedure. Although not statistically significant, the estimated magnitude of the association between neuraxial anesthesia and hospital survival after adjusting for ASA classification was similar to that found from the unadjusted analysis (OR, 3.0; 95% CI, 0.9–10.3), which is also similar in magnitude to the effect found if the analysis is restricted to nonemergency procedures (OR, 3.3; 95% CI, 1.0–10.9). If the analysis was restricted to neuraxial block patients who required chest compressions (rather than hemodynamic instability accompanied by loss of consciousness, according to the definition of cardiac arrest), patients who arrested during neuraxial anesthesia were still found to have significantly better hospital survival compared with those who arrested during general anesthesia (56% vs 24%; P = 0.039; OR, 4.0; 95% CI, 1.1–14.7).
In addition to the 26 arrests during neuraxial anesthesia, we also identified 1 cardiac arrest during the 20-yr study period that occurred during the performance of a peripheral (interscalene) block. Immediately after injection of 20 mL of 1% etidocaine plus 1:200,000 epinephrine, the patient seized and became asystolic. The patient was successfully resuscitated within 5 min without negative long-term sequelae.
Initial case reports from the 1940s described cardiac arrest as an inexplicable complication of spinal anesthesia (14–16). The pathophysiology for cardiac collapse was often not considered, despite the patients’ young age, healthy medical status, and initially uneventful intraoperative course; neuropsychiatric effects of hypoxic brain injury were of greater interest. Similarly, although large series of spinal anesthesia often included cardiac arrest as an observation, the significance of such a major complication was not discussed (6,10). In the last decade, reports of cardiac arrest associated with epidural block have been reported, although the overall frequency has reportedly decreased (2,7,8). Our study evaluated the association of preexisting medical conditions and intraoperative events with survival in patients experiencing a cardiac arrest during neuraxial compared with general anesthesia over 20 years at a single institution.
Several series have suggested that cardiac arrest during spinal and epidural anesthesia is not uncommon. Auroy et al. (2) reported 32 cardiac arrests among 103,730 regional anesthetics performed over a 5-month period, 7 of which were fatal. The incidence of cardiac arrest was significantly more frequent with spinal (6.4 per 10,000; 95% CI, 3.9–8.9) than with epidural (1.0 per 10,000; 95% CI, 0.2–2.9) anesthesia (P < 0.05) or peripheral nerve blocks (1.4 per 10,000; 95% CI, 0.3–4.1) (P < 0.05). Twenty (77%) of 26 patients who arrested during spinal anesthesia survived. Importantly, the size of the study by Auroy et al. (2) allowed analysis of potential variables associated with arrest and survival after arrest. For example, significant blood loss was reported at the time of cardiac arrest in nine patients, and three arrests occurred during cementing of the femur during THA. The time between the onset of spinal blockade and the occurrence of cardiac arrest was longer in nonsurvivors than in survivors (42 ± 19 minutes versus 17 ± 16 minutes) (P < 0.05). The average age of survivors was 57 ± 20 years, versus 82 ± 7 years for nonsurvivors (P < 0.05). In addition, the risk of death after cardiac arrest was increased with ASA classification and THA surgery (P < 0.05). Sedation was not noted to be the cause of any cardiac arrests, and bradycardia preceded all events. All patients who survived the cardiac arrest had complete neurologic recovery. In a subsequent investigation, Auroy et al. (3) reviewed 158,083 regional anesthetics, including 78,104 neuraxial blocks performed between August 1998 and May 1999. All 10 cardiac arrests occurred during spinal anesthesia (2.5 per 10,000; 95% CI, 0–5.1). Similar to the previous study, all arrests occurred more than 40 minutes after the intrathecal injection and were preceded by bradycardia. The three deaths occurred in elderly patients (>80 years) who were undergoing hip surgery. Although the most recent series reported less-frequent cardiac arrests, the incidence of death was actually more than the earlier review (30% versus 26%) (2,3).
Our results are similar to those of Auroy et al. (2,3). The overall rates of cardiac arrest are comparable, as is the increased frequency associated with spinal compared with epidural anesthesia. Hip replacement was the most commonly implicated surgical procedure in all series. As with previous surveys, we noted that the arrest often occurred well after establishment of the neuraxial block and was frequently associated with an intraoperative event, such as significant blood loss or cement placement during an orthopedic procedure. The likelihood of survival without neurologic sequelae was high and ranged from 65% to 74%. Although not statistically significant, our results are directionally consistent regarding patient-related variables associated with fatal cardiac arrest—increased age and higher ASA physical status. Contrary to Auroy et al. (2), we did not note a difference in survival when the arrest occurred during hip surgery compared with other surgical procedures. In addition, the presenting cardiac rhythm was asystole (not severe bradycardia) in most of our cases; asystole was associated with improved hospital survival compared with other presenting rhythms. Although, overall, it does not seem unreasonable to predict that elderly patients with multiple comorbidities undergoing a major surgical procedure would be at increased risk for fatal cardiac arrest, these results markedly differ from the cases included in the ASA Closed Claims Project.
During the initial review of the ASA Closed Claims database in 1988, Caplan et al. (13) discovered 14 cases of sudden cardiac arrest in healthy patients who had received spinal anesthesia for relatively minor surgery. The cases were similar in that the patients were young (36 ± 15 years) and healthy (ASA status I and II), the event was unexpected, and the outcome was poor despite apparently appropriate care. Anesthetic care had been in progress for an average of 36 ± 18 minutes at the time of the arrest. Bradycardia, hypotension, and/or cyanosis frequently preceded the arrest. The authors concluded that undiagnosed respiratory insufficiency, high sympathetic blockade, or both may have contributed to occurrence or outcome and recommended 1) that pulse oximetry be used whenever sedatives are administered or the patient’s ability to communicate is impaired; 2) that epinephrine be administered early in cases of sudden bradycardia, hypotension, or both; and 3) that a full resuscitative dose of epinephrine be given immediately upon cardiac arrest. The efficacy of early and aggressive pharmacologic intervention has subsequently been confirmed clinically and in animal models (17).
While our series noted a significant decrease in the frequency of cardiac arrests after 1988, particularly during spinal anesthesia, it is unclear whether the outcome has been affected by these recommendations (5,18). In 2001, the Closed Claims database contained 181 claims involving cardiac arrest during spinal or epidural anesthesia.1 The outcome in 161 (89%) of the claims was brain damage or death. These 161 cases accounted for 14% of all regional anesthesia-related claims in the database. Consistent with the earlier Closed Claims analyses, most cases involved ASA status I or II patients (53%) undergoing nonemergency surgery (70%) with spinal (56%) or epidural (44%) anesthesia. Importantly, prompt resuscitative efforts did not improve outcome; only 16% of cases were considered preventable with better monitoring.
Our review included four cardiac arrests during neuraxial block in patients younger than 50 years of age. Three of these patients were similar demographically to those in the ASA Closed Claims database: healthy and undergoing minor surgery with documented sensory levels above T4 after the initiation of spinal (two patients) or epidural (one patient) block. All three patients were successfully resuscitated without neurologic sequelae. The fourth patient, a healthy 38-year-old parturient with an indwelling epidural catheter and a T8 sensory level, experienced a fatal amniotic fluid embolus during rupture of her amniotic membranes. The survival rate among these patients was similar to the overall rate for our series and suggests that neither the patient population nor the neurologic outcome of patients included in the ASA Closed Claims Project is representative of those who arrest during neuraxial block.
Theories regarding the mechanism by which neuraxial block contributes to cardiac arrest involve a circulatory etiology. The evidence for an underlying respiratory source etiology is sparse. For example, sensory levels up to T4 do not result in hypoventilation. Likewise, although excessive sedation was speculated to have contributed to many of the early arrests during spinal anesthesia that occurred before the widespread use of pulse oximetry, several series report oxygen saturations of >than 90% at the time of arrest; often the patients had not received sedatives before the event (2,7,8). Although this is speculative, it is likely that the decrease in preload associated with neuraxial block results in a shift in cardiac autonomic balance toward the parasympathetic system. This secondarily results in bradycardia. At least three mechanisms have been proposed, including activation of the low-pressure baroreceptors in the right atrium, the receptors within the myocardial pacemaker cells, and mechanoreceptors in the left ventricle (stimulating a paradoxical Bezold-Jarisch response). In addition, a high sympathetic level may directly favor vagal tone; sedation, hypoxemia, hypercarbia, and chronic medications (such as β-adrenergic antagonists) may contribute to the development and severity of bradycardia (7). Intravascular fluid administration, the administration of mixed α- and β-agonists, and vagolytic therapy have all been advocated to decrease the frequency of and improve the survival associated with cardiac arrest during neuraxial block (5). Among the 17 patients in our series who arrested during neuraxial block and were successfully resuscitated, restoration of hemodynamic stability was accomplished early after diagnosing the cardiac arrest. Only two patients required more than five minutes of resuscitative efforts; intensive and prolonged resuscitations were less likely to be successful.
The physiologic factors that contribute to cardiac arrest during neuraxial block remain incompletely defined. More relevant is patient survival after cardiac arrest under neuraxial versus general anesthesia. Laboratory and clinical series report conflicting outcomes with respect to anesthetic technique. In a canine model, animals under spinal anesthesia did not exhibit normal increases in catecholamine release after cardiac arrest. As a result, the coronary perfusion pressure was less than that needed for successful resuscitation. Only the administration of exogenous epinephrine increased coronary perfusion pressure above the critical threshold (19). Likewise, clinical studies have reported conflicting rates of immediate survival related to anesthetic technique. Biboulet et al. (20) reported 11 cardiac arrests among 101,769 anesthetics. The incidence of cardiac arrest was more frequent during neuraxial (6 per 10,000) compared with general (0.8 per 10,000) anesthesia; 4 of 6 patients survived cardiac arrest in the general anesthesia group, compared with only 1 of 5 in the spinal anesthesia group. In a more recent study involving 518,294 anesthetics over an 11-year period, Sprung et al. (1) reported 223 cardiac arrests, noting a decreased frequency of cardiac arrests during regional compared with general anesthesia (1.5 per 10,000 and 5.5 per 10,000, respectively). The relatively low frequency among the regional anesthesia group precluded meaningful comparisons with the general anesthesia group with respect to predictors of survival.
Our series, conducted over a 20-year period, controlled for surgical procedure to compare survival after cardiac arrest during neuraxial versus general anesthesia. Univariately, patients who arrested during neuraxial block were more likely to survive than those who arrested during general anesthesia: 69% vs 38% for immediate survival and 65% vs 31% for hospital survival, respectively. The ORs corresponded to approximately a fourfold increase in the likelihood of surviving if the arrest occurred during a neuraxial block. The increased likelihood of survival persisted after adjusting for all patient/procedural characteristics, with the exception of ASA physical status and emergency surgery. However, after adjusting for ASA classification and emergency status, the magnitude of the association between improved survival and neuraxial anesthesia was similar but no longer statistically significant. Conservatively, this suggests that the likelihood of survival after cardiac arrest during neuraxial anesthesia is equal to or more than the likelihood of survival after cardiac arrest during general anesthesia.
In any study involving perioperative mortality, it is important to consider the source and integrity of the database. Our institution is a tertiary referral center. However, 90% of patients live within a 500-mile radius and receive their primary care here. Despite the advantage of consistent data collection, the results may not be applicable to the general population. Likewise, the statistical power for assessing the association of a specific patient or procedural factor with survival is dependent on the prevalence of the risk factor among patients who experienced an arrest. As a result, nonstatistically significant findings should be interpreted with caution when the number of patients with and without the risk factor is small.
In summary, this retrospective study evaluated the frequency of cardiac arrest and predisposing factors associated with survival between 1983 and 2002. The frequency of cardiac arrest during neuraxial anesthesia decreased significantly over the study interval and was more frequent with a spinal compared with an epidural technique. Importantly, 65% of patients in our series who arrested during neuraxial anesthesia were resuscitated without neurologic sequelae; few cases resembled the cardiac arrests included in the ASA Closed Claims Project with respect to demographics or outcome. After controlling for patient and procedural variables, we conclude that a cardiac arrest during neuraxial anesthesia is associated with an equal or greater likelihood of survival compared with cardiac arrest during general anesthesia.
1. Sprung J, Warner ME, Contreras MG, et al. Predictors of survival following cardiac arrest in patients undergoing noncardiac surgery: a study of 518,294 patients at a tertiary referral center. Anesthesiology 2003;99:259–69.
2. Auroy Y, Narchi P, Messiah A, et al. Serious complications related to regional anesthesia. Anesthesiology 1997;87:479–86.
3. Auroy Y, Benhamou D, Bargues L, et al. Major complications of regional anesthesia in France. Anesthesiology 2002;97:1274–80.
4. Liu SS, McDonald SB. Current issues in spinal anesthesia. Anesthesiology 2001;94:888–906.
5. Pollard JB. Cardiac arrest during spinal anesthesia: common mechanisms and strategies for prevention. Anesth Analg 2001;92:252–6.
6. Phillips OC, Ebner H, Nelson AT, Black MH. Neurologic complications following spinal anesthesia with lidocaine: a prospective review of 10,440 cases. Anesthesiology 1969;30:284–9.
7. Liguori GA, Sharrock NE. Asystole and severe bradycardia during epidural anesthesia in orthopedic patients. Anesthesiology 1997;86:250–7.
8. Geffin B, Sharpiro L. Sinus bradycardia and asystole during spinal and epidural anesthesia: a report of 13 cases. J Clin Anesth 1998;10:278–85.
9. Tarkkila PJ, Kaukinen S. Complications during spinal anesthesia: a prospective study. Reg Anesth 1991;16:101–6.
10. Moore DC, Bridenbaugh LD. Spinal (subarachnoid) block: a review of 11,574 cases. JAMA 1966;195:907–12.
11. Carpenter RL, Caplan RA, Brown DL, et al. Incidence and risk factors for side effects of spinal anesthesia. Anesthesiology 1992;76:906–16.
12. Olsson GL, Hallen B. Cardiac arrest during anaesthesia: a computer aided study in 250,543 anaesthetics. Acta Anaesthesiol Scand 1988;32:653–64.
13. Caplan RA, Ward RJ, Posner K, Cheney FW. Unexpected cardiac arrest during spinal anesthesia: a closed claims analysis of predisposing factors. Anesthesiology 1988;68:5–11.
14. Howkins J, McLaughlin CR, Daniel P. Neuronal damage from temporary cardiac arrest. Lancet 1946;1:488–92.
15. Kral VA. Neuropsychiatric sequelae of cardiac arrest during spinal anaesthesia. Can Med Assoc J 1951;64:138–42.
16. Noble AB. Cerebral anoxia complicating spinal anesthesia. Can Med Assoc J 1946;54:378–9.
17. Cheney FW. The American Society of Anesthesiologists Closed Claims Project: what have we learned, how has it affected practice, and how will it affect practice in the future? Anesthesiology 1999;91:552–6.
18. Aromaa U, Lahdensuu M, Cozanitus DA. Severe complications associated with epidural and spinal anaesthesias in Finland 1987–1993: a study based on patient insurance claims. Acta Anaesthesiol Scand 1997;41:445–52.
19. Rosenberg JM, Wortsman J, Wahr JA, et al. Impaired neuroendocrine response mediates refractoriness to cardiopulmonary resuscitation in spinal anesthesia. Crit Care Med 1998;26:533–7.
20. Biboulet P, Aubas P, Dubourdieu J, et al. Fatal and non fatal cardiac arrests related to anesthesia. Can J Anaesth 2001;48:326–32.
1 Caplan RA, unpublished data, presented at the ASRA Conference on Local Anesthetic Toxicity, November 18, 2001.© 2005 International Anesthesia Research Society