| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the Departments of *Anesthesiology, Perioperative and Pain Medicine, Cardiac Surgery, and Cardiology, Children's Hospital Boston, Harvard Medical School, Boston, Massachusetts.
Address correspondence and reprint requests to Kirsten C. Odegard, MD, Cardiac Anesthesia Service, Children's Hospital Boston, 300 Longwood Avenue, Boston, MA 02115. Address e-mail to kirsten.odegard{at}childrens.harvard.edu.
Abstract
BACKGROUND: The frequency of anesthesia-related cardiac arrests during pediatric anesthesia has been reported between 1.4 and 4.6 per 10,000 anesthetics. ASA physical status >III and younger age are risk factors. Patients with congenital cardiac disease may also be at increased risk. Therefore, in this study, we evaluated the frequency of cardiac arrest in patients with congenital heart disease undergoing cardiac surgery at a large pediatric tertiary referral center.
METHODS: Using an established data registry, all cardiac arrests from January 2000 through December 2005 occurring in the cardiac operating rooms were reviewed. A cardiac arrest was defined as any event requiring external or internal chest compressions, with or without direct cardioversion. Events determined to be anesthesia-related were classified as likely related or possibly related.
RESULTS: There were 41 cardiac arrests in 40 patients (median age, 2.9 mo; range, 2 days to 23 yr) during 5213 anesthetics over the time period, for an overall frequency of 0.79%; 78% were open procedures requiring cardiopulmonary bypass and 22% closed procedures not requiring cardiopulmonary bypass. Eleven cardiac arrests (26.8%) were classified as either likely (n = 6) or possibly related (n = 5) to anesthesia, (21.1 per 10,000 anesthetics) but with no mortality; 30 were categorized as procedure-related. The incidence of anesthesia-related and procedure-related cardiac arrests was highest in neonates (P < 0.001). There was no association with year of event or experience of the anesthesiologist.
CONCLUSION: The frequency of anesthesia-related cardiac arrest in patients undergoing cardiac surgery is increased, but is not associated with an increase in mortality. Neonates and infants are at higher risk. Careful preparation and anticipation is important to ensure timely and effective resuscitation.
Despite advances in perioperative care, including monitoring and drugs, unexpected cardiac arrest remains a significant hazard during anesthesia (1–5). Anesthesia-related morbidity and mortality is more frequent in children than in adults, and is more frequent in infants and younger children than in older children (1,3,4,6–10). The incidence of intraoperative cardiac arrest may vary, depending on the study era, whether cardiac arrest was directly caused by anesthesia or whether anesthesia was a contributing factor (11).
Besides age, additional risk factors contributing to cardiac arrest during pediatric anesthesia include higher ASA physical status, emergency surgery, and comorbid conditions. The risk for cardiac arrest during pediatric cardiac surgery has not been systematically evaluated. As a possible benchmark for the incidence of cardiac arrest during pediatric anesthesia, the Pediatric Perioperative Cardiac Arrest (POCA) Registry reported an incidence of 1.4 ± 0.45 per 10,000 anesthetics, using data compiled between 1994 and 1997, with a mortality rate of 26% for those patients who sustained an arrest (3). They defined anesthesia- related events as those events in which anesthesia personnel or the anesthesia process played at least some role in the cause of the cardiac arrest. However, this is a voluntary reporting registry and under-reporting of events is likely. Further, the registry did not focus on patients with cardiac disease (only 17% in the initial report). In two recent large, single-institution audits of adverse events related to pediatric anesthesia, the incidence of anesthesia-related cardiac arrest was 3.3 and 4.6 per 10,000 anesthetics, respectively, with no mortality (8,10). The risk for anesthesia-related cardiac arrest in pediatric patients with congenital heart disease (CHD) undergoing cardiac surgery was also not determined in either of these studies.
In a detailed audit of all anesthesia cases performed at Children's Hospital Boston from January 2000 through end December 2004, the frequency of anesthesia-related cardiac arrests and deaths was 2.7 per 10,000 anesthetics, with five cases of death or permanent disability related to these arrests. This audit included patients managed in the cardiac operating rooms (ORs) but not the catheterization laboratory. In a multivariate analysis, the predicting risk of cardiac arrest had two patient-related factors and one practitioner-related risk factor; ASA physical status 
III, patient age 
1 yr and OR time 40% (12).
As an independent quality assurance measure, the Cardiac Anesthesia Service (CAS) at Children's Hospital Boston has maintained a complete registry of all patients with CHD anesthetized in the cardiac OR and catheterization laboratory since January 2000. Using this registry, the purpose of this study was to examine the frequency of anesthesia-related cardiac arrests in patients with CHD undergoing cardiac surgery at a large pediatric tertiary referral center.
METHODS
After obtaining IRB approval, all intraoperative cardiac arrests occurring during pediatric cardiac surgery at Children's Hospital Boston, from January 1, 2000 through December 31, 2005 were reviewed.
Since January 2000, the CAS has maintained a registry for all anesthetics delivered in the cardiac ORs, cardiac catheterization laboratories, and off-site areas such as the cardiac magnetic resonance imaging suite. This registry was developed as a quality improvement tool for the CAS. A data form was included with each anesthesia record, and after completion by the anesthesia staff or fellows, the form and copy of the anesthetic record were filed and cross-referenced with the daily schedule and billing records. Information on the data forms included patient demographics, diagnoses, staff anesthesiologist and surgeon, nature of procedure, anesthetic technique, details of cardiopulmonary bypass (CPB), the disposition of the patient at completion of the procedure, and documentation of any events or complications that occurred during the procedure.
A cardiac arrest was defined as any event requiring external chest compression or internal cardiac massage with or without cardioversion. The charts, including the anesthesia and surgical records, of each patient who suffered a cardiac arrest while under anesthesia were reviewed initially by three of the authors (KCO, JDN, PCL), then at a joint meeting of all cardiac anesthesia faculty and, finally, with one of the cardiac surgery staff (EB). The cardiac anesthesia and surgical staff were asked to determine which factors could have contributed to the cardiac arrest, including anesthesia technique, the patient's premorbid condition, the surgical procedure, and other possible contributing causes.
The surgical procedures were divided into two groups; 1) open procedures requiring CPB, and 2) closed procedures not requiring CPB. The cardiac arrests that occurred in the open procedure group were further divided into pre-CPB events, which were defined as occurring from induction of anesthesia until aortic cannulation, and post-CPB events, occurring from the time of decannulation to transfer from the cardiac OR to the cardiac intensive care unit. Patients who were unable to be weaned from CPB or experienced a cardiac arrest after transfer to the cardiac intensive care unit were not included.
The cardiac arrests were categorized as 1) anesthesia-related, according to the definition used by the POCA registry, or 2) not anesthesia-related or procedure-related. Arrests determined to be anesthesia-related were sub classified as likely related or possibly related and are defined in Table 1.
|
Once the cardiac arrest events had been identified from the CAS registry, the anesthetic record and patient chart were examined to review each event in detail. Information sought included patient characteristics of ASA physical status classification and age (grouped as <1 mo, >1 mo but <12 mo, 1–5 yr, 6–18 yr, and >18 yr), anesthesia practitioner-related characteristics including years of experience and occurrence of events over time, and procedural characteristics including cardiac anatomy and surgical details. As with many other pediatric centers performing congenital cardiac surgery, many adults with CHD are now managed in our institution because of the need for specialized care. To be complete with our experience over the time frame, we included these patients in our analysis of events. Neither the patients' families nor patients who had a cardiac arrest were contacted, but a review of all subsequent clinic and admission notes was performed for possible longer-term sequelae, in particular neurological injury.
To be consistent with commonly used descriptions, the frequency of perioperative cardiac arrest over the 6-yr period was calculated per 10,000 anesthetics and per 100 surgical procedures. Descriptive analyses of the patients who had cardiac arrests are presented. It was not possible to determine the risk for cardiac arrest based on ASA status, as the majority of patients were classified as either status III or IV. 
2 analysis was used to evaluate whether anesthesia or nonanesthesia-related events were associated with patient age group, and anesthesiologist case volume and years of experience. Since the number of events was low, it was not possible to undertake a stratified analysis considering the effects of patient age and case volume together. The Mantel Haenszel test for linear association was used to look for a linear trend for increasing or decreasing events per year. In view of the relatively small number of anesthesia-related events, a conservative significant P < 0.001 was chosen. Data analysis was conducted using SPSS (v 14.0, SPSS, Chicago, IL) statistical software.
RESULTS
Between January 1, 2000 and December 31, 2005, 5213 anesthetics were administrated to patients undergoing cardiac surgery for 4069 (78%) open procedures and 1144 (22%) closed procedures. Over this time period, the early mortality for all surgical patients to the time of discharge from hospital was 1.8%. Three patients died in the OR because they could not be weaned from CPB, and 30 patients were placed directly onto extracorporeal membrane oxygenation support (ECMO) from CPB; five of these patients had been transported to the OR already on ECMO before the surgical procedure. None of these 33 patients received resuscitation for cardiac arrest after surgery and are not included in the analysis.
There were 41 cardiac arrests in 40 patients during this 6-yr period, for an overall frequency of 0.79%; one patient experienced a cardiac arrest pre- and post-CPB. Of the 40 patients, 27 were male, the median age was 2.9 mo (range, 2 days to 23 yr) and median weight 4 kg (range, 2.0–80 kg); 14 patients were in ASA class III (35%) and 26 patients in class IV (65%).
Surgical Procedures
Details of the surgical procedures are shown in Table 2. The overall frequency of a cardiac arrest was 0.83 per 100 closed procedures and 0.7 per 100 open procedure (0.16 for patients undergoing a septal defect repair, 0.8 for cavopulmonary connection, 1.2 for systemic ventricle outflow reconstruction, and 1.2 for pulmonary ventricle outflow reconstruction). When analyzed by specific procedures, the highest risk procedures for cardiac arrest were truncus arteriosus repair (n = 23, frequency 17.4 per 100 procedures), modified Blalock-Taussig shunt operation for pulmonary atresia with intact ventricular septum (n = 28, frequency 3.6 per 100 procedures), neonates with coarctation or interruption of the aorta with ventricular septal defect repair (n = 57, frequency 3.5 per 100 procedures), and Stage I palliation for hypoplastic left heart syndrome (n = 45, frequency 2.2 per 100 procedures).
|
Of the 41 cardiac arrests, 11 (26.8%) were classified as either likely-related or possibly-related to anesthesia, for an incidence of 21.1 per 10,000 anesthetics (0.21%); the other 30 cardiac arrests in 29 patients were categorized as surgery or procedure-related, for an incidence of 0.58 per 100 procedures.
Anesthesia-Related Events
The demographics and age range of patients who had an arrest either likely related or possibly related to anesthesia are shown in Tables 3 and 4. The median age for patients with an anesthesia-related event was 122 days, and for procedure-related events was 68 days. For both anesthesia-related and procedure-related events, the risk was significantly higher in neonates compared with that in other age groups (P < 0.001). In patients >18 yr, the frequency of a cardiac arrest was 1 per 100 procedures, and all three events were determined to be procedure-related after surgical review.
|
|
|
Of the 5213 anesthetics, an inhaled induction was used in 1018 patients (19.5%) with one anesthesia-related arrest (9.8 per 10,000 anesthetics); the remaining 10 anesthetic-related events occurred in patients who had received an IV induction (23.8 per 10,000 anesthetics). Patients receiving an inhaled induction were older (median age, 2.8 yr; range, 5 days to 18 yr) and larger (median weight, 12.7 kg; range, 2.2–75 kg) than the median age and weight for the entire cohort.
There were no differences in the frequency of events over time (Table 6), nor a relationship to the experience of individual anesthesia staff (P = 0.84, data not shown). A central venous line (CVL) was placed in 1821 patients (35% of total cases); there was one arrest secondary to VF induced by passing of the guide wire, for an incidence of 0.05 per 100 CVL procedures.
|
Nonanesthesia-Related Cardiac Arrests
Thirty of the 41 cardiac arrests (73.2%) were categorized as being procedure-related (Table 7); 24 occurred during open procedures (0.59 per 100 open procedures) and 6 during closed procedures (0.52 per 100 closed procedures). Of the 24 cardiac arrests occurring in the open procedure group, 15 cardiac arrests (62.5%) occurred before CPB; 13 patients had brief episodes of VF after sternotomy due to direct surgical stimulation of the heart during dissection, and two patients received cardiopulmonary resuscitation because of severe hemorrhage after direct injury to the heart during sternotomy. All patients who had a procedure-related arrest before CPB had successful completion of surgery and were discharged home. Nine patients (27.5%) experienced a cardiac arrest after weaning from CPB; three of these patients were placed on ECMO and one died before discharge.
|
Of the six patients in the closed group who had a procedure-related cardiac arrest, two patients were placed on ECMO and one of these patients died. Two patients with pulmonary atresia, intact ventricular septum, and right ventricle dependent coronary circulation developed significant myocardial ischemia and VF during placement of a modified Blalock-Taussig shunt; both these patients died, one patient died in the OR and the other patient was placed on ECMO but died later in the ICU. The remaining four patients completed their surgery and were discharged home.
The median hospital stay for all patients was 10 days (range, 4–118 days); one patient died in the OR and five patients were placed on ECMO after the cardiac arrest (three survivors). Overall, 37 of 40 (92.5%) patients who had a cardiac arrest survived to discharge. All but one of the 40 cardiac arrests occurred during normal weekdays and working hours (7:00 am–7:00 pm); one nonanesthesia-related cardiac arrest occurred out of hours (weekend morning).
DISCUSSION
This single-center review of cardiac arrests classified as either likely related or possibly related to anesthesia in patients with congenital heart disease undergoing cardiac surgery, demonstrates a 4.6–15 fold higher frequency of anesthesia-related cardiac arrest when compared with multi-institutional and single-center reports of cardiac arrest during pediatric anesthesia. However, there was no anesthesia-related morbidity or mortality in our study.
Factors contributing to cardiac arrest during pediatric anesthesia may be related to patient, practitioner, procedural, and hospital characteristics. Younger age (<1 yr), higher ASA status, and emergency surgery have been identified as patient-related factors (3,12). The increased risk for anesthesia-related events in neonates and infants has also been identified in other reports (1,4,8,13), and was also the case in our series for patients undergoing cardiac surgery.
Patients with CHD typically have a higher ASA physical status, but as the vast majority of patients in our cardiac anesthesia registry were classified as either ASA III or IV, we were unable to evaluate ASA status as a specific risk factor. Keenan and Boyan reported that cardiac arrest was three times as frequent and death twice as frequent in ASA physical class III or IV patients when compared to patients with a lower ASA physical status (2). Tiret et al. also demonstrated an increased risk for complications with higher ASA physical class (13). In the POCA study, death after anesthesia-related cardiac arrest was predicted strongly by the ASA physical class (37% ASA III and IV versus 4% ASA I and II) (3). Despite the fact that our patients were at higher risk based on ASA classification, there was no anesthesia-related mortality or morbidity; in all cases, the event was immediately recognized and resuscitation was brief and effective.
Information regarding the anesthetic risk associated with pediatric cardiac surgery is limited and not contemporary. Strong et al. (14) reported a 3% anesthetic mortality in 100 infants younger than 1-yr-old undergoing cardiac surgery in 1966. Hickey et al. (15) in 1984 reported an overall anesthetic complication rate of 2%, with no mortality in 500 consecutive patients undergoing cardiac surgery. There was no mention of intraoperative cardiac arrest in this article, but there were nine intraoperative deaths deemed unrelated to anesthesia.
It can be difficult to distinguish between anesthesia-related factors and procedural-related factors contributing to cardiac arrest in a patient with underlying cardiac disease, but the possible association between altered coronary perfusion and myocardial ischemia and cardiac arrest, which can be inferred in 7 (63.3%) of our anesthesia-related cardiac arrests is noteworthy. Coronary perfusion is reduced in patients who have uncontrolled or continuous run-off of blood flow from the systemic to pulmonary circulation, and therefore low aortic root diastolic pressure. As noted in Tables 5 and 7, there was this substrate in our cohort of patients with a diagnosis of truncus arteriosus, and patients with a ductus dependent systemic circulation, such as hypoplastic left heart syndrome and interruption of the aortic arch or coarctation with ventricular septal defect. Patients with altered coronary blood flow, such as those with pulmonary atresia with intact ventricular septum and a right ventricle dependent coronary circulation from fistulae, are also at increased risk for ischemia.
Myocardial ischemia and cardiac arrest may occur suddenly after induction of anesthesia and introduction of positive pressure ventilation in the group of patients described above. Reduced coronary perfusion pressure may be an important contributing factor, and can be caused by a decrease in preload and a reduction in aortic root pressure, such as occurs if pulmonary vascular resistance decreases and pulmonary run-off is increased, thereby decreasing systemic perfusion. Subendocardial perfusion is particularly tenuous if diastolic perfusion time is reduced by concomitant tachycardia. These patients also have a limited ability to increase coronary blood flow when myocardial oxygen demand is increased, such as secondary to increased contractility or wall stress in response to a surgical stimulus if there is an inadequate depth of anesthesia to blunt a stress response (16–18). While we categorized the majority of pre-CPB procedure-related events as occurring secondary to mechanical stimulation, such as electrocautery, it is possible some of these patients had marginal coronary blood flow that increased their susceptibility to arrhythmia, and VF in particular. Finally, the importance of maintaining diastolic pressure and coronary perfusion and avoiding tachycardia is important in the setting of severe left ventricle hypertrophy, as seen in two patients who had anesthesia-related arrests (Williams–Beurin Syndrome and hypertrophic cardiomyopathy).
Interestingly, our six likely anesthesia-related cardiac arrests, five patients had an event around the time of induction of general anesthesia; two were airway-related events, two ischemia-related events, and one during a CVL placement. The induction of general anesthesia and establishing monitoring and vascular access is a critical phase, and it is therefore important that distractions are minimized and staffs are focused prepared for sudden adverse events.
The most common causes of anesthesia-related cardiac arrest have been reported due to pharmacologic interactions, drug overdose, and cardiovascular instability (3). More recently, the POCA registry reported on 163 anesthesia-related cardiac arrests between 1998 and 2003. Although the incidence per 10,000 anesthetics was not reported, there was a decrease in the proportion of infants suffering cardiac arrest and in the number of medication-related cardiac arrests (19). As might be expected due to the presence of significant underlying cardiac disease, the causes of cardiac arrest in our patients are different from those reported by the POCA registry. In our series, we had one anesthesia procedure-related event (CVL guidewire), one medication-related event (neostigmine), two events caused by loss of the airway, and seven that were determined to be secondary to myocardial ischemia. In our patients, general anesthesia was most commonly maintained with IV drugs, in part explaining the absence of events related to myocardial depression from inhaled anesthetics. None of the anesthesia-related events occurred during rapid transfusion or intravascular volume shifts, nor were they related to biochemical or metabolic perturbations as reported in the POCA registry.
Practitioner-related factors did not appear to contribute to cardiac arrest in our series, and there was no relation to year of event or experience of the anesthesiologist, compared with a recent report of all anesthesia cases performed at Children's Hospital Boston from January 2000 through end December 2004, where one of the predicting risks of cardiac arrest was OR time 40%. (12) Keenan et al. (20) reported that the presence of a pediatric anesthesiologist substantially decreased the frequency of anesthetic-related cardiac arrest in infants. The CAS at Children's Hospital Boston is a separate division within the Department of Anesthesiology, Pain and Perioperative Medicine. There are currently 10 full-time academic attending staffs in the CAS, all of whom have additional training and experience in a wide range of areas including pediatric anesthesia, cardiac anesthesia, critical care medicine, general pediatrics, internal medicine, and pediatric cardiology. The combined years of experience for all cardiac anesthesia attending staff at the end of 2005 was 95 yr, with a median of 8 yr (range, 1–25 yr). The 11 anesthesia-related arrests involved seven different attending staff, and there was no relationship with years of experience or case volume.
Operations for CHD carry a significant risk of death for all but the simplest lesions. Using a risk adjusted method (RACHS-1), Jenkins reported a decreased mortality rate when complex pediatric cardiac surgery was performed in high volume when compared with that in low volume centers (21). The mortality rates for each RACHS category were derived from 1994 to 1996 data. In a more recent review of mortality rates drawn from the Congenital Heart Surgeon's Society between 2002 and 2004, significant decreases in the mortality rates for RACHS-1 categories were reported (22). This improvement reflects the advances in the management of congenital heart surgery over the past decade and, of note, there was no association demonstrated in this article with surgical volume; rather, mortality was influenced by case-mix. The authors concluded that risk-stratification based on mortality was less useful and that other factors determined outcome in high-quality institutions.
Silber et al. (23) evaluated patient and hospital characteristics associated with death after two commonly performed surgical procedures in adults, and noted that an "adverse event occurrence" (for example, cardiac arrest in our patients) was primarily associated with patient-related characteristics, while the mortality related to a specific adverse event or "failure to rescue" (such as inability to successfully resuscitate from a cardiac arrest in our patients) was associated primarily with hospital characteristics. Specific hospital characteristics included in this model were: number of hospital beds, daily census, total number of physicians, experience of surgical and anesthesia staff and nursing ratios. The structure of our pediatric cardiovascular program, including a dedicated CAS, may be an important hospital characteristic that contributed to the very low mortality (or "failure to rescue" rate) in our patients. There are no data to indicate whether there are differences in cardiac anesthesia-related morbidity and mortality between high- and low-volume centers, and although speculative, it is likely the systems and process of care remain strong determinants of outcome. The ORs, equipment, vasoactive infusions, and resuscitation drugs are set-up identically for every case, and there is one-to-one cover with an attending anesthesiologist assigned for each case. Surgical house staffs are immediately available in the ORs at the time of anesthesia induction. Further, there are dedicated and experienced nursing staffs in the cardiac ORs who work closely to assist the cardiac anesthesia staff during induction and during all critical phases of the procedures.
Possible methodological weaknesses associated with our study include the possibility of incomplete reporting. Although we tried to maintain a complete registry with completion of a data form for each case, we still relied on self-reporting of adverse events. Rigorous follow-up of anesthesia records each day by office staff and cross-referencing with our business office helped minimize missing data. In addition, our report does not quantify "near misses" during cardiac anesthesia, such as hypotension or rhythm changes that require immediate intervention to prevent circulatory collapse and arrest.
Our study is the first to evaluate the risk for cardiac arrest in children undergoing cardiac surgery. The frequency is increased compared with that published for cardiac arrest during noncardiac pediatric anesthesia, and infants appear to be at the highest risk. This is perhaps not surprising, given the underlying pathophysiology and complex surgical procedures in patients with CHD. Despite the increased frequency of events, there was no failure to rescue after the arrests, which supports the notion, that these patients may be optimally managed by a dedicated team of cardiac anesthesiologists working closely with the cardiac surgical and OR nursing staff to promptly manage critical events.
Footnotes
Accepted for publication April 10, 2007.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  P. J. Davis When Assessing What We Know We Don't Know Is Not Enough: Another Perspective on Pediatric Outcomes Anesth. Analg., August 1, 2007; 105(2): 301 - 303. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
