Risks of Noncardiac Operations and Other Procedures in Children With Complex Congenital Heart Disease

Transcription

1 Risks of Noncardiac Operations and Other Procedures in Children With Complex Congenital Heart Disease Scott C. Watkins, MD, Brent S. McNew, MD, and Brian S. Donahue, MD, PhD Department of Anesthesiology, Vanderbilt University Medical Center, Monroe Carell Jr. Children s Hospital at Vanderbilt, Nashville, Tennessee Background. Children with complex congenital heart disease entail risk when undergoing noncardiac operations and other procedures requiring general anesthesia. To address concerns regarding intraoperative instability, need for postoperative mechanical ventilation, and postoperative hospital length of stay (LOS), we present our 5-year experience with 71 patients with complex congenital heart disease who underwent 252 surgical procedures. Methods. We reviewed the records of all patients from July 2006 to January 2011 who underwent a cardiac procedure with a Risk Adjustment for Congenital Heart Surgery-1 score of 6, and included all who underwent noncardiac procedures during this interval. Perioperative data were gathered to identify patients at risk for induction and maintenance instability, need for postoperative mechanical ventilation, and postoperative hospital LOS. Univariate predictors of these outcome variables were evaluated and entered into stepwise regression algorithms to determine independent variables. Results. We identified 252 procedures that were performed on 71 patients during the study interval. These procedures were performed under 173 general anesthesias. Using each general anesthesia as a case, induction instability was independently associated with stage of palliation before cavopulmonary shunt, case complexity, and preoperative use of angiotensin-converting enzyme inhibitor in a multivariate logistic regression. Maintenance instability was independently associated with case complexity and preoperative use of digoxin and inotropes. Among the 145 cases where the patient was not intubated before the procedure, postoperative need for mechanical ventilation was associated only with preoperative hospital LOS exceeding 14 days. Finally, the resulting linear regression model showed postoperative hospital LOS was independently associated with preoperative hospital LOS exceeding 14 days, presence of moderate ventricular dysfunction, preoperative use of an inotrope, and negatively associated with use of digoxin. Conclusions. Within this population, we have identified independent risk factors for specific clinical outcomes. before stage II palliation, undergoing more invasive procedures, and receiving inotropes, angiotensin-converting enzyme inhibitors, or digoxin appear to be at risk for intraoperative hemodynamic instability. with preoperative hospital LOS exceeding 14 days appear to be at greater risk for requiring postoperative mechanical ventilation. with preoperative LOS exceeding 14 days, with ventricular dysfunction, receiving inotropes, and not receiving digoxin appear to be at risk for protracted hospitalization. Application of these results should assist clinicians in assessing perioperative risk. (Ann Thorac Surg 2013;95:204 11) 2013 by The Society of Thoracic Surgeons Children with congenital heart disease (CHD) often undergo noncardiac operations, or other procedures, for reasons related to their cardiac condition or for other manifestations of congenital syndromes. Children with CHD are at greater perioperative risk than children without CHD [1 3], but current reports only examine death or cardiac arrest and do not stratify the CHD population. Perioperative management of children with complex CHD is therefore based largely on retrospective case reports, case series, and institutional practice culture. This makes it difficult to extrapolate published Accepted for publication Sept 7, Address correspondence to Dr Donahue, Vanderbilt University Medical Center, Department of Anesthesiology, 2200 Children s Way, Nashville, TN 37232; brian.donahue@vanderbilt.edu. findings to the general population. Practitioners are therefore uncertain whether these patients can safely undergo elective surgical procedures, whether they require invasive monitoring or intensive care unit admission, the relative safety of laparoscopic or thoracoscopic procedures, the type of anesthesia and qualifications of the anesthesia team required, and whether they should be referred to specialty centers. The literature identifies death and cardiac arrest as risks associated with CHD in pediatric surgical patients [1 3]. We recently reported results from a population of patients with hypoplastic left heart syndrome undergoing noncardiac procedures at our institution [4], describing risk factors and a practice plan. Here, we expand the population using the Risk Adjustment for Congenital Heart Surgery (RACHS-1) scoring system [5] to identify 2013 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc

2 Ann Thorac Surg WATKINS ET AL 2013;95: SURGICAL RISK FOR COMPLEX CHD 205 children having undergone RACHS-1 category 6 procedures because this is a recognized high-risk population [6 8]. We have observed that when these patients present for noncardiac operations or for noninvasive procedures involving general anesthesia, physicians are keenly interested in several specific issues: identifying which patients are likely to develop hemodynamic instability during anesthetic induction or maintenance, which patients will require postoperative mechanical ventilation, and what factors affect postoperative hospital length of stay (LOS). Using data from a 4.5-year cohort at our institution, we used univariate and multivariate approaches to test whether specific preoperative variables were associated with these important outcomes. Material and Methods After Institutional Review Board approval, we queried The Society of Thoracic Surgeons database for our institution between July 2006 and January This database provides a comprehensive list of all patients who have undergone cardiac operations at our institution. We searched for all patients who underwent a RACHS-1 category 6 procedure during the study interval and identified 113 patients. We reviewed the electronic medical record for each patient and collected data for each noncardiac procedure performed on these patients. A total of 252 noncardiac procedures were performed on 71 of these patients. Perioperative data were extracted from patient records using the following definitions. Preoperative hospital LOS was classified as greater than 14 days or 14 days or less. Echocardiographic evidence of ventricular dysfunction was classified as none or mild or greater than mild, as stated on the echocardiogram report within 1 week before the operation. The stage of palliation was defined as: prestage I palliation (pre-sip, unrepaired), prestage II palliation (those destined for SIIP, but not requiring SIP, unrepaired), post-sip/pre-siip (interstage single ventricles), post-siip/pre-siiip (bidirectional Glenn patients), and post-siiip (post-fontan). Stage of palliation before cavopulmonary shunt therefore included pre-sip, pre-siip, and post-sip/pre-siip patients. s were defined as those that the patient was receiving at the time of the preoperative evaluation. Intraoperative hemodynamic instability was defined as a greater than 25% change in blood pressure from baseline for more than 10 minutes or significant rhythm change or major arrhythmia while the patient was in the operating room (OR). Surgical times were defined as follows: induction was the time from OR entry to skin incision, maintenance was the time from incision to skin closure, and emergence was the time from skin closure to OR exit. Cardiac arrest was defined as requirement for chest compressions, electrical defibrillation or cardioversion, or initiation of pediatric life-support pharmacotherapy. Case Classification We classified the complexity of each case by the invasiveness of the procedures performed. Noninvasive procedures consisted of imaging studies, auditory testing, or procedures where the most invasive component consisted only of peripheral intravenous placement (such as venography). Airway or gastrointestinal endoscopy, central catheter placement, dental procedures, tracheostomy, minor urologic procedures, abscess drainage, and other surgical procedures that did not involve opening of the abdominal, cranial, or thoracic spaces were classified as minimally invasive. Gastrostomy tube placement, fundoplication, enterostomy, craniofacial, and other procedures that did open the abdominal, cranial, or thoracic spaces were considered invasive. The 252 surgical procedures were performed during 173 general anesthesias. Because clinicians are concerned with risks surrounding sedation and general anesthesia, we considered each general anesthesia to be one case. The complexity of each anesthetic case was assigned a complexity score, defined as the complexity of the most invasive procedure performed: 0 for noninvasive, 1 for minimally invasive, and 2 for invasive procedures. For example, if a fundoplication and peripherally inserted central catheter placement were performed during the same general anesthesia, the case complexity was 2 (invasive). If an imaging study and a bronchoscopy were performed during the same general anesthesia, the case complexity was 1 (minimally invasive). Statistical Analysis Statistical analysis involved association of outcome measures (induction and maintenance instability, need for postoperative mechanical ventilation, and postoperative hospital LOS) with the perioperative variables of age at date of operation, preoperative intubation, palliation status before cavopulmonary shunt, case complexity, preoperative hospital LOS exceeding 14 days, greater than mild ventricular dysfunction, use of regional anesthesia, and preoperative use of angiotensin-converting enzyme (ACE) inhibitors, diuretics, digoxin, or inotropes. Univariate comparisons were performed using SPSS software (SPSS Inc, Chicago, IL). The Fisher exact test was used for discrete data, the Mann-Whitney U test or Kruskal-Wallis test for continuous data, and the Spearman correlation for association between two continuous variables. Because risk factors may be interdependent, we used multivariate regression to determine which factors were independent contributors to outcomes. Multivariate regression used the following stepwise process: All risk factors were entered into the model, and the nonsignificant variables were eliminated stepwise until the resulting model only contained significant variables. Results Study Population Our search yielded 113 patients who had undergone RACHS-1 category 6 procedures at our institution from

3 206 WATKINS ET AL Ann Thorac Surg SURGICAL RISK FOR COMPLEX CHD 2013;95: Table 1. Noncardiac Procedures Procedure No. (%) Gastrostomy 69 (27.4) Fundoplication/fundoplasty 59 (23.4) PICC procedure 32 (12.7) Airway endoscopy 19 (7.5) Central catheter placement/replacement 14 (5.6) MRI/CT 14 (5.6) Various ear, nose, and throat 7 (2.8) Exploratory laparotomy 6 (2.4) Enterostomy/enterostomy closure 5 (2.0) Minor urologic 4 (1.6) Upper or lower GI endoscopy 4 (1.6) Bowel resection 3 (1.2) Various other 16 (6.3) Total 252 (100) CT computed tomography; GI gastrointestinal; MRI magnetic resonance imaging; PICC peripherally inserted central catheter. July 2006 to January During the study interval, 71 of these patients underwent at least one noncardiac procedure; all of these had single-ventricle hearts or were destined for a single-ventricle palliative approach. The most common primary cardiac diagnoses for these 71 patients were hypoplastic left heart syndrome (70.4%), unbalanced atrioventricular septal defect (12.7%), and double-inlet left ventricle (5.6%). These patients underwent 252 noncardiac procedures, during 173 anesthesias, with some undergoing multiple procedures during the same general anesthesia. During the study interval, 61.1% received one general anesthesia, 15.9% received two, and 23% had three or more. The 252 noncardiac surgical procedures are summarized in Table 1. The most common were gastrostomy and fundoplication, which together accounted for 50.8% of all procedures. Vascular access and airway endoscopies together comprised another 25.8% of procedures. As reported in Table 2, most of the 173 anesthetic cases were performed on patients aged younger than 2 years (160 patients [92.5%]). Most of these cases were in patients younger than 6 months (123 cases). More than half the cases (61.5%) were between SIP and SIIP at the time of operation. The next most common group was those in the post-siip/pre-siiip (post-glenn/pre-fontan, 28.9%) group, followed by the post-siiip (6.9%) population. General anesthesia was supplemented with a regional technique in 31 cases (17.9%). Standard anesthesia monitoring (pulse oximetry, temperature, noninvasive blood pressure, electrocardiogram) was used in all cases, and invasive monitoring was used in 16 (9.2%). Preoperative hospital LOS exceeded 14 days for 125 patients (72.3%). The anesthetic case complexity was noninvasive (complexity 0) for 16 cases, with the remaining 157 almost evenly divided between minimally invasive (complexity 1) and invasive (complexity 2). Table 2. Data for 173 Anesthetic Cases Variable No. (%) Stage at operation Prestage I palliation 3 (1.7) Prestage II palliation 1 (0.6) Between stages I and II 107 (61.8) Between stages II and III 50 (28.9) Poststage III palliation 12 (6.9) Age at operation 31 days 24 (13.9) days 99 (57.2) 181 days 2 years 37 (21.4) 2 years 13 (7.5) Supplemental oxygen 74 (43.5) Regional anesthesia 31 (17.9) Preoperative intubation 28 (16.2) Ventricular dysfunction 63 (36.4) ACE inhibitor 124 (71.7) Diuretic 158 (91.3) Digoxin 67 (38.7) Inotrope 29 (16.8) Preoperative hospital LOS 14 days 125 (72.3) Use of invasive monitoring 16 (9.2) Case complexity 0 Noninvasive 16 (9.2) MRI/CT imaging 13 Venography 2 Auditory evaluation 1 1 Minimally invasive 78 (45.0) Central catheter placement 45 Airway endoscopy 16 Gastrointestinal endoscopy 3 Tracheostomy 2 Minor ear, nose, and throat 2 Abscess drainage 2 Various other 8 2 Invasive 79 (45.7) Esophagogastric fundoplasty 53 Gastrostomy tube 12 Exploratory laparotomy 6 Ostomy closure and takedown 4 Various other 4 Primary cardiac diagnosis Hypoplastic left heart 131 (73.6) Unbalanced AVSD 19 (11.0) Double-inlet left ventricle 8 (4.6) Tricuspid atresia 5 (2.9) Interrupted aortic arch 5 (2.9) Double-outlet right ventricle 4 (2.3) Shone s complex 1 (0.6) ACE angiotensin-converting enzyme; AVSD atrioventricular septal defect; CT computed tomography; LOS length of stay; MRI magnetic resonance imaging.

4 Ann Thorac Surg WATKINS ET AL 2013;95: SURGICAL RISK FOR COMPLEX CHD 207 Table 3. Univariate Analysis of Factors Associated With Induction Instability Perioperative Factor a Induction Instability With (n 73) Without (n 100) p Value Age, mean (SD) days 140 (198) 263 (298) Preoperative intubation 9 (12.3) 19 (19.0) Palliation stage before 58 (79.5) 53 (53.0) CP shunt Complexity 0 Noninvasive 1 (1.4) 15 (15.0) Minimally invasive 24 (32.9) 54 (54.0) 2 Invasive 48 (65.8) 31 (31.0) Preoperative LOS (82.2) 65 (65.0) days Moderate ventricular 27 (37.5) 36 (36.0) dysfunction present Use of regional 24 (32.9) 7 (7.0) anesthesia ACE inhibitor 60 (82.2) 64 (64.0) 0.01 Diuretic 67 (91.8) 91 (91.0) Digoxin 37 (50.7) 30 (30.0) Inotrope 8 (11.0) 21 (21.0) 0.1 a Data are shown as number (%) unless indicated otherwise. ACE angiotensin-converting enzyme; CP cavopulmonary; LOS length of stay; SD standard deviation. Instability on Induction of Anesthesia Table 3 lists univariate associations of instability during induction with the proposed perioperative variables. Induction instability was associated with younger age, palliation stage before cavopulmonary shunt, increased case complexity, longer preoperative hospital LOS, use of regional anesthesia, ACE inhibitor use, and digoxin use. Because many of these factors overlap and are not necessarily independent of each other, we performed a logistic regression, with the resulting logistic model summarized in Table 4. The model contained three independent variables: palliation stage before cavopulmonary shunt, case complexity score, and preoperative ACE inhibitor. These three variables explained 19.1% of the variability in the dependent variable in this model. Use of Table 4. Logistic Regression Model of Induction Instability Variable OR (95% CI) p Value Palliation stage before CP shunt 2.41 ( ) Complexity score 2.89 ( ) Preoperative ACE inhibitor 2.77 ( ) Constant a 0.04 ( ) a Constant refers to a parameter included in linear and logistic regression to represent the outcome when no risk factors are present. ACE angiotensin-converting enzyme; CI confidence interval; CP cavopulmonary; OR odds ratio. Table 5. Univariate Analysis of Factors Associated With Maintenance Instability Perioperative Factor a regional anesthesia was marginally significant before exclusion from the model (odds ratio, 2.7; p 0.06). Instability During Maintenance of Anesthesia Table 5 reports a univariate analysis of the perioperative factors associated with maintenance instability. Maintenance instability was associated with younger age at operation, palliation stage before cavopulmonary shunt, complexity score, use of regional anesthesia, and preoperative digoxin. Again, because of overlap and dependence of variables among each other, a multivariable logistic regression model for maintenance instability was constructed using the same stepwise approach as above for induction instability. The resulting model showed complexity score, preoperative use of digoxin, and preoperative use of inotropic agents were statistically signif- Table 6. Logistic Regression Model of Maintenance Instability Variable OR (95% CI) p Value Complexity score 3.06 ( ) Preoperative digoxin 2.05 ( ) Preoperative inotrope 3.31 ( ) Constant 0.13 ( ) CI confidence interval; With (n 88) OR odds ratio. Maintenance Instability Without (n 85) p Value Age, mean (SD) days 166 (218) 258 (304) Preoperative intubation 16 (18.2) 12 (14.1) Palliation stage before CP 66 (75.0) 45 (52.9) shunt Complexity 0 Noninvasive 4 (4.5) 12 (14.1) Minimally invasive 31 (35.2) 47 (55.3) 2 Invasive 53 (60.2) 26 (30.6) Preoperative LOS (77.3) 57 (67.1) days Moderate ventricular 37 (43.0) 26 (31.0) dysfunction present Use of regional 21 (23.9) 10 (11.8) anesthesia use ACE inhibitor 61 (69.3) 63 (74.1) Diuretic 81 (92.0) 77 (90.6) Digoxin 41 (46.6) 26 (30.6) Inotrope 18 (20.5) 11 (12.9) a Data are shown as number (%) unless indicated otherwise. ACE angiotensin converting enzyme; CP cavopulmonary; LOS length of stay; SD standard deviation.

5 208 WATKINS ET AL Ann Thorac Surg SURGICAL RISK FOR COMPLEX CHD 2013;95: Table 7. Univariate Analysis of Factors Associated With Postoperative Mechanical Ventilation Perioperative Factor a With (n 18) Postoperative Mechanical Ventilation Without (n 127) p Value Age, mean (SD) days 200 (252) 221 (283) Palliation stage before CP 14 (77.8) 77 (60.60) shunt Complexity 0 Noninvasive 1 (5.6) 13 (10.2) Minimally invasive 6 (33.3) 56 (44.1) 2 Invasive 11 (61.1) 58 (45.7) Preoperative LOS (94.4) 54 (68.5) days Moderate ventricular 7 (41.2) 45 (36) 0.79 dysfunction present Use of regional anesthesia 3 (16.7) 27 (21.3) ACE inhibitor 14 (77.8) 104 (81.9) Diuretic 17 (94.4) 116 (91.3) Digoxin 4 (22.2) 60 (47.2) Inotrope 4 (22.2) 4 (3.1) a Data are shown as number (%) unless indicated otherwise. ACE angiotensin converting enzyme; CP cavopulmonary; LOS length of stay; SD standard deviation. icant and positively associated with instability during the maintenance period (Table 6). These factors explained 18.9% of the variability in the dependent variable. Postoperative Mechanical Ventilation Of the 28 cases where the patient was intubated and mechanically ventilated at the time of the operation, 26 remained intubated and mechanically ventilated postoperatively. These patients were excluded from analysis of risk factors for postoperative mechanical ventilation because extubation and avoidance of postoperative mechanical ventilation is a more important goal for patients who are not intubated at the time of the operation, and efforts were not specifically made to attempt to extubate patients who were already intubated at the date of their operation. Table 7 reports a univariate analysis of the factors associated with postoperative mechanical ventilation for those 145 cases not intubated at the time of the operation. The only significant factor was preoperative hospital stay exceeding 14 days (odds ratio, 7.70; range, 1.01 to 58.8; p 0.05). No factors were significant in the multivariate logistic regression model. Factors Associated With Postoperative Hospital LOS In Table 8 we report our evaluation of the same preoperative risk factors for association with postoperative hospital LOS. Univariate analysis showed preoperative intubation and duration of preoperative hospital LOS exceeding 14 days were associated with postoperative hospital LOS. Again, because of overlap between variables, we used the stepwise method previously described to construct a linear regression model in an attempt to determine independent variables associated with postoperative hospital LOS. This resulting model (Table 9) shows preoperative hospital LOS exceeding 14 days, ventricular dysfunction, preoperative use of digoxin, and preoperative use of inotropic agent were independently associated with postoperative hospital LOS. These four variables accounted for 30.5% of the variability in the dependent variable. Death Perioperative cardiac arrest occurred in 3 patients (1.7%) within our cohort. One patient responded to standard Table 8. Univariate Analysis of Factors Associated With Postoperative Hospital Length of Stay Perioperative Factor Spearman Correlation Postoperative Hospital LOS, Mean (SD) In Presence of Factor In Absence of Factor p Value Age, days Preoperative intubation 65 (49) 33 (49) Palliation stage before CP shunt 14 (77.8) 77 (60.60) Complexity 0 Noninvasive 20.6 (36.6) Minimally invasive 47.0 (57.6) 2 Invasive 33.3 (43.3) Preoperative LOS 14 days 17 (94.4) 54 (68.5) Moderate ventricular dysfunction present 7 (41.2) 45 (36) 0.79 Use of regional anesthesia 3 (16.7) 27 (21.3) ACE inhibitor 14 (77.8) 104 (81.9) Diuretic 17 (94.4) 116 (91.3) Digoxin 4 (22.2) 60 (47.2) Inotrope 29 (44) 85 (55) ACE angiotensin-converting enzyme; CP cavopulmonary; LOS length of stay; SD standard deviation.

6 Ann Thorac Surg WATKINS ET AL 2013;95: SURGICAL RISK FOR COMPLEX CHD 209 Table 9. Linear Regression Model of Postoperative Hospital Length of Stay Variable Coefficient (SE) p Value Preoperative hospital 31.9 (7.3) LOS 14 days Moderate ventricular 21.6 (6.8) dysfunction present Preoperative digoxin 16.2 (6.9) Preoperative inotrope 43.1 (9.0) Constant 6.7 (7.1) LOS length of stay; therapy and resolved. Two patients progressed to extracorporeal membrane oxygenation cannulation and eventually recovered. No patients died within 24 hours of the procedure. One patient (0.4%) died within 7 days. The 30-day mortality rate for our cohort was 1.4% (n 3). The study was therefore underpowered to make statistical inferences regarding risk factors and death. Comment SE standard error. Children with CHD represent a high-risk population for anesthesia and surgical intervention. Awareness of this risk has prompted investigators to create large registries addressing the risk for perioperative cardiac arrest [2, 3]. In one of these reports, single ventricle was the most common congenital heart lesion associated with perioperative cardiac arrest [3]. Here, we further investigate this high-risk population by evaluating preoperative risk factors associated with less critical events: intraoperative instability (during induction and maintenance), need for postoperative mechanical ventilation, and postoperative hospital LOS. Questions regarding these outcomes are frequently the grounds for discussion when procedures on these patients are being considered [9, 10]. These data will assist clinicians in better defining the at-risk population. The need for noncardiac procedures involving general anesthesia in RACHS-1 category 6 patients is common at our institution: 63% required at least one procedure under general anesthesia, and almost 1 in 4 (23%) required three or more general anesthesias. Most patients in our cohort were aged younger than 2 years, most of them younger than 6 months, at the time of their noncardiac procedures. General operations, vascular access, airway endoscopies, and diagnostic imaging studies represented the bulk of the procedures; a similar mix of cases has been reported in other studies [3, 11]. The periods of anesthesia induction and maintenance are of concern to clinicians because the immediate presence of vasoactive drugs and positive-pressure ventilation represents a challenge to these patients physiology. We observed that induction instability was independently associated with complexity score, stage of palliation before cavopulmonary anastomosis, and preoperative ACE inhibitor use. Use of regional anesthesia, which consisted of epidural or single-shot caudal, was the last factor to be excluded from the model, probably because its use was entirely confined to patients undergoing invasive abdominal procedures. The additional univariate risks of younger age, preoperative hospital stay, and digoxin use were no longer significant after the multivariate risks were accounted for. These findings are consistent with our understanding of single-ventricle physiology [9, 12, 13]: General anesthetics and induction agents produce changes in systemic vascular resistance that can decrease pulmonary flow in patients with shunt physiology [10]. Reports in adults have shown that ACE inhibition is associated with perioperative hypotension [14 16], although other reports did not find this association [17, 18]. As a result, some authors [14, 16] advocate avoiding ACE inhibitor therapy before elective operations, although this recommendation is not universal [15]. Important for the practitioner is the observation that the three risk factors are independent, such that a patient with shunt physiology, receiving an ACE inhibitor, and undergoing an invasive procedure is at highest risk for induction instability. Instability during maintenance of anesthesia was independently associated with the case complexity score and use of digoxin and inotropes. The latter two of these risk factors may indicate a specific population with impaired ability to respond to the decreased systemic vascular resistance during general anesthesia. Also, more complex cases require deeper levels of anesthesia, which causes a further decrease in systemic vascular resistance, along with positive-pressure ventilation, which decreases venous return and preload. Of note, ACE inhibitor use was only a risk factor for instability during induction (which is chiefly hypotension), possibly because fluid administration is commonly used after anesthesia induction to compensate for the hypotension. Once this is done, hypotension during maintenance may be a marker for patients with impaired ventricular function; namely, those receiving digoxin and inotropic drugs. Need for mechanical ventilation, or failure to extubate, is an important outcome for patients who are not intubated at the time of their operation. Extubation failure portends a prolonged hospital LOS for pediatric intensive care unit patients with CHD [19], but specific data are lacking for extubation failure after elective noncardiac operations in this population. In our cohort, 18 patients required postoperative mechanical ventilation, representing 12.4% of the 145 patients who were not intubated when they presented for their operations. The only risk factor associated with need for postoperative mechanical ventilation was preoperative LOS greater than 14 days (104 patients). Children with more complex CHD, such as those in our study, are at higher risk for prolonged hospitalization [7, 20]. Chronically hospitalized pediatric patients require more invasive procedures and are at risk for respiratory complications [21]. What we demonstrate here is that within the population of patients who are not already intubated, preoperative hospital LOS appears to override other risk factors for postoperative mechanical ventilation. We observed that postoperative hospital LOS was

7 210 WATKINS ET AL Ann Thorac Surg SURGICAL RISK FOR COMPLEX CHD 2013;95: independently associated with preoperative hospital LOS, ventricular dysfunction, and preoperative inotrope use, and negatively associated with preoperative digoxin use. Longer preoperative LOS and preoperative inotrope use are indirect markers of more severe patient morbidity. As mentioned, our evidence exists for postoperative mechanical ventilation and postoperative hospital LOS. Digoxin use was negatively associated with postoperative hospital LOS for reasons that are unclear and have not been previously reported. It is possible that digoxin serves as a marker for a currently unidentified risk factor. Our population lacked statistical power to analyze cardiac arrest and mortality rates, which have been reported in larger multicenter studies. The Perioperative Cardiac Arrest registry reported that one of three cardiac arrests involved children with CHD [3]. In this report, patients with CHD accounted for 22% of the cardiac arrests that occurred during noncardiac procedures. Half of the cardiac arrests involving patients with CHD occurred outside the cardiac OR or catheterization laboratory (ie, during noncardiac procedures). Most patients were aged younger than 2 years, with half aged younger than 6 months. The most frequent procedures associated with cardiac arrests in patients with CHD were gastrointestinal, fundoplication, and otolaryngology procedures, and vascular access. The most common defects associated with cardiac arrest were single-ventricle variants (24%), followed by left-to-right shunts (18%) and outflow obstructive lesions (16%). In our population, although using the anesthesia cases to define the population overrepresents those patients who have extended hospital LOS because they undergo multiple procedures, it does provide an accurate representation of the cases observed in a surgical setting across time. Therefore, our conclusions apply to the population of cases (not necessarily patients) that appear in an OR setting over time. This study is subject to the limitations inherent in any retrospective analysis of risk factors in a mixed population. Associations of risk could be a result of other associated factors, and association does not necessarily imply causality. For these reasons, multivariate regression was used to determine independence of risk factors. However, our reported outcomes are also dependent on practice patterns and protocols that may be specific to our institution, which limits applicability of our findings to other practice populations. Reports of similar associations from other institutions will therefore strengthen the conclusions drawn here. In conclusion, the purpose of this study was to identify risk factors for adverse outcomes, to assist clinicians with clinical decision making when a high-risk cardiac patient presents for an elective procedure. The need for noncardiac operations or other procedures is common in patients with CHD and advanced RACHS-1 score. Authors who have recently reviewed the literature have expressed concern regarding the lack of data to assist in defining the at-risk population [9, 10, 12]. To that end, we have provided risk factors associated with instability during induction and maintenance, need for postoperative mechanical ventilation, and duration of postoperative hospital LOS for a population of children with RACHS-1 category 6 univentricular heart disease undergoing procedures involving general anesthesia. Use of this information should foster improved care for this high-risk population. References 1. Baum VC, Barton DM, Gutgesell HP. Influence of congenital heart disease on mortality after noncardiac surgery in hospitalized children. Pediatrics 2000;105: Flick RP, Sprung J, Harrison TE, et al. Perioperative cardiac arrests in children between 1988 and 2005 at a tertiary referral center: a study of 92,881 patients. Anesthesiology 2007;106:226 37; quiz Ramamoorthy C, Haberkern CM, Bhananker SM, et al. Anesthesia-related cardiac arrest in children with heart disease: data from the Pediatric Perioperative Cardiac Arrest (POCA) registry. Anesth Analg 2010;110: Watkins S, Morrow SE, McNew BS, Donahue BS. Perioperative management of infants undergoing fundoplication and gastrostomy after stage I palliation of hypoplastic left heart syndrome. Pediatr Cardiol 2012;33: Jenkins KJ. Risk adjustment for congenital heart surgery: the RACHS-1 method. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu 2004;7: Jenkins KJ, Gauvreau K, Newburger JW, Spray TL, Moller JH, Iezzoni LI. Consensus-based method for risk adjustment for surgery for congenital heart disease. J Thorac Cardiovasc Surg 2002;123: Larsen SH, Pedersen J, Jacobsen J, Johnsen SP, Hansen OK, Hjortdal V. The RACHS-1 risk categories reflect mortality and length of stay in a Danish population of children operated for congenital heart disease. Eur J Cardiothorac Surg 2005;28: Jenkins KJ, Gauvreau K. Center-specific differences in mortality: preliminary analyses using the Risk Adjustment in Congenital Heart Surgery (RACHS-1) method. J Thorac Cardiovasc Surg 2002;124: Sumpelmann R, Osthaus WA. The pediatric cardiac patient presenting for noncardiac surgery. Curr Opin Anaesthesiol 2007;20: White MC. Approach to managing children with heart disease for noncardiac surgery. Paediatr Anaesth 2010;21: Torres A, DiLiberti J, Pearl RH, et al. Noncardiac surgery in children with hypoplastic left heart syndrome. J Pediatr Surg 2002;37: Cannesson M, Collange V, Lehot JJ. Anesthesia in adult patients with congenital heart disease. Curr Opin Anaesthesiol 2009;22: Warner MA, Lunn RJ, O Leary PW, Schroeder DR. Outcomes of noncardiac surgical procedures in children and adults with congenital heart disease. Mayo Perioperative Outcomes Group. Mayo Clin Proc 1998;73: Comfere T, Sprung J, Kumar MM, et al. Angiotensin system inhibitors in a general surgical population. Anesth Analg 2005;100: Pigott DW, Nagle C, Allman K, Westaby S, Evans RD. Effect of omitting regular ACE inhibitor medication before cardiac surgery on haemodynamic variables and vasoactive drug requirements. Br J Anaesth 1999;83: Colson P, Ryckwaert F, Coriat P. Renin angiotensin system antagonists and anesthesia. Anesth Analg 1999;89: Reich DL, Hossain S, Krol M, et al. Predictors of hypotension after induction of general anesthesia. Anesth Analg 2005;101:

8 Ann Thorac Surg WATKINS ET AL 2013;95: SURGICAL RISK FOR COMPLEX CHD Ryckwaert F, Colson P. Hemodynamic effects of anesthesia in patients with ischemic heart failure chronically treated with angiotensin-converting enzyme inhibitors. Anesth Analg 1997;84: Baisch SD, Wheeler WB, Kurachek SC, Cornfield DN. Extubation failure in pediatric intensive care incidence and outcomes. Pediatr Crit Care Med 2005;6: INVITED COMMENTARY Complex congenital heart disease (CHD) incrementally increases the risk of noncardiac procedures to the greatest degree in neonates and infants. Although the mortality of palliative operations for single-ventricle physiology has steadily declined, the mortality of noncardiac operations in this population, surprisingly, has not [1]. One or more noncardiac procedures such as gastrostomy, fundoplication, airway endoscopy, vascular access, and imaging may occur within the period of achieving serial circulation in patients with single-ventricle physiology. Unfortunately, this persistent higher mortality in noncardiac procedures has not been explained because of the low incidence of complex CHD in the population. In this issue, Watkins and colleagues [2] have expanded the preoperative risk assessment of noncardiac cases in this patient population with the potential to improve intraoperative care. Perioperative data were retrospectively collected for 71 (63%) category 6 RACHS-1 cardiac surgical patients who had undergone at least one noncardiac surgical procedure over a 5-year period at one institution. Through the personal observation by clinicians involved with complex CHD and noncardiac procedures, the authors concluded that certain clinical events such as hemodynamic instability during induction and maintenance of anesthesia, postoperative mechanical ventilation, and hospital length of stay were the most commonly requested aspects of patient information for preoperative assessment. Multivariate analysis with each of the five clinical events produced seven perioperative factors, not including age, that significantly favored their occurrence with noncardiac procedures. The strength of this study is the selection of these five specific noncritical clinical events for analysis with perioperative data. The finding of a strong association between the complexity of the procedure and hemodynamic instability during both induction and maintenance of anesthesia improves the planning for intraoperative care. Invasive monitoring may be more 20. Pagowska-Klimek I, Pychynska-Pokorska M, Krajewski W, Moll JJ. Predictors of long intensive care unit stay following cardiac surgery in children. Eur J Cardiothorac Surg 2011;40: Peterson-Carmichael SL, Cheifetz IM. The chronically critically ill patient: pediatric considerations. Respir Care 2012;57: strongly considered for these noncardiac procedures in this population in the future, based on an incidence of only 9.2% in the current study. A weakness of the study is an uncertainty about whether the findings may be generalized to other neonates and infants with complex CHD having noncardiac procedures apart from the authors institution. Only four types of noncardiac procedures accounted for 75% of the total in the study; consequently, several pediatric procedures will not be applicable to the findings, limiting the benefit to other patients. Because neonates and infants undergo noncardiac procedures after cardiac procedures for complex CHD, there is a need to explain this greater mortality. Although prospective trials would serve this purpose more efficiently and reliably, they are unlikely to be performed because of inadequate power and the heterogeneous nature of complex CHD. However, additional retrospective studies that develop risk stratification of perioperative factors may allow for better preparation for noncardiac surgical procedures and improve outcome. William C. Oliver, Jr, MD Department of Anesthesiology College of Medicine, Mayo Clinic 200 First St SW Rochester, MN oliver.william@mayo.edu References 1. Torres A, DiLiberti JD, Pearl RH, et al. Noncardiac surgery in children with hypoplastic left heart syndrome. J Pediatr Surg 2002;37: Watkins SC, McNew BS, Donahue BS. Risks of noncardiac operations and other procedures in children with complex congenital heart disease. Ann Thorac Surg 2013;95: by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc