Risk Factors and Early Outcomes of Multiple Reoperations in Adults With Congenital Heart Disease

Transcription

1 Risk Factors and Early Outcomes of Multiple Reoperations in Adults With Congenital Heart Disease Kimberly A. Holst, BS, Joseph A. Dearani, MD, Harold M. Burkhart, MD, Heidi M. Connolly, MD, Carole A. Warnes, MD, Zhuo Li, MS, and Hartzell V. Schaff, MD Division of Cardiovascular Surgery, Mayo Clinic and Foundation, Rochester, Minnesota Background. Advances in treatment of congenital heart disease (CHD) have resulted in most patients surviving to adulthood. Despite surgical correction, the need for reoperation(s) persists, and there are few outcome data. This study examined early postoperative results to determine risk factors for cardiac injury and early death in adults with CHD undergoing repeat median sternotomy. Methods. Data from the most recent median sternotomy of 984 adults (49% male) with CHD were analyzed. Mean age at operation was 36.4 years. Diagnoses were conotruncal anomaly, 361 (37%); Ebstein/Tricuspid valve, 174 (18%); pulmonary stenosis/right ventricular outflow tract obstruction, 92 (9%); single ventricle, 71 (7%); atrioventricular septal defect, 64 (7%); subaortic stenosis, 62 (6%); aortic arch abnormalities, 23 (2%); anomalous pulmonary vein, 21 (2%); Marfan syndrome, 14 (1%); and other, 102 (10%). Results. Overall early mortality was 3.6%: including 2%, 6%, 7%, and 0% at sternotomy 2 (n 597), 3 (n 284), 4(n 72), and 5 (n 31), respectively. Cardiac injury occurred in 6%. Independent predictors of cardiac injury were single-ventricle diagnosis and increased number of prior sternotomies. Increased time from previous sternotomy decreased the incidence of cardiac injury. Independent risk factors for early death were urgent operation, single-ventricle diagnosis, and longer bypass time. Increased preoperative ejection fraction decreased early mortality. Conclusions. Subsequent sternotomy showed increased early mortality, yet neither sternotomy number nor cardiac injury was an independent predictor of early death. Two variables were protective: early mortality was reduced with increased ejection fraction and cardiac injury was less likely with increased interval from the previous sternotomy. (Ann Thorac Surg 2011;92:122 30) 2011 by The Society of Thoracic Surgeons Advances in diagnosis and treatment of children with congenital heart disease (CHD) have greatly improved survival and quality of life. Currently, more than 95% of individuals with CHD survive to adulthood [1]. An estimate of adults with congenital heart disease (ACHD) in the United States was 800,000 in 2000 [2], growing by approximately 8960 adult cases annually [3]. Surgical intervention for most anomalies is corrective in establishing normal or nearly normal physiology. Intracardiac shunts are closed, abnormal valves are repaired whenever possible, and a biventricular circulation is the general goal. These corrective operations are considered palliative because many patients require multiple operations or interventions, or both, throughout their lifetime [1, 4]. Most reoperations are related to valve deterioration, residual or recurrent intracardiac lesions, or arrhythmias. Accepted for publication March 22, Presented at the Fifty-seventh Annual Meeting of the Southern Thoracic Surgical Association, Orlando, FL, Nov Address correspondence to Dr Dearani, Division of Cardiovascular Surgery, Mayo Clinic, 200 First St SW, Rochester, MN 55905; dearani. joseph@mayo.edu. The number of reoperations in ACHD will continue to rise as the population increases. Evidence-based data to guide management of this population are limited [5]. Our objective was to determine risk factors for early death and cardiac injury in a large, singlecenter experience of ACHD undergoing repeat median sternotomy. Surgical strategies to approach the ACHD patients undergoing repeat operation will be addressed. Patients and Methods After approval by the Mayo Clinic Institutional Review Board (July 9, 2009), we searched our database for ACHD who were aged older than 18 years at the repeat median sternotomy at Mayo Clinic Rochester from January 1, 1993, to December 31, A retrospective record investigation examined demographic characteristics, surgical procedures, comorbidities, and early results. The medical and surgical teams involved were trained in CHD; consequently, the algorithms and protocols used were uniform, with minor variation in practice styles by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg HOLST ET AL 2011;92: OUTCOMES AFTER MULTIPLE REOPERATIONS IN CHD Abbreviations and Acronyms ACHD adults with congenital heart disease AHD acquired heart disease ASD atrial septal defect CHD congenital heart disease CI confidence interval CPB cardiopulmonary bypass DORV double-outlet right ventricle EC Extracardiac HCM hypertrophic cardiomyopathy ICD implantable cardioverter-defibrillator PA pulmonary artery PFO patent foramen ovale RV right ventricle RVOTO right ventricular outflow tract obstruction TOF Tetralogy of Fallot VSD ventricular septal defect Table 2. Diagnoses Diagnosis No. (%) 123 Conotruncal anomalies a 361 (36.7) Ebstein/tricuspid valve disease 174 (17.7) Pulmonary stenosis, RVOTO 92 (9.4) Single ventricle 71 (7.2) Atrioventricular septal defect 64 (6.5) Subaortic stenosis, HCM 62 (6.3) Coarctation, interrupted arch 23 (2.3) Anomalous pulmonary vein 21 (2.1) Marfan syndrome 14 (1.4) Other 102 (10.4) Total 984 (100) a Includes double-outlet right ventricle, pulmonary atresia, transposition of the great arteries, tetralogy of Fallot, truncus arteriosus. HCM hypertrophic cardiomyopathy; RVOTO right ventricular outflow tract obstruction. ADULT CARDIAC Patient Population During this 15-year study period, 7700 operations were performed in 6831 patients of all ages with CHD at Mayo Clinic Rochester. The patient cohort for this review included 984 consecutive ACHD (484 male, 500 female) undergoing repeat median sternotomy. The study excluded patients with isolated bicuspid aortic valve, isolated patent foramen ovale, and those undergoing, or with history of, cardiac transplantation. The most recent repeat sternotomy was analyzed for patients with more than one repeat median sternotomy during the study period. Repeat sternotomy was defined as occurring beyond 14 days from the previous sternotomy. We documented 597 second sternotomies, 284 third, 72 fourth, and 31 fifth or more. The patients were a mean age of years (range, 18 to 83 years). Demographic and preoperative clinical data are reported in Table 1. An interesting finding was that age and time from previous sternotomy decreased as the sternotomy number increased. Average time since the previous sternotomy was years (range, 18 days to 52 years). Diagnoses are outlined in Table 2. Operative Technique All patients underwent repeat median sternotomy. Cannulation techniques included: Peripheral cannulation was through the right (left, less common) groin as determined by preopera- Table 1. Demographic and Preoperative Data Sternotomy Number Variable a All (n 597) (n 284) (n 72) (n 31) (n 984) Male sex 294 (49) 137 (48) 36 (50) 17 (55) 484 (49) Age, years (18 83) Body mass index, kg/m ( ) Diabetes 35 (6) 9 (3) 7 (10) 1 (3) 52 (5) Hypertension 106 (18) 29 (10) 6 (8) 1 (3) 142 (14) Dyslipidemia 90 (15) 30 (11) 1 (1) 2 (6) 123 (13) Coronary artery disease 43 (7) 13 (5) 2 (3) 1 (3) 59 (6) Smoking history 135 (23) 58 (20) 10 (14) 11 (35) 214 (22) History of stroke 47 (8) 24 (8) 7 (10) 3 (10) 81 (8) Hematocrit ( ) Creatinine 2 mg/dl 9 (2) 5 (2) 4 (6) 1 (3) 19 (2) Ejection fraction ( ) Arrhythmia 66 (11) 31 (11) 7 (10) 2 (6) 106 (11) Endocarditis 32 (5) 27 (10) 15 (21) 7 (23) 81 (8) Years from last sternotomy (18 d 51.7 yrs) a Data are presented as number (%) or mean standard deviation (range).

3 124 HOLST ET AL Ann Thorac Surg OUTCOMES AFTER MULTIPLE REOPERATIONS IN CHD 2011;92: tive ultrasound evaluation. Vessels were exposed by small cutdown incisions. Percutaneous groin cannulation was not used because direct repair of the femoral arteriotomy and venotomy at the end of the procedure was preferred. Femoral arterial cannulation was usually with an 8-mm chimney graft; direct cannulation of the femoral artery (Medtronic Biomedicus, 17F, 19F, 21F) through a cutdown with the Seldinger technique was used if the artery was large enough to allow distal extremity perfusion around the arterial cannula. The femoral vein was cannulated with a 22- or 25-mm thin-walled cannula (Edwards Lifesciences Inc, Irvine, CA) advanced into the right atrium, also using the open Seldinger technique and transesophageal echocardiogram guidance. The right internal jugular vein was cannulated percutaneously with a 16- or 18-mm cannula (Edwards Lifesciences Inc) advanced into the superior vena cava. Vacuum assist was used cautiously to augment venous return and was avoided when intracardiac shunts were present to avoid potential air embolism from unexpected cardiotomy during dissection. When aortic arch operations were being performed, right axillary artery cannulation with a chimney graft was preferred to allow selective cerebral perfusion during arch reconstruction. The decision to convert to central cannulation after mediastinal dissection was individualized. An intraoperative cardiac injury was defined as entry into a cardiac chamber or vascular structure during resternotomy or mediastinal dissection that required initiation of cardiopulmonary bypass (CPB) for control or repair. A cardiac injury may be anticipated (ie, expected) or unexpected. An injury was anticipated when there was evidence of aortic or conduit adherence or erosion into the sternum as noted on preoperative computed tomography or magnetic resonance imaging. Other anticipated injuries included a hypertensive, enlarged right ventricle (RV), or a dilated or aneurysmal coronary artery or button close to the sternum. When a cardiac injury was anticipated, CPB was intentionally initiated before the sternotomy to decompress the anatomic structure of concern. When an unexpected injury occurred, CPB was initiated. If a right-sided structure was injured, mild systemic hypothermia (34 C) or normothermia was used; more profound hypothermia was used for injury to left-sided structures. Deep hypothermia, with or without intermittent periods of circulatory arrest, was used for expected injury to the dilated aorta or pulmonary venous (right) atrium. Statistical Analysis Descriptive statistics for the 984 patients are reported. Categoric variables are presented as frequency and Table 3. Operative Procedures Procedure No. a Pulmonary valve 317 Repair 5 Replacement 312 Aortic valve 238 Repair 19 Replacement 219 Tricuspid valve 295 Repair 160 Replacement 135 Mitral valve 176 Repair 67 Replacement 135 Ventricular septal defect 30 Fontan 49 Aortic aneurysm 76 Ascending 74 Arch 15 Descending 3 Coronary artery bypass graft 22 Arrhythmia 108 ICD, permanent pacemaker 22 Myectomy 26 Cavopulmonary anastomosis 19 Pericardiectomy 15 Systemic-PA shunt 32 Other 58 a Multiple procedures may be performed in one patient; multiple-valve procedures: 251 (26%). ICD implantable cardioverter-defibrillator; PA pulmonary artery. percentage, and continuous variables are presented as mean standard deviation or median (range), as appropriate. To test the association between the sternotomy number and outcome, including cardiac injury and early mortality, the last sternotomy data of each of the 984 patients was used for statistical modeling. Early mortality was defined as death within 30 days or before hospital discharge. Logistic regression models were used to identify predictors for cardiac injury and early death. Multivariable models considered significant univariate variables (p 0.05) with model selection using the stepwise method (backward and forward methods resulted in the same model). All statistical tests were two-sided with the level set at 0.05 for statistical significance. Results Early Outcome Operative procedures are summarized in Table 3. CPB was initiated before completion of the median sternotomy in 9% of patients, most commonly in patients with aortic dilatation (conotruncal anomalies and Marfan syndrome), and single-ventricle diagnoses (13% and 18% of

4 Ann Thorac Surg HOLST ET AL 2011;92: OUTCOMES AFTER MULTIPLE REOPERATIONS IN CHD Table 4. Perioperative Data Sternotomy Number All Variable a (n 597) (n 284) (n 72) (n 31) (n 984) 125 ADULT CARDIAC Urgent operation 22 (4) 10 (4) 6 (8) 3 (10) 41 (4) Bypass time, min (0 620) Circulatory arrest 54 (9) 28 (10) 11 (15) 6 (19) 99 (10) Circulatory arrest time, min (0 142) Cross-clamp 534 (89) 239 (84) 56 (78) 22 (71) 851 (86) Cross-clamp time, min (0 285) Cardiac injury 24 (4) 20 (7) 7 (10) 4 (13) 55 (6) Mechanical support 21 (4) 13 (5) 5 (7) 4 (13) 43 (4) Postoperative transfusion 310 (52) 160 (56) 54 (75) 27 (87) 551 (56) a Categoric data are expressed as number (%) and continuous data as mean standard deviation (range). these patients, respectively). Femoral artery cannulation occurred in 15% of patients and femoral vein cannulation in 9%. Patients with atrioventricular septal defect and Ebstein/tricuspid valve anomaly had relatively low incidences of groin cannulation (arterial, 3% and 7%; venous, 2% and 3%, respectively). When peripheral cannulation was initiated, conversion to central aortic cannulation occurred in 8% and to central venous (bicaval or right atrial) cannulation in 55%. Perioperative data are summarized in Table 4. Blood transfusion was defined as transfusion of any blood product component that occurred after the operation and during the hospital stay. For the purpose of data analysis in this review, a patient was classified by whether a transfusion occurred (yes or no). The frequency of postoperative transfusion increased as sternotomy number increased. Location and incidence of cardiac injury by diagnostic category is reported in Table 5. Significant multivariate predictors of cardiac injury were single-ventricle diagnosis (odds ratio, 2.86; 95% confidence interval, 1.22 to 6.75; p 0.016) and increased number of previous sternotomies (odds ratio, 1.40; 95% confidence interval, 1.02 to 1.92; p 0.037). Increased time from previous sternotomy independently decreased the risk of cardiac injury (odds ratio, 0.99 for every 180 days; 95% confidence interval, 0.97 to 1; p 0.043). A summary of early results is reported in Table 6. No early deaths occurred in the 5 sternotomy group, which may reflect the low number of 31 patients. Significant risk factors for early death are reported in Table 7. Independent predictors were urgent operation, single ventricle, and increased CPB time. An interesting finding was that sternotomy number was not an independent predictor of early death. Although each subsequent sternotomy resulted in an increase in the early mortality rate, neither sternotomy number nor cardiac injury was an independent predictor of early death. Two variables were protective: univariate analysis showed increased ejection fraction decreased early death and multivariate analysis showed increased time from previous sternotomy decreased the incidence of cardiac injury. Table 5. Location of Cardiac Injury by Diagnosis Diagnosis a Pts Cardiac Right Right EC Innominate Injuries Ventricle Atrium Aorta Conduit Vein Other Conotruncal anomalies (6) 1 (5) 2 (9) 4 (18) 12 (55) 3 (14) Ebstein/tricuspid valve (2) 1 (33) 1 (33) 1 (33) Pulmonary stenosis, RVOTO 92 6 (7) 1 (17) 2 (33) 2 (33) 1 (17) Single ventricle 71 8 (11) 3 (42) 1 (14) 1 (14) 1 (14) 2 (28) Atrioventricular septal defect 64 1 (2) 1 (100) Subaortic stenosis, HCM 62 3 (5) 1 (33) 1 (33) 1 (33) Coarctation, interrupted arch 23 2 (9) 1 (50) 1 (50) Anomalous pulmonary vein 21 2 (10) 2 (100) Marfan syndrome 14 1 (7) 1 (100) Other (7) 1 (14) 1 (14) 5 (71) Total population (6) 5 (9) 8 (15) 15 (27) 15 (27) 9 (16) 3 (5) a Data are presented as number (%). EC extracardiac; HCM hypertrophic cardiomyopathy; RVOTO right ventricular outflow tract obstruction.

5 126 HOLST ET AL Ann Thorac Surg OUTCOMES AFTER MULTIPLE REOPERATIONS IN CHD 2011;92: Table 6. Early Results Variable a Sternotomy Number All (n 597) (n 284) (n 72) (n 31) (n 984) Early death 14 (2.3) 16 (5.6) 5 (6.9) 0 35 (3.6) Ventilation 5 days 29 (5) 18 (6) 5 (7) 5 (16) 57 (6) Arrhythmia 46 (8) 20 (7) 4 (6) 3 (10) 73 (7) Permanent pacemaker 22 (4) 13 (5) 3 (4) 0 38 (4) Reoperation (bleeding) 23 (4) 6 (2) 5 (7) 0 34 (3) Stroke 5 (1) 6 (2) 1 (1) 0 12 (1) Postoperative dialysis 16 (3) 10 (4) 4 (6) 1 (3) 31 (3) Sternal wound infection 11 (2) 2 (1) 5 (7) 1 (3) 19 (2) Hospital stay, days a Categoric data are expressed as number (%) and continuous data as mean standard deviation. Comment Extended survival of patients with CHD and increased frequency of procedures leads to complex anatomy, diverse clinical conditions, and complicated surgical intervention(s). Anatomic landmarks, particularly coronary arteries, may be obscured, extracardiac conduits or dilated ascending aortas may invade the chest wall, and increased adhesions make reoperation difficult. Reduced myocardial function, increased arrhythmias, decreased vascular compliance, increased rheology, other organ system dysfunction, and age-associated comorbidities contribute to outcome [6, 7]. Although cardiac reoperation has been less problematic in children [8, 9], reports of reoperation in adults are overwhelmingly limited to acquired heart disease (AHD) and show varied results with a wide range of conclusions [5, 10 15]. There are currently no data addressing risk and outcome of reoperation in large numbers of ACHD. Many significant univariate predictors of early death were identified in our series of reoperation in ACHD. Multivariate analysis, however, was limited to three independent predictors: increased CPB time, urgent operation, and single ventricle. Our study identified increased CPB time as an independent predictor of early death in reoperation for ACHD. Prolonged circulatory arrest and cross-clamp times were also related to worse outcome. This is in line with the review by Kirshbom and colleagues [16] of reoperation in CHD, which demonstrated increased early mortality with increased procedure time. These findings emphasize the importance of operative efficiency to optimize early outcome. Urgent operation is a logical risk factor for poor outcome and has been found to be a predictor of early death in AHD [10, 11, 15]. Patients with single-ventricle anomalies are known to have worse outcomes in adulthood [17] and were at greater risk for early death compared with other diagnostic groups in our series. Although common acquired comorbidities were present in our population (Table 1), previous stroke and renal insufficiency (creatinine 2 mg/dl) were the only risk factors associated with early death. Although this study correlated death with blood transfusion in ACHD, it is not a new phenomenon in cardiac surgery [18] and is an important modifiable risk factor. This finding supports Table 7. Risk Factors for Early Death Variable Univariate Multivariate OR (95% CI) p Value OR (95% CI) p Value History of stroke 3.56 ( ) Creatinine 2 mg/dl ( ) Ejection fraction (per 10) 0.54 ( ) Single ventricle 4.95 ( ) ( ) Urgent operation 9.92 ( ) ( ) Bypass time (per 10 min) 1.14 ( ) ( ) Circulatory arrest time (per 10 min) 1.54 ( ) Cross-clamp time (per 10 min) 1.13 ( ) Cardiac injury 2.75 ( ) Postoperative transfusion 3.95 ( ) Sternotomy # ( ) Sternotomy # ( ) CI confidence interval; OR odds ratio.

6 Ann Thorac Surg HOLST ET AL 2011;92: OUTCOMES AFTER MULTIPLE REOPERATIONS IN CHD 127 regular use of transfusion algorithms, which have been shown to reduce blood product usage in cardiac operations [19, 20]. Our analysis found increased early mortality with increasing sternotomy number, but this was not a risk factor by multivariable analysis. Other recent publications of reoperation in CHD have found varied associations between repeat sternotomy and early death [8, 16]; these discrepancies could reflect a wide range of possibilities and require further investigation. Cardiac injury did not have a significant effect on early death in our study, which is congruent with previous findings in CHD [8, 16]. Increased ejection fraction was the only risk factor protective for early death in our series. Early postoperative results and morbidity (Table 6) were not included in the analysis to predict early death due to low incidence. Stroke did occur in 12 patients, consisting of embolic or ischemic in 8, hemorrhagic in 3, and unknown in 1, and deserves comment. In the 8 patients with embolic or ischemic strokes, 2 were receiving mechanical support and 2 had left-sided prosthetic valves. Two patients in this group had cardiac injuries, one to the extracardiac conduit and the other to the right atrium-rv connection; importantly, neither stroke was due to air embolism. The three hemorrhagic strokes all occurred in patients who were receiving anticoagulation. Of the 12 patients who had stroke, 6 survived and 6 died. Many important technical issues must be considered at resternotomy in ACHD. Preoperative imaging with computed tomography or magnetic resonance imaging is helpful and is performed liberally in our practice for almost all diagnoses. Ultrasound evaluation of peripheral vasculature groin vessels, internal jugular vein, and axillary vessels is performed preoperatively. The decision to expose or cannulate the groin vessels and the timing of initiation of CPB before sternotomy is individualized; preoperative imaging aids in decision making. Despite careful surgical planning, cardiac injury during resternotomy may be inevitable or expected in some situations. The concerns are different with regard to a hazardous reentry in adults with AHD vs CHD. Difficulties with resternotomy in AHD are often related to patent coronary bypass graft injury, particularly in a graftdependent circulation [10, 15], whereas cardiac injury in CHD is more likely related to a rent in an extracardiac conduit, dilated aorta, or dilated RV or right atrium. Prior coronary bypass grafting is rare in ACHD; and in our series, only 4.5% had prior bypass grafting. A recent study of AHD reported 60% of adults without congenital lesions had prior bypass grafting at resternotomy [10]. Other differences include the presence of left ventricular systolic dysfunction, and aortic/peripheral atherosclerotic disease being common in acquired heart disease. In ACHD, systolic systemic ventricular function (morphologic left ventricle is most common) is usually normal and the pulmonary ventricle (usually a morphologic RV) is often dilated and dysfunctional. Our analysis could not predict the location of cardiac injury for a given diagnosis (Table 5); however, in our experience, structures frequently at risk include: (1) RVto-pulmonary artery conduit with truncus and transposition; (2) Aorta with conotruncal anomalies or Marfan syndrome; (3) RV with Ebstein or any CHD diagnosis with RV dilatation (eg, tetralogy of Fallot); (4) Right atrium with Ebstein or after an atrial switch procedure (i.e., pulmonary venous atrium), any diagnosis with tricuspid regurgitation, or after atriopulmonary Fontan connection; (5) Native coronary arteries; for example, anomalous left anterior descending across a dilated RV outflow tract (eg, tetralogy of Fallot), or a dilated/ aneurysmal right coronary button in a prior aortic root replacement (eg, Ross procedure).; and (6) Injury to the innominate vein is not uncommon regardless of diagnosis, and when it occurs it can result in catastrophic bleeding. Injury may be related to high venous pressure, decreased vein pliability from a long-term indwelling pacemaker, defibrillator leads that tear with sternal separation, or vein adherence to the retrosternum due to lack of tissue (adipose, thymic, etc.) from removal at a prior operation. To minimize tearing of atrial, ventricular, or innominate vein tissues with mediastinal dissection, we routinely enter both pleural spaces to allow the mediastinum to fall away from the chest wall and reduce tension and tearing of these structures at sternal separation. Important technical issues and pearls for the surgeon include the following: First, it is essential to know that when a cardiac rent occurs (left-sided or right-sided) on CPB, there is a risk of air embolism and appropriate precautions should be taken. If an intracardiac shunt is present (atrial or ventricular septal defect) and an injury occurs to a right-sided structure, dissection can proceed with CPB and the laceration can be repaired after tension in the injured area has been relieved, but central venous pressure must be positive ( 5 mm Hg) to avoid air entry into the right heart and subsequent systemic air embolism across the shunt. If a left-sided rent occurs (aorta or pulmonary venous atrium in prior atrial switch procedure), CPB with deep hypothermia is often necessary before completion of the sternotomy to avoid air embolism. The second point relates to injury of a dilated ascending aorta in the setting of aortic regurgitation. Significant aortic injury may require deep hypothermia before resternotomy for proper control and repair. When ventricular fibrillation occurs during the cooling process, left ventricular distention and myocardial injury will occur in the absence of left heart decompression. When significant aortic regurgitation is present, left heart venting can be accomplished through a small concomitant left anterior thoracotomy before sternal division for apical vent placement. Third, although CPB can facilitate a safe resternotomy, longer CPB times also result in increased morbidity and death, as seen in this review. As a consequence, the operation should be orchestrated so that the CPB time is as short, but safe, as possible. This can be accomplished by using peripheral CPB to facilitate resternotomy, then separating from CPB after sternal division to complete mediastinal dissection whenever possible. ADULT CARDIAC

7 128 HOLST ET AL Ann Thorac Surg OUTCOMES AFTER MULTIPLE REOPERATIONS IN CHD 2011;92: Fourth, important factors to avoid lower extremity compartment syndrome when peripheral cannulation is used relate to (1) inadequate arterial perfusion to the leg (ischemia) and (2) inadequate venous drainage of the leg because a large femoral venous cannula causes obstruction in the presence of ipsilateral femoral artery antegrade perfusion. When either of these two issues is present, we convert to central cannulation. Fifth, intraoperative transesophageal echocardiography is excellent at identifying an intracardiac shunt, but it is not perfect. We routinely perform an intraoperative bubble study with transesophageal echocardiogram evaluation. We also maintain mild Trendelenburg position during the operation, use CO 2 in the operative field, and place an aortic tack vent on aspiration before right-sided cardiotomy to minimize potential air embolism from an unrecognized intracardiac shunt. Finally, although some have advocated the use of a synthetic mediastinal barrier at sternal closure to facilitate subsequent sternotomy [21, 22], others have identified increased adhesions and cardiac injury during resternotomy [16]. We use a barrier selectively, and in our experience, polytetrafluoroethylene pericardial membrane seems to work well as an effective barrier without causing problematic adhesions. Study limitations include that the current series is a retrospective review and lacks a primary sternotomy comparison group. Many of the univariate risk factors identified may have a greater effect on late survival; however, multivariate analysis was limited by the low incidence of early death and could only assess three variables. There is great diversity in the ACHD population, and to make statistical analysis feasible, it was necessary to group patients with similar characteristics, thus limiting the ability to analyze the data for a specific diagnosis or individual. Despite these limitations, analysis of this large number of patients who had a focused and consistent management strategy contributes greatly to the paucity of literature available about reoperation in ACHD. In conclusion, reoperations in ACHD are complex procedures and require thoughtful surgical planning. Independent risk factors for early death were urgent operation, increased CPB time, and single-ventricle diagnosis. Risk of cardiac injury was increased with a singleventricle diagnosis and increased number of prior sternotomies. Subsequent sternotomy showed increased early death, yet neither sternotomy number nor cardiac injury was an independent predictor of early death. Two variables were protective: early death was reduced with increased ejection fraction and cardiac injury was less likely with an increased time interval from the previous sternotomy. We acknowledge the contributions and mentorship of Drs Gordon Danielson and Francisco Puga, and Judy Lenoch for data acquisition and analysis. References 1. Warnes CA. The adult with congenital heart disease: born to be bad? J Am Coll Cardiol 2005;46: American College of Cardiology. 32nd Bethesda Conference. Care of the adult with congenital heart disease. Bethesda, MD, Oct 2 3, Warnes CA, Liberthson R, Danielson GK, et al. Task force 1: the changing profile of congenital heart disease in adult life. J Am Coll Cardiol 2001;37: Berdat PA, Immer F, Pfammatter JP, Carrel T. Reoperations in adults with congenital heart disease: analysis of early outcome. Int J Cardiol 2004;93: Giamberti A, Chessa M, Abella R, et al. Morbidity and mortality risk factors in adults with congenital heart disease undergoing cardiac reoperations. Ann Thorac Surg 2009;88: Bhat AH, Sahn DJ. Congenital heart disease never goes away, even when it has been treated : the adult with congenital heart disease. Curr Opin Pediatr 2004;16: Perloff JK, Warnes CA. Challenges posed by adults with repaired congenital heart disease. Circulation 2001;103: Morales DL, Zafar F, Arrinton K, et al. Repeat sternotomy in congenital heart surgery: no longer a risk factor. Ann Thorac Surg 2008;86: Russell JL, LeBlanc JG, Sett SS, Potts JE. Risks of repeat sternotomy in pediatric cardiac operations. Ann Thorac Surg 1998;66: Park CB, Suri RM, Burkhart HM, et al. Identifying patients at particular risk of injury during repeat sternotomy: analysis of 2555 cardiac reoperations. J Thorac Cardiovasc Surg 2010;140: Sabik JF 3rd, Blackstone EH, Houghtaling PL, Walts PA, Lytle BW. Is reoperation still a risk factor in coronary artery bypass surgery? Ann Thorac Surg 2005;80: Elahi M, Dhannapuneni R, Firmin R, Hickey M. Direct complications of repeat median sternotomy in adults. Asian Cardiovasc Thorac Ann 2005;13: Potter DD, Sundt TM, Zehr KJ, et al. Operative risk of reoperative aortic valve replacement. J Thorac Cardiovasc Surg 2005;129: Potter DD, Sundt TM, Zehr KJ, et al. Risk of repeat mitral valve replacement for failed mitral valve prostheses. Ann Thorac Surg 2004:78: Ellman PI, Smith RL, Girotti ME, et al. Cardiac injury during resternotomy does not affect perioperative mortality. J Am Coll Surg 2008;206: Kirshbom PM, Myung RJ, Simsic JM, et al. One thousand repeat sternotomies for congenital cardiac surgery: risk factors for reentry injury. Ann Thorac Surg 2009;88: Wauthy P, Massaut J, Sanoussi A, et al. Ten-year experience with surgical treatment of adults with congenital cardiac disease. Cardiol Young 2011;21: Murphy GJ, Barnaby CR, Rogers CA, et al. Increased mortality, postoperative morbidity, and cost after red blood cell transfusion in patients having cardiac surgery. Circulation 2007;116: Nuttall GA, Oliver WC, Santrach PJ, et al. Efficacy of a simple intraoperative transfusion algorithm for nonerythrocyte component utilization after cardiopulmonary bypass. Anesthesiology 2001;94: Avidan MS, Alcock EL, Da Fonseca J, et al. Comparison of structured use of routine laboratory tests or near-patient assessment with clinical judgment in the management of bleeding after cardiac surgery. Br J Anaesth 2004;92: Walther T, Rastan A, Dähnert I, et al. A novel adhesion barrier facilitates reoperations in complex congenital cardiac surgery. J Thorac Cardiovasc Surg 2005;129: Lahtinen J, Satta J, Lähde S, et al. Computed tomographic evaluation of retrosternal adhesions after pericardial substitution. Ann Thorac Surg 1998;66:

8 Ann Thorac Surg HOLST ET AL 2011;92: OUTCOMES AFTER MULTIPLE REOPERATIONS IN CHD DISCUSSION DR KIRK R. KANTER (Atlanta, GA): Joe, first of all, I want to congratulate you and your coauthors for getting a medical student interested enough to do all of the hard work that was necessary for this huge study. This is the future of our profession, to have bright, young medical students presenting at our meetings and developing interest in our profession. This is a remarkable paper and a huge series. I have just a few questions. The incidence of unplanned institution of bypass was about 6%, but this was not a risk factor for mortality. This suggests that you are doing something so that you are prepared for these catastrophes. How often do you routinely cannulate the groin or put a wire in the groin or do you use an alternative incision other than a redo sternotomy for these patients? Are there certain subsets of patients such as those with an enlarged right heart, Ebstein s anomaly with a huge right atrium, or old Fontans, for whom you would be more likely to prophylactically go on bypass rather than waiting for an injury? Even though unplanned initiation of bypass was not a risk factor for mortality, was it a risk factor for stroke since you had a certain incidence of stroke? In anticipation of an eventual redo sternotomy, do you do anything at the first operation such as put in pericardial substitutes, close the pericardium, or bring the lung over the heart? I think that all of us could benefit from your expertise in this area. One point that was made in the manuscript that I think is worthwhile bringing out for this audience is that you strongly suggested that these operations should be performed by congenital heart surgeons. I just wanted to know if you still believe that. DR DEARANI: Thanks, Dr Kanter. The 6% incidence of cardiac injury is not a reflection of an uncontrollable situation in the operating room. This is a manner with which we wanted to accumulate anything from a rent in the innominate vein to a laceration of the ascending aorta. So it was all-inclusive. In this particular group of patients, of course, the concern by the surgeon was from the very, very beginning. So many patients were cannulated. We do not use wires. We open and expose the groin, we percutaneously cannulate the superior vena cava via the internal jugular vein on the right side, and we expose, in general, the right groin or the left groin depending upon patency of peripheral vessels. In some of these patients the femoral artery is quite small, and what we do to minimize compartment syndrome is usually a chimney graft on the femoral artery, unless the femoral artery is very, very large. We use a smaller femoral venous cannula with vacuum assist. And so most of these patients were either cannulated or exposed or the sternotomy was even being done on bypass because of the concern. Stroke. I think that the risk of stroke is a reflection of the number of valve patients, specifically left-sided valves. I don t know the exact number of tissue vs mechanical valves. I know on the right side they are almost predominantly tissue; on the left side there is going to be a smattering depending upon number of sternotomy and circumstances with previous tissue prostheses. Your question about mediastinal barriers, we have mixed feelings about that. If we are fortunate enough to do the primary sternotomy, we try to close the pericardium. If we have not been there for the primary sternotomy, and you know as well as any of the congenital surgeons that have been practicing for 20 or more years, the pericardium, a lot of it was removed the first time around for a variety of reasons. So when you have that, we 129 selectively use Gore-Tex pericardial membrane ( W.L. Gore and Associates, Flagstaff, Arizona). That seems to be the best in our experience. Some of these other biologic substances can really cause a profound inflammatory reaction and it has been quite mixed to the point that we don t use them anymore. And, finally, on the surface it appears that the vast majority of operations are valve-related. To me, that is not a free pass for an adult cardiac surgeon to do a lot of these cases. The exposure of a systemic atrioventricular valve in a dextrocardia patient or with corrected transposition, the knowledge of all of the anomalous coronary patterns that can coexist with these patients, and just the experience of doing it day after day, it is my belief that I think congenital surgeons are the most suited to do these cases. Thanks. DR CONSTANTINE MAVROUDIS (Cleveland, OH): That was a great presentation, Kimberly, and considering all the work that you did on this project, it would seem appropriate for you to answer this question. I am interested in those patients that had a major cavitary entry, whether atrial or ventricular, which complicated the operation. Did you discover how many of those had debilitating strokes, and if you did, were you able to assess the intraoperative hemodynamics during the dissection and subsequent commencement of cardiopulmonary bypass? In other words, did a cavitary entry result in untoward hemodynamic problems and did any of those patients suffer a stroke from that event? MS HOLST: Thank you, Dr Mavroudis. Your point is well taken; however, the objective of this manuscript was limited to primarily focus on risk factors for mortality and cardiac injury. While we do have some information on incidence of stroke and corresponding relationship to mortality, we have not analyzed the impact of cardiac injury, other intraoperative variables, or additional specific risk factors for neurologic morbidity. Further insight into this relationship could be obtained from our patient cohort and would be an excellent study objective. Thank you. DR MAVROUDIS: Thank you very much for your answer and excellent presentation. DR JOSEPH E. BAVARIA (Philadelphia, PA): I would like to make a comment about the provocative question about an adult congenital heart surgeon doing this. I think it is a little simplistic myself, and I appreciate your comments. At Penn, which is the CHOP Penn adult congenital program, we have multiple surgeons involved with that program. If there are certain kinds of cases that would require, for example, an aortic specialist, especially maybe one who could do really good valve-sparing procedures in an adult congenital case, that would be something that that person would do. If it is something intracardiac and is a tough place, maybe the congenital cardiac surgeon would do it. My point is that I would not advocate any kind of a simplistic approach to that particular question, and maybe a really good place like the Mayo or Boston or Penn or something, we would have multiple surgeons involved in the adult congenital clinic. DR DEARANI: Well, I think there is room for collaboration. The data is present, and the outcome of surgery in adults with congenital heart disease is significantly reduced when they are taken care of by congenital personnel, and the cost of hospitalization and stay is also reduced, and it has to do with knowledge ADULT CARDIAC

9 130 HOLST ET AL Ann Thorac Surg OUTCOMES AFTER MULTIPLE REOPERATIONS IN CHD 2011;92: and expertise. That is not to say that there is not room for collaboration, but I think, in general, adult congenital operations should be done by congenital surgeons. DR JOHN EDMUND MAYER, JR (Boston, MA): Did you analyze for the presence of any residual intracardiac shunt lesions, patent foramen ovale (PFO) or atrial septal defect, in terms of the incidence of postoperative neurologic events? Would you comment on your strategies for minimizing the risk of air embolism, particularly if you are on bypass, the heart is beating, and you have a cavitary entry? DR DEARANI: We have addressed this in the manuscript. To the best of my knowledge, stroke was not related to inadvertent air embolism, but your point is well taken. This is what we do. We exclude the presence of a PFO in the operating room, and sometimes that involves agitated saline injection in the right femoral vein, not just the right internal jugular or left internal jugular vein or wherever the anesthesia monitor line is, because of the nature of the flap of the PFO. If there is a shunt present in the heart, then we in those circumstances are more actually in a hazardous reentry to actually do the sternotomy on bypass to avoid an inadvertent entry into the right ventricle or the right atrium or a conduit, and then after we complete the sternotomy, we then perform some of the dissection off bypass to try to trim the bypass time down a little bit. In the event that there is an entry into the right heart in the presence of a shunt, it is a loud conversation with the perfusionist about keeping the central venous pressure positive so that blood is coming out of the rent and air is not going into the rent. And my final comment would be is after the mediastinal dissection is done, even something as simple as a pulmonary valve replacement, we always put a tack vent in the ascending aorta before a cardiotomy in the right heart just because every once in a while a shunt can be present that was not identified. The Society of Thoracic Surgeons Policy Action Center The Society of Thoracic Surgeons (STS) is pleased to announce a new member benefit the STS Policy Action Center, a website that allows STS members to participate in change in Washington, DC. This easy, interactive, hassle-free site allows members to: Personally contact legislators with one s input on key issues relevant to cardiothoracic surgery Write and send an editorial opinion to one s local media senators and representatives about upcoming medical liability reform legislation Track congressional campaigns in one s district and become involved Research the proposed policies that help or hurt one s practice Take action on behalf of cardiothoracic surgery This website is now available at by The Society of Thoracic Surgeons Ann Thorac Surg 2011;92: /$36.00 Published by Elsevier Inc