Total Anomalous Pulmonary Venous Connection: Factors Associated With Mortality and Recurrent Pulmonary Venous Obstruction

Transcription

1 Total Anomalous Pulmonary Venous Connection: Factors Associated With Mortality and Recurrent Pulmonary Venous Obstruction S. Adil Husain, MD, Elaine Maldonado, MD, Debbie Rasch, MD, Joel Michalek, MD, Richard Taylor, MD, Christopher Curzon, MD, Steve Neish, MD, and John H. Calhoon, MD Departments of Cardiothoracic Surgery, Pediatrics, Division of Cardiology, Pediatrics, Division of Critical Care, Anesthesiology, and Epidemiology and Biostatistics, University of Texas Health Sciences Center San Antonio, San Antonio, Texas Background. Surgical repair of total anomalous pulmonary venous connection (TAPVC) is associated with high rates of mortality and need for reintervention. The purpose of this study was to identify variables associated with surgical mortality and, in particular, to define predictors of recurrent pulmonary venous obstruction. Methods. All patients who underwent surgical repair for TAPVC from 2005 to 2010 at a single institution were included in our analysis. Hospital course, operative data, and outpatient records were reviewed. Results. Fifty-one patients were available for review and all were included in the analysis. Anatomic TAPVC subtypes included supracardiac 26 (51%), intracardiac 10 (19.6%), infracardiac 9 (17.6%), and mixed 6 (11.8%). Pulmonary venous obstruction was present at initial operation in 13 (25.5%) patients. Median age at repair was 18 days and median weight was 3.6 kg. Single-ventricle physiology was present in 9 (17.6%), with a diagnosis of heterotaxy syndrome in 7 (13.7%). There were 5 (9.8%) operative and 2 late deaths. Recurrent pulmonary venous obstruction requiring reintervention was found in 8 (15.7%) patients with median time to reintervention of 220 days. Obstructed TAPVC was found to be associated with surgical mortality (p 0.01). Cardiopulmonary bypass (p 0.02) and aortic cross-clamp times (p 0.03) were found to be associated with increased risk for reintervention. Intraoperative transesophageal echocardiography findings of a mean confluence gradient 2 mm Hg or greater was found to be markedly associated with recurrent pulmonary venous obstruction requiring reintervention (p < 0.001). Conclusions. Mortality after repair of TAPVC is highest in patients presenting with obstruction at time of repair. Longer cardiopulmonary bypass and cross-clamp times are associated with recurrent pulmonary venous obstruction requiring reintervention. The strongest association with need for reintervention was in patients with intraoperative transesophageal echocardiography Doppler evidence of pulmonary venous obstruction. (Ann Thorac Surg 2012;94:825 32) 2012 by The Society of Thoracic Surgeons Total anomalous pulmonary venous connection (TAPVC) is a rare cardiac anomaly, often discussed in the surgical literature in regard to variables impacting outcomes as well as recurrent pulmonary venous obstruction. Its incidence has been cited to be approximately 1% to 3% of all children born with congenital heart defects and is characterized by abnormal pulmonary venous drainage into the systemic venous circulation [1]. Metrics of surgical outcome for TAPVC are most often associated with operative mortality rates and incidence of recurrent pulmonary venous obstruction. Early historical studies noted a surgical mortality rate as high as 37%, Accepted for publication April 6, Presented at the Fifty-eighth Annual Meeting of the Southern Thoracic Surgical Association, San Antonio, TX, Nov 9 12, Address correspondence to Dr Husain, Division of Pediatric Cardiothoracic Surgery, University of Texas Health Sciences Center San Antonio, 7703 Floyd Curl Dr, Mail Code 7841, San Antonio, TX ; husain@uthscsa.edu. with a rate of 57% for children operated upon within their first year of life [2]. Mortality rates for surgical intervention have improved steadily with advances in diagnosis, medical stabilization, and finally surgical treatment. However, diverse surgical mortality rates are cited in the literature, ranging from 10% to 30% and have been associated with variables such as single-ventricle physiology, heterotaxy syndrome, other coexisting cardiac anomalies, and preoperative evidence of pulmonary venous obstruction [3 5]. Wide ranges in rates of reintervention are also described and are noted to be associated with presenting anatomic subtypes and obstructive features, as well as specific surgical techniques employed at time of initial repair [6]. The purpose of this study is to evaluate our institution s experience at the University of Texas Health Sciences Center San Antonio (UTHSCSA) and CHRISTUS Santa Rosa Children s Hospital (CSRCH), with surgical management for TAPVC. The study period is relatively short and examines results within the most current 2012 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc

2 826 HUSAIN ET AL Ann Thorac Surg TOTAL ANOMALOUS PULMONARY VENOUS CONNECTION 2012;94: surgical era. In particular, we wish to identify variables associated with surgical mortality as well as define predictors of recurrent pulmonary venous obstruction requiring surgical reintervention. Patients and Methods Table 1. Patient Demographics A review of the pediatric cardiothoracic surgical database at our institution between January 2005 and December 2010 identified 51 patients who underwent surgical repair for TAPVC. The CHRISTUS Santa Rosa Children s Hospital Ethics Committee provided approval for this study and the need for individual consent was waived. Surgical mortality was defined as death within 30 days of operation or during primary hospitalization. Reintervention was defined as any operation performed secondary to recurrent pulmonary venous obstruction. When evaluating specific surgical technique, our study included 4 surgeons. A variety of techniques were employed in regard to approach taken for repair and specific location for confluence to atrium anastomosis. These included configuration of the confluence into the dome of the L atrium, the L atrial appendage, and also through an opening into the R atrium, followed by incision of the intra-atrial septum and reconstruction. No repairs were undertaken with use of absorbable sutures such as polydioxanone suture. This has been cited previously as a possible variable in decreasing the incidence of recurrent pulmonary vein stenosis at the area of the confluence [6]. All reinterventions were addressed by a sutureless technique of repair by creating a pericardial well and avoiding placement of sutures within the pulmonary vein intimal tissue. Potential variables associated with surgical as well as late mortality and predictors associated with need for surgical reintervention were examined. These included age and weight at time of initial repair, gender, anatomic subtype of TAPVC, preoperative evidence of obstruction, need for additional surgical procedures, single-ventricle physiology, heterotaxy syndrome, cardiopulmonary bypass (CPB) times, aortic cross-clamp (ACC) times, blood product usage (including a specific analysis of packed red blood cells, fresh frozen plasma, platelets, and cryoprecipitate), antifibrinolytic management strategies (aprotinin versus aminocaproic acid [Amicar; American Reagent, Shirley, NY]) and intraoperative transesophageal echocardiography (TEE) findings after separation from CPB. All TEE studies were performed and reviewed by a single pediatric echocardiographer. Mean gradients were measured at the level of the pulmonary venous confluence anastomosis to the left atrium. The TEE studies were not able to be located for 2 patients in the cohort and both came from early in the study period. Both of these patients survived surgical intervention and neither have required reintervention to date. Statistical analysis of the data was performed by the UTHSCSA Department of Biostatistics and Epidemiology. Wilcoxon and Fisher exact tests were employed to determine statistical significance regarding variables associated with surgical mortality, late mortality, and need for surgical reintervention. Results Total Age (days) Mean (SD) (981.3) Median 18 Gender, n (%) Female 17 (33.3) Male 34 (66.7) Total 51 Weight (kg) Number 51 Mean (SD) 5.3 (6) Median 3.6 TAPVC type, n (%) Supracardiac 26 (51) Intracardiac 10 (19.6) Infracardiac 9 (17.6) Mixed 6 (11.8) Total 51 TAPVC total anomalous pulmonary venous connection. Between 2005 and 2010, 51 patients underwent surgical intervention for TAPVC. Patient demographics are as noted (Table 1). Of the 51 patients, 34 were male (66.7%) and 17 female (33.3%). Male predominance was found to be statistically significant (p 0.04). Median age at time of surgery was 8 days (range 0 to 316 days) with median weight 3.6 kg (range 2.0 to 13.4 kg). Regarding anatomic subtypes encountered, 26 patients (51%) presented with supracardiac, 10 patients (19.6%) with intracardiac, 9 patients (17.6%) with infracardiac, and 6 patients (11.8%) with mixed TAPVC. At time of presentation, 13 patients (25.5%) had evidence of obstruction. In patients who presented with obstruction, 9 (69.2%) had evidence of infracardiac, 3 (23.1%) of supracardiac, and 1 (9.7%) of cardiac TAPVC. The incidence of obstructed TAPVC was thus found most commonly in infracardiac TAPVC (p ). Single-ventricle physiology and associated palliative surgical approach was noted in 9 (17.6%) patients, and there were 7 patients (13.7%) with heterotaxy syndrome (Table 2). In 13 patients (25.5%), additional procedures were performed at time of initial TAPVC repair, including creation of a systemic to pulmonary artery shunt (4 patients), pulmonary artery banding (3 patients), bidirectional Glenn procedure (2 patients), tricuspid valve repair (2 patients), ventricular septal defect closure (1 patient), and aortic arch reconstruction (1 patient). Follow-up for the entire study cohort ranged from 4 to 112 months with a median of 60 months. To the present date, no new late deaths or reintervention for pulmonary venous obstruction have occurred within the entire study cohort.

3 Ann Thorac Surg HUSAIN ET AL 2012;94: TOTAL ANOMALOUS PULMONARY VENOUS CONNECTION 827 Table 2. Operative Variables Total CPB (minutes) Mean (SD) 89.4 (34.3) Median 82 Minimum, maximum 51, 204 ACC (minutes) Mean (SD) 41.2 (19.2) Median 36 Minimum, maximum 18, 122 Antifibrinolytic, n (%) Aprotinin 31 (60.8) Amicar 20 (39.2) Mortality, n (%) No 46 (90.2) Yes 5 (9.8) Reintervention, n (%) No 43 (84.3) Yes 8 (15.7) Obstruction Supracardiac 3 (23.1) Intracardiac 1 (7.7) Infracardiac 9 (69.2) Mixed 0 (0) Total 13 (25.5) Single ventricle Supracardiac 5 (55.6) Intracardiac 0 (0) Infracardiac 3 (33.3) Mixed 1 (11.1) Total 9 (17.6) Heterotaxy Supracardiac 4 (57.1) Intracardiac 0 (0) Infracardiac 2 (28.6) Mixed 1 (14.3) Total 7 (13.7) ACC aortic cross clamp; CPB cardiopulmonary bypass. Operative Technique Surgery was performed with standard CPB. Ischemic arrest with placement of an ACC was performed in all cases with administration of antegrade cardioplegia for myocardial protection. The use of deep hypothermic circulatory arrest was limited to only 4 patients (7.8%) and was associated with surgeon preference and not anatomic findings. As such, this variable and its impact upon mortality or rates of reintervention were not examined in the study. Median CPB time was 82 minutes (range 51 to 204 minutes) while median ACC time was 36 minutes (range 18 to 122 minutes). The use of antifibrinolytics was era dependent with aprotinin used for the first 31 cases (62%) and Amicar employed for the subsequent 20 cases (38%) (Table 2). All blood product use was noted and statistical analysis was employed to evaluate its possible association with mortality and rates of reintervention. Reconstruction of the pulmonary venous confluence at time of initial intervention was found to have been performed on 5 patients, all of whom had mean TEE gradient findings of 2 mm Hg or greater. In 2 of these patients, the gradient improved to less than 2 mm Hg and neither required reintervention. Of the 3 in whom the gradient was not below 2 mm Hg, all underwent future reintervention. Univariate Analysis As 4 surgeons were involved with this study cohort, this variable was analyzed to determine any correlation with mortality or need for reintervention. Using the Fisher exact test, surgeon of record was not statistically significant in regard to a risk factor for operative mortality (p 0.47), late deaths (p 0.82), or need for reintervention (p 0.94). Operative and Late Mortality There were 5 operative mortalities (9.8%) and 2 late deaths (3.9%). Preoperative pulmonary venous obstruction was noted to be most correlative with operative mortality (p 0.01) and odds ratio with 95% confidence interval of 16.4 (1.6 to 165.5). Surgical mortality was also found to be most prevalent in the infracardiac group (3 of 9, 33.3%), which was statistically significant (p 0.04) (Table 3). There were 2 late deaths in the study cohort. As a result, total mortality in the study was 7 of 51 (13.8%). When including late mortality within the statistical analysis of investigated variables, several additional statistically significant correlations were observed (Table 4). These included need for additional procedures at time of initial repair (p 0.008), heterotaxy syndrome (p 0.05), single-ventricle cohort (p 0.01), CPB time (p 0.01), and evidence of pulmonary venous obstruction at time of initial operative intervention (p 0.008). Due to the short duration of our study, it is important to note that 1 surgical death was in a patient who had undergone reintervention for pulmonary venous obstruction at time of stage II bidirectional Glenn procedure. This was a child with an unbalanced atrioventricular canal defect, heterotaxy syndrome, and an interrupted inferior vena cava. As such, this child s statistical analysis transcends many data variables in our study, namely mortality and reintervention as a single-ventricle patient undergoing a staged palliative and electively timed reintervention. Reoperation for Pulmonary Venous Stenosis There were 8 patients (15.7%) who required surgical reintervention for pulmonary venous obstruction. Median number of days to reintervention was 220 (range 67 to 1,166 days). Several studied variables were noted to be significantly associated with recurrent pulmonary venous stenosis requiring operative reintervention (Table 5). These included gender (p 0.04), CPB time (p 0.02), ACC time (p 0.03) and mean intraoperative TEE gradient at area of confluence 2.0 mm Hg or greater (p 001), and odds ratio with 95% confidence interval 26.4 (2.9 to 243.6). A review of the operative records revealed that 7 patients underwent revision of the pulmonary venous confluence at time of initial operation due to either

4 828 HUSAIN ET AL Ann Thorac Surg TOTAL ANOMALOUS PULMONARY VENOUS CONNECTION 2012;94: Table 3. Variables Associated With Surgical Mortality Number Yes p Value TAPVC type, n (%) Supracardiac 25 (54.3) 1 (20) Intracardiac 10 (21.7) 0 (0) Infracardiac 6 (13) 3 (60) Mixed 5 (10.9) 1 (20) Gender, n (%) Female 15 (32.6) 2 (40) Male 31 (67.4) 3 (60) Additional procedures, n (%) No 36 (78.3) 2 (40) Yes 10 (21.7) 3 (60) Heterotaxy, n (%) No 40 (87) 4 (80) Yes 6 (13) 1 (20) Ventricles, n (%) Single ventricle 7 (15.2) 2 (40) Biventricular 39 (84.8) 3 (60) Total 46 5 Obstruction, n (%) No 37 (80.4) 1 (20) Yes 9 (19.6) 4 (80) Antifibrinolytic, n (%) Aprotinin 28 (60.9) 3 (60) Amicar 18 (39.1) 2 (40) Gradient (mm HG), n (%) 2 36 (80) 4 (100) 2 9 (20) 0 (0) Age (days) Median Weight (kg) Median CPB (minutes) Number 45 4 Median ACC (minutes) Number 45 4 Median PRBC FFP Cryoprecipitate Median 0 1 Platelets Median a a b 0.08 b 0.11 b 0.86 b 0.18 b 0.15 b 0.72 b 0.39 b Table 4. Variables Associated With All Mortality (Including Late Deaths) No Yes p Value TAPVC type, n (%) 0.14 a Supracardiac 23 (88.5) 3 (11.5) Intracardiac 10 (100) 0 (0) Infracardiac 6 (66.7) 3 (33.3) Mixed 5 (83.3) 1 (16.7) Gender, n (%) Female 15 (88.2) 2 (11.8) Male 29 (85.3) 5 (14.7) Additional procedures, n (%) a No 36 (94.7) 2 (5.3) Yes 8 (61.5) 5 (38.5) Heterotaxy, n (%) 0.05 a No 40 (90.9) 4 (9.1) Yes 4 (57.1) 3 (42.9) Ventricles, n (%) 0.0 Single ventricle 5 (55.6) 4 (44.4) Biventricular 39 (92.9) 3 (7.1) Obstruction, n (%) a No 36 (94.7) 2 (5.3) Yes 8 (61.5) 5 (38.5) Antifibrinolytic, n (%) Aprotinin 27 (87.1) 4 (12.9) Amicar 17 (85) 3 (15) Gradient (mm Hg), n (%) 0.58 a 2 34 (85) 6 (15) 2 9 (100) 0 (0) Age (days) 0.34 b Median Weight (kg) 0.31 b Median CBP (minutes) 0.01 b Number 43 6 Median ACC (minutes) 0.65 b Number 43 6 Median PRBC 0.41 b FFP Cryoprecipitate 0.74 b Median 0 0 Platelets 0.45 b Median 1 0 a Fisher exact test. b Wilcoxon test. a Fisher exact test. b Wilcoxon test. ACC aortic cross clamp; CPB cardiopulmonary bypass; FFP fresh frozen plasma; PRBC packed red blood cells; TAPVC total anomalous pulmonary venous connection. ACC aortic cross clamp; CPB cardiopulmonary bypass; FFP fresh frozen plasma; PRBC packed red blood cells; TAPVC total anomalous pulmonary venous connection.

5 Ann Thorac Surg HUSAIN ET AL 2012;94: TOTAL ANOMALOUS PULMONARY VENOUS CONNECTION 829 Table 5. Variables Associated With Recurrent Pulmonary Venous Stenosis, Need for Reintervention No. Yes p Value TAPVC type, n (%) Supracardiac 23 (53.5) 3 (37.5) Intracardiac 8 (18.6) 2 (25) Infracardiac 7 (16.3) 2 (25) Mixed 5 (11.6) 1 (12.5) Gender, n (%) Female 17 (39.5) 0 (0) Male 26 (60.5) 8 (100) Additional procedures, n (%) No 33 (76.7) 5 (62.5) Yes 10 (23.3) 3 (37.5) Heterotaxy, n (%) No 38 (88.4) 6 (75) Yes 5 (11.6) 2 (25) Ventricles, n (%) Single ventricle 6 (14) 3 (37.5) Biventricular 37 (86) 5 (62.5) Obstruction, n (%) No 32 (74.4) 6 (75) Yes 11 (25.6) 2 (25) Antifibrinolytic, n (%) Aprotinin 27 (62.8) 4 (50) Amicar 16 (37.2) 4 (50) Total 43 8 Gradient (mm Hg), n (%) 2 38 (92.7) 2 (25) 2 3 (7.3) 6 (75) Age (days) Median Weight (kg) Median CPB (minutes) Number 41 8 Median ACC (minutes) Number 41 8 Median PRBC FFP Cryoprecipitate Median Platelets Median 0 2 a Fisher exact test. b Wilcoxon test a 0.04 a 0.4 a 0.3 a 0.14 a 0.7 a b 0.7 b 0.02 b 0.03 b 0.87 b 1 b 0.66 b 0.15 b ACC aortic cross clamp; CPB cardiopulmonary bypass; FFP fresh frozen plasma; PRBC packed red blood cells; TAPVC total anomalous pulmonary venous connection. anatomic concerns or TEE findings. The records were not uniformly consistent in documentation of the TEE gradient findings, which led to the decision for revision. Due to the short time span of our study, it is important to note that 2 patients in the reintervention arm were patients who required a third intervention secondary to recurrent pulmonary venous obstruction. Both of these patients underwent sutureless technique at time of initial reintervention and both were also noted to have very diminutive pulmonary veins extending distally. One of these patients died 8 months after third intervention and thus fell within the analysis of late deaths. In addition, follow-up is ongoing in our patient cohort and the need for reintervention in patients most recently operated upon may change our statistical analysis and findings with time. Multivariate Analysis In addition to the univariate analysis, we did perform multivariate logistic regression to evaluate the outcomes of operative and late mortality as well as need for reintervention for pulmonary venous stenosis. s evaluated included weight, single-ventricle physiology, anatomic type of TAPVC, presence of heterotaxy, presence of obstructed TAPVC, CPB and ACC times, antifibrinolytic choice, as well as TEE findings of confluence gradient. For operative deaths, none of these covariates were independently associated with death adjusted for the other covariates. Using forward and backward conditional model building, however, only obstructed pulmonary veins at time of operative intervention seemed to contribute significantly to the ability of the model to predict death (p 0.02). For overall deaths, the occurrence of obstructed veins as significantly associated with death independent of other variables (p 0.03). When analyzing the risk for reintervention due to pulmonary venous stenosis, no covariates emerged as significant when controlled for others. However, the duration of CPB as well as the TEE findings of gradient 2.0 mm Hg or greater emerged in the model building to add significantly to the predictive ability of the model and were significantly associated with need for reintervention, independent of each other (p for CPB time and p for gradient 2.0 mm Hg). The multivariate analysis was limited by the small sample size and rarity of outcomes, yielding low power to identify independent associations. Comment We reviewed our patients undergoing intervention for TAPVC at a single institution in the modern era from January 2005 through December Our approach and interest was secondary to several factors. The UTHSCSA moved its congenital heart program to CSRCH in We were interested in evaluating our experience with a higher risk disease entity after our relocation. In addition, it was our impression that the incidence of TAPVC was higher in our region in comparison with other geographic areas on a national level. During this study

6 830 HUSAIN ET AL Ann Thorac Surg TOTAL ANOMALOUS PULMONARY VENOUS CONNECTION 2012;94: period, our program performed 984 open cardiac surgical cases and thus this cohort comprised 51 of 984 (5.2%) of our open surgical case volume. Our experience was thus somewhat unique in that it involved a large number of patients with a specific disease entity (TAPVC) who were treated at a new surgical site, and operated upon within a relatively short amount of time. Due to a lack of organized data collection prior to 2005, no comparison studies were performed regarding the program prior to its relocation. We analyzed the cohort with an interest in defining risk factors for surgical mortality and recurrent pulmonary venous stenosis requiring operative reintervention. Mortality Our study cohort and techniques employed for initial repair were associated with a surgical mortality of 9.8% and a late mortality of 3.9%. These mortality rates favor comparably with other reports [7 9]. Of all variables analyzed in the series, only TAPVC type and evidence of obstruction at time of repair were noted to be associated with surgical mortality. When including late mortality into our statistical analysis, heterotaxy syndrome, singleventricle physiology, additional surgical procedures, longer cardiopulmonary bypass times, and initial findings of pulmonary venous obstruction were all found to be significant variables of association. These findings are well described in other studies [10]. In particular, a recent report of 100 patients repaired at Children s Memorial Hospital in Chicago, IL described a 47% mortality (8 of 17 patients) in single-ventricle patients who required TAPVC repair [11]. This group also cited a 29% rate of reintervention (5 of 17 patients) in the single-ventricle physiology cohort. The study by Morales and colleagues [12], at Texas Children s Hospital, reveals an exception to this general finding. Their group reported a 5-year survival for patients with heterotaxy syndrome of 79% compared with 89% in patients without heterotaxy. As such, they do show a 10% survival difference in these patients and it is worth note that a majority of their patients requiring reintervention had heterotaxy syndrome. Need for Reintervention Our study cohort and techniques employed for initial repair were associated with a reintervention rate of 13.8%. No female patients in the cohort required reintervention. Although other authors have found correlations between the single-ventricle and heterotaxy population with the need for reintervention, none of statistical significance was seen in our analysis [10]. The primary technique employed to repair patients with recurrent pulmonary venous stenosis was a sutureless pericardial technique as described by Caldarone and colleagues [13] as well as Lacour-Gayet and colleagues [14, 15]. Our group paid particular attention to the intraoperative TEE findings. In particular, we evaluated the median gradient noted at the anastomosis between the confluence of the pulmonary veins and the left atrium. All gradients were noted as a whole number. A single pediatric echocardiographer either performed the study, or evaluated the studies that had been obtained prior to her arrival to our program in September Of all patients in the study, 2 intraoperative TEE studies were unable to be located. Neither of these patients had a surgical nor late mortality. Neither patient has required reintervention for pulmonary vein stenosis. Statistical analysis was performed on mean gradients to determine an association as a risk factor for reintervention. This was found to be a gradient 2 mm Hg or greater. As such, we would recommend that a high degree of importance be placed upon the intraoperative TEE findings regarding both decision making for further operative revision at time of initial intervention, and as a marker for family and multidisciplinary team discussions regarding postoperative and follow-up care as well as the suspicion of need for future reintervention. Intraoperative TEE findings can be challenging to interpret as a tool for acute decision making. Limitations regarding TEE data interpretation and the clinical significance of this statistical finding may be questioned. Uniformity in technique regarding where the mean gradient is actually evaluated and the impact of diminutive distal pulmonary venous architecture and thus a color flow acceleration appearing prior to the confluence can create difficulty in gradient interpretation. As such, anastomosis size and laminar flow noted at the confluence also can be of significance in addition to the numeric mean gradient finding. Adequate filling of the heart and optimization of hemodynamic parameters often alter the gradients noted. In addition, as the left-sided structures become accustomed to receiving pulmonary venous return, significant lability is often noted regarding inotropic needs, myocardial function, and overall parameters of perfusion. Predictable findings of pulmonary hypertension can also create additional challenges. It is not uncommon in our practice to note a varying and often times decreasing mean gradient as we continue our process of modified ultrafiltration, decannulation, and closure of the chest. The technical challenges of the initial attempt at repair, as well as the patient s clinical picture, will often dictate our decision to reattempt creation of an adequate confluence anastomosis. In addition, preoperative diagnostic studies employed to identify pulmonary venous architecture distally into the lungs can provide information of clinical significance. As such, the TEE findings provide a valuable additional tool for intraoperative decision making; however, other factors clearly impact the algorithms that we may follow to decide upon the need for further attempts at repair. As such, we currently employ a post repair TEE gradient of 2 mm Hg or greater as a marker for concern but not necessarily an algorithmic determiner of further intraoperative revision of the confluence. Prior descriptions of intraoperative TEE data and the impact of their findings upon surgical outcomes for TAPVC are limited. Kelle and colleagues [11] did describe follow-up echocardiographic evaluations (transthoracic) on 65 of the 100 patients described in their study. Of these 65 patients, 45 had numerical measure-

7 Ann Thorac Surg HUSAIN ET AL 2012;94: TOTAL ANOMALOUS PULMONARY VENOUS CONNECTION 831 ments of a mean gradient available. They defined a gradient 2.5 mm Hg or less as having no obstruction and created a classification system of no, mild, moderate, and severe obstruction. All patients with severe obstruction (mean gradient 10 mm Hg) required reintervention and died of complex medical issues. No data were provided for patients with mild (2.5 to 5.0 mm Hg) or moderate (5.0 to 10.0 mm Hg) regarding need for reintervention [11]. Although the limitations of this diagnostic tool are accepted, our study provides powerful data as to the impact that any intraoperative suspicion of any residual obstruction has upon surgical outcomes. Our group has determined surgical outcomes and associated risk factors in a single institution, modern era series of infants with TAPVC. Our mortality as well as reintervention rates favor comparably with that of many published series. Surgical mortality was found to be most associated with evidence of obstruction at time of initial intervention. Late mortality was found to be associated with obstruction at presentation, in addition to factors such as single-ventricle physiology and heterotaxy syndrome. Rates for reintervention were most associated with intraoperative TEE findings of a mean confluence to left atrium gradient 2 mm Hg or greater. The clinical versus statistical findings associated with this particular statistical finding is worthy of further discussion. Multivariate analysis was limited by small sample size and rarity of outcomes, and thus produces a clear limitation to the study power. These data and their analysis continue to assist our efforts of educating families and other members of our multidisciplinary team as to factors that are most associated with key metrics of outcome. References 1. Karamlou T, Gurofsky R, Al Sukhni E, et al. Factors associated with mortality and reoperation in 377 children with total anomalous pulmonary venous connection. Circulation 2007;115: Wuasch DC, Deutsch M, Reul GJ, Hallman GL, Cooley DA. Total anomalous pulmonary venous return. Review of 125 treated surgically. Ann Thorac Surg 1975;19: Yong MS, d Udekem Y, Robertson T, et al. Outcomes of surgery for simple total anomalous pulmonary venous drainage in neonates. Ann Thorac Surg 2011;91: Kirshbom PM, Myung RJ, Gaynor JW, et al. Preoperative pulmonary venous obstruction affects long term outcome for survivors of total anomalous pulmonary venous connection repair. Ann Thorac Surg 2002;74: Hancock Friesen CL, Zurakowski D, Thiagarajan RR, et al. Total anomalous pulmonary venous connection: an analysis of current management strategies in a single institution. Ann Thorac Surg 2005;79: Hawkins JA, Minich LL, Tani LY, Ruttenberg HD, Sturtevant JE, McGough EC. Absorbable polydioxanone suture and results in total anomalous pulmonary venous connection. Ann Thorac Surg 1995;60: Caldarone CA, Najm HK, Kadletz M, et al. Surgical management of total anomalous pulmonary venous drainage: impact of coexisting cardiac anomalies. Ann Thorac Surg 1998;66: Michielon G, Di Donato RM, Pasquini L, et al. Total anomalous pulmonary venous connection: long-term appraisal with evolving technical solutions. Eur J Cardiothorac Surg 2002;22: Hyde JA, Stümper O, Barth MJ, et al. Total anomalous pulmonary venous connection: outcome of surgical correction and management of recurrent venous obstruction. Eur J Cardiothorac Surg 1999;15: Gaynor JW, Collins MH, Rychik J, Gaughan JP, Spray TL. Long-term outcome of infants with single ventricle and total anomalous pulmonary venous connection. J Thorac Cardiovasc Surg 1999;117: Kelle AM, Backer CL, Gossett JG, Kaushal S, Mavroudis C. Total anomalous pulmonary venous connection: Results of surgical repair of 100 patients at a single institution. J Thorac Cardiovasc Surg 2010;139: Morales DLS, Braud BE, Booth JH, et al. Heterotaxy patients with total anomalous pulmonary venous return: Improving surgical results. Ann Thorac Surg 2006;82: Caldarone CA, Najm HK, Kadletz M, et al. Relentless pulmonary venous stenosis after repair of total anomalous pulmonary venous drainage. Ann Thorac Surg 1998;66: Lacour-Gayet F, Zoghbi J, Serraf AE, et al. Surgical management of progressive pulmonary venous obstruction after repair of total anomalous pulmonary venous connection. J Thorac Cardiovasc Surg 1999;117: Lacour-Gayet F, Rey C, Planche C. Pulmonary vein stenosis: description of a sutureless surgical technique using the pericardium in situ. [Article in French] Arch Mal Coeur Vaiss 1996;89: DISCUSSION DR JAMES A. QUINTESSENZA (St. Petersburg, FL): Let me commend you on a very nice presentation and thanks for providing the manuscript a few days ago. The group from San Antonio present their series of 51 patients with TAPVR [total anomalous pulmonary venous return] over a five-year period. Single ventricles comprised 17% of the group and obstruction was present in 25% upon presentation. They had an overall mortality of 13.8%, which is consistent with contemporary reports. The STS [society of thoracic surgeons] congenital heart database has a 15.8% discharge mortality for neonates with TAPVR, so this report is slightly better than that. Postoperative obstruction developed in 15%. As we have heard, in this study mortality was related to preoperative pulmonary vein obstruction, which is an expected association. Postoperative pulmonary vein obstruction and reintervention correlated with increased cardiopulmonary bypass and cross-clamp times, which may be a surrogate for more complex or difficult anatomy, which I think would lend itself to that finding. The TEE [transesophageal echocardiography] findings of a mean gradient of 2 mm Hg or more were strongly associated with the need for reintervention. And this brings me to two questions. The abnormality on the TEE finding was very predictive and is kind of a harbinger of problems to come. What exactly do you do once you get this information in the operating room to decide that something else may or may not need to be done right then and there? And secondly, of the patients who recur with pulmonary vein obstruction, the diffuse pulmonary vein stenosis patients are really the most difficult ones to deal with. I m not sure how many of those patients you had, but is there anything in your data set that might be predictive of this subgroup? I enjoyed the presentation and congratulations.

8 832 HUSAIN ET AL Ann Thorac Surg TOTAL ANOMALOUS PULMONARY VENOUS CONNECTION 2012;94: DR HUSAIN: Thank you. I will try to answer the questions in order. I think this is a limitation of the study, and clearly, your perspective as a surgeon and what you know about the preoperative anatomy of the distal architecture, as well as your comfort level with the confluence anastomosis, plays a significant role in determining the need for further intervention. Scenarios where perhaps I am uncomfortable about the anastomosis and conduct of the operation, the confluence size is in question by our echocardiographer, or I don t see laminar flow through the confluence itself, raises concern. In some instances, we note a gradual improvement in the echocardiographic mean gradient as we improve our hemodynamics, stabilize our volume status, and complete the process of hemostasis. DR QUINTESSENZA: One point I would like to make is in a small baby just the physical presence of the probe has been shown to cause hemodynamic instability, presumably because of a compression on the veins. We actually had a probe that I was suspicious was causing a problem. I took a handheld epicardial probe, pulled the TEE out, and the pulmonary vein velocity normalized. So, one thing to be thinking about is that the physical presence of the TEE probe might affect your findings. DR HUSAIN: Absolutely. I think other subtle things that we noted, which can impact hemodynamics and echocardiographic findings, include removing stay sutures and taking out the sternal retractor. Clearly, external variables which impact anatomy and geometry even slightly can have a huge impact upon physiologic parameters and findings. One of the reasons we became interested in the intraoperative TEE findings as a variable of importance was the lack of data in the literature regarding its intraoperative use. The group from Chicago, Children s Memorial, did use postoperative transthoracic echocardiographic findings in their study and they came up with a gradient of 2.5 as their cutoff for having no obstruction, and patients that had a gradient greater than 2.5 were associated with an increased incidence of reintervention. Again, this data was on follow-up transthoracic studies in about 40% of their patient cohort. As such, the use of the intraoperative findings has not been reported in the literature. DR CHRISTOPHER BAIRD (Boston, MA): I just wanted to follow up on the last comment. In Boston, we have actually gone away from doing TEEs on many of the smaller neonates and infants, TAPVR s and some of the arch reconstructions, due to concerns of compression and/or distortion. We are in the process of determining scores for not only technical operative performance but also for the intraoperative imaging and comparing them to discharge imaging and longer term outcomes. DR ANDREW C. FIORE (St. Louis, MO): We switched over to epicardial only for TAPVR. Southern Thoracic Surgical Association: Fifty-Ninth Annual Meeting Make plans now to attend the Fifty-Ninth Annual Meeting of the Southern Thoracic Surgical Association (STSA) on November 7 10, 2012, at the Naples Grande Beach Resort in Naples, FL. Please visit to make hotel reservations and to learn more about the schedule of events by The Society of Thoracic Surgeons Ann Thorac Surg 2012;94: /$36.00 Published by Elsevier Inc