Inferior Vena Cava Oxygen Saturation Monitoring After the Norwood Procedure

Transcription

1 Inferior Vena Cava Oxygen Saturation Monitoring After the Norwood Procedure Robert J. Dabal, MD, Leslie A. Rhodes, MD, Santiago Borasino, MD, MPH, Mark A. Law, MD, Stephen M. Robert, MD, and Jeffrey A. Alten, MD Division of Cardiothoracic Surgery, Department of Surgery, Department of Pediatrics, and Section of Cardiac Critical Care, Division of Cardiology, University of Alabama at Birmingham, Birmingham, Alabama Background. Superior vena cava oxygen saturation monitoring in the early postoperative period after the Norwood procedure (NP) has been associated with improved survival and decreased adverse events (AE). There is no data describing inferior vena cava saturation (SIVO 2 ) monitoring after NP. We sought to investigate the utility of intermittent SIVO 2 monitoring after NP and to assess the correlation of SIVO 2 with renal near-infrared spectroscopy (rnirs). We hypothesized failure to achieve SIVO 2 greater than 45% within the first 4 hours after NP is predictive of AE, and that rnirs correlates with SIVO 2. Methods. A retrospective study of 26 consecutive NP patients who received postoperative management with SIVO 2 monitoring according to a strict protocol was conducted. Primary outcome was AE, defined as cardiopulmonary resuscitation, extracorporeal membrane oxygenation, death before discharge, or residual surgical defects. Results. Ten (38%) patients had one or more AE; mortality was 23%. On admission to the cardiac intensive care unit, patients with AE had lower SIVO 2 (45% ± 9.4% versus 62% ± 12.0%; p < 0.001) and lower rnirs (56 ± 6.5 versus 77 ± 7.2; p < 0.001). At 4 hours, 90% of AE patients had an SIVO 2 less than 45% versus 6% of non- AE patients. Both SIVO 2 and rnirs were highly predictive of AE: the area under the receiver-operating characteristic curve was greater than 0.86 and 0.95, respectively. Two hours after admission, an SIVO 2 less than 45% predicted AE with a specificity of 93%, a sensitivity of 70%, and a positive predictive value of 82%. The SIVO 2 was strongly correlated with rnirs (r [ 0.81). Conclusions. Intermittent SIVO 2 can be used to guide early postoperative NP management; rnirs is an accurate continuous, noninvasive surrogate for SIVO 2. An SIVO 2 of less than 45% in the first 4 hours after the NP is predictive of AE. (Ann Thorac Surg 2013;95: ) Ó 2013 by The Society of Thoracic Surgeons The high-risk postoperative period after the Norwood procedure with right ventricle to pulmonary artery conduit (NP) has been well documented. The deleterious effects of cardiopulmonary bypass (CPB), ventriculotomy, and ischemia reperfusion lead to dysfunction of the single right ventricle that is responsible for supplying cardiac output (CO) simultaneously to both the pulmonary ( Qp) _ and systemic ( Qs) _ circulations. There is minimal total CO reserve, so even small changes in oxygen delivery or consumption can lead to anaerobic metabolism, lactic acidosis, and eventual cardiovascular collapse. This volatile physiology is exacerbated by the labile vascular resistances inherent in the neonatal period with resultant imbalances in Qp/ _ Qs. _ A management goal for NP patients encompasses early identification of compromised oxygen delivery such that interventions can be made before clinical deterioration. Superior vena cava Accepted for publication Jan 29, Presented at the Fifty-ninth Annual Meeting of the Southern Thoracic Surgical Association, Naples, FL, Nov 7 10, Address correspondence to Dr Alten, Department of Pediatrics, University of Alabama at Birmingham, th Ave S ACC504, Birmingham, AL 35233; jalten@peds.uab.edu. oxygen saturation (SSVO 2 ) monitoring is commonly used as an estimate of systemic oxygen delivery in the early postoperative period, and identifies shock before changes in traditional surrogates of CO, such as blood pressure, arterial oxygen saturations, and serum lactate [1, 2]. Inability to optimize SSVO 2 in the early postoperative period is associated with increased risk of morbidity and mortality [3 5]. Recently, the use of near-infrared spectroscopy (NIRS) to provide a continuous, noninvasive estimate of SSVO 2 has gained prominence and identifies patients at risk for adverse events (AE) [6, 7]. The utility of monitoring inferior vena cava saturation (SIVO 2 ) after complex neonatal cardiac surgery has not been described. A fall in SIVO 2 may be an earlier indicator of limited CO and decreased oxygen delivery when compared with SSVO 2. In patients with normal physiology, there is redistribution of blood flow away from splanchnic and renal beds during early shock, while perfusion to the cerebral and coronary circulations is preserved [8]. This leads to increased oxygen extraction in the abdominal organs, which decreases SIVO 2 but not SSVO 2,asbloodflow to the brain will be maintained until later stages of shock. The goal of this study was to determine whether SIVO 2 monitoring or renal NIRS (rnirs) could identify those Ó 2013 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc

2 Ann Thorac Surg DABAL ET AL 2013;95: IVC OXIMETRY MONITORING AFTER NORWOOD 2115 Abbreviations and Acronyms AE = adverse event AUC = area under the curve CICU = cardiac intensive care unit cnirs = cerebral near-infrared spectroscopy CO = cardiac output CPB = cardiopulmonary bypass CPR = cardiopulmonary resuscitation CVL = central venous line ECMO = extracorporeal membrane oxygenation NIRS = near-infrared spectroscopy NP = Norwood procedure POD = postoperative day _Qp = pulmonary circulation _Qs = systemic circulation rnirs = renal near-infrared spectroscopy SIVO 2 = inferior vena cava oxygen saturation = superior vena cava oxygen saturation SSVO 2 patients at highest risk for AE in the early postoperative period. Hypotheses to be tested were (1) an inability to achieve target SIVO 2 of greater than 45% in the first 4 hours after cardiac intensive care unit (CICU) admission is associated with increased incidence of AE; (2) rnirs monitoring in the first 4 hours can discriminate patients at risk for AE; and (3) rnirs and SIVO 2 are correlated, enabling rnirs to be used as an effective continuous, noninvasive surrogate for oxygen transport balance. Patients and Methods Patient Selection and Data Collection The study was approved by the Institutional Review Board of the University of Alabama at Birmingham. This is a retrospective study of 26 consecutive patients who underwent NP from March 2010 to November We use only femoral central venous lines (CVL) in all singleventricle patients to avoid the consequences of upper central vein thrombosis. Patients were eligible for inclusion into this study if they had a femoral CVL and were managed according to the postoperative NP protocol. Patients were excluded if they were placed on extracorporeal membrane oxygenation (ECMO) in the operating room. Hemodynamic and oximetry data at CICU admission and at 2 and 4 hours (Table 1) was extracted from bedside flow sheets that were prospectively completed by nurses as part of the NP protocol. All other data, including AE, were collected from our CICU clinical database. Adverse event was defined a priori as death before discharge, cardiopulmonary resuscitation, emergent ECMO, or residual surgical defects. Operative Procedure After median sternotomy, CPB was established with arterial cannulas in the base of the innominate artery and in the ductus arteriosus along with bicaval venous cannulation. All patients received a single dose of del Nido Table 1. Comparison of Hemodynamic and Oxygen Transport Variables a Variable Adverse Events (n ¼ 10) No Adverse Events (n ¼ 16) p Value Femoral venous oxygen saturation (%) Admission < hours after admission < hours after admission <0.001 Renal NIRS (%) Admission < hours after admission < hours after admission Cerebral NIRS (%) Admission hours after admission hours after admission Arterial oxygen saturation (%) Admission hours after admission hours after admission Pulse oximetry (%) Admission hours after admission hours after admission Mean arterial pressure (mm Hg) Admission hours after admission hours after admission Pulse pressure (mm Hg) Admission < hours after admission < hours after admission <0.001 Inotropic agent score Admission hours after admission hours after admission Lactic acid (mmol/l) Admission hours after admission hours after admission ph Admission hours after admission hours after admission a All numbers are presented as mean standard deviation. NIRS ¼ near-infrared spectroscopy. cardioplegia solution and were cooled to 22 C. Continuous low-flow cerebral perfusion was used for all cases during arch reconstruction. The aortic arch reconstruction was completed using a patch of bovine pericardium for augmentation after coarctectomy. A 5-mm or 6-mm (based on weight) ringed Gore-Tex (W,L, Gore & Assoc, Flagstaff, AZ) shunt was placed from the right ventricle to the pulmonary artery bifurcation, with patch

3 2116 DABAL ET AL Ann Thorac Surg IVC OXIMETRY MONITORING AFTER NORWOOD 2013;95: augmentation of the bifurcation using bovine pericardium. Cerebral NIRS (cnirs) and rnirs monitoring was used. Ten milligrams of methylprednisolone per kilogram was given 8 hours and 1 hour before surgery; no intraoperative steroids were given. All patients received zero-balance ultrafiltration during CPB and single-pass ultrafiltration after CPB. Afterload reduction was achieved with high-dose milrinone. Delayed sternal closure was used for all cases. Postoperative Management All patients were managed with the guidance of a structured NP protocol (Fig 1). All children were started on protocol-standardized inotropic agents in the operating room, consisting of milrinone 0.75 mg $ kg 1 $ min 1, epinephrine 0.03 mg $ kg 1 $ min 1, vasopressin 0.02 U $ kg 1 $ h 1, and calcium gluconate 30 mg $ kg 1 $ h 1. These infusions and other CICU support were titrated to achieve the following goals: SIVO 2 greater than 45%, arteriovenous oxygenation difference less than 25%, arterial oxygen saturation 70% to 80%, central venous pressure 8 to 12 cm H 2 O, mean arterial pressure 45 to 55 mm Hg, heart rate less than 170 beats/min, ph 7.35 to 7.45, CO 2 35 to 45 mm Hg, rnirs greater than 60%, cnirs greater than 50%, hemoglobin greater than 14 g/dl, and serum lactate 50% of admission value within 12 hours. To minimize oxygen consumption, all patients received fentanyl, midazolam, and cisatracurium infusions as well as temperature maintenance between 36 and 37 C and judicious weaning of catecholamines. Cisatracurium was discontinued on postoperative day (POD) 1, whereas fentanyl and midazolam were continued until sternal closure. Starting in January 2011, all patients received prophylactic peritoneal dialysis within 6 hours of admission to the CICU (median, 2.5 hours); all other patients were started on furosemide infusions on POD 1 and received passive peritoneal drainage. Simultaneous sampling of arterial and femoral venous blood was performed on admission and at least every 2 hours through POD 1. Arteriovenous oxygen difference and _ Qp/ _ Qs was calculated by bedside nurse and documented on a flow sheet along with concomitant hemodynamic variables recorded at the time of blood draws. All patients received heparin 10 U $ kg 1 $ h 1. Statistics SPSS version 20 (IBM, Chicago, IL) was used for all analysis. Continuous variables not normally distributed were summarized as medians with interquartile range; group comparison was made using the Wilcoxon s rank sum test. Continuous variables with normal distribution were summarized as means with standard deviations and compared using unpaired Student s t test. Categorical data was compared using Fisher s exact test. Probability values less than 0.05 were considered statistically significant. Prognostic data were displayed graphically using area under the receiver-operating characteristic curve (AUC) analysis. Spearman s rank correlation was used to determine relationship between SIVO 2 and cnirs or rnirs. All statistical tests were two-tailed. Results Of the 29 patients who underwent the NP during the study period, 26 met inclusion criteria. Three were excluded from the analysis for the following reasons: 1 had failure to wean from CPB, and 2 did not have femoral CVL. Demographic characteristics, potential surgical risk factors, and anatomic diagnosis variables are compared in Table 2. Median age at surgery was 6 days (range, 3 to 33 days), mean weight was kg, mean gestational age was 38 weeks, and 50% were black and 50% white. Extracardiac diagnoses included the following: (1) Jacobsen s syndrome, (2) dwarfism and skeletal anomalies, (3) unknown syndrome including skeletal anomalies and absent corpus callosum, (4) unknown syndrome with skeletal, facial, and genitourinary anomalies, (5) primary immunodeficiency, and (6) partial duplication of chromosome 10 with congenital thrombocytopenia. Additional cardiac diagnoses included intact atrial septum (n ¼ 2) and total anomalous pulmonary venous return (n ¼ 1). Ten (38%) patients had one or more AE: cardiopulmonary resuscitation (CPR; n ¼ 9), death (n ¼ 6), ECMO (n ¼ 4), or reoperation (n ¼ 2). In-hospital mortality was 23%. Causes of death included 2 patients with primary cardiac failure on POD 1 and 7, multiorgan dysfunction and pulmonary hypertension on POD 13, sudden death on POD 23, and 2 with multiorgan dysfunction and sepsis POD 12 and 19. Six of 9 patients who had CPR died. Causes of CPR were primary cardiac dysfunction (n ¼ 5), respiratory acidosis (n ¼ 2), and tamponade (n ¼ 2). Outcome of CPR patients was as follows: 4 patients were placed on ECMO (1 survivor), 1 had a reoperation (survived), 1 could not be resuscitated, 2 were late deaths, and 1 had no further AE. Residual surgical lesions included kinked shunt and neoaortic obstruction. Comparison of hemodynamic and oxygen transport variables between patients with and without AE at admission and at 2 and 4 hours is shown in Table 1. At every point, AE patients had worse SIVO 2, rnirs, and lactate. The arterial oxygen saturation, mean arterial pressure, ph, inotropic agent score, and cnirs became different and clinically important only at later times. The pulse oxygen saturation was not different between groups. At 4 hours, 90% of AE patients had an SIVO 2 less than 45% versus 6% of non-ae patients (p ¼ ). Both SIVO 2 and rnirs were highly predictive of AE by receiver-operating characteristic analysis: AUC was greater than 0.86 and 0.95, respectively, for all three times (Fig 2; p < 0.001). Arterial hypotension and low cnirs were poorly predictive of AE in the first 2 hours with AUC less than 0.7, but by 4 hours mean arterial pressure became a good predictor of AE (AUC, 0.88). Two hours after admission, an SIVO 2 less than 45% predicted AE with a specificity of 93%, a sensitivity of 70%, and a positive predictive value of 82%. Admission lactate and pulse pressure were also strong predictors of AE: AUC, 0.92 and 0.98 respectively (p < 0.001). The SIVO 2 was strongly correlated with rnirs in the first 4 hours of admission (r ¼ 0.81; p < 0.001; Fig 3).

4 Fig 1. Postoperative Norwood management protocol. (ABG ¼ arterial blood gas; A VO 2 ¼ arteriovenous oxygen saturation difference; cnirs ¼ cerebral near-infrared spectroscopy; CO ¼ cardiac output; CVP ¼ central venous pressure; ECMO ¼ extracorporeal membrane oxygenation; FIO 2 ¼ fraction of inspired oxygen; HgB ¼ hemoglobin; PRBC ¼ packed red blood cells; rnirs ¼ renal near-infrared spectroscopy; SIVO 2 ¼ femoral venous saturation; Temp ¼ temperature; VBG ¼ venous blood gas.) Ann Thorac Surg DABAL ET AL 2013;95: IVC OXIMETRY MONITORING AFTER NORWOOD 2117

5 2118 DABAL ET AL Ann Thorac Surg IVC OXIMETRY MONITORING AFTER NORWOOD 2013;95: Table 2. Demographic Variables and Surgical Risk Factors a Variable Total (n ¼ 26) Adverse Event (n ¼ 10) No Adverse Event (n ¼ 16) p Value Male sex 19 (73) 6 (60) 13 (81) 0.37 Gestational age <37 wk 6 (23) 3 (30) 3 (19) 0.64 Weight <2.5 kg 5 (19) 3 (30) 2 (13) 0.34 Ascending aorta <2.5 mm 12 (46) 5 (50) 7 (44) 1.0 Additional cardiac diagnosis 3 (12) 1 (10) 2 (13) 1.0 Extracardiac diagnosis 6 (23) 5 (50) 1 (6) 0.02 Intubated preoperatively 4 (15) 3 (30) 1 (6) 0.60 Preoperative inotropic agents 3 (12) 3 (30) 0 (0) 0.05 Median CPB (min) 176 (154, 199) 191 (164.5, 237) 171 (148, 194.5) 0.21 Median ACC (min) 77 (65.5, 90.5) 80.5 (75, 91) 70.5 (61.5, 90) 0.12 Diagnosis MA/AA 12 (46) 5 (50) 7 (44) 1.0 MS/AA 6 (23) 2 (20) 4 (25) 1.0 MS/AS 4 (15) 2 (20) 2 (13) 0.63 Other 4 (15) 1 (10) 3 (19) 1.0 a All results are presented as number (%) or median (interquartile range). AA ¼ aortic atresia; ACC ¼ aorta cross-clamp time; AS ¼ aortic stenosis; CPB ¼ cardiopulmonary bypass time; MA ¼ mitral atresia; MS ¼ mitral stenosis. Lactate (r ¼ 0.4; p ¼ 0.01), cnirs (r ¼ 0.36; p ¼ 0.002), mean arterial pressure (r ¼ 0.15; p ¼ 0.2), and arterial oxygen saturation (r ¼ 0.5; p ¼ 0.001) were more weakly correlated with SIVO 2. Comment In this retrospective study we are the first to demonstrate that intermittent SIVO 2 monitoring can be used as a target Fig 2. Receiver-operating characteristic for discriminating adverse events versus no adverse events for inferior vena cava oxygen saturation (SIVO 2 ) and renal near-infrared spectroscopy (rnirs) 2 hours after cardiac intensive care unit (CICU) admission. Area under the receiver-operating characteristic curve is 0.92 and 0.95, respectively (p < 0.001). of goal-directed management after NP, while helping identify patients at risk for AE. Failure to achieve an SIVO 2 greater than 45% within the first 4 hours after CICU admission in our population was strongly associated with cardiac arrest, emergent ECMO initiation, presence of residual surgical defects, or death. In addition we demonstrate the utility of rnirs monitoring as an accurate continuous noninvasive surrogate of SIVO 2 in the early postoperative period. Decreases in SIVO 2 and rnirs identify patients at risk for AE on admission to the CICU, while preceding changes in most other hemodynamic variables. Early identification of at-risk patients affords the opportunity to intervene during the period when patients are most vulnerable for cardiovascular collapse and during which interventions may have the strongest influence on outcomes [9, 10]. Introduction of goal-directed SSVO 2 management in the early postoperative period has been associated with improved early survival after the NP [3, 5, 11]. This widely accepted strategy is based on data obtained solely from the superior vena cava as a surrogate for CO, with postoperative management targeting SSVO 2 greater than 55% to 60% [3]. As revealed by a recent multicenter survey, approximately 20% of institutions frequently use the femoral vein for postoperative CVL access [12]. However, it is not known whether SIVO 2 can be used as an alternative to SSVO 2 as an early indicator of oxygen transport imbalance. In critically ill populations with normal physiology, inferior vena cava saturations fall before superior vena cava saturations in the early stages of shock [8, 13], making it feasible that a similar pattern may occur in patients with single-ventricle physiology, with SIVO 2 demonstrating reduced oxygen delivery even earlier than SSVO 2 after the NP. We chose lower SIVO 2 (<45%) as our target in goal-directed management based on the theory that splanchnic and renal perfusion

6 Ann Thorac Surg DABAL ET AL 2013;95: IVC OXIMETRY MONITORING AFTER NORWOOD 2119 Fig 3. (A) Relationship between renal nearinfrared spectroscopy (rnirs) and inferior vena cava oxygen saturation (SIVO 2 ) during the first 4 hours of cardiac intensive care unit (CICU) admission (r ¼ 0.81; p < 0.001). (B) Relationship between cerebral near-infrared spectroscopy (cnirs) and inferior vena cava oxygen saturation during the first 4 hours of cardiac intensive care unit admission (r ¼ 0.36; p ¼ 0.02). (n ¼ 26, all with three data points [admission, and 2 and 4 hours]). decrease more in states of stress than cerebral perfusion does. Failure to achieve an SIVO 2 greater than 45% with our postoperative goal-directed NP protocol predicts patients who had an AE with 82% accuracy. The absolute difference in SIVO 2 between patients with and without AE was large (17% to 26%), compared with only 6% difference in studies using superior vena cava saturations to predict death or ECMO in postoperative NP patients [3]. This large clinical difference in SIVO 2 may allow easier clinical identification of patients at risk for AE compared with SSVO 2. The SIVO 2 is significantly lower in patients with AE on admission to the CICU and continues to drop at 2 and 4 hours, contrary to the SIVO 2 in non-ae patients, which remains constant near 60%. The decrease in SIVO 2 was present before changes in other hemodynamic measures including heart rate, blood pressure, inotropic agent score, and cnirs, similar to what has been described for SSVO 2 measurements [3, 5]. It is tempting to speculate that moreaggressive early management in the patients with low SIVO 2 may have prevented some AE. Perhaps reduction in SIVO 2 warrants urgent evaluation for residual surgical defects, tamponade, and respiratory abnormalities, followed by aggressive attempts to improve oxygen transport balance. Inability to improve the SIVO 2,especiallyinthe setting of serum lactate that is not decreasing, should prompt consideration of early ECMO initiation for hemodynamic stabilization before cardiovascular collapse. Although our ECMO incidence after NP was similar to that of a recent large multicenter study (10%) [14], all ECMO initiation in our population was performed emergently after CPR. The survival rate of patients who received CPR in our series was 33%, below that reported for single-ventricle ECMO survival [15 17]. Earlyinstitu- tion of ECMO in neonates with low SIVO 2 and lactic acidosis may decrease the incidence of cardiac arrest and subsequent organ injury, potentially increasing survival. Renal NIRS has recently been shown to correlate with SVC oximetry and is able to predict adverse outcomes such as kidney injury and lactic acidosis after neonatal cardiac surgery (NP), especially when used in conjunction with cnirs [7, 18 21]. Similar to findings in one pediatric cardiac catheterization study [22], we showed a strong correlation between rnirs and SIVO 2. Although there was a 20% rnirs saturation difference between AE and non-ae patients, cnirs was only slightly lower in AE patients, with differences insignificant until 4 hours. This supports our belief that SIVO 2 is an earlier indicator of oxygen transport imbalance than SSVO 2, as cnirs has been found to correlate well with SSVO 2 [21, 23]. Although clearly useful, SIVO 2 monitoring does require an invasive procedure and indwelling catheter. A high incidence of CVL-related thrombosis is found in pediatric cardiac patients [24]. Monitoring of SSVO 2 carries the potential risk of superior vena cava or innominate vein thrombosis and an increased incidence of chylothoraces [25], which can complicate the subsequent stages of a Fontan palliation. In contrast, thrombosis of the femoral vein has not been shown to be of significant clinical significance after NP, and insertion of femoral CVL with ultrasound guidance results in a very low likelihood of complications [26]. For these reasons, femoral venous catheterization may be preferable in this particular population. Limitations The main limitations of this study are related its retrospective nature and small sample size. Despite a wellestablished clinical NP protocol, we cannot discount potential clinician management variability with regard to timing and type of resuscitation in the first few hours of admission. Comparing a continuous variable (rnirs) with an intermittent value (SIVO 2 ) leads to the possibility that SIVO 2 does not accurately reflect the dynamic changes often seen with NIRS monitoring. The incidence of extracardiac anomalies was 23% in our cohort, much higher than 8% recently reported in patients with hypoplastic left heart [27]. This has been shown to be a significant risk factor for AE after NP and may have contributed to AE in our patients as well as making results obtained in our high-risk population in this study less generalizable to other centers. Despite limitations, SIVO 2 and rnirs were consistently, with little variability, different between patients with and without AE, including those without extracardiac anomalies. Conclusions Intermittent SIVO 2 may be used in the postoperative management of NP patients. An SIVO 2 less than 45% in the first 4 hours after the NP is associated with increased incidence of AE, and rnirs is an accurate continuous, noninvasive surrogate for SIVO 2. Further research is necessary to determine whether changes in operative technique or

7 2120 DABAL ET AL Ann Thorac Surg IVC OXIMETRY MONITORING AFTER NORWOOD 2013;95: postoperative management aimed at increasing SIVO 2 will improve outcomes after NP. It also remains to be determined whether early ECMO initiation in patients failing to meet SIVO 2 goals will reduce organ injury and mortality. References 1. Feinstein JA, Benson DW, Dubin AM, et al. Hypoplastic left heart syndrome: current consideration and expectations. J Am Coll Cardiol 2012;59(1 Suppl):S Tweddell JS, George MH, Fedderly RT, et al. Phenoxybenzamine improves systemic oxygen delivery after the Norwood procedure. Ann Thorac Surg 1999;67: Tweddell JS, Ghanayem NS, Mussatto KA, et al. Mixed venous oxygen saturation monitoring after stage 1 palliation for hypoplastic left heart syndrome. Ann Thorac Surg 2007;84: Tweddell JS, Hoffman GM, Mussatto KA, et al. Improved survival of patients undergoing palliation of hypoplastic left heart syndrome: lessons learned from 115 consecutive patients. Circulation 2002;106(12 Suppl 1):I Bradley SM, Atz AM. Postoperative management: the role of mixed venous oxygen saturation monitoring. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu 2005: Tortoriello TA, Stayer SA, Mott AR, et al. A noninvasive estimation of mixed venous oxygen saturation using nearinfrared spectroscopy by cerebral oximetry in pediatric cardiac surgery patients. Paediatr Anaesth 2005;15: Chakravarti SB, Mittnacht JC, Katz JC, Nguyen K, Joashi U, Srivastava S. Multisite near-infrared spectroscopy predicts elevated blood lactate level in children after cardiac surgery. J Cardiothorac Vasc Anesth 2009;23: Martin J, Shekerdemian LS. The monitoring of venous saturations of oxygen in children with congenitally malformed hearts. Cardiol Young 2009;19: Hoffman TM, Wernovsky G, Atz AM, et al. Efficacy and safety of milrinone in preventing low cardiac output syndrome in infants and children after corrective surgery for congenital heart disease. Circulation 2003;107: Azakie T, Merklinger SL, McCrindle BW, et al. Evolving strategies and improving outcomes of the modified Norwood procedure: a 10-year single-institution experience. Ann Thorac Surg 2001;72: Rossi AF, Sommer RJ, Lotvin A, et al. Usefulness of intermittent monitoring of mixed venous oxygen saturation after stage I palliation for hypoplastic left heart syndrome. Am J Cardiol 1994;73: Wernovsky G, Ghanayem N, Ohye RG, et al. Hypoplastic left heart syndrome: consensus and controversies in Cardiol Young 2007;17(Suppl 2): Marx G, Reinhart K. Venous oximetry. Curr Opin Crit Care 2006;12: Ohye RG, Sleeper LA, Mahony L, et al. Comparison of shunt types in the Norwood procedure for single-ventricle lesions. N Engl J Med 2010;362: Aharon AS, Drinkwater DC Jr, Churchwell KB, et al. Extracorporeal membrane oxygenation in children after repair of congenital cardiac lesions. Ann Thorac Surg 2001;72: Allan CK, Thiagarajan RR, del Nido PJ, Roth SJ, Almodovar MC, Laussen PC. Indication for initiation of mechanical circulatory support impacts survival of infants with shunted single-ventricle circulation supported with extracorporeal membrane oxygenation. J Thorac Cardiovasc Surg 2007;133: Debrunner MG, Porayette P, Breinholt JP 3rd, Turrentine MW, Cordes TM. Midterm survival of infants requiring postoperative extracorporeal membrane oxygenation after Norwood palliation. Pediatr Cardiol 2012 Sept 25 [Epub ahead of print]. 18. Owens GE, King K, Gurney JG, Charpie JR. Low renal oximetry correlates with acute kidney injury after infant cardiac surgery. Pediatr Cardiol 2011;32: Kaufman J, Almodovar MC, Zuk J, Friesen RH. Correlation of abdominal site near-infrared spectroscopy with gastric tonometry in infants following surgery for congenital heart disease. Pediatr Crit Care Med 2008;9: Phelps HM, Mahle WT, Kim D, et al. Postoperative cerebral oxygenation in hypoplastic left heart syndrome after the Norwood procedure. Ann Thorac Surg 2009;87: Ranucci M, Isgro G, de la Torre T, Romitti F, Conti D, Carlucci C. Near-infrared spectroscopy correlates with continuous superior vena cava oxygen saturation in pediatric cardiac surgery patients. Paediatr Anaesth 2008;18: Ortmann LA, Fontenot EE, Seib PM, Eble BK, Brown R, Bhutta AT. Use of near-infrared spectroscopy for estimation of renal oxygenation in children with heart disease. Pediatr Cardiol 2001;32: Nagdyman N, Ewert P, Peters B, Miera O, Fleck T, Berger F. Comparison of different near-infrared spectroscopic cerebral oxygenation indices with central venous and jugular venous oxygenation saturation in children. Paediatr Anaesth 2008;18: Manlhoit C, Menjak IB, Brand~ao LR, et al. Risk, clinical features, and outcomes of thrombosis associated with pediatric cardiac surgery. Circulation 2011;124: Borasino S, El Masri K, Diaz F, Smith KS, Dabal RJ, Alten JA. Central venous line location is a risk factor for chylothorax in infants after cardiac surgery. Abstract presented at the 9th Annual Conference of the Pediatric Cardiac Intensive Care Society Miami Beach, FL, December Alten JA, Borasino S, Gurley WQ, Law MA, Toms R, Dabal RJ. Ultrasound-guided femoral vein catheterization in neonates with cardiac disease. Pediatr Crit Care Med 2012;13: Patel A, Hickey E, Mavroudis C, et al. Impact of noncardiac congenital and genetic abnormalities on outcomes in hypoplastic left heart syndrome. Ann Thorac Surg 2010;89: DISCUSSION DR BRIAN KOGON (Atlanta, GA): Did you measure the inferior vena cava SvO 2 (venous oxygen saturation) before surgery with a local line or by some other means to identify patients going into surgery in a suboptimal condition? DR DABAL: We did not. I think there certainly is the potential for that. If you look at the kids in the group that were on preoperative inotropes, those patients all ended up in the group with adverse events. So that would suggest that probably if we had looked at it preoperatively, it probably was low before we even got started. DR ERLE AUSTIN (Louisville, KY): That was a nice presentation. Of course it s good to be able to predict an adverse event. Have you developed an algorithm? If the number you get happens to be less than 45, what is the algorithm? What are you going to do to try and make it better? DR DABAL: Well, that s obviously the bottom line of the study and that s whatwehavebeentryingtodo.ithinkthat you can start with standard medical management. We haven t employed ECMO (extracorporeal membrane oxygenation)

8 Ann Thorac Surg DABAL ET AL 2013;95: IVC OXIMETRY MONITORING AFTER NORWOOD 2121 a lot, and I think that that probably is something that we need to be a little more aggressive about doing, employing ECMO in these kids who are coming out of the operating room and already are in the category of being at high risk of adverse events. And so optimizing medical management would be the first option, and if that doesn t succeed rapidly, then I think moving on to ECMO would be a good alternative. DR EDWARD L. BOVE (Ann Arbor, MI): I enjoyed your presentation. Can I ask you a little bit about pulse pressure. I m sure you said it and I missed it. What kind of shunts were these? Were they RV-PA (right ventricle to pulmonary artery) shunts? DR DABAL: They were. DR BOVE: And rather than looking at pulse pressure only, is there any more information you can glean from that data, specifically diastolic pressure itself? DR DABAL: I don t know that we specifically looked at diastolic pressure. We spent a lot of time looking at pulse pressure, and it s pretty clear that the kids who come out of the operating room with a pulse pressure greater than about 20 do significantly better than those patients who have a pulse pressure less than 20 when we leave the operating room. With regard to absolute diastolic pressures, we didn t analyze those results.