- Email: [email protected]
Postoperative Cerebral Oxygenation in Hypoplastic Left Heart Syndrome After the Norwood Procedure
PEDIATRIC CARDIAC
Postoperative Cerebral Oxygenation in Hypoplastic Left Heart Syndrome After the Norwood Procedure Heather M. Phelps, DO, William T. Mahle, MD, Dennis Kim, MD, PhD, Janet M. Simsic, MD, Paul M. Kirshbom, MD, Kirk R. Kanter, MD, and Kevin O. Maher, MD Sibley Heart Center Cardiology, Department of Pediatrics, and Children’s Healthcare of Atlanta, Department of Cardiothoracic Surgery, Emory University School of Medicine, Atlanta, Georgia
Background. Cerebral near-infrared spectroscopy (NIRS) is being used with increasing frequency in the care of pediatric patients after surgery for congenital heart disease. Near-infrared spectroscopy provides a means of evaluating regional cerebral oxygen saturation (cSaO2) noninvasively, with correlations to cardiac output and central venous saturation. Prior studies have demonstrated that systemic venous saturation can predict outcome after the Norwood procedure. With this in mind, we sought to determine whether regional cSaO2 by NIRS technology could predict risk of adverse outcome after the Norwood procedure. Methods. We reviewed the first 48 hours of postoperative hemodynamic data on 50 patients with hypoplastic left heart syndrome at our institution who underwent the Norwood procedure. Cerebral oxygen saturation data within 48 hours of surgery were analyzed for association with subsequent adverse outcome, which was defined as intensive care unit length of stay greater than 30 days, need for extracorporeal membrane oxygenation, or hospital death after 48 hours.
Results. There were 18 adverse events among the 50 subjects. The mean cSaO2 for the entire cohort at 1 hour, 4 hours, and 48 hours after surgery was 51% ⴞ 7.5%, 50% ⴞ 9.4%, and 59% ⴞ 8.1%, respectively. Mean cSaO2 for the first 48 postoperative hours of less than 56% was a risk factor for subsequent adverse outcome (odds ratio 11.9, 95% confidence interval: 2.5 to 55.8). Mean cerebral NIRs of less than 56% over the first 48 hours after surgery yielded a sensitivity of 75.0% and a specificity of 79.4% to predict those at risk for subsequent adverse events. Conclusions. Low regional cerebral oxygen saturation by NIRS in the first 48 hours after the Norwood procedure has a strong association with subsequent adverse outcome. Monitoring of cerebral saturation can serve as a valuable monitoring tool and can identify patients at risk for poor outcome.
N
between cSaO 2 and outcomes after the Norwood procedure.
ear-infrared spectroscopy (NIRS) was first used more than 30 years ago for evaluation of cerebral oxygen saturation (cSaO2) [1]. Light in the near-infrared range (700 to 1100 nm) is capable of penetrating further when compared with visible light, allowing for the evaluation of deeper tissue. Cerebral NIRS monitoring is a method of evaluating cSaO2 that has become increasingly common in the preoperative, intraoperative, and postoperative management of patients with complex congenital heart disease [2–5]. Although this type of monitoring is frequently used, it is unclear whether changes in cSaO2 have a relationship to postoperative outcome. At our institution, cSaO2 monitoring with NIRS is routinely used in the preoperative and postoperative management of newborns with hypoplastic left heart syndrome. The relationship between NIRS data and subsequent outcome after complex newborn surgical procedures such as the Norwood procedure is poorly understood. We sought to investigate the relationship Accepted for publication Jan 30, 2009. Address correspondence to Dr Maher, Sibley Heart Center Cardiology, McGill Building, 2835 Brandywine, Suite 300, Atlanta, GA 30341; e-mail: [email protected].
© 2009 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2009;87:1490 – 4) © 2009 by The Society of Thoracic Surgeons
Patients and Methods Institutional Review Board approval was obtained for this study, including a waiver of parental consent for study participation. We retrospectively reviewed 50 consecutive newborns with hypoplastic left heart syndrome at Children’s Healthcare of Atlanta who underwent the Norwood procedure between February 2005 and March 2007. There were patients in our study who were also enrolled in the National Institutes of Health–funded multicenter randomized Single Ventricle Reconstruction Trial. These patients were randomly assigned to a Blalock-Taussig shunt or a right ventricle to pulmonary artery conduit for pulmonary blood supply as part of the Norwood procedure. Data were obtained from the surgical database, hospital charts, and nursing flow sheets. Cerebral saturation was measured using Somanetics INVOS probes and monitors (INVOS Somanetics, Troy, MI). The probe transmits light at wavelengths of 730 and 810 nm, providing a continuous representation of re0003-4975/09/$36.00 doi:10.1016/j.athoracsur.2009.01.071
PHELPS ET AL POSTOPERATIVE NORWOOD CEREBRAL OXYGENATION
1491
Fig 1. Cerebral saturation as measured by near-infrared spectroscopy (NIRS) monitoring versus postoperative hour after the Norwood procedure. Graph represents data from all 50 patients (good outcome and adverse outcome).
gional hemoglobin saturation by evaluating the absorption and reflection characteristics of hemoglobin in the cerebral cortex. At our institution, probes are placed on both sides of the forehead and results are displayed on bedside NIRS monitors. Cerebral saturation was recorded every 30 minutes by the bedside nurse. The first 48 postoperative hours of cSaO2 data were reviewed. Additionally, lactic acid, mean blood pressure, oxygen saturation, heart rate, intensive care unit time, and ventilatory time were recorded. Cerebral SaO2 was analyzed for association with subsequent adverse outcome. Adverse outcome was defined as hospital death, need for extracorporeal membrane oxygenation, or cardiac intensive care unit length of stay greater than 30 days. Review of cSaO2 results from left- versus right-sided NIRS probes was completed, as was comparison of cSaO2 results based on the source of pulmonary blood flow.
Operative Procedure After initiation of general anesthesia and cardiopulmonary bypass, 0.2 mg/kg phentolamine is administered. Regional arterial inflow is provided through a polytetrafluoroethylene (PTFE) shunt anastomosed to the innominate artery and a second cannula in the patent ductus arteriosus at the surgeon’s discretion. The patient is cooled to a core temperature of 18°C over a minimum of 20 minutes using pH stat blood gas management. After cardioplegic arrest, selective cerebral perfusion is maintained through the innominate artery shunt at 25 to 30 mL · kg⫺1 · min⫺1 during the aortic reconstruction, which is typically performed with a pulmonary homograft
patch. An atrial septectomy is performed, and pulmonary blood flow is established with a 5-mm to 6-mm ringed PTFE right ventricle to pulmonary artery (RV-PA) conduit leading to the pulmonary bifurcation or with a 3.5 mm PTFE modified Blalock-Taussig (BT) shunt. Choice of source of pulmonary blood flow is made based on randomization due to participation in the the National Institutes of Health–funded multicenter randomized Single Ventricle Reconstruction Trial or at the discretion of the surgeon. Sternal closure is performed in the operating room at the discretion of the surgeon at the end of the case.
Postoperative Management Patients are admitted to a pediatric cardiac intensive care unit. Inotropic support is typically milrinone (0.5 to 1.0 g · kg⫺1 · min⫺1), and low-dose dopamine is used. Volume ventilation is adjusted to produce pCO2 in the 35 to 45 range: FiO2 is 0.21 to 0.5 for the majority of patients, keeping systemic saturations 75% to 85%. The hematocrit is maintained between 40% and 50%. All patients were on mechanical ventilation during the first 48 hours of postoperative care. Patients do not routinely receive neuromuscular blockade.
Statistics Data are presented as mean and standard deviation or median and range, where appropriate. Low cSaO2 was defined as any measurement of less than 56%. Statistical analysis was completed using Fisher’s exact test, Student’s t test, and Pearson’s correlation coefficient. StatisFig 2. Cerebral saturation by near-infrared spectroscopy (NIRS) separated by good and adverse outcome groups versus postoperative hours. Patients with good outcomes (dotted line) had a higher mean cSaO2 during the first 48 hours postoperatively than patients with adverse outcomes (dashed line).
PEDIATRIC CARDIAC
Ann Thorac Surg 2009;87:1490 – 4
1492
PHELPS ET AL POSTOPERATIVE NORWOOD CEREBRAL OXYGENATION
PEDIATRIC CARDIAC
tical significance was defined as p less than 0.05. Analysis was performed with STATA 6.0 (STATA Corp, College Station, TX). A general linear model repeated-measures analysis of variance was used to compare cerebral NIRs values between patients with and without subsequent adverse events. We used receiver operating characteristic (ROC) curve analysis to identify optimal cutpoint that maximized the probability of correctly identifying those at greatest risk of decompensation, using dichotomous adverse event as the classifier. The ROC analysis is a graphical and quantitative technique which, for a given continuous criterion variable X, can determine an optimal cutpoint c for a classified decision. Here, the criterion variable X is mean cerebral NIRs in the first 48 hours after surgery. We used a bootstrap approach to build confidence intervals for optimizing c and its corresponding optimal sensitivity and specificity. A greater degree of separation between the distributions of those with and without adverse events will result in a higher area under the ROC curve. Paradoxically, this can result in a wider confidence interval for the cutpoint because of the flatness of the objective function. For the form of our rule, increasing the cutpoint cSaO2 will increase sensitivity and decrease specificity for identifying those with adverse events.
Results Anatomic pathology of the patient population is as follows: aortic atresia/mitral atresia (27), aortic atresia/ mitral stenosis (13), aortic stenosis/mitral stenosis (7), aortic stenosis/mitral stenosis with ventricular septal defect (1), severe mitral stenosis (1), and unbalanced atrioventricular canal with coarctation (1). The mean length of stay in the intensive care unit for the entire cohort was 17.3 days (4.7 to 67.3); mean ventilatory time was 9.9 days (2.2 to 47.9). A total of 18 adverse events occurred in 11 patients. The overall hospital survival was 88%. There were 6 deaths, 4 patients who required initiation of extracorporeal membrane oxygenation and 8 patients who stayed in the cardiac intensive care unit for more than 30 days. The mean cSaO2 for the entire cohort was 51%, 50%, 63%, and 59% at postoperative hours 1, 4, 24, and 48, respectively. Figure 1 demonstrates the mean Table 1. Patient Characteristics Mean (range) weight at surgery Mean (range) age at surgery Pulmonary artery blood source BT shunt RV to PA conduit Open sternum Mean lactate (range), mmol/L Mean intensive care unit days (range) Mean ventilatory days (range) BT ⫽ Blalock-Taussig; ventricle.
3.157 kg (2–4.1) 5.4 days (2–21)
PA ⫽ pulmonary artery;
9/50 41/50 29/50 3.53 (1.15–8.23) 17.3 (4.7–67.3) 9.9 (2.2–47.9) RV ⫽ right
Ann Thorac Surg 2009;87:1490 – 4
Table 2. Mean Cerebral Saturation by Near-Infrared Spectroscopy Measurementa Postoperative hours 1 hour 4 hours 48 hours Type of aortopulmonary connection Blalock-Taussig shunt RV-PA conduit Outcome Good outcome Adverse outcome
51 ⫾ 7.5 50 ⫾ 9.4 59 ⫾ 8.1 54.58 ⫾ 4.11 59.98 ⫾ 7.84 60.77 ⫾ 5.91 52.75 ⫾ 9.93
a Cerebral saturation data, mean ⫾ SD, by postoperative time, source of pulmonary blood flow, and by postoperative outcome, good versus adverse.
PA ⫽ pulmonary artery;
RV ⫽ right ventricle.
cSaO2 for all patients over the first 48 postoperative hours. There was a statistically significant difference between the mean cSaO2 in the adverse outcome group and the rest of the sample. The adverse outcome group had a mean cSaO2 over the first 48 postoperative hours of 52.8% compared with 60.8% in the nonadverse outcome cohort (p ⬍ 0.001). Figure 2 demonstrates the mean cSaO2 over the first 48 hours for the two cohorts (adverse outcomes versus no adverse outcomes). Low cerebral NIRs over the first 48 hours was associated with subsequent adverse events (p ⬍ 0.01) Mean cerebral NIRs of less than 56% over the first 48 hours after surgery yielded a sensitivity of 75.0% and a specificity of 79.4% to predict those at risk for subsequent adverse events. There was a trend toward higher mean lactic acid associated with subsequent adverse outcome, (p ⫽ 0.068); however, there was no correlation found between lactic acid and cSaO2 in the sample (r ⫽ – 0.315) or in the adverse (r ⫽ – 0.22) or nonadverse (r ⫽ – 0.37) groups. There was no statistically-significant difference in cSaO2 between patients who had delayed sternal closure and those who had primary sternal closure (p ⫽ 0.10). During the study period, patients underwent the Norwood procedure utilizing either a BT shunt or a RV-PA conduit for pulmonary arterial blood supply. In this cohort, 9 patients had BT shunts and 41 had RV-PA conduits (Table 1). There was a trend toward lower cerebral saturation in patients with a BT shunt (54.6%) as compared with those with a RV-PA conduit (60%; p ⫽ 0.054). A comparison of cSaO2 from the right versus left NIRS probes from all patients revealed no significant difference of left versus right cSaO2: mean left-sided cSaO2 was 58.9%, mean right-sided cSaO2 was 56.6% (p ⫽ 0.18). For the 9 patients who had a right BT shunt Norwood, there was also no difference between left- and right-sided cSaO2.: mean right-sided cSaO2 was 53.7%, mean left-sided cSaO2 was 54.0% (p ⫽ 0.27). Cerebral saturation data are summarized in Table 2.
Comment In 1980, Norwood and colleagues [6] first reported their experience with first-stage palliation for hypoplastic left heart syndrome. Ongoing advances in the medical and surgical management of these patients have resulted in a dramatic improvement in survival. Despite these improvements, stage 1 palliation for hypoplastic left heart syndrome continues to carry a persistent morbidity and mortality associated with this disease [7–9]. The ability of caregivers to recognize patients at risk for excessive morbidity and mortality in this patient population is compelling, allowing for potential interventions and improved outcomes. Tweddell and colleagues [10, 11] have advocated continuous monitoring of systemic venous saturation (SvO2) with indwelling catheters as part of the postoperative management of the Norwood procedure. These investigators demonstrated that SvO2 was strongly correlated with clinical outcome and may be predictive of later neurodevelopmental status [10, 11]. Complications were predicted by higher systemic oxygen saturation, lower heart rate, higher inspired oxygen requirement, lower pH, younger age, and lower SvO2. However, SvO2 was the only marker that was associated with mortality. In a study by Bhutta and coworkers [12], NIRS estimation of cSaO2 correlated well with mixed venous oxygenation in children with biventricular anatomy. In a recent study by Kirshbom and colleagues [13] of our program, NIRS was compared with superior vena cava saturation obtained during catheterization of single-ventricle patients. They demonstrated that cSaO2 is closely correlated with superior vena cava saturation in patients with this type of anatomy [13]. As SvO2 has been shown to correlate with survival as well as with the NIRS estimate of cSaO2, one might predict that cSaO2 would also be predictive of outcome from the Norwood procedure. The NIRS estimation of cSaO2 has the advantage of being noninvasive; it can be used in both the inpatient and outpatient setting, and will not put venous vessels at risk of thrombosis. In addition, no known infectious complications are associated with its use. Patients with a BT shunt Norwood tended to have a lower mean cSaO2 as compared with the RV-PA conduit patients, with a p value of 0.054. Although the small sample size (only 9 patients in the BT shunt group) limits the power of this study to clearly delineate the association between source of pulmonary blood flow and postoperative cSaO2 in Norwood patients, this is an interesting trend that merits further study. The reasons for a lower cSaO2 may be explained by known differences in pulmonary and systemic blood flow (Qp and Qs) for these patients. Maher and coworkers [14] and others [15, 16] have previously demonstrated a higher ratio of pulmonary to systemic blood flow for the BT shunt Norwood in comparison with the RV- PA conduit patients. This increase in systemic blood flow for the RV-PA conduit patients may account for the increased cSaO2. Several types of monitoring are performed in the neonate after the Norwood procedure to assess overall
PHELPS ET AL POSTOPERATIVE NORWOOD CEREBRAL OXYGENATION
1493
stability and cardiovascular hemodynamics. These include sVO2, cSaO2, oxygen saturation, arterial pressure, intracardiac pressures, lactic acid level, arterial blood gas, and base deficit. Li and colleagues [4] recently demonstrated the strong correlation of cSaO2 with systolic blood pressure, systemic saturation, systemic blood flow and oxygen delivery. Many of the determinants of cSaO2 for these patients are individually followed postoperatively after the Norwood procedure. Determination of cSaO2 by NIRS may provide an overall assessment of cardiovascular status of these patients, with the result being based on a number of important physiologic variables. Perhaps more importantly is the ability to have a continuous evaluation of cSaO2, allowing for the monitoring of trends in cSaO2. This type of monitoring can also alert the clinician to additional causes of cardiovascular compromise. In one case report, when a pericardial effusion developed several days after the Norwood procedure, the cSaO2 dropped significantly while the other noninvasive and invasive measures of cardiac hemodynamic status remained stable, leading to echocardiography and diagnosis of pericardial effusion [17]. A lower cerebral saturation in the early postoperative period is associated with a higher risk for adverse events. Knowing the hourly normative values can assist in the continuous postoperative evaluation of these patients. Data from the current study as well as prior research demonstrate that cSaO2 is typically in the low 50s upon arrival from the operating room and gradually rises into the high 50s and low 60s in the first 2 postoperative days [4]. A persistently low cSaO2 or a decrease in the cSaO2 should cause further evaluation and intervention for these patients. The ability to observe these changes early and monitor continuously may improve outcomes for these high risk patients in a similar fashion as SvO2 measurement. The extent to which caregivers can improve or modify a low cSaO2 is not well understood. Interventions such as transfusion of red blood cells, addition of inotropic support, or sedation are strategies that may increase cSaO2 in this patient population.
Study Limitations The retrospective nature of this investigation and relatively small sample population may limit generalizations of these findings. Further research is required to determine whether changes in operative technique or postoperative management aimed at increasing cSaO2 will improve outcome for patients undergoing the Norwood procedure. In conclusion, low cerebral oxygen saturation as measured by NIRS is predictive of subsequent adverse outcome in patients with hypoplastic left heart syndrome after the Norwood procedure. Monitoring of the cerebral saturation after the Norwood procedure provides additional hemodynamic information that can be used as a part of the postoperative management for these patients.
PEDIATRIC CARDIAC
Ann Thorac Surg 2009;87:1490 – 4
1494
PHELPS ET AL POSTOPERATIVE NORWOOD CEREBRAL OXYGENATION
Ann Thorac Surg 2009;87:1490 – 4
References
PEDIATRIC CARDIAC
1. Jobsis FF. Noninvasive, infrared monitoring of cerebral and myocardial oxygen sufficiency and circulatory parameters. Science 1977;198:1264 –7. 2. Hayashida M, Kin N, Tomioka T, et al. Cerebral ischaemia during cardiac surgery in children detected by combined monitoring of BIS and near-infrared spectroscopy. Br J Anaesth 2004;92:662–9. 3. Nollert G, Jonas R, Reichart B. Optimizing cerebral oxygenation during cardiac surgery: a review of experimental and clinical investigations with near infrared spectrophotometry. Thorac Cardiovasc Surg 2000;48:247–53. 4. Li J, Van Arsdell G, Zhang G, et al. Assessment of the relationship between cerebral and splanchnic oxygen saturations measured by near-infrared spectroscopy and direct measurements of systemic haemodynamic variables and oxygen transport after the Norwood procedure. Heart 2006;92:1678 – 85. 5. Hoffman G, Stuth E, Jaquiss R, et al. Changes in cerebral and somatic oxygenation during stage 1 palliation of hypoplastic left heart syndrome using continuous regional cerebral perfusion. J Thorac Cardiovasc Surg 2004;127:223–33. 6. Norwood WI, Kirklin JK, Sanders SP. Hypoplastic left heart syndrome: experience with palliative surgery. Am J Cardiol 1980;45:87–91. 7. Jacobs JP, O’Brien SM, Chai PJ, et al. Management of 239 patients with hypoplastic left heart syndrome and related malformations from 1993 to 2007. Ann Thorac Surg 2008;85: 1691–97. 8. Nillson B, Mellander M, Sudow G, Berggren H. Results of staged palliation for hypoplastic left heart syndrome: a complete population-based series. Acta Pediatrica 2006;95: 594 – 600. 9. Tibbals J, Kawahira Y, Carter BG, Donath S, Brizard C, Wilkonson J. Outcomes of surgical treatment of infants with
10.
11.
12. 13.
14.
15.
16.
17.
hypoplastic left heart syndrome: an institutional experience 1983-2004. J Paediatr Child Health 2007;43:746 –51. 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; 84:1301–11. Hoffman GM, Mussatto KA, Brosig CL, et al. Systemic venous oxygen saturation after the Norwood procedure and childhood neurodevelopmental outcome. J Thorac Cardiovasc Surg 2005;130:1094 –100. Bhutta AT, Ford JW, Parker JG, et al. Noninvasive cerebral oximeter as a surrogate for mixed venous saturation in children. Pediatr Cardiol 2007;28:34 – 41. Kirshbom PM, Forbess JM, Kogon BE, et al. Cerebral near infrared spectroscopy is a reliable marker of systemic perfusion in awake single ventricle children. Pediatr Cardiol 2007;28:42–5. Maher KO, Pizarro C, Gidding SS, Malec E, Norwood WI, Murphy JD. Hemodynamic profile after the Norwood procedure with right ventricle to pulmonary artery conduit. Circulation 2003;108:782– 4. Bradley SM, Simsic JM, McQuinn TC, Habib DM, Shirali GS, Atz AM. Hemodynamic status after the Norwood procedure: a comparison of right ventricle to pulmonary artery connection versus modified Blalock-Taussig shunt. Ann Thorac Surg 2004;78:933– 41. Ballweg JA, Dominguez TE, Ravishankar C, et al. A contemporary comparison of the effect of shunt type in hypoplastic left heart syndrome on the hemodynamics and outcome at stage 2 reconstruction. J Thorac Cardiovasc Surg 2007;134: 297–303. Maher KO, Phelps HM, Kirshbom PM. Near infrared spectroscopy changes with impeding pericardial tamponade. Pediatr Crit Care Med 2009;10:e13–5.
INVITED COMMENTARY The article by Phelps and colleagues [1] adds to a growing body of literature regarding the clinical application of cerebral oximetric monitoring to the perioperative care of both pediatric and adult cardiac surgical patients. For a monitoring technology to be useful, it must be accurate and reproducible, and it should prompt correlated interventions that impact clinical outcome. Debate continues over whether current cerebral oximetry technology satisfies these requirements. Experience with devices that render absolute values for cerebral oxygen content is limited, and relative values can be affected by various factors, such as cerebral blood volume, hematocrit, acid-base status, and cranial anatomy. Areas of the brain remote from the monitored frontal cortex are at risk of undetected injury, potentially limiting the usefulness of the technology in avoiding significant central nervous system insult. There is significant inter-patient variability of rSO2 values, and a lack of a well-defined normative range. This is particularly true among children with congenital heart disease in which arterial desaturation and parallel circulatory arrangements further complicate our understanding of monitored values. These limitations notwithstanding, the association between lower mean rSO2 values and adverse events after Norwood palliation described by the authors are consistent with a large number of prior reports that impute clinical utility to continuous cerebral oximetry monitoring. In light of our increasing awareness of perioperative cerebral injury © 2009 by The Society of Thoracic Surgeons Published by Elsevier Inc
and its subsequent neuro-developmental impact, mechanisms to accurately surveil the cerebral environment are needed. Systemic measures of oxygen delivery such as SvO2 or lactate are poor predictors of cerebral insult. The clinical benefit of cerebral oximetric monitoring is intuitive, and the supporting evidence is compelling. Although weaknesses of the present design and understanding of the technology require the clinician to place rSO2 values in proper context, further design improvements and clinical understanding of its role represent a significant opportunity to improve care and outcomes. David M. Overman, MD Division of Cardiac Surgery The Children’s Heart Clinic Children’s Hospitals and Clinics of Minnesota 2545 Chicago Ave, South Suite 106 Minneapolis, MN 55404 e-mail: [email protected]
Reference 1. Phelps HM, Mahle WT, Kim D, et al. Postoperative cerebral oxygenation in hypoplastic left heart syndrome after the Norwood procedure. Ann Thorac Surg 2009;87:1490 – 4.
0003-4975/09/$36.00 doi:10.1016/j.athoracsur.2009.03.015
Postoperative Cerebral Oxygenation in Hypoplastic Left Heart Syndrome After the Norwood Procedure Heather M. Phelps, DO, William T. Mahle, MD, Dennis Kim, MD, PhD, Janet M. Simsic, MD, Paul M. Kirshbom, MD, Kirk R. Kanter, MD, and Kevin O. Maher, MD Sibley Heart Center Cardiology, Department of Pediatrics, and Children’s Healthcare of Atlanta, Department of Cardiothoracic Surgery, Emory University School of Medicine, Atlanta, Georgia
Background. Cerebral near-infrared spectroscopy (NIRS) is being used with increasing frequency in the care of pediatric patients after surgery for congenital heart disease. Near-infrared spectroscopy provides a means of evaluating regional cerebral oxygen saturation (cSaO2) noninvasively, with correlations to cardiac output and central venous saturation. Prior studies have demonstrated that systemic venous saturation can predict outcome after the Norwood procedure. With this in mind, we sought to determine whether regional cSaO2 by NIRS technology could predict risk of adverse outcome after the Norwood procedure. Methods. We reviewed the first 48 hours of postoperative hemodynamic data on 50 patients with hypoplastic left heart syndrome at our institution who underwent the Norwood procedure. Cerebral oxygen saturation data within 48 hours of surgery were analyzed for association with subsequent adverse outcome, which was defined as intensive care unit length of stay greater than 30 days, need for extracorporeal membrane oxygenation, or hospital death after 48 hours.
Results. There were 18 adverse events among the 50 subjects. The mean cSaO2 for the entire cohort at 1 hour, 4 hours, and 48 hours after surgery was 51% ⴞ 7.5%, 50% ⴞ 9.4%, and 59% ⴞ 8.1%, respectively. Mean cSaO2 for the first 48 postoperative hours of less than 56% was a risk factor for subsequent adverse outcome (odds ratio 11.9, 95% confidence interval: 2.5 to 55.8). Mean cerebral NIRs of less than 56% over the first 48 hours after surgery yielded a sensitivity of 75.0% and a specificity of 79.4% to predict those at risk for subsequent adverse events. Conclusions. Low regional cerebral oxygen saturation by NIRS in the first 48 hours after the Norwood procedure has a strong association with subsequent adverse outcome. Monitoring of cerebral saturation can serve as a valuable monitoring tool and can identify patients at risk for poor outcome.
N
between cSaO 2 and outcomes after the Norwood procedure.
ear-infrared spectroscopy (NIRS) was first used more than 30 years ago for evaluation of cerebral oxygen saturation (cSaO2) [1]. Light in the near-infrared range (700 to 1100 nm) is capable of penetrating further when compared with visible light, allowing for the evaluation of deeper tissue. Cerebral NIRS monitoring is a method of evaluating cSaO2 that has become increasingly common in the preoperative, intraoperative, and postoperative management of patients with complex congenital heart disease [2–5]. Although this type of monitoring is frequently used, it is unclear whether changes in cSaO2 have a relationship to postoperative outcome. At our institution, cSaO2 monitoring with NIRS is routinely used in the preoperative and postoperative management of newborns with hypoplastic left heart syndrome. The relationship between NIRS data and subsequent outcome after complex newborn surgical procedures such as the Norwood procedure is poorly understood. We sought to investigate the relationship Accepted for publication Jan 30, 2009. Address correspondence to Dr Maher, Sibley Heart Center Cardiology, McGill Building, 2835 Brandywine, Suite 300, Atlanta, GA 30341; e-mail: [email protected].
© 2009 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2009;87:1490 – 4) © 2009 by The Society of Thoracic Surgeons
Patients and Methods Institutional Review Board approval was obtained for this study, including a waiver of parental consent for study participation. We retrospectively reviewed 50 consecutive newborns with hypoplastic left heart syndrome at Children’s Healthcare of Atlanta who underwent the Norwood procedure between February 2005 and March 2007. There were patients in our study who were also enrolled in the National Institutes of Health–funded multicenter randomized Single Ventricle Reconstruction Trial. These patients were randomly assigned to a Blalock-Taussig shunt or a right ventricle to pulmonary artery conduit for pulmonary blood supply as part of the Norwood procedure. Data were obtained from the surgical database, hospital charts, and nursing flow sheets. Cerebral saturation was measured using Somanetics INVOS probes and monitors (INVOS Somanetics, Troy, MI). The probe transmits light at wavelengths of 730 and 810 nm, providing a continuous representation of re0003-4975/09/$36.00 doi:10.1016/j.athoracsur.2009.01.071
PHELPS ET AL POSTOPERATIVE NORWOOD CEREBRAL OXYGENATION
1491
Fig 1. Cerebral saturation as measured by near-infrared spectroscopy (NIRS) monitoring versus postoperative hour after the Norwood procedure. Graph represents data from all 50 patients (good outcome and adverse outcome).
gional hemoglobin saturation by evaluating the absorption and reflection characteristics of hemoglobin in the cerebral cortex. At our institution, probes are placed on both sides of the forehead and results are displayed on bedside NIRS monitors. Cerebral saturation was recorded every 30 minutes by the bedside nurse. The first 48 postoperative hours of cSaO2 data were reviewed. Additionally, lactic acid, mean blood pressure, oxygen saturation, heart rate, intensive care unit time, and ventilatory time were recorded. Cerebral SaO2 was analyzed for association with subsequent adverse outcome. Adverse outcome was defined as hospital death, need for extracorporeal membrane oxygenation, or cardiac intensive care unit length of stay greater than 30 days. Review of cSaO2 results from left- versus right-sided NIRS probes was completed, as was comparison of cSaO2 results based on the source of pulmonary blood flow.
Operative Procedure After initiation of general anesthesia and cardiopulmonary bypass, 0.2 mg/kg phentolamine is administered. Regional arterial inflow is provided through a polytetrafluoroethylene (PTFE) shunt anastomosed to the innominate artery and a second cannula in the patent ductus arteriosus at the surgeon’s discretion. The patient is cooled to a core temperature of 18°C over a minimum of 20 minutes using pH stat blood gas management. After cardioplegic arrest, selective cerebral perfusion is maintained through the innominate artery shunt at 25 to 30 mL · kg⫺1 · min⫺1 during the aortic reconstruction, which is typically performed with a pulmonary homograft
patch. An atrial septectomy is performed, and pulmonary blood flow is established with a 5-mm to 6-mm ringed PTFE right ventricle to pulmonary artery (RV-PA) conduit leading to the pulmonary bifurcation or with a 3.5 mm PTFE modified Blalock-Taussig (BT) shunt. Choice of source of pulmonary blood flow is made based on randomization due to participation in the the National Institutes of Health–funded multicenter randomized Single Ventricle Reconstruction Trial or at the discretion of the surgeon. Sternal closure is performed in the operating room at the discretion of the surgeon at the end of the case.
Postoperative Management Patients are admitted to a pediatric cardiac intensive care unit. Inotropic support is typically milrinone (0.5 to 1.0 g · kg⫺1 · min⫺1), and low-dose dopamine is used. Volume ventilation is adjusted to produce pCO2 in the 35 to 45 range: FiO2 is 0.21 to 0.5 for the majority of patients, keeping systemic saturations 75% to 85%. The hematocrit is maintained between 40% and 50%. All patients were on mechanical ventilation during the first 48 hours of postoperative care. Patients do not routinely receive neuromuscular blockade.
Statistics Data are presented as mean and standard deviation or median and range, where appropriate. Low cSaO2 was defined as any measurement of less than 56%. Statistical analysis was completed using Fisher’s exact test, Student’s t test, and Pearson’s correlation coefficient. StatisFig 2. Cerebral saturation by near-infrared spectroscopy (NIRS) separated by good and adverse outcome groups versus postoperative hours. Patients with good outcomes (dotted line) had a higher mean cSaO2 during the first 48 hours postoperatively than patients with adverse outcomes (dashed line).
PEDIATRIC CARDIAC
Ann Thorac Surg 2009;87:1490 – 4
1492
PHELPS ET AL POSTOPERATIVE NORWOOD CEREBRAL OXYGENATION
PEDIATRIC CARDIAC
tical significance was defined as p less than 0.05. Analysis was performed with STATA 6.0 (STATA Corp, College Station, TX). A general linear model repeated-measures analysis of variance was used to compare cerebral NIRs values between patients with and without subsequent adverse events. We used receiver operating characteristic (ROC) curve analysis to identify optimal cutpoint that maximized the probability of correctly identifying those at greatest risk of decompensation, using dichotomous adverse event as the classifier. The ROC analysis is a graphical and quantitative technique which, for a given continuous criterion variable X, can determine an optimal cutpoint c for a classified decision. Here, the criterion variable X is mean cerebral NIRs in the first 48 hours after surgery. We used a bootstrap approach to build confidence intervals for optimizing c and its corresponding optimal sensitivity and specificity. A greater degree of separation between the distributions of those with and without adverse events will result in a higher area under the ROC curve. Paradoxically, this can result in a wider confidence interval for the cutpoint because of the flatness of the objective function. For the form of our rule, increasing the cutpoint cSaO2 will increase sensitivity and decrease specificity for identifying those with adverse events.
Results Anatomic pathology of the patient population is as follows: aortic atresia/mitral atresia (27), aortic atresia/ mitral stenosis (13), aortic stenosis/mitral stenosis (7), aortic stenosis/mitral stenosis with ventricular septal defect (1), severe mitral stenosis (1), and unbalanced atrioventricular canal with coarctation (1). The mean length of stay in the intensive care unit for the entire cohort was 17.3 days (4.7 to 67.3); mean ventilatory time was 9.9 days (2.2 to 47.9). A total of 18 adverse events occurred in 11 patients. The overall hospital survival was 88%. There were 6 deaths, 4 patients who required initiation of extracorporeal membrane oxygenation and 8 patients who stayed in the cardiac intensive care unit for more than 30 days. The mean cSaO2 for the entire cohort was 51%, 50%, 63%, and 59% at postoperative hours 1, 4, 24, and 48, respectively. Figure 1 demonstrates the mean Table 1. Patient Characteristics Mean (range) weight at surgery Mean (range) age at surgery Pulmonary artery blood source BT shunt RV to PA conduit Open sternum Mean lactate (range), mmol/L Mean intensive care unit days (range) Mean ventilatory days (range) BT ⫽ Blalock-Taussig; ventricle.
3.157 kg (2–4.1) 5.4 days (2–21)
PA ⫽ pulmonary artery;
9/50 41/50 29/50 3.53 (1.15–8.23) 17.3 (4.7–67.3) 9.9 (2.2–47.9) RV ⫽ right
Ann Thorac Surg 2009;87:1490 – 4
Table 2. Mean Cerebral Saturation by Near-Infrared Spectroscopy Measurementa Postoperative hours 1 hour 4 hours 48 hours Type of aortopulmonary connection Blalock-Taussig shunt RV-PA conduit Outcome Good outcome Adverse outcome
51 ⫾ 7.5 50 ⫾ 9.4 59 ⫾ 8.1 54.58 ⫾ 4.11 59.98 ⫾ 7.84 60.77 ⫾ 5.91 52.75 ⫾ 9.93
a Cerebral saturation data, mean ⫾ SD, by postoperative time, source of pulmonary blood flow, and by postoperative outcome, good versus adverse.
PA ⫽ pulmonary artery;
RV ⫽ right ventricle.
cSaO2 for all patients over the first 48 postoperative hours. There was a statistically significant difference between the mean cSaO2 in the adverse outcome group and the rest of the sample. The adverse outcome group had a mean cSaO2 over the first 48 postoperative hours of 52.8% compared with 60.8% in the nonadverse outcome cohort (p ⬍ 0.001). Figure 2 demonstrates the mean cSaO2 over the first 48 hours for the two cohorts (adverse outcomes versus no adverse outcomes). Low cerebral NIRs over the first 48 hours was associated with subsequent adverse events (p ⬍ 0.01) Mean cerebral NIRs of less than 56% over the first 48 hours after surgery yielded a sensitivity of 75.0% and a specificity of 79.4% to predict those at risk for subsequent adverse events. There was a trend toward higher mean lactic acid associated with subsequent adverse outcome, (p ⫽ 0.068); however, there was no correlation found between lactic acid and cSaO2 in the sample (r ⫽ – 0.315) or in the adverse (r ⫽ – 0.22) or nonadverse (r ⫽ – 0.37) groups. There was no statistically-significant difference in cSaO2 between patients who had delayed sternal closure and those who had primary sternal closure (p ⫽ 0.10). During the study period, patients underwent the Norwood procedure utilizing either a BT shunt or a RV-PA conduit for pulmonary arterial blood supply. In this cohort, 9 patients had BT shunts and 41 had RV-PA conduits (Table 1). There was a trend toward lower cerebral saturation in patients with a BT shunt (54.6%) as compared with those with a RV-PA conduit (60%; p ⫽ 0.054). A comparison of cSaO2 from the right versus left NIRS probes from all patients revealed no significant difference of left versus right cSaO2: mean left-sided cSaO2 was 58.9%, mean right-sided cSaO2 was 56.6% (p ⫽ 0.18). For the 9 patients who had a right BT shunt Norwood, there was also no difference between left- and right-sided cSaO2.: mean right-sided cSaO2 was 53.7%, mean left-sided cSaO2 was 54.0% (p ⫽ 0.27). Cerebral saturation data are summarized in Table 2.
Comment In 1980, Norwood and colleagues [6] first reported their experience with first-stage palliation for hypoplastic left heart syndrome. Ongoing advances in the medical and surgical management of these patients have resulted in a dramatic improvement in survival. Despite these improvements, stage 1 palliation for hypoplastic left heart syndrome continues to carry a persistent morbidity and mortality associated with this disease [7–9]. The ability of caregivers to recognize patients at risk for excessive morbidity and mortality in this patient population is compelling, allowing for potential interventions and improved outcomes. Tweddell and colleagues [10, 11] have advocated continuous monitoring of systemic venous saturation (SvO2) with indwelling catheters as part of the postoperative management of the Norwood procedure. These investigators demonstrated that SvO2 was strongly correlated with clinical outcome and may be predictive of later neurodevelopmental status [10, 11]. Complications were predicted by higher systemic oxygen saturation, lower heart rate, higher inspired oxygen requirement, lower pH, younger age, and lower SvO2. However, SvO2 was the only marker that was associated with mortality. In a study by Bhutta and coworkers [12], NIRS estimation of cSaO2 correlated well with mixed venous oxygenation in children with biventricular anatomy. In a recent study by Kirshbom and colleagues [13] of our program, NIRS was compared with superior vena cava saturation obtained during catheterization of single-ventricle patients. They demonstrated that cSaO2 is closely correlated with superior vena cava saturation in patients with this type of anatomy [13]. As SvO2 has been shown to correlate with survival as well as with the NIRS estimate of cSaO2, one might predict that cSaO2 would also be predictive of outcome from the Norwood procedure. The NIRS estimation of cSaO2 has the advantage of being noninvasive; it can be used in both the inpatient and outpatient setting, and will not put venous vessels at risk of thrombosis. In addition, no known infectious complications are associated with its use. Patients with a BT shunt Norwood tended to have a lower mean cSaO2 as compared with the RV-PA conduit patients, with a p value of 0.054. Although the small sample size (only 9 patients in the BT shunt group) limits the power of this study to clearly delineate the association between source of pulmonary blood flow and postoperative cSaO2 in Norwood patients, this is an interesting trend that merits further study. The reasons for a lower cSaO2 may be explained by known differences in pulmonary and systemic blood flow (Qp and Qs) for these patients. Maher and coworkers [14] and others [15, 16] have previously demonstrated a higher ratio of pulmonary to systemic blood flow for the BT shunt Norwood in comparison with the RV- PA conduit patients. This increase in systemic blood flow for the RV-PA conduit patients may account for the increased cSaO2. Several types of monitoring are performed in the neonate after the Norwood procedure to assess overall
PHELPS ET AL POSTOPERATIVE NORWOOD CEREBRAL OXYGENATION
1493
stability and cardiovascular hemodynamics. These include sVO2, cSaO2, oxygen saturation, arterial pressure, intracardiac pressures, lactic acid level, arterial blood gas, and base deficit. Li and colleagues [4] recently demonstrated the strong correlation of cSaO2 with systolic blood pressure, systemic saturation, systemic blood flow and oxygen delivery. Many of the determinants of cSaO2 for these patients are individually followed postoperatively after the Norwood procedure. Determination of cSaO2 by NIRS may provide an overall assessment of cardiovascular status of these patients, with the result being based on a number of important physiologic variables. Perhaps more importantly is the ability to have a continuous evaluation of cSaO2, allowing for the monitoring of trends in cSaO2. This type of monitoring can also alert the clinician to additional causes of cardiovascular compromise. In one case report, when a pericardial effusion developed several days after the Norwood procedure, the cSaO2 dropped significantly while the other noninvasive and invasive measures of cardiac hemodynamic status remained stable, leading to echocardiography and diagnosis of pericardial effusion [17]. A lower cerebral saturation in the early postoperative period is associated with a higher risk for adverse events. Knowing the hourly normative values can assist in the continuous postoperative evaluation of these patients. Data from the current study as well as prior research demonstrate that cSaO2 is typically in the low 50s upon arrival from the operating room and gradually rises into the high 50s and low 60s in the first 2 postoperative days [4]. A persistently low cSaO2 or a decrease in the cSaO2 should cause further evaluation and intervention for these patients. The ability to observe these changes early and monitor continuously may improve outcomes for these high risk patients in a similar fashion as SvO2 measurement. The extent to which caregivers can improve or modify a low cSaO2 is not well understood. Interventions such as transfusion of red blood cells, addition of inotropic support, or sedation are strategies that may increase cSaO2 in this patient population.
Study Limitations The retrospective nature of this investigation and relatively small sample population may limit generalizations of these findings. Further research is required to determine whether changes in operative technique or postoperative management aimed at increasing cSaO2 will improve outcome for patients undergoing the Norwood procedure. In conclusion, low cerebral oxygen saturation as measured by NIRS is predictive of subsequent adverse outcome in patients with hypoplastic left heart syndrome after the Norwood procedure. Monitoring of the cerebral saturation after the Norwood procedure provides additional hemodynamic information that can be used as a part of the postoperative management for these patients.
PEDIATRIC CARDIAC
Ann Thorac Surg 2009;87:1490 – 4
1494
PHELPS ET AL POSTOPERATIVE NORWOOD CEREBRAL OXYGENATION
Ann Thorac Surg 2009;87:1490 – 4
References
PEDIATRIC CARDIAC
1. Jobsis FF. Noninvasive, infrared monitoring of cerebral and myocardial oxygen sufficiency and circulatory parameters. Science 1977;198:1264 –7. 2. Hayashida M, Kin N, Tomioka T, et al. Cerebral ischaemia during cardiac surgery in children detected by combined monitoring of BIS and near-infrared spectroscopy. Br J Anaesth 2004;92:662–9. 3. Nollert G, Jonas R, Reichart B. Optimizing cerebral oxygenation during cardiac surgery: a review of experimental and clinical investigations with near infrared spectrophotometry. Thorac Cardiovasc Surg 2000;48:247–53. 4. Li J, Van Arsdell G, Zhang G, et al. Assessment of the relationship between cerebral and splanchnic oxygen saturations measured by near-infrared spectroscopy and direct measurements of systemic haemodynamic variables and oxygen transport after the Norwood procedure. Heart 2006;92:1678 – 85. 5. Hoffman G, Stuth E, Jaquiss R, et al. Changes in cerebral and somatic oxygenation during stage 1 palliation of hypoplastic left heart syndrome using continuous regional cerebral perfusion. J Thorac Cardiovasc Surg 2004;127:223–33. 6. Norwood WI, Kirklin JK, Sanders SP. Hypoplastic left heart syndrome: experience with palliative surgery. Am J Cardiol 1980;45:87–91. 7. Jacobs JP, O’Brien SM, Chai PJ, et al. Management of 239 patients with hypoplastic left heart syndrome and related malformations from 1993 to 2007. Ann Thorac Surg 2008;85: 1691–97. 8. Nillson B, Mellander M, Sudow G, Berggren H. Results of staged palliation for hypoplastic left heart syndrome: a complete population-based series. Acta Pediatrica 2006;95: 594 – 600. 9. Tibbals J, Kawahira Y, Carter BG, Donath S, Brizard C, Wilkonson J. Outcomes of surgical treatment of infants with
10.
11.
12. 13.
14.
15.
16.
17.
hypoplastic left heart syndrome: an institutional experience 1983-2004. J Paediatr Child Health 2007;43:746 –51. 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; 84:1301–11. Hoffman GM, Mussatto KA, Brosig CL, et al. Systemic venous oxygen saturation after the Norwood procedure and childhood neurodevelopmental outcome. J Thorac Cardiovasc Surg 2005;130:1094 –100. Bhutta AT, Ford JW, Parker JG, et al. Noninvasive cerebral oximeter as a surrogate for mixed venous saturation in children. Pediatr Cardiol 2007;28:34 – 41. Kirshbom PM, Forbess JM, Kogon BE, et al. Cerebral near infrared spectroscopy is a reliable marker of systemic perfusion in awake single ventricle children. Pediatr Cardiol 2007;28:42–5. Maher KO, Pizarro C, Gidding SS, Malec E, Norwood WI, Murphy JD. Hemodynamic profile after the Norwood procedure with right ventricle to pulmonary artery conduit. Circulation 2003;108:782– 4. Bradley SM, Simsic JM, McQuinn TC, Habib DM, Shirali GS, Atz AM. Hemodynamic status after the Norwood procedure: a comparison of right ventricle to pulmonary artery connection versus modified Blalock-Taussig shunt. Ann Thorac Surg 2004;78:933– 41. Ballweg JA, Dominguez TE, Ravishankar C, et al. A contemporary comparison of the effect of shunt type in hypoplastic left heart syndrome on the hemodynamics and outcome at stage 2 reconstruction. J Thorac Cardiovasc Surg 2007;134: 297–303. Maher KO, Phelps HM, Kirshbom PM. Near infrared spectroscopy changes with impeding pericardial tamponade. Pediatr Crit Care Med 2009;10:e13–5.
INVITED COMMENTARY The article by Phelps and colleagues [1] adds to a growing body of literature regarding the clinical application of cerebral oximetric monitoring to the perioperative care of both pediatric and adult cardiac surgical patients. For a monitoring technology to be useful, it must be accurate and reproducible, and it should prompt correlated interventions that impact clinical outcome. Debate continues over whether current cerebral oximetry technology satisfies these requirements. Experience with devices that render absolute values for cerebral oxygen content is limited, and relative values can be affected by various factors, such as cerebral blood volume, hematocrit, acid-base status, and cranial anatomy. Areas of the brain remote from the monitored frontal cortex are at risk of undetected injury, potentially limiting the usefulness of the technology in avoiding significant central nervous system insult. There is significant inter-patient variability of rSO2 values, and a lack of a well-defined normative range. This is particularly true among children with congenital heart disease in which arterial desaturation and parallel circulatory arrangements further complicate our understanding of monitored values. These limitations notwithstanding, the association between lower mean rSO2 values and adverse events after Norwood palliation described by the authors are consistent with a large number of prior reports that impute clinical utility to continuous cerebral oximetry monitoring. In light of our increasing awareness of perioperative cerebral injury © 2009 by The Society of Thoracic Surgeons Published by Elsevier Inc
and its subsequent neuro-developmental impact, mechanisms to accurately surveil the cerebral environment are needed. Systemic measures of oxygen delivery such as SvO2 or lactate are poor predictors of cerebral insult. The clinical benefit of cerebral oximetric monitoring is intuitive, and the supporting evidence is compelling. Although weaknesses of the present design and understanding of the technology require the clinician to place rSO2 values in proper context, further design improvements and clinical understanding of its role represent a significant opportunity to improve care and outcomes. David M. Overman, MD Division of Cardiac Surgery The Children’s Heart Clinic Children’s Hospitals and Clinics of Minnesota 2545 Chicago Ave, South Suite 106 Minneapolis, MN 55404 e-mail: [email protected]
Reference 1. Phelps HM, Mahle WT, Kim D, et al. Postoperative cerebral oxygenation in hypoplastic left heart syndrome after the Norwood procedure. Ann Thorac Surg 2009;87:1490 – 4.
0003-4975/09/$36.00 doi:10.1016/j.athoracsur.2009.03.015