| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2005;130:1523-1530
© 2005 The American Association for Thoracic Surgery
Surgery for Congenital Heart Disease |
a Departments of Pediatrics (Divisions of Cardiology
b Neurology
c Epidemiology/Biostatistics)
d Surgery (Division of Cardiothoracic Surgery)
e Radiology
f Anesthesiology, Cincinnati Children's Hospital Medical Center and the University of Cincinnati College of Medicine, Cincinnati, Ohio
* Address for reprints: Catherine Dent, MD, Cincinnati Children's Hospital Medical Center, 3333 Burnet Ave, MLC 2003, Cincinnati, OH 45229-3039 (Email: catherine.dent{at}cchmc.org).
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
METHODS: Twenty-two term neonates with hypoplastic left heart syndrome were studied with brain magnetic resonance imaging before and at a median of 9.5 days after the Norwood operation. Results were compared with preoperative, intraoperative, and postoperative risk factors to identify predictors of neurologic injury.
RESULTS: Preoperative magnetic resonance imaging (n = 22) demonstrated ischemic lesions in 23% of patients. Postoperative magnetic resonance imaging (n = 15) demonstrated new or worsened ischemic lesions in 73% of patients, with periventricular leukomalacia and focal ischemic lesions occurring most commonly. Prolonged low postoperative cerebral oximetry (<45% for >180 minutes) was associated with the development of new or worsened ischemia on postoperative magnetic resonance imaging (P = .029).
CONCLUSIONS: Ischemic lesions occur commonly in neonates with hypoplastic left heart syndrome before surgical intervention. Despite the adoption of regional cerebral perfusion, postoperative cerebral ischemic lesions are frequent, occurring in the majority of infants after the Norwood operation. Long-term follow-up is necessary to assess the functional effect of these lesions.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
Although the cause of brain injury in these infants is likely multifactorial, with contributions from preoperative, intraoperative and postoperative events, the use of deep hypothermic circulatory arrest (DHCA) appears to play a role in poor neurologic and developmental outcome.
6-10
Ischemic brain lesions have been documented on pathologic brain specimens
11
and on magnetic resonance imaging (MRI)
12
in neonates after surgical palliation with DHCA.
Because of these concerns, many centers have adopted the technique of regional low-flow cerebral perfusion (RLFP) during aortic arch reconstruction in lieu of DHCA. RLFP decreases the period of cerebral ischemia by limiting decreases in cerebral blood volume and oxygen saturation.
13
Although RLFP has been associated with better neurologic outcome in animal models,
14
such studies have not been performed in human subjects. In the present study we compare preoperative and postoperative brain MRI findings in neonates undergoing the Norwood procedure with RLFP.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
36 weeks) and an intention to undergo surgical intervention (Norwood operation with aortic arch reconstruction) with RLFP. Infants were excluded if they had (1) a history of birth asphyxia (5-minute Apgar score <5), (2) a genetic anomaly associated with neurodevelopmental abnormalities, or (3) preoperative cardiac arrest. MRI scans of the brain were performed preoperatively (day of the operation) and in the early postoperative period. The scans were performed on a Signa LX 1.5-T scanner (GE Medical, Milwaukee, Wis). The following sequences were performed: (1) sagittal T1-weighted spin-echo images, (2) axial T1-weighted inversion-recovery images, (3) axial and coronal T2-weighted fast spin-echo images, (4) axial diffusion-weighted images, and (5) short echo proton magnetic resonance spectroscopy in the basal ganglia. All MRI scans were reviewed by a single neuroradiologist blinded to the subjects' clinical status. MRI scans were reviewed for congenital and acquired lesions, including general or focal atrophy, periventricular leukomalacia (PVL), cerebral edema, delayed myelination, intraparenchymal hemorrhage, intraventricular hemorrhage, and infarction. All lesions were classified as mild, moderate, or severe. PVL, infarction, and intraparenchymal hemorrhage were considered to represent ischemia.
Clinical Management
Preoperative clinical management was provided in the CICU. Infants were maintained on a continuous prostaglandin infusion. Pulmonary overcirculation was managed with administration of subambient oxygen (fraction of inspired oxygen, 0.17-0.20) or addition of inhaled CO2. Inotropic medication was administered at the discretion of the attending physician. As an assessment of overall status, a preoperative inotropic score was calculated as the sum of all inotrope doses, correcting for potency.
15,16
Preoperative neurologic evaluation included electroencephalography (EEG) and examination by a pediatric neurologist directed at level of consciousness, motor tone, response to stimuli, and deep tendon reflexes. Regional cerebral oxygen saturation (rSO
2) was monitored continuously by using near-infrared spectroscopy (NIRS; Somanetics INVOS 5100A, Troy, Mich), with the probe placed on the right side of the patient's forehead. Monitoring commenced 12 hours before the operation. Data were recorded at 1-minute intervals. An rSO
2 value of less than 45% was considered to represent cerebral desaturation. Cumulative time spent with rSO
2 values of less than 45% was recorded. Management was not altered on the basis of cerebral oximetry readings.
For the preoperative MRI, anesthesia was induced with fentanyl (5 µg/kg), midazolam (0.1 mg/kg), and vecuronium (0.2 mg/kg). In patients not already mechanically ventilated, nasotracheal intubation was performed. Patients were monitored during the MRI with continuous pulse oximetry, capnography, electrocardiography, and blood pressure measurements. After the MRI, patients were transported to the operating suite for cardiac surgery. Surgical repair consisted of aortic arch reconstruction, ascending aorta–to–pulmonary artery anastomosis, and creation of an unrestrictive atrial septal communication. Pulmonary blood flow was provided by either a systemic–to–pulmonary artery shunt or a right ventricle–to–pulmonary artery conduit.
Cardiopulmonary bypass (CPB) and surgical management followed our usual institutional practice. Whole blood was added to the primer to yield a goal hematocrit value of 28% to 30% during CPB. Arterial blood gas-pH management followed the alpha-stat strategy on CPB initiation, with switch to pH-stat strategy during cooling. The alpha-stat strategy was resumed during rewarming. All patients were cooled to deep hypothermia (18°C). During reconstruction of the aortic arch, continuous RLFP was provided through the innominate shunt at 30 mL·kg–1 ·min–1. Before separation from CPB, patients received a loading dose of milrinone (37.5 µg/kg). Dopamine and epinephrine infusions were instituted and titrated to achieve adequate blood pressure and systemic vascular resistance.
Postoperative management was provided in the CICU. Inotropic medication was adjusted at the discretion of the attending physician. Sodium nitroprusside was added as tolerated for afterload reduction and to improve cardiac output. Management targets included mean arterial pressure of 45 mm Hg or greater, mixed venous oxygen saturation (SvO 2) of 50% or greater, arterial oxygen saturation (SaO 2) of 70% or greater, and a hematocrit value of greater than 40%. Blood for SvO 2 measurement was sampled from a catheter placed intraoperatively in the superior vena cava. SvO 2 was measured on admission and every 4 hours. Cerebral rSO 2 monitoring continued for 48 hours after the operation. Postoperative inotrope scores were calculated every 6 hours.
The postoperative MRI was performed when the patients were deemed suitable for transport, generally between 5 and 14 days after the operation. Sedation was provided with oral pentobarbital (5 mg/kg) in spontaneously breathing patients and with inhaled anesthesia in intubated patients. Patients were monitored in similar fashion as the preoperative scan. Postoperative EEG and neurologic examination by a pediatric neurologist were performed when the patient was no longer receiving sedation.
Data Collection and Statistical Analysis
Primary outcome measures were (1) ischemia on preoperative MRI and (2) development of new or worsened parenchymal lesions on postoperative MRI when compared with the preoperative study. Data collected preoperatively included patient demographics, birth history, Apgar scores, birth weight and head circumference, indirect measures of cardiac output (acid-base status [pH and base deficit], serum lactate level, mean arterial pressure, and SaO
2], inotropic score, and rSO
2 value. Operative events (including type of shunt placed for pulmonary blood flow and duration of CPB, RLFP, and DHCA) and intraoperative variables (acid-base status, serum lactate level, hematocrit value, and rSO
2 value) were included in the analysis. Postoperative variables analyzed included acid-base status (pH and base deficit), lactate level, mean arterial pressure, diastolic arterial pressure, urine output, inotropic scores, SaO
2, SvO
2, and rSO
2 value. Cerebral desaturation was assessed both as a continuous variable (cumulative minutes with rSO2 <45%) and a categoric variable (>180 or <180 cumulative minutes of cerebral desaturation).
Distributions of variables were examined for normality, and appropriate adjustments were done, either by using variance stabilizing transformations or nonparametric analyses when necessary. Associations between the outcome variables and various risk factors were analyzed with Fisher exact tests for categoric outcome and risk factor variables, Student t tests and logistic regression for binary outcomes, and linear regression for continuous variables. Analysis was performed with SAS statistical software (SAS Institute, Cary, NC).
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
Prenatal diagnosis of congenital heart disease (CHD) had been made in 17 (77%) of the 22 patients.
|
|
Neurologic Evaluation and MRI Results
Preoperative studies
Table 3
summarizes preoperative results. Preoperative neurologic examination was performed in 17 of the 22 patients. Eleven patients had normal examination result, 5 patients had diminished level of consciousness caused by sedative medications but had otherwise normal examination results, and 1 patient had increased muscle tone. Of the 15 patients with preoperative EEG, 12 (80%) had normal EEG pattern, and 3 (20%) had mild diffuse slowing. These 3 patients had all had normal neurologic examination results and were not receiving sedative medication. No seizures were observed clinically or by means of EEG before the operation. Preoperative MRI was performed on the day of the operation in all 22 patients. No congenital structural abnormalities were identified in any patient. Ischemic lesions were seen in 5 (23%) of 22 patients. One infant had multifocal diffusion abnormalities in the parietal white matter consistent with mild PVL, 1 had focal infarction in the right frontal lobe, and 3 had focal ischemic lesions in the left frontal lobe. Extra-axial hemorrhage was identified in 2 patients who had bilateral small hemorrhagic collections in the posterior cerebral and occipital regions. In 8 (36%) of the 22 patients, 5 of whom had otherwise normal studies, mild increases in T1 signal of unclear significance occurred in the basal ganglia–thalamic region. Ten patients had completely normal study results. Magnetic resonance spectroscopy was performed in 20 of the 22 patients, including the 5 with ischemic lesions, and demonstrated normal pattern without lactate increase in all patients.
|
Postoperative studies
Table 4
summarizes postoperative study results. Of the 21 patients who underwent surgical intervention with RLFP, postoperative MRI was performed in 15 patients. Five patients died early after the operation, and MRI was deferred in 1 patient because of the need for cardiac pacing. The median duration between surgical intervention and postoperative MRI was 9.5 days (range, 4-112 days). Two patients had follow-up MRI performed more than 3 weeks after the operation because of clinical instability.
|
New or worsened ischemic lesions were observed in 11 (73%) of 15 patients. These lesions included PVL in 7 (47%) patients and focal ischemic or hemorrhagic-ischemic lesions in 8 (53%) patients. The PVL occurred primarily in the frontoparietal region and occurred bilaterally in 6 of the 7 patients (Figure 1). All patients with preoperative ischemia had new or worsened lesions on postoperative study. Mild extra-axial (subdural or choroid plexus) or intraventricular hemorrhage not requiring intervention was observed in 7 (47%) patients. One patient had significant hemorrhage into an area of ischemia in the left frontal area (seen preoperatively), which was surgically drained (Figure 2). Magnetic resonance spectroscopy demonstrated lactate increase in 3 of the 15 patients, all with ischemic lesions on MRI. Eight (73%) of 11 patients with conventional BT shunts and 3 (75%) of 4 patients with right ventricle–to–pulmonary artery conduits had new or worsened ischemia on postoperative MRI.
|
|
Risk Factor Analysis
Analysis of preoperative variables demonstrated a significantly higher peak preoperative base deficit in patients with ischemic lesions on preoperative MRI compared with those not showing ischemic lesions on preoperative MRI (P = .024). No other preoperative variables, including other indirect measures of cardiac output, were associated with preoperative ischemia.
Analysis of risk factors for the presence of new or worsened area of ischemia on postoperative MRI demonstrated a significant association between prolonged postoperative low rSO 2 values (>180 cumulative minutes with an rSO 2 <45%) and the development of MRI lesions (P = .029, Fisher exact test). Sensitivity and specificity were calculated at 82% and 75%, respectively, with a positive predictive value of 90% and a negative predictive value of 60%. The nadir of rSO 2, indirect measures of cardiac output (eg, SvO 2), and inotrope scores in the postoperative period were not related to the development of lesions. Similarly, intraoperative variables, including type of pulmonary blood flow (BT shunt or right ventricle–to–pulmonary artery conduit) and durations of CPB and RLFP, did not relate to the development of MRI ischemic lesions. Patients with evidence of preoperative ischemia were more likely to have new postoperative ischemia (P = .08), but this did not reach statistical significance.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
Risk factors for preoperative ischemia have not been well defined. In our study, although peak preoperative base deficit, one indicator of tissue perfusion, was associated with the presence of preoperative ischemia, other postnatal clinical indicators of overall cardiac output (eg, pH and serum lactate) were not associated with lesions on preoperative MRI. Our study also demonstrates that preoperative rSO
2 is not associated with preoperative MRI ischemic lesions. A possible explanation for this finding is that cerebral ischemia occurred before the time period of NIRS monitoring, either in utero or after birth but before the monitoring period. Although our NIRS monitoring did not encompass the entire preoperative time period, no patient had a significant change in hemodynamics or SaO
2 values during the preoperative period, and therefore the measurements obtained during our monitoring period are likely to be representative of the entire preoperative state. Nevertheless, because rSO
2 measurements are being performed in a small (regional) area of the brain, they might not be a global representation of cerebral perfusion or might miss areas of regional ischemia and could therefore be falsely reassuring. A recent study found that low cerebral blood flow is common in patients with CHD and is associated with MRI findings of PVL.
18
There is also evidence that ischemia might occur in these infants in utero. Abnormalities in cerebral vascular blood flow dynamics have been found in fetuses with HLHS, similar to those seen with other forms of fetal hypoxia.
19,20
In contrast to a previous study
21
that demonstrated very low preoperative cerebral rSO
2 values in patients with HLHS, we found cerebral desaturation to be relatively uncommon in our preoperative patients, with only 4 of the 22 patients having any preoperative rSO
2 values of less than 45% and only 2 with more than 30 minutes at an rSO
2 value of less than 45%. Interestingly, all 4 of these patients had normal preoperative MRI results.
Postoperative Findings
Despite modification in surgical technique with RLFP, new or worsened ischemic brain lesions occurred in the majority (73%) of our patients after the Norwood procedure in a similar incidence to that seen with DHCA.
12
Prolonged low rSO
2 values (>180 minutes with rSO
2 
45%) were associated with the presence of lesions on postoperative MRI. Intraoperative factors, such as CPB and RLFP time, were not associated with MRI lesions nor were indirect measures of overall cardiac output, such as SvO
2 value, serum lactate level, or acid-base status. These data are interesting for 2 reasons. First, they suggest that the postoperative rather than intraoperative period might now be the critical period during which neurologic injury occurs or progresses. Second, they suggest that the usual clinical measurements followed in the CICU to assess global cardiac output might not reflect cerebral perfusion, particularly in the early postoperative period, perhaps as a result of loss of cerebral autoregulation.
The modification of the Norwood procedure to adopt the use of RLFP has been made in many centers to avoid longer periods of DHCA. The association of longer periods of DHCA with impaired neurologic outcome has been well documented,
1,7,8,22,23
but until recently, DHCA has been believed to be a necessary component of arch reconstruction. RLFP limits the period of cerebral ischemia, improves cerebral oxygenation, and has been shown to be associated with better neurologic outcome in piglets.
14
It is unknown, however, whether RLFP is associated with better neurologic outcome after the Norwood procedure. Our study demonstrates that ischemic lesions on MRI are no less frequent in patients who have had operations with RLFP versus those who underwent operations with DHCA.
12
Hoffman and colleagues
24
have noted that although RLFP provides consistent brain perfusion associated with high rSO
2 values, cerebral desaturation occurs rapidly on removal from CPB. Furthermore, they noted that cerebral rSO
2 values decreased to much lower levels after CPB than somatic rSO
2 values (measured in the splanchnic bed), suggesting that cerebral vascular resistance is higher than somatic resistance in this time period. We noted an identical rapid decrease in cerebral rSO
2 values in our patients after separation from CPB, which persisted in the early postoperative period.
Animal models of cerebral hypoxia-ischemia have demonstrated an association between low rSO
2 values and neuronal dysfunction, cell death, and poor neurologic outcome. In a piglet model of graded hypoxia-ischemia to determine thresholds for neurologic injury, brain tissue lactate accumulation began when rSO
2 values decreased to less than 45%.
25
Our findings suggest that the rSO
2 threshold for neuronal injury is similar in neonates and that maneuvers to improve cerebral perfusion during the postoperative period might lessen the incidence of postoperative ischemia. There are both animal and human data associating low rSO
2 values and neurologic outcome.
26,27
The clinical relevance of our MRI findings is unknown. In the preterm neonate, however, MRI findings of PVL have been associated with long-term neurocognitive impairment, including visuospatial and visuomotor abnormalities, attention deficit disorders, and developmental delay.
28,29
Given the frequency of neurodevelopmental abnormalities in patients with HLHS, it seems likely that these lesions are clinically relevant as well.
Limitations
A limitation of this study is that neurodevelopmental evaluation of these children is not yet available. Although an association between MRI findings and neurodevelopmental outcome has been established in preterm infants, this association has not yet been found for infants with CHD. A second limitation is the timing of the postoperative MRI, which was performed at a median of 9.5 days after surgical intervention. Although the patients were not clinically suitable for transport and MRI scanning until this time, we cannot definitively pinpoint the timing of the ischemia (operative vs postoperative). Third, our patients had just one postoperative MRI scan. In one study resolution of MRI lesions occurred on later MRI in some patients,
12
making clinical follow-up even more crucial. Lastly, we used previous studies in which DHCA was used for the Norwood procedure as a comparison group for our patients rather than a control DHCA group within our own institution, and therefore we cannot eliminate the possibility of institutional differences accounting for some of our findings.
|
Conclusions |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
|
Acknowledgments |
|---|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. Chen, R. A. Zimmerman, G. P. Jarvik, A. S. Nord, R. R. Clancy, G. Wernovsky, L. M. Montenegro, D. M. Hartman, S. C. Nicolson, T. L. Spray, et al. Perioperative stroke in infants undergoing open heart operations for congenital heart disease. Ann. Thorac. Surg., September 1, 2009; 88(3): 823 - 829. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Tyree, M. Tyree, and R. DiGeronimo Correlation of brain tissue oxygen tension with cerebral near-infrared spectroscopy and mixed venous oxygen saturation during extracorporeal membrane oxygenation Perfusion, September 1, 2009; 24(5): 325 - 331. [Abstract] [PDF]
|
|
|
|
 |
|
 W. T. Mahle, J. W. Newburger, G. P. Matherne, F. C. Smith, T. R. Hoke, R. Koppel, S. S. Gidding, R. H. Beekman III, S. D. Grosse, and on behalf of the American Heart Association Congen Role of Pulse Oximetry in Examining Newborns for Congenital Heart Disease: A Scientific Statement From the American Heart Association and American Academy of Pediatrics Circulation, August 4, 2009; 120(5): 447 - 458. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. T. Mahle, J. W. Newburger, G. P. Matherne, F. C. Smith, T. R. Hoke, R. Koppel, S. S. Gidding, R. H. Beekman III, S. D. Grosse, and on behalf of the American Heart Association Congen Role of Pulse Oximetry in Examining Newborns for Congenital Heart Disease: A Scientific Statement from the AHA and AAP Pediatrics, August 1, 2009; 124(2): 823 - 836. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. D. Salazar, R. D. Coleman, S. Griffith, J. D. McNeil, M. Steigelman, H. Young, B. Hensler, P. Dixon, J. Calhoon, F. Serrano, et al. Selective cerebral perfusion: real-time evidence of brain oxygen and energy metabolism preservation. Ann. Thorac. Surg., July 1, 2009; 88(1): 162 - 169. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. D. Kussman, D. Wypij, J. A. DiNardo, J. W. Newburger, J. E. Mayer Jr, P. J. del Nido, E. A. Bacha, F. Pigula, E. McGrath, P. C. Laussen, et al. Cerebral Oximetry During Infant Cardiac Surgery: Evaluation and Relationship to Early Postoperative Outcome Anesth. Analg., April 1, 2009; 108(4): 1122 - 1131. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. D. Kurth, J. C. McCann, J. Wu, L. Miles, and A. W. Loepke Cerebral Oxygen Saturation-Time Threshold for Hypoxic-Ischemic Injury in Piglets Anesth. Analg., April 1, 2009; 108(4): 1268 - 1277. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. J. Licht, D. M. Shera, R. R. Clancy, G. Wernovsky, L. M. Montenegro, S. C. Nicolson, R. A. Zimmerman, T. L. Spray, J. W. Gaynor, and A. Vossough Brain maturation is delayed in infants with complex congenital heart defects. J. Thorac. Cardiovasc. Surg., March 1, 2009; 137(3): 529 - 536. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
