| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
#¶#
**¶¶||¶
Departments of *Anesthesiology, Perioperative, and Pain Medicine, Clinical Research Program, Cardiology and Cardiovascular Surgery, Children’s Hospital Boston; Departments of ||Anesthesia, ¶Cardiology, #Pediatrics, and **Surgery, Harvard Medical School; and Department of Biostatistics, Harvard School of Public Health, Boston, Massachusetts
Address correspondence and reprint requests to Barry D. Kussman, MBBCh, Department of Anesthesiology, Perioperative and Pain Medicine, Children’s Hospital, 300 Longwood Ave., Boston, MA 02115. Address e-mail to barry.kussman{at}childrens.harvard.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
2 percentage points/absolute scale units) before, during, and after cardiopulmonary bypass, irrespective of the use of deep hypothermic circulatory arrest. Further longitudinal neurological outcome studies are required to determine whether uni- or bi-hemispheric monitoring is required in this patient population. |
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Cerebral oximetry using near-infrared spectroscopy (NIRS) is a developing technology for noninvasive assessment of cerebral oxygenation (11–14). The INVOS 5100 (Somanetics, Troy, MI) is the only cerebral oximeter commercially available in the United States and has been approved for use as a trend monitor of cerebral oxygenation. The INVOS 5100 expresses cerebral oxygen saturation as the regional cerebral oxygen saturation index (rSO2i). The rSO2i is accordingly a measure of the oxygen saturation of blood in the gas-exchanging circulation (arterioles, capillaries, and venules) of brain tissue. Bi-hemispheric measurement of cerebral oxygen saturation is recommended by the manufacturer, but correlation of the left and right rSO2i in infants undergoing congenital cardiac surgery without aortic arch reconstruction has not been reported. The aim of this study was to evaluate differences in bi-hemispheric measurement of cerebral oxygenation using NIRS during pediatric cardiac surgery. The hypothesis was that the difference in left and right cerebral oxygen saturation would be <10 percentage points/absolute scale units (%).
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Anesthetic technique was not specifically controlled but was conducted according to our institutional practice. Large-dose opioid anesthesia (fentanyl 100 µg/kg) was supplemented with midazolam and isoflurane, as tolerated, and neuromuscular blockade was achieved with pancuronium. The head was turned to just off the midline to prevent pressure or movement on the endotracheal tube by the surgical team while avoiding the possible effects of extremes of lateral head position on cerebral blood flow and venous drainage. Standard monitoring was used, including a radial or femoral artery catheter for measurement of systemic arterial blood pressure and intermittent blood sampling, as well as tympanic, esophageal, and rectal temperatures.
Surface cooling was initiated with low ambient room temperature, a cooling mattress, and ice packs applied to the head. Core cooling was begun with the initiation of CPB. The bypass circuit was primed with Plasmalyte-A (Baxter, Deerfield, IL) and blood to achieve the desired hematocrit during cooling. A nonpulsatile roller pump with a membrane oxygenator (D 901 Lilliput 1 Open System; Dideco, Mirandola, Italy) was used. Standard pump flow rates of 150–200 mL · kg–1 · min–1 for full flow and 50 mL · kg–1 · min–1 for LF were used. Perfusion pressure was determined by CPB flow rate, with pressures in the range of 30–40 mm Hg being acceptable at full flow. When the rectal temperature reached 18°C or lower, and after at least 20 min of cooling, DHCA or continuous LF bypass was begun. A pH-stat acid-base management strategy was used in all patients during hypothermia. Methylprednisolone (30 mg/kg), phentolamine (0.2 mg/kg), and furosemide (0.25 mg/kg) were given at the initiation of CPB to all patients. At the onset of rewarming, mannitol (0.5 g/kg) and phentolamine (0.2 mg/kg) were given to all patients. Patients were rewarmed for at least 30 min to a rectal temperature of 34°C; hemofiltration was performed during rewarming.
Cerebral oxygen saturation was measured with the INVOS 5100. Cerebral oximetry relies on the penetration of the scalp, skull, and cranial contents by near-infrared light and measures the ratio of oxyhemoglobin to total hemoglobin to derive a hemoglobin oxygen saturation. The INVOS 5100 is a continuous wave spectrometer and uses two wavelengths of near-infrared light (730 and 810 nm) to measure the ratio of oxyhemoglobin to total hemoglobin. The sensor contains a light-emitting diode (LED) and 2 detectors located 30 and 40 mm from the LED, allowing removal of the extracranial contribution of absorbed and scattered light by the application of a subtraction algorithm. Because the volume of blood is predominantly venous, it reflects the balance between oxygen supply and demand in the brain. The scale unit for rSO2i is percent (%) but is referred to as an index because it is not easily validated in vivo, is regional in nature, and may not reflect events occurring at a distance from the sensor. After the induction of anesthesia, one Pediatric SomaSensor (Somanetics) was placed on the right and one on the left forehead, according to the manufacturer’s guidelines. After an accommodation period, data collection was begun and downloaded to a storage disk every 10 s throughout the case for further analysis. The INVOS 5100 collects light absorption data 15 times a second, and when 50 samples have been collected, they are averaged to determine a new rSO2i value, which is displayed on the screen (approximately every 4 s).
Intraoperative data were collected and analyzed at the following time points: after induction, initiation of CPB, 10 min after start of cooling, onset of LF, onset of DHCA, resumption of LF, start of rewarming, 10 min after start of rewarming, warm flow (35°C), immediately off CPB, 60 min after CPB, and 6 h after CPB. Longer term (1 yr) neurodevelopmental outcome and assessment of brain injury with magnetic resonance imaging (MRI) is continuing in the patients enrolled in this study, and because we are still blinded to randomization, we could not analyze the rSO2i values in respect to hematocrit strategy.
Data analyses were performed with SAS Version 9.00 (SAS Institute, Inc, Cary, NC). Descriptive statistics for demographic, intraoperative, physiologic, and rSO2i variables were based on the mean and standard deviations, medians and ranges, or frequencies. Comparisons of left versus right rSO2i values were based on paired t-tests and Bland-Altman analyses (15). Physiologic and saturation data are expressed as mean ± sd. The primary outcome variable was the mean difference in all rSO2i values between cerebral hemispheres at each time point. Although the mean difference in left and right rSO2i is unknown, we chose a mean difference of 10% (absolute scale units) because more than 10 scale unit left-right rSO2i differences have been reported as highly unusual, and decreases (relative) of 20% from baseline have been associated with possible neurologic symptoms (16–18). Because this is an observational secondary analysis, a power calculation was not performed in advance. Nevertheless, a sample size of 60 paired differences, assuming a common standard deviation of 16 and a correlation of 0.75 between paired left and right measurements, provides 92% power to detect mean differences of 5% between left and right measurements.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Demographic and intraoperative data for these 60 infants are shown in Table 1. After the induction of deep hypothermia, 35 infants (58%) underwent continuous LF and 25 infants (42%) underwent DHCA. Physiologic data are shown in Table 2.
|
|
Regional cerebral oxygen saturation at each time point is shown in Table 3. The rSO2i increased during cooling, decreased during the period of DHCA (when occurring), and was restored to baseline levels by 6 h after CPB. Left and right rSO2i were similar before, during, and after CPB (mean difference, 2 absolute percentage points/scale units at each time point). These findings were the same regardless of whether DHCA was used. The nadir rSO2i at the end of DHCA (median, 26; range, 1–59 min) was left 64 ± 17 and right 65 ± 14 (P = 0.97). For those infants who had continuous LF, the nadir rSO2i at the end of the LF CPB period was left 81 ± 13 and right 82 ± 13 (P = 0.67).
|
Graphical displays of individual subject data showed similar trends between left and right rSO2i measurements (results not shown). Representative Bland-Altman plots assessing agreement between left and right rSO2i measurements at postinduction and 10 min after cooling suggest that there is variability within patients (Fig. 1). Of 634 paired left versus right measurements on subjects, 77 (12%) were more than 10 percentage points apart and were evenly divided between left and right. In 40 measurements, the left side was >10 percentage points higher than the right side, and in 37 measurements, the right side was >10 percentage points higher (P = 0.82). These 77 measurements came from 28 of 60 (47%) patients; of these measurements, 1 subject had 12, 1 had 7, 2 had 5, 4 had 4, 2 had 3, 8 had 2, and 10 had 1. These differences >10 scale points were distributed across many perfusion phases, being more frequent from 10 min after start of cooling on CPB to immediately off CPB.
|
There were six patients who were more than 10 percentage points apart at baseline. For the 71 paired left versus right measurements that were more than 10 percentage points apart after baseline, the left and right rSO2i were compared with their respective postinduction value. In six (8%) of the paired measurements, corresponding to five patients, either the left or the right rSO2i had decreased more than 20% (relative) less than baseline. The time points at which this occurred were on CPB (1), resume LF (3), at start rewarming (1), and 10 min after start rewarming (1). In 27 (38%) paired measurements in 11 patients, both the left and right rSO2i were within 20% (relative) of baseline. Of the remaining 44 paired measurements, in 38 measurements (86%), from 19 patients, the left, right, or both rSO2i had increased more than 20% (relative) above baseline.
Of the 634 paired left versus right measurements on subjects, 13 (2%) were more than 20 percentage points apart. This included four measurements for which the left side was >20 percentage points higher than the right side and nine measurements for which the right side was >20 percentage points higher (P = 0.27). These 13 measurements came from 4 of 60 (7%) subjects; of these measurements, 1 subject had 9, 1 had 2, and 2 had 1 each.
The largest difference in any patient was 31 percentage points at 2 time points—resume LF (L = 40; r = 71) and start rewarming (L = 51, r = 82)—i.e., immediately after DHCA. This neonate, undergoing repair of D-TGA with VSD and whose postinduction rSO2i values were left 55% and right 79% (arterial oxygen saturation, 95%), accounted for 12 of the 77 (16%) paired left versus right measurements that were more than 10 percentage points apart and 9 of the 13 (69%) paired measurements that were more than 20 percentage points apart. There were no postoperative neurologic complications in this patient, and a brain MRI study, with spectroscopy, performed at 16 mo of age, was normal. All patients survived to discharge, and there were no clinical signs of neurologic injury.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Despite the availability of NIRS for several years, it has not been widely used during congenital cardiac surgery. Explanations include limitations of the technology, insufficient outcome data to determine sensitivity and specificity, and cost of the instrumentation. NIRS instruments differ in the technology and algorithm used to measure hemoglobin saturation, and because large inter- and intraindividual differences in saturation have been found among instruments, critical values of cerebral oxygenation determined by one instrument cannot be applied to values measured by another (19–22).
Defining a normative range for cerebral oxygen saturation in pediatric patients has been problematic. In adults, mean rSO2i values of 71% ± 6% and 67% ± 10% have been reported for healthy volunteers and adult cardiac surgical patients, respectively (16,23). Using a frequency domain cerebral oximeter (the INVOS 5100 is a continuous wave oximeter), Kurth et al. (22) found cerebral O2 saturation (ScO2) values of 68% ± 10% in healthy children. Mean rSO2i values of 78% ± 8% were reported for healthy children breathing air at an altitude of 1610 m (24). Large interpatient variance of cerebral oxygen saturation has been shown for neonates without CHD breathing room air (25). In CHD, cerebral oxygenation has been shown to vary with anatomy and arterial saturation (22). ScO2 values similar to healthy subjects (68% ± 10%) were found in patients with VSD, aortic coarctation, and single ventricle after Fontan operation, whereas ScO2 was significantly decreased in patients with patent ductus arteriosus (53% ± 8%), TOF (57% ± 12%), hypoplastic left heart syndrome (46% ± 8%), pulmonary atresia (38% ± 6%), and single ventricle after aortopulmonary shunt (50% ± 7%) or bidirectional Glenn operation (43% ± 6%). The administration of sedative medication or general anesthesia, which typically causes a decrease in cerebral oxygen consumption, is a confounding factor when attempting to establish normative ranges in children.
Baseline rSO2i measurements in the present study were symmetrical. Although numerically similar to those reported by Kurth et al. (22), rSO2i and ScO2 do not necessarily represent the same level of cerebral oxygenation. As demonstrated by the standard deviation, both at baseline and thereafter, there is a wide range with interindividual variability of cerebral oxygenation. Baseline (or developing) cerebral oxygen saturation asymmetry has been attributed to carotid or intracranial arterial stenosis, intracranial space-occupying lesion, old infarction, extracranial lesions, such as hemangioma, excessive frontal sinus fluid, or a skull defect, subclavian steal, and interference from an infrared-emitting device or ambient light (16). During cardiac surgery, cerebral oxygen saturation asymmetry may also occur after complications related to aortic cannulation (26,27). Obstruction of the superior vena cava during venous cannulation or on bypass can be associated with alterations in cerebral oxygen saturation (28–30).
Although multiple factors influence neurologic outcome during cardiac surgery, mechanisms of cerebral injury are quite different between adults and young children. Risk factors for cerebrovascular disease in children differ significantly from those in adults, in whom arteriosclerosis is the leading cause (31). Fatty streaks are present in infants and children, but they do not significantly narrow the lumina of blood vessels (32). Fibrous or atheromatous plaques that result in narrowing of the lumen usually begin in the late teens and early twenties (33). The most common identifiable risk factor for childhood ischemic stroke is complex CHD (31), including related factors of coagulation disorders, dehydration, and infection. Other causes of childhood ischemic stroke include vasculitis, dissection, cancer, metabolic disorders, moyamoya disease, sickle cell anemia, and perinatal complications. Variable developmental patterns in the Circle of Willis are unlikely to account for large differences between left and right cerebral oxygen saturation. An anatomical study of 350 normal brains found a normal configuration of the circle of Willis in 52.3%, with the anterior and posterior halves of the polygon varying in respect to the type of anomaly (34). The most frequent anomaly (27.4%) was a reduced size of one of the component vessels. In no case was there absence of an anatomical connection between the right and left anterior cerebral artery. Accessory vessels (duplications and triplications) were found in 18.9%, the majority in the anterior portion of the circle. A truly incomplete circle was found in 0.6% with absence of the left posterior communicating artery. No abnormalities of the middle cerebral artery were reported in this series.
It is important to stress that the INVOS 5100 is approved by the Food and Drug Administration as a trend monitor of cerebral oxygenation. Although studies in adults suggest that cerebral ischemia may occur with (a) a decrease in rSO2i of more than 10 percentage points/absolute scale units, (b) more than 20% decrease (relative) from baseline, or (c) an absolute value of <50% (18,35–37), the critical rSO2i or degree of relative change from an individual established baseline in children that results in cerebral hypoxic-ischemic damage is uncertain. Cerebral oximetry is thus a trend monitor of greatest value in situations in which cerebral oxygen saturation could dangerously change when changes in systemic hemodynamics and oxygenation would not predict that change (38). The similarity of left and right rSO2i measurements in the present study raises the question of whether unilateral monitoring of cerebral oxygen saturation is sufficient for many patients undergoing infant cardiac surgery. Cerebral oximetry (unilateral) as a component of multimodality neurophysiologic monitoring has been shown to be potentially beneficial in pediatric cardiac surgery (17). In a prospective cohort study, an interventional algorithm was used during surgery to detect and correct specific deficiencies in cerebral oxygenation or perfusion, as measured by cerebral oximetry, transcranial Doppler sonography, and electroencephalography. Limitations of this study included nonrandomization to an intervention versus nonintervention strategy, no longer-term evaluation of neurologic outcome, and no determination of absolute level of cerebral oxygen saturation or transcranial Doppler flow, at which their needs to be an intervention. The restricted space on the neonatal and infant forehead may only allow for unilateral placement of the Pediatric SomaSensor during multimodality neurophysiologic monitoring. Although bilateral monitoring is recommended during aortic arch reconstruction using regional LF perfusion (10) or in situations where venous drainage is altered, a recent review suggested that bilateral monitoring was not required when both carotid arteries were perfused under normal circumstances (39). In our study, six patients had baseline rSO2i asymmetry more than 10 percentage points, but in only one patient did this difference persist throughout the procedure and up to 18 hours after surgery (after which rSO2i monitoring was discontinued). Despite this, the patient did not have postoperative neurologic complications and had a normal follow-up brain MRI study. In patients with left-right differences more than 10%, these were more frequent from 10 minutes after start of cooling on CPB to off bypass, i.e., at the higher end of the rSO2i scale where differences between sides would presumably not have the same adverse implications as at the lower end of the scale. Comparing these asymmetric pairs to their respective baseline values, in only 8% of these patients did the left or right rSO2i decrease more than 20% from baseline; in the other 92% of patients, the change in rSO2i was to increase more than the baseline value.
When advocating routine use of cerebral oximetry, the significant cost of the disposable sensors may limit widespread introduction of this form of regional monitoring. In the United States, the retail price of the INVOS 5100B is $25,000, and the disposable Somasensors are between $110 and $140, with the final cost being determined by the contract between the hospital and Somanetics (information from Somanetics). Although an estimated break-even analysis may justify a hospital expenditure for neurophysiologic monitoring of $2,142 per case (17), many practitioners are hesitant to incur additional intraoperative costs for monitoring without large-scale prospective clinical trials showing that it improves outcome. Although the potential of NIRS for preventing or predicting neurologic injury during pediatric cardiac surgery can only be established by large scale, prospective clinical trials, reducing the cost of cerebral oximetry by using a single sensor, may promote more widespread use of the technology.
In conclusion, bi-hemispheric measurements of regional cerebral oxygen saturation using NIRS were similar in neonates and infants undergoing biventricular repair without aortic arch reconstruction. Further longitudinal neurological outcome studies are required to determine whether uni- or bi-hemispheric monitoring is required during pediatric cardiac surgery.
We thank Gene Walters for monitoring and data collection, Ludmila Kyn for database and statistical programming, Donna Donati and Donna Duva for data management, and Kathy Alexander for project coordination.
|
Footnotes |
|---|
Accepted for publication May 5, 2005.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Dworschak Adequacy of Brain Oxygen Saturation for Goal-Directed Therapy Anesth. Analg., August 1, 2007; 105(2): 537 - 538. [Full Text] [PDF]
|
|
|
|
 |
|
 K. N. Fenton, K. Lessman, K. Glogowski, S. Fogg, and K. F. Duncan Cerebral Oxygen Saturation Does Not Normalize Until After Stage 2 Single Ventricle Palliation Ann. Thorac. Surg., April 1, 2007; 83(4): 1431 - 1436. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. D. Williams and C. Ramamoorthy Brain monitoring and protection during pediatric cardiac surgery. Seminars in Cardiothoracic and Vascular Anesthesia, March 1, 2007; 11(1): 23 - 33. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
