Neurological Monitoring for Congenital Heart Surgery : Anesthesia & Analgesia

Secondary Logo

Journal Logo

Pediatric Anesthesia: Review Article

Neurological Monitoring for Congenital Heart Surgery

Andropoulos, Dean B. MD; Stayer, Stephen A. MD; Diaz, Laura K. MD; Ramamoorthy, Chandra MB, BS, FFA (UK)

Author Information
Anesthesia & Analgesia 99(5):p 1365-1375, November 2004. | DOI: 10.1213/01.ANE.0000134808.52676.4D
  • Free

The reported incidence of neurological complications after heart surgery in children ranges from 2% to 25% (1–4). A retrospective report of 706 children undergoing heart surgery found that 2.3% had acute neurological complications (3). In contrast to adults undergoing heart surgery with cardiopulmonary bypass (CPB), who develop postoperative neurological sequelae that are largely embolic in nature, the etiology of neurological dysfunction in children is multifactorial (5–8). Techniques such as deep hypothermic circulatory arrest (DHCA) and low-flow bypass, which have allowed successful correction of complex cardiac defects in neonates and infants, may themselves contribute to neurological damage in this vulnerable population (9,10). Furthermore, the details of the conduct of CPB contribute to potential brain injury after CPB (7). Such details include arterial blood gas (ABG) management: α-stat (not correcting ABG for temperature) versus pH stat (correcting ABG for temperature), hematocrit on bypass, and the rate and extent of cooling and rewarming.

When assessing neurological outcome in children after open-heart surgery, clinicians and investigators often assume that these children are neurologically normal before surgery. However, central nervous system (CNS) malformations are more frequent in patients with congenital heart disease, specifically those with hypoplastic left heart syndrome (HLHS), where brain dysgenesis may approach 30% (6,11). In addition, children with chromosomal defects, particularly those with microdeletions of chromosome 22, have a more frequent incidence of CNS abnormalities, as do neonates with coarctation of aorta (12,13). Hence, these developmental brain disturbances contribute to the incidence of brain injury in the perioperative setting.

Although vital organs are routinely monitored during congenital heart surgery with CPB, the brain is not typically monitored. Any strategy for prevention or rescue from adverse neurological events during CPB must start with the routine use of neurological monitoring systems that allow easy, reliable, and reproducible detection of these events. Despite the availability of several modalities of monitoring for almost 20 years, in our opinion neurological monitoring during CPB remains in its infancy. In this article, we review the currently available modalities for neurological monitoring during congenital heart surgery in children and present preliminary evidence that such monitoring improves neurological outcome in this high-risk population.

Electroencephalographic Monitors

The electroencephalogram (EEG) during cardiac surgery in infants with CPB has been well characterized (14). The standard EEG using between 2–16 channels has been used in congenital heart surgery (4). The major disadvantage of using scalp EEG is the requirement for a dedicated technician to place the electrodes and interpret the EEG, which is neither practical nor cost-effective in busy congenital heart surgery centers. The EEG provides a rough guide to anesthetic depth and can be used to determine electrocerebral silence before using DHCA (15). The EEG is affected by several factors, including anesthetics, temperature, and CPB. Impracticalities of the use of an intraoperative EEG include electrical signal interference, complexity of placement and difficulty in interpretation. Newer devices using processed EEG technology are more user-friendly.

The Bispectral Index (BIS) monitor (Aspect Medical Systems, Newton, MA) is one device approved by the United States Food and Drug Administration (FDA) currently used to assess the depth of anesthesia. BIS sensor electrodes are applied to the forehead and temple, producing a frontal-temporal montage that connects to a processing unit. The device is easy to use; electrode placement is not difficult, and the monitor requires no calibration or warm up time. Via a proprietary algorithm of Aspect Corporation based on the normal adult EEG, BIS uses Fourier transformation and bispectral analysis of a one-channel processed EEG pattern to compute a single number, the BIS (16). This index ranges from 0 (isoelectric EEG) to 100 (awake) with mean awake values in the 90–100 range in adults, infants, and children (17). The real-time unprocessed EEG wave form of the BIS can be used to recognize EEG burst suppression or electrical silence, which could be useful in alerting the anesthesiologist of the need to continue cooling before the initiation of DHCA. Even this use has limitations; rarely, for example, an isoelectric EEG in the newborn infant may not correlate with a lack of all cortical activity (18). The BIS is subject to motion artifact, electromyographic activity, and radiofrequency interference from electrical equipment in the operating room. BIS values are anesthetic-specific; for example, the BIS value in children at 1 MAC halothane is 56, versus 36 for 1 MAC isoflurane, although at awakening the values are identical (19).

During CPB, hemodilution and temperature changes alter pharmacokinetics and pharmacodynamics of opioid anesthetics, which can lead to awareness under anesthesia (20). The overall incidence of awareness in adults undergoing cardiac surgery varies from 1.1% (21) to 23%, which is more than that reported for general surgical procedures (22,23). Although there are no documented reports of awareness under anesthesia in children undergoing heart surgery, BIS or other EEG monitoring is potentially useful to detect a level of anesthetic depth associated with awareness. In a cohort of children undergoing open-heart surgery with an anesthetic tailored for “fast-tracking,” BIS increased during rewarming, a period considered at risk for awareness under anesthesia (24). However, in this study, and in a similar study in infants <1 year of age, BIS did not correlate with stress hormone levels, a surrogate for light levels of anesthesia, nor with plasma fentanyl levels (25). Additional studies are needed to determine the utility of the BIS monitor in infants and children undergoing CPB with or without DHCA.

The Patient State Index (PSI) (Physiometrix, Inc, N. Billerica, MA) is a FDA-approved 4-channel processed EEG monitor also based on the adult EEG, which may have less dependence on the anesthetic used than the BIS (26). Comparison of the BIS and PSI in adults revealed a wide variation in individual values associated with loss of consciousness during induction of anesthesia, and there was a significant incidence of erroneous classification of the patient being unconscious when they could follow commands (27). The BIS and PSI demonstrated only about 70% accuracy in predicting loss or return of consciousness. Other drawbacks of the PSI include the larger area necessary for sensor placement on the front of the forehead compared with the BIS, which competes with other neurological monitors, such as near-infrared cerebral oximeter.

It is difficult to recommend EEG monitoring for routine use in infants and children undergoing pediatric CPB at this time. Reasons include the paucity of pediatric studies, particularly with the PSI, and the finding that EEG changes were responsible for only 5% of abnormal neurological monitor findings in a large pediatric study (4), and that these abnormal readings on the EEG did not appear to correlate with adverse postoperative outcomes.

Monitors of Cerebral Oxygenation

Near-Infrared Spectroscopy

Near-infrared spectroscopy (NIRS) is a noninvasive optical technique used to monitor brain tissue oxygenation. Most devices use 2–4 wavelengths of near-infrared light at 700–1000 nm, where the iron-porphyrin complexes of oxygenated and deoxygenated hemoglobin have distinct absorption spectra (28–30). Commercially available devices measure the concentrations of oxyhemoglobin and de-oxyhemoglobin and determine cerebral oxygen saturation. The cerebral oximeter probe is placed on the forehead below the hairline. A light emitting diode emits near-infrared light, which passes through a banana-shaped tissue volume of approximately 10 mL in the frontal cerebral cortex to 2 or 3 detectors placed 3–5 cm from the emitter. The terminology used to describe cerebral oxygen saturation as measured by NIRS varies, depending on the method and device used (2 versus 4 wavelengths, spatially resolved versus frequency domain spectroscopy). These terms include ScO2 for cerebral cortical oxygen saturation, TOI for tissue oxygenation index, HbO2 for oxygenated hemoglobin signal, and rSO2i for regional cerebral saturation index. In this review, rSO2i will be used to designate cerebral oxygen saturation for clarity and also because it is the designation used for the only FDA-approved device. Anatomical models predict that 75% of the cerebral blood volume in the light path is venous and 25% is arterial. Watzman et al. (31) attempted to verify this index in children with congenital heart disease by measuring jugular venous bulb and arterial saturations and comparing them with cerebral saturation measured with frequency-domain NIRS. The actual ratio in patients varied widely but averaged 85:15 venous:arterial blood ratio.

The proximal detector in the Somanetics INVOS system (Somanetics, Inc. Troy, MI) detects light absorbed by extracranial tissues and is subtracted from the total signal (detected by the distal electrode), leaving only the intracranial contribution. The pediatric model (INVOS 5100) is designed for patients weighing 4–40 kg and uses a different algorithm that takes into account the thinner skull and extracranial tissues of infants and children (32). This device is FDA-approved for use in children, is compact, simple to use, and requires no warm up. It uses two near-infrared wavelengths: 730 and 810 nm. The INVOS processor displays a numerical value; the rSO2i, which is the ratio of oxyhemoglobin to total hemoglobin in the light path. The rSO2i is reported as a percentage on a scale from 15% to 95%. A cerebral blood volume index, representing the total hemoglobin in the light path, can also be calculated. This measurement may be used as an estimate of cerebral blood volume, although currently it can be used for research purposes only because it is not FDA approved owing to insufficient data correlating it with other measurements of cerebral blood volume.

The NIRO 300 (Hamamatsu Photonics, Hamamatsu, Japan) uses spatially resolved spectrophotometry to calculate absolute concentrations of oxygenated hemoglobin and total hemoglobin rather than hemoglobin saturations. Theoretically, this could lead to improved accuracy; however, this device is not FDA approved. Prototype cerebral oximeters using frequency-domain technology are under development and have the potential to measure absolute rather than calculated rSO2i (28).

Direct comparison of the INVOS 4100 and NIRO 300 in healthy anesthetized adults during normocapnia, hypocapnia, and hypercapnia revealed a positive correlation for data points in both absolute values, and change from baseline values (r = 0.58 and 0.85, respectively) (25). Bland-Altman comparison of the 2 devices demonstrated that although the correlation was good when all 60 values were combined, there may be large differences by individual comparison (33). The INVOS 5100 pediatric sensor tends to read significantly higher (by 14% ± 8%) than the NIRO 300 4 cm interoptode sensor (pediatric equivalent) (31). Regardless of the device used, it is important to note that all devices measure combined arterial and venous blood oxygen saturation and the values obtained will therefore not be identical to measured jugular venous bulb oxygen saturation (SjvO2). Therefore, maneuvers that increase arterial oxygen saturation, e.g., increasing Fio2, will increase cerebral oxygenation as measured by these devices, but may not change the SjvO2. These devices are best used as trend monitors, with each patient serving as his or her own control, and a 20% relative decrease from baseline may be a clinically important change in the Somanetics system (4). Newer software versions continuously calculate and display the relative change from baseline and allow labeling of significant events, such as the start of CPB or DHCA.

In an attempt to validate the noninvasive measurement of rSO2i in children with congenital heart disease SjvO2 and rSO2i have been compared. In one study of 40 infants and children undergoing congenital heart surgery or cardiac catheterization, the correlation for paired measurements was poor except for infants less than 1 year of age (34). In another study of 30 infants and children undergoing cardiac catheterization, an excellent correlation (r = 0.93) was found (35). Greeley et al. (36) derived cerebral metabolic rate for oxygen consumption (CMRO2) from cerebral blood flow measurements calculated using 133Xe washout technology and jugular venous and arterial oxygen saturation measurements in infants and children undergoing congenital heart surgery. They compared CMRO2 to changes in oxygenated hemoglobin (HbO2) signal measured by NIRS and demonstrated important parallels. After DHCA, cerebral blood flow and CMRO2 were depressed, whereas HbO2 was increased when compared with patients undergoing continuous bypass without DHCA. Recovery from DHCA produced low CMRO2 and a limited ability to use oxygen, resulting in an increased HbO2 signal.

Near-infrared light is also absorbed by Cu++ atoms of cytochrome a,a3, making NIRS a potentially useful monitor of cerebral intraneuronal oxygen delivery by assessing the redox state of cytochrome molecules in the electron chain of the mitochondria. The Hamamatsu NIRO 300 can measure cytochrome redox state. However, the light absorption by hemoglobin is an order of magnitude greater than that of cytochrome a,a3, and the cytochrome a,a3 absorption signal is highly dependent on hematocrit, bringing into question the validity of these readings (37).

Animal Models with Relevance to Pediatric Cardiac Surgery

Animal models have well known limitations; however, important data have been generated using such models that support the use of NIRS as a monitor that measures meaningful changes in brain oxygenation that reflect disturbances in neuronal functioning. Low rSO2i can lead to cell dysfunction, cell death, and by extension, poor neurological outcome in these models. These studies cannot be done in humans for obvious reasons. In a model of graded cerebral hypoxemia at normothermia in neonatal piglets with a baseline rSO2i of 68%, Kurth et al. (38) determined that the threshold for half of the animals to experience an increased cerebral lactate concentration was a rSO2i of 44%. Minor EEG changes occurred in half of the animals at an rSO2i of 42%, major EEG change at an rSO2i of 37%, and decreased brain adenosine triphosphate in half of the animals at an rSO2i of 33%. rSO2i also strongly correlated with sagittal sinus oxygen saturation, cerebral blood flow measured by laser Doppler flowmetry, and Pao2.

In piglet models of DHCA, Sakamoto et al. (39,40) used NIRS under varying conditions: differing hematocrits (20% versus 30%) and temperatures (15°C versus 25°C), pH-stat versus α-stat blood gas management, and varying lengths of DHCA (60–100 minutes). They determined that the nadir of rSO2i occurred sooner at higher temperatures, lower hematocrits, and with α-stat blood gas management. The time spent at the nadir of rSO2i correlated with neurological outcome determined both behaviorally and on neuropathologic examination: all animals with rSO2i nadir time <25 minutes were free of any evidence of neurological injury.

In another piglet model, NIRS detected cerebral desaturation when the superior vena cava (SVC) was partially or totally occluded during CPB. No other measurements were affected by this maneuver: arterial blood pressure, heart rate, SVC pressure measurements in the case of partial occlusion, or mixed venous oxygen saturation measured in the bypass circuit. Cerebral desaturation may develop in infants if they have SVC obstruction during SVC cannulation for bypass or in patients undergoing cavopulmonary anastomosis, where the SVC is often partially occluded (41).

Clinical Data in Pediatric Cardiac Surgery

Baseline preoperative rSO2i as measured by a frequency-domain oximeter varies with different cardiac lesions (29). The baseline rSO2i breathing room air is approximately 70% in acyanotic patients without large left to right intracardiac shunts; rSO2i for cyanotic patients is usually 40%–60%. Ramamoorthy et al. (42) exposed neonates with unrepaired HLHS to two different inspired gas mixtures to control pulmonary overcirculation. Those treated with 17% inspired oxygen had rSO2i averaging 53%, and those who received 21% oxygen and 3% inspired CO2 had rSO2i averaging 68%; however, there were no significant differences in the arterial saturation between groups. Significant decreases in rSO2i occur during periods of hemodynamic instability or arterial desaturation (43). Increases in rSO2i occur during cooling and vary with the rate of temperature change. Improvements in cardiac output and oxygen delivery result in an increase in rSO2i. Characteristic changes in cerebral oxygenation occur during CPB in children with or without DHCA (Fig. 1) (43).

Figure 1.:
Typical changes in regional cerebral oxygen saturation index (rSO2i) in congenital heart surgery with cardiopulmonary bypass (CPB) with a) mild hypothermia, b) moderate hypothermia, c) deep hypothermia with regional low-flow cerebral perfusion (RLFP), and d) deep hypothermic circulatory arrest (DHCA). pH stat blood gas management was used in all cases. Cases b–d had slow cooling over 20–30 min, and hematocrit 25%–30% when the lowest temperature was reached. Hemofiltration and transfusion were used to achieve hematocrit 40%–45% before separation from bypass for cases b–d. Figure 1a: rSO2i in a 13-yr-old undergoing right ventricle to pulmonary artery conduit replacement at 32°C with hematocrit 22% on bypass. With the onset of bypass, acute hemodilution from hematocrit of 35% to 22% resulted in a decrease in rSO2i from 74% to 45%. Cooling to 32°C increased rSO2i to >60%. b, rSO2i in a 7-mo-old undergoing repair of complete atrioventricular canal with moderate hypothermic bypass to 28°C. A brief period (60 s) of circulatory arrest was used when the aortic cross-clamp (Ao xcl) was applied at 31°C, resulting in a decrease in rSO2i from 75% to 48%. c, typical changes in right and left-sided rSO2i during cardiac surgery with CPB, regional cerebral low flow perfusion, and DHCA at 18°C, in a neonate undergoing the Norwood Stage I palliation for hypoplastic left heart syndrome. There is a brief period of DHCA before and after the period of RLFP, and the left-sided rSO2i is lower than the right during RLFP. d, rSO2i in a neonate with unbalanced, left-ventricle dominant complete atrioventricular canal and infradiaphragmatic total anomalous pulmonary venous return. This patient underwent four periods of DHCA at 18°C of 3, 12, 26, and 1 min, for a total of 42 min.

Predictably, rSO2i decreases during DHCA to a nadir approximately 60%–70% (relative change) less than baseline values obtained prebypass and the nadir is reached at 20–40 minutes, after which there is no further decrease (44). At this point it appears that there is no additional oxygen uptake by the brain. Interestingly, this time period correlates with clinical and experimental studies suggesting that 45 minutes is the safe duration for circulatory arrest (7). Reperfusion results in an increase in rSO2i to levels seen at full bypass flow before DHCA. Andropoulos et al. (45) used bilateral NIRS to assess changes in rSO2i during regional low-flow cerebral perfusion (RLFP) in neonatal patients undergoing aortic arch reconstruction. RLFP is a technique in which the brain is perfused via the right innominate artery, providing unilateral cerebral perfusion. During periods of perfusion via both carotid arteries (i.e., before and after RLFP, whether on or off bypass), bilateral rSO2i values agreed closely, whereas during RLFP the mean rSO2i was 6.3% lower on the left side, and about half of the patients experienced an absolute difference of more than 10% in rSO2i values during RLFP, suggesting bilateral monitoring is useful during unilateral perfusion techniques, but not necessary when both carotid arteries are perfused under normal circumstances.

Recent studies using α-stat blood gas management for simpler surgeries (i.e., ventricular septal defect or tetralogy of Fallot repair) and mild or moderate hypothermia of 25°C–35°C have demonstrated a lower rSO2i in the patients undergoing bypass with colder temperatures (25°C–28°C) versus warmer bypass at 32°C–35°C (46,47). This suggests that the cerebral vasoconstriction from decreased Paco2 at the lower temperature decreases oxygen delivery more than the effect of cooling to decrease CMRO2, resulting in lower rSO2i. These studies demonstrate that NIRS may play a role in more moderate hypothermia because approximately 30% of patients in these studies experienced significant cerebral oxygen desaturation (absolute rSO2i <50%, or more than 20% relative decrease from baseline). Finally, in a study of cyanotic patients with aortopulmonary collaterals undergoing hypothermic CPB and randomized to α-stat versus pH stat blood gas management, rSO2i was significantly lower in the α-stat group (48).

Relationship Between Low rSO2i and Adverse Neurological Outcome

Clinical evidence suggests a correlation between low rSO2i and adverse neurological outcome. A study of 26 infants and children undergoing surgery with bypass and DHCA found that the 3 patients with low rSO2i had acute postoperative neurological changes: seizures in 1 and prolonged coma in 2 (44). In these 3 patients the increase in rSO2i was much less after the onset of CPB (average 3% relative increase versus 33% increase in patients without neurological deficit) and the duration of cooling before DHCA was shorter. Austin et al. (4) studied 250 pediatric patients undergoing cardiac surgery with bypass and found that 41% of patients experienced prolonged periods of relative cerebral oxygen desaturation, defined as more than 20% decrease less than prebypass baseline. One-quarter of these patients with cerebral desaturation had postoperative adverse neurological events consisting of prolonged coma, seizures, or hemiparesis.

It would be intuitive to conclude that low rSO2i as measured by NIRS leads to adverse neurological outcomes and therefore should be monitored and treated. The studies cited above support this notion. However, additional prospective outcome data in infants and children using this modality are necessary. An additional limitation of the NIRS monitor is that it measures rSO2i in only a small tissue volume of the frontal cerebral cortex; other areas of the cortex and deeper structures in the brain are not assessed by this technique. It would be useful to attempt to correlate abnormal rSO2i with the frequency of changes detected by magnetic resonance imaging after CPB in newborns, specifically periventricular leukomalacia (49).

Other Uses of NIRS

Other important areas of investigation include correlating the rSO2i with mixed venous oxygen saturation (SvO2) to assess whether it can be used as a noninvasive measure of SvO2. Preliminary data suggest that measured SVC HbO2 saturation correlates well in infants with congenital heart disease after cardiac surgery (50). Other potential uses of this technology include measuring HbO2 saturation in subdiaprhagmatic viscera (e.g., kidney and intestines). There is evidence that these values decline significantly during aortic cross-clamping for coarctation repair (51) and are preserved during cerebral RLFP; this may have implications for end organ function (52,53).

Transcranial Doppler Ultrasound

Transcranial Doppler ultrasound (TCD) is a sensitive, real-time monitor of cerebral blood flow velocity (CBFV) and emboli during congenital heart surgery. Currently available instruments use pulsed-wave ultrasound at 2 MHz frequency that is range-gated, emits a power of 100 mW, and has a sample volume length of up to 15 mm. A display of the frequency spectrum of Doppler signals is easily interpreted, and peak systolic and mean flow velocities, in cm/s, are displayed, as well as a pulsatility index that is equal to the peak velocity minus the end-diastolic velocity, divided by the mean velocity. As with cardiac ultrasound, the advantage of pulsed-wave Doppler ultrasound is that a precise sample volume can be selected which insonates only the arteries of interest without contamination from other sources.

The most consistent and reproducible technique for clinical use in patients of all ages is to monitor the middle cerebral artery (MCA) through the temporal window, which can usually be found just above the zygoma and just anterior to the tragus of the ear (54). Several transducer probes are available, ranging from very small disk probes suitable for infants and children, to larger, heavier probes for adolescents and adults. The depth of the sample volume and angle of insonation is adjusted until the bifurcation of the MCA and the anterior cerebral artery (ACA) is detected. This is heralded by a maximal antegrade signal (positive deflection, toward the transducer) from the MCA, accompanied by a retrograde signal from the ACA (negative deflection, away from the transducer) of the same or very similar velocity and wave form as the MCA flow.

In infants, an alternative site for monitoring is through the anterior fontanelle, using a hand-held pencil-type probe, placing the probe over the lateral edge of the fontanelle, and aiming caudally, at a greater depth than for the temporal window, at the internal carotid artery. The depth of measurement and normal flow velocities for the MCA through the temporal window are listed in Table 1 (55). These normal velocities were determined in awake children without cardiovascular disease, under perfect examination conditions. Lesions producing large diastolic runoff (e.g., large patent ductus arteriosus) will decrease diastolic blood flow to the brain. Hemodynamic instability, less than optimal probe positioning, and general anesthesia may reduce these velocities in clinical practice. Typically the clinician must accept a stable baseline for the individual patient and use it as the basis for comparison. Fixation of the probe in the optimal position is problematic. There are several commercial devices available for fixation in larger patients (Table 2), but in practice it is often simpler to affix a small disk Doppler probe with a clear adhesive dressing and clear tape over the temporal window of infants and children, assuring access and some mobility to the probe for the frequent adjustments that may be necessary. Currently available FDA approved TCD manufacturers are listed in Table 2.

Table 1:
Normal Transcranial Doppler Velocities for Infants and Children
Table 2:
Transcranial Doppler Manufacturers

TCD has been used extensively in pediatric cardiac surgical research to examine cerebral physiology in response to CPB, hypothermia, low-flow bypass, RLFP to the brain, and DHCA. Hillier et al. (56) used TCD to study cerebrovascular hemodynamics during hypothermic bypass with DHCA in 10 infants. CBFV did not return to baseline levels after DHCA. Calculated cerebral vascular resistance (mean arterial blood pressure − central venous pressure/CBFV) was increased immediately after DHCA and remained increased until the end of bypass. The observed decrease in CBFV during cooling was presumed to be attributable to decreased metabolic demand by the brain and thus less blood flow. In this study an α-stat blood gas strategy was used, which can produce relative cerebral vasoconstriction during cooling in smaller arterioles downstream to the MCA and ACA (57). TCD of the MCA through the temporal window was used to describe the cerebral pressure-flow velocity relationship during hypothermic bypass in 25 neonates and infants <9 months old. CBFV was examined over a wide range of cerebral perfusion pressure varying from 6 to 90 mm Hg at 3 temperatures: normothermia (36°C–37°C), moderate hypothermia (23°C–25°C), and profound hypothermia (14°C–20°C). Cerebral pressure flow autoregulation was preserved with normothermia, partially affected with moderate hypothermia, and totally lost with profound hypothermia; these results agree with previous research performed using xenon washout to quantitate cerebral blood flow (58).

TCD has also been used to determine the threshold of detectable cerebral perfusion during low-flow CPB. Zimmerman et al. (59) studied 28 neonates during bypass undergoing the arterial switch operation with α-stat blood gas management. At 14°C–15°C, bypass flow was sequentially reduced to 0 mL · kg−1 · min−1. All patients had detectable CBFV more than 20 mL · kg−1 · min−1, although one had no perfusion at 20 mL · kg−1 · min−1 and 8 had none at 10 mL · kg−1 · min−1, leading the authors to conclude that 30 mL · kg−1 · min−1 was the minimum acceptable flow in this population. Andropoulos et al. (60) used TCD of the MCA to determine the level of bypass flow necessary during RLFP for neonatal aortic arch reconstruction. They studied 34 neonates and established a baseline mean CBFV under full flow bypass (150 mL · kg−1 · min−1) using pH stat management at 17°C–22°C: a mean of 22 cm/s. These authors then used the TCD signal to determine the necessary bypass flow during RLFP, a mean of 63 mL · kg−1 · min−1. Interestingly, this level of bypass flow did not correlate with mean arterial blood pressure in the radial artery or rSO2i measured by NIRS. The necessary CPB flow as determined by TCD varied widely, leading the authors to conclude that TCD was a valuable monitor to ensure adequate but not excessive cerebral blood flow during this complicated perfusion technique. The TCD is a useful monitor to detect acute decreases in CBFV; this can allow for adjustment of bypass cannulae, which might avert a neurological disaster (61,62).

Isolated cerebral emboli are a frequent occurrence during open-heart surgery in children. Emboli are easily detected by TCD, although false positive artifacts from electrocautery and physical contact with the ultrasound transducer occur (63). True emboli have characteristic audio and visual signals and are designated as high intensity transient signals (HITS) that can be counted by the TCD software. The HITS counter can be an accurate gauge of the number of emboli detected in the artery being monitored; however, it does not distinguish false positive artifacts, and if a large number of emboli are detected in a short period of time, the number counted is meaningless. In actual practice, the anesthesiologist can observe the monitor intermittently and decide whether the number of emboli detected is clinically significant and initiate treatment. Despite the potential utility of this monitor to detect emboli, one study found that the number of emboli detected in the carotid artery during pediatric congenital heart surgery does not appear to correlate with acute postoperative neurological deficits (63). Another limitation of emboli detection is that it only occurs after the fact; however, limitation of further emboli may be possible with this monitor.

One caveat when usng TCD clinically is that this device measures CBFV rather than blood flow. CBFV is dependent on the diameter of the blood vessel, whereas cerebral blood flow depends on cerebral vascular resistance, which changes in response to changes in CO2, temperature, cerebral perfusion pressure, and bypass flow. Thus, changes in CBFV often correlate well with changes in cerebral blood flow in the individual patient, particularly at deep hypothermia when autoregulation is lost and the caliber of the blood vessels is unchanged. However, the clinician must always estimate the state of the patient’s cerebral vascular resistance to translate TCD into meaningful information for clinical decision-making.

Multimodality Neurological Monitoring

Simultaneous neurological monitoring—NIRS, TCD, and processed EEG—may hold the greatest promise in detecting and correcting neurological abnormalities during congenital heart surgery, just as pulse oximetry combined with capnography is more effective at preventing morbidity from ventilation mishaps than either modality alone (64). With combined use of TCD to measure blood flow in the MCA/ACA and NIRS to measure rSO2i of blood in the frontal lobe, it is possible to monitor up to 70% of the blood flow distribution to a cerebral hemisphere. A study by Austin et al. (4) showed that 70% (176 of 250) of pediatric patients experienced a significant change in one or more variables when multimodality neurological monitoring—NIRS, TCD, and four-channel qualitative EEG—was used during cardiac surgery. Abnormal neurological monitor values were not treated during the initial observational phase of this study, when multimodality monitoring was of unproven benefit and 26% of patients experienced adverse early neurological outcomes (seizures, prolonged coma, or hemiparesis). As experience was gained with multimodality monitoring and abnormal values appeared to correlate with adverse outcomes, a treatment algorithm was developed. The investigators in this study originally intended to randomize patients to treatment according to the algorithm or to receive standard care without monitoring. However, given the correlation of adverse outcomes with abnormal neurological monitor values, the investigators felt that it would be unethical to leave the obvious changes they observed untreated. They then began the next phase of the study, which treated patients according to the algorithm. When the abnormal monitoring values were treated, the incidence of adverse neurological outcomes decreased to 7%. This decreased incidence was very similar to that in the group that did not experience abnormal values on their neurological monitors (6%).

In this study EEG changes were responsible for only 5% of the monitoring abnormalities, with NIRS changes responsible for 58% and TCD 37%. The 26% incidence of acute neurological complications in the pretreatment phase of this study is more than most other reports of neurological outcomes after congenital heart surgery (3). Although this study was prospective, it was neither randomized nor controlled. However, it represents the best evidence that multimodality neurological monitoring, particularly NIRS and TCD, in conjunction with a treatment algorithm can reduce adverse neurological outcomes during congenital heart surgery. A modification of the treatment algorithm used in the Austin et al. study (4) is presented in Table 3. These modifications account for the anesthesiologist being the interpreter of the monitors (instead of a dedicated technician as used in the Austin et al. study) and the use of these monitors in the pre- and post-CPB periods. Eliminating the need for a separate technician obviates the problems of technician availability and third-party reimbursement for placement and interpretation of the neurological monitors. It should be noted that although low rSO2i values have been correlated with acute gross neurological outcomes, there are no data correlating low rSO2i with either more subtle neuropsychological dysfunction or long-term neurodevelopmental outcomes in congenital heart disease.

Table 3:
Neurological Monitoring Treatment Algorithm for Congenital Heart Surgery

In patients younger than 2 years of age the BIS is not used not only because of space limitations on the forehead but also because of insufficient data confirming its utility in this age group. Anesthesiologists can learn to place and interpret these monitors and treat the patient according to the algorithm presented in this review or another algorithm. For the NIRS monitor, a relative decrease of rSO2i of more than 20% from a stable baseline obtained preincision is the primary indication for treatment. Other critical values are an absolute value of rSO2i <30% or an absolute value of 95% accompanied by a significant increase in CBFV. It should be noted that the recommendation to reperfuse the patient at a NIRS nadir of 30 minutes will not be possible in some cases; i.e., the Norwood Stage I palliation where the aorta is opened during reconstruction. In our institutions, low rSO2i values are responsible for approximately 90% of the abnormal monitor values that result in treatment and frequently correlate with decreases in CBFV values. The TCD alone is responsible for approximately 10% of interventions, either as the result of cerebral emboli or significant increases in CBFV during deep hypothermia when the rSO2i is at its maximum value of 95%, placing the patient at risk for overperfusion to the brain. The BIS values are rarely responsible for initiating treatment. BIS values of >80 are frequently observed during rewarming on CPB, but BIS is typically not altered by benzodiazepines and opioids administered into the bypass circuit. It is only after bypass when small-dose volatile anesthetic gases are administered that the BIS value declines, again emphasizing the dependence of the BIS values on anesthetic. We have not recognized a case of awareness under anesthesia during rewarming with BIS >80.

Adverse Effects of Neurological Monitoring

No adverse effects from intraoperative neurological monitoring have been reported. In clinical practice of up to 12 hours use, there may be minor skin irritation associated with the adhesives used for the NIRS and BIS monitors and fixation devices for the TCD, but these are self-limited. The TCD should be used at the lowest power (i.e., scale setting) possible and should not be directed near the orbit because of potential effects on the eye (65).


The anesthesiologist can now monitor the brain during congenital heart surgery with commonly available clinical monitors, potentially improving neurological outcome with appropriate treatment strategies. More careful outcome studies, assessing both acute and long-term neurodevelopmental changes, including the common neuropsychological dysfunction seen after bypass, are clearly needed. These outcome studies will be complicated by the frequent occurrence of congenital neurological abnormalities and by neurological events occurring in the perioperative period outside of the operating room. Other problems with outcome studies include the fact that gross early neurological insults are far less common in the modern era in experienced centers (3). This improvement in acute outcomes has occurred in the absence of the use of neurological monitoring in most centers. The study size necessary to prove a significant decrease from the current 1%–2% incidence of early adverse events will be difficult to achieve. It may be that the greatest benefit will be in the subtler, long-term neurodevelopmental outcomes, which are difficult to quantify but of obvious importance to the patients’ quality of life. Isolated case reports of the prevention of neurological disasters during CPB (66) or of changing the surgical approach (67) add to the evidence of the benefit of using neurological monitoring. All of this evidence taken together suggests that neurological monitoring should become routine during congenital heart surgery. If the anesthesiologist places the monitors and interprets the monitoring values, the cost savings in preventing a small number of neurological complications with routine monitoring will more than pay for the cost of the monitoring hardware and disposable sensors (4). The vulnerable immediate postoperative period in high-risk patients should also be studied. Because a high survival rate after complex heart disease in children is no longer in question, the goal should be to improve the quality of life of these children with complex heart disease through better detection of neurological events and their expeditious treatment.


1. Fallon P, Aparicio JM, Elliott MJ, Kirkham FJ. Incidence of neurological complications of surgery for congenital heart disease. Arch Dis Child 1995;72:418–22.
2. Ferry PC. Neurologic sequelae of cardiac surgery in children. Am J Dis Child 1987;141:309–12.
3. Menache CC, du Plessis AJ, Wessel DL, et al. Current incidence of acute neurologic complications after open-heart operations in children. Ann Thorac Surg 2002;73:1752–8.
4. Austin EH III, Edmonds HL Jr, Auden SM, et al. Benefit of neurophysiologic monitoring for pediatric cardiac surgery. J Thorac Cardiovasc Surg 1997;114:707–15, 717.
5. Murkin JM. Etiology and incidence of brain dysfunction after cardiac surgery. J Cardiothorac Vasc Anesth 1999;13:12–7.
6. Kirkham FJ. Recognition and prevention of neurological complications in pediatric cardiac surgery. Pediatr Cardiol 1998;19:331–45.
7. du Plessis AJ. Mechanisms of brain injury during infant cardiac surgery. Semin Pediatr Neurol 1999;6:32–47.
8. Wernovsky G, Newburger J. Neurologic and developmental morbidity in children with complex congenital heart disease. J Pediatr 2003;142:6–8.
9. Newburger JW, Jonas RA, Wernovsky G, et al. A comparison of the perioperative neurologic effects of hypothermic circulatory arrest versus low-flow cardiopulmonary bypass in infant heart surgery. N Engl J Med 1993;329:1057–64.
10. Bellinger DC, Jonas RA, Rappaport LA, et al. Developmental and neurologic status of children after heart surgery with hypothermic circulatory arrest or low-flow cardiopulmonary bypass. N Engl J Med 1995;332:549–55.
11. Glauser TA, Rorke LB, Weinberg PM, Clancy RR. Congenital brain anomalies associated with the hypoplastic left heart syndrome. Pediatrics 1990;85:984–90.
12. Bingham PM, Zimmerman RA, McDonald-McGinn D, et al. Enlarged Sylvian fissures in infants with interstitial deletion of chromosome 22q11. Am J Med Genet 1997;74:538–43.
13. Young RS, Liberthson RR, Zalneraitis EL. Cerebral hemorrhage in neonates with coarctation of the aorta. Stroke 1982;13:491–4.
14. Miller G, Rodichok LD, Baylen BG, Myers JL. EEG changes during open heart surgery on infants aged 6 months or less: relationship to early neurologic morbidity. Pediatr Neurol 1994;10:124–30.
15. Akiyama T, Kobayashi K, Nakahori T, et al. Electroencephalographic changes and their regional differences during pediatric cardiovascular surgery with hypothermia. Brain Dev 2001;23:115–21.
16. Sigl JC, Chamoun NG. An introduction to bispectral analysis for the electroencephalogram. J Clin Monit 1994;10:392–404.
17. Denman WT, Swanson EL, Rosow D, et al. Pediatric evaluation of the bispectral index (BIS) monitor and correlation of BIS with end-tidal sevoflurane concentration in infants and children. Anesth Analg 2000;90:872–7.
18. Ashwal S. Brain death in the newborn. Current perspectives. Clin Perinatol 1997;24:859–82.
19. Davidson AJ, Czarnecki C. The Bispectral Index in children: comparing isoflurane and halothane. Br J Anaesth 2004;92:14–7.
20. Koren G, Crean P, Klein J, et al. Sequestration of fentanyl by the cardiopulmonary bypass (CPBP). Eur J Clin Pharmacol 1984;27:51–6.
21. Phillips AA, McLean RF, Devitt JH, Harrington EM. Recall of intraoperative events after general anaesthesia and cardiopulmonary bypass. Can J Anaesth 1993;40:922–6.
22. Goldmann L, Shah MV, Hebden MW. Memory of cardiac anaesthesia. Psychological sequelae in cardiac patients of intra-operative suggestion and operating room conversation. Anaesthesia 1987;42:596–603.
23. Dowd NP, Cheng DC, Karski JM, et al. Intraoperative awareness in fast-track cardiac anesthesia. Anesthesiology 1998;89:1068–73.
24. Laussen PC, Murphy JA, Zurakowski D, et al. Bispectral index monitoring in children undergoing mild hypothermic cardiopulmonary bypass. Paediatr Anaesth 2001;11:567–73.
25. Kussman BD, Gruber EM, Zurakowski D, et al. Bispectral index monitoring during infant cardiac surgery: relationship of BIS to the stress response and plasma fentanyl levels. Paediatr Anaesth 2001;11:663–9.
26. Prichep LS, Gugino LD, John ER, et al. The Patient State Index as an indicator of the level of hypnosis under general anaesthesia. Br J Anaesth 2004;92:393–9.
27. Schneider G, Gelb AW, Schmeller B, et al. Detection of awareness in surgical patients with EEG-based indices–bispectral index and patient state index. Br J Anaesth 2003;91:329–35.
28. Yoshitani K, Kawaguchi M, Tatsumi K, et al. A comparison of the INVOS 4100 and the NIRO 300 near-infrared spectrophotometers. Anesth Analg 2002;94:586–90.
29. Kurth CD, Steven JM, Nicolson SC, et al. Kinetics of cerebral deoxygenation during deep hypothermic circulatory arrest in neonates. Anesthesiology 1992;77:656–61.
30. Kurth CD, Steven JL, Montenegro LM, et al. Cerebral oxygen saturation before congenital heart surgery. Ann Thorac Surg 2001;72:187–92.
31. Watzman HM, Kurth CD, Montenegro LM, et al. Arterial and venous contributions to near-infrared cerebral oximetry. Anesthesiology 2000;93:947–53.
32. Dullenkopf A, Frey B, Baenziger O, et al. Measurement of cerebral oxygenation state in anaesthetized children using the INVOS 5100 cerebral oximeter. Paediatr Anaesth 2003;13:384–91.
33. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8:135–60.
34. Daubeney PE, Pilkington SN, Janke E, et al. Cerebral oxygenation measured by near-infrared spectroscopy: comparison with jugular bulb oximetry. Ann Thorac Surg 1996;61:930–4.
35. Abdul-Khaliq H, Troitzsch D, Berger F, Lange PE. Regional transcranial oximetry with near infrared spectroscopy (NIRS) in comparison with measuring oxygen saturation in the jugular bulb in infants and children for monitoring cerebral oxygenation. Biomed Tech (Berl) 2000;45:328–32.
36. Greeley WJ, Bracey VA, Ungerleider RM, et al. Recovery of cerebral metabolism and mitochondrial oxidation state is delayed after hypothermic circulatory arrest. Circulation 1991;84:400–6.
37. Sakamoto T, Jonas RA, Stock UA, et al. Utility and limitations of near-infrared spectroscopy during cardiopulmonary bypass in a piglet model. Pediatr Res 2001;49:770–6.
38. Kurth CD, Levy WJ, McCann J. Near-infrared spectroscopy cerebral oxygen saturation thresholds for hypoxia-ischemia in piglets. J Cereb Blood Flow Metab 2002;22:335–41.
39. Sakamoto T, Zurakowski D, Duebener LF, et al. Combination of alpha-stat strategy and hemodilution exacerbates neurologic injury in a survival piglet model with deep hypothermic circulatory arrest. Ann Thorac Surg 2002;73:180–9.
40. Sakamoto T, Hatsuoka S, Stock UA, et al. Prediction of safe duration of hypothermic circulatory arrest by near-infrared spectroscopy. J Thorac Cardiovasc Surg 2001;122:339–50.
41. Sakamoto T, Duebener LF, Tayler G, et al. Obstructed SVC cannula may result in cerebral ischemia which can be detected by near-infrared spectroscopy during CPB [abstract]. Anesth Analg 2002;93:SCA 82.
42. Ramamoorthy C, Tabbutt S, Kurth CD, et al. Effects of inspired hypoxic and hypercapnic gas mixtures on cerebral oxygen saturation in neonates with univentricular heart defects. Anesthesiology 2002;96:283–8.
43. Daubeney PE, Smith DC, Pilkington SN, et al. Cerebral oxygenation during paediatric cardiac surgery: identification of vulnerable periods using near infrared spectroscopy. Eur J Cardiothorac Surg 1998;13:370–7.
44. Kurth CD, Steven JM, Nicolson SC. Cerebral oxygenation during pediatric cardiac surgery using deep hypothermic circulatory arrest. Anesthesiology 1995;82:74–82.
45. Andropoulos DB, Diaz LK, Fraser CD Jr, et al. Is bilateral monitoring of cerebral oxygen saturation necessary during neonatal aortic arch reconstruction? Anesth Analg 2004;98:1267–72.
46. Morimoto Y, Niida Y, Hisano K, et al. Changes in cerebral oxygenation in children undergoing surgical repair of ventricular septal defects. Anaesthesia 2003;58:77–83.
47. Shaaban AM, Harmer M, Elliott M, et al. A pilot study of evaluation of cerebral function by S100 beta protein and near-infrared spectroscopy during cold and warm cardiopulmonary bypass in infants and children undergoing open-heart surgery. Anaesthesia 2004;59:20–6.
48. Sakamoto T, Kurosawa H, Shin’oka T, et al. The influence of pH strategy on cerebral and collateral circulation during hypothermic cardiopulmonary bypass in cyanotic patients with heart disease: results of a randomized trial and real-time monitoring. J Thorac Cardiovasc Surg 2004;127:12–9.
49. Mahle WT, Tavani F, Zimmerman RA, et al. An MRI study of neurological injury before and after congenital heart surgery. Circulation 2002;106:109–14.
50. Nagdyman N, Fleck T, Barth S, et al. Relation of cerebral tissue oxygenation index to central venous oxygen saturation in children. Intensive Care Med 2004;30:468–71.
51. Berens RJ, Hoffman GM, Robertson FA, et al. Near infrared oximetry (NIRS) allows noninvasive assessment of regional oxygenation during pediatric coarctation repair [abstract]. Anesthesiology 2003;99:A1386.
52. Pigula FA, Gandhi SK, Siewers RD, et al. Regional low-flow perfusion provides somatic circulatory support during neonatal aortic arch surgery. Ann Thorac Surg 2001;72:401–6.
53. Hoffman GM, Stuth EA, Jaquiss RD, 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.
54. Fischer AQ, Truemper EJ. Applications in the neonate and child. In: Babikian VL, Wechlser LR, eds. Transcranial Doppler ultrasonography. St. Louis: Mosby, 1993.
55. Truemper EJ, Fischer AZ. Cerebrovascular developmental anatomy and physiology in the infant and child. In: Babikian VL, Wechlser LR, eds. Transcranial Doppler ultrasonography. St. Louis: Mosby, 1993:355–75.
56. Hillier SC, Burrows FA, Bissonnette B, Taylor RH. Cerebral hemodynamics in neonates and infants undergoing cardiopulmonary bypass and profound hypothermic circulatory arrest: assessment by transcranial Doppler sonography. Anesth Analg 1991;72:723–8.
57. Huber P, Handa J. Effect of contrast material, hypercapnia, hyperventilation, hypertonic glucose and papaverine on the diameter of the cerebral arteries: angiographic determination in man. Invest Radiol 1967;2:17–32.
58. Greeley WJ, Ungerleider RM, Kern FH, et al. Effects of cardiopulmonary bypass on cerebral blood flow in neonates, infants, and children. Circulation 1989;80:I209–15.
59. Zimmerman AA, Burrows FA, Jonas RA, Hickey PR. The limits of detectable cerebral perfusion by transcranial Doppler sonography in neonates undergoing deep hypothermic low-flow cardiopulmonary bypass. J Thorac Cardiovasc Surg 1997;114:594–600.
60. Andropoulos DB, Stayer SA, McKenzie ED, Fraser CD Jr. Novel cerebral physiologic monitoring to guide low-flow cerebral perfusion during neonatal aortic arch reconstruction. J Thorac Cardiovasc Surg 2003;125:491–9.
61. Rodriguez RA, Cornel G, Semelhago L, et al. Cerebral effects in superior vena caval cannula obstruction: the role of brain monitoring. Ann Thorac Surg 1997;64:1820–2.
62. Rodriguez RA, Cornel G, Splinter WM, et al. Cerebral vascular effects of aortovenous cannulations for pediatric cardiopulmonary bypass. Ann Thorac Surg 2000;69:1229–35.
63. O’Brien JJ, Butterworth J, Hammon JW, et al. Cerebral emboli during cardiac surgery in children. Anesthesiology 1997;87:1063–9.
64. Cote CJ, Rolf N, Liu LM, et al. A single-blind study of combined pulse oximetry and capnography in children. Anesthesiology 1991;74:980–7.
65. Office of Science and Technology Annual Report. Ultrasound bioeffects and temperature rises in the eye during ultrasonic insonation. Washington, DC: Food and Drug Administration, Center for Devices and Radiological Health.
66. Janelle GM, Mnookin S, Gravenstein N, et al. Unilateral cerebral oxygen desaturation during emergent repair of a DeBakey type 1 aortic dissection: potential aversion of a major catastrophe. Anesthesiology 2002;96:1263–5.
67. Di Bardino D, Heinle JS, Andropoulos DB, et al. Interrupted aortic arch in a newborn diagnosed by neurophysiologic monitoring. Ann Thorac Surg. In press.
© 2004 International Anesthesia Research Society