Asian Cardiovasc Thorac Ann 2007;15:194-199
© 2007 Asia Publishing EXchange Ltd
Near-Infrared Spectroscopy Monitoring of Brain Oxygen in Infant Cardiac Surgery
Ji-Hong Huang, MD,
Zhao-Kang Su, MD,
Shun-Min Wang, MD
Department of Pediatric Thoracic and Cardiovascular Surgery, Xinhua Hospital, Shanghai Children’s Medical Center, Shanghai Jiaotong University Medical College, Shanghai, China
For reprint information contact: Zhao-Kang Su, MD Tel: 86 21 5873 2020 Ext. 3283 Fax: 86 21 5839 3915 Email: zhaokang_su{at}yahoo.com.cn, Department of Pediatric Thoracic and Cardiovascular Surgery, Xinhua Hospital, Shanghai Children’s Medical Center, Shanghai Jiaotong University Medical College, 1678, Dongfang Road, Shanghai 200127, China.
 |
ABSTRACT
|
|---|
The use of near-infrared spectroscopy for monitoring cerebral oxygenation during different types of cardiopulmonary bypass was evaluated in 24 patients aged 5 to 13 months. They underwent open-heart surgery under cardiopulmonary bypass with moderate hypothermia, deep hypothermia with low flow, or deep hypothermia with circulatory arrest. Near-infrared spectroscopy data were compared with electroencephalography and biochemical indicators (neuron-specific enolase, lactate). Near-infrared spectroscopy data showed no correlation with biochemical indicators in patients undergoing cardiopulmonary bypass with moderate hypothermia or deep hypothermia with low flow. In the deep hypothermia with circulatory arrest group, the oxygenated hemoglobin signal declined to a nadir during circulatory arrest. The period from reaching the nadir until reperfusion and the minimum values of oxygenated hemoglobin correlated closely with increases in neuron-specific enolase and lactate. All patients with an oxygenated hemoglobin-signal nadir time < 35 min were free from behavioral evidence of brain injury. The oxygenated hemoglobin-signal nadir time may be useful in predicting the safe duration of circulatory arrest.
|
INTRODUCTION
|
|---|
Neurological damage is a frequent and potentially devastating complication after cardiopulmonary bypass (CPB), especially under deep hypothermia with circulatory arrest (DHCA). Therefore, monitoring of cerebral oxygenation can be useful. In this study, near-infrared spectroscopy (NIRS) was used for brain oxygen monitoring during open-heart operations in infants, and compared with other indicators of brain injury to determine the safe duration of DHCA.
|
PATIENTS AND METHODS
|
|---|
After local institutional review board approval and written informed consent were obtained, 24 patients with ventricular septal defect and pulmonary hypertension undergoing open-heart surgery from June 2003 through November 2003 were studied. There were 14 males and 10 females. Age ranged from 5 to 13 months, with a median age of 8 months. Body weight ranged from 4.5 to 9 kg, with a median of 6.5 kg. Patients were randomly assigned to one of 3 groups of 8 patients each undergoing different modes of CPB: moderately hypothermic CPB (MHCPB), deep hypothermic low flow (DHLF), and deep hypothermic circulatory arrest (DHCA). In the MHCPB group, anesthesia was induced with intravenous fentanyl 15 µg·kg–1 and pancuronium 0.1 mg·kg–1, and maintained with inhaled isoflurane. After induction of anesthesia, all patients were ventilated with a fraction of inspired oxygen of 0.5, and the right jugular bulb was catheterized in a retrograde manner. Esophageal and rectal temperatures were recorded continuously. A membrane oxygenator (Minimax; Medtronic, Inc., Anaheim, CA, USA) with an arterial filter (Olson Medical Sales, Inc., Ashland, MA, USA) was used for CPB. Core cooling was started as soon as the perfusion cannulas were in place, and stopped when the nasopharyngeal temperature reached 28°C–30°C. Perfusion was maintained at a flow rate of 100 to 125 mL·kg–1·min–1 and blood pressure at 30–50 mm Hg. Alpha-stat strategy was applied throughout the operation in this group. In the DHLF group, body surface and core cooling were used. Thiopental was added to the heart-lung machine when the nasopharyngeal temperature reached 30°C. When it reached 25°C, the perfusion rate was reduced to 50 mL·kg–1·min–1. At 18°C, the perfusion rate was reduced to 30 mL·kg–1·min–1 until rewarming. It was raised to 150 mL·kg–1·min–1 and maintained at this level during the rewarming phase. In this group, pH-stat strategy was used in the cooling phase and alpha-stat in the other phases. In the DHCA group, we also applied body surface and core cooling. The pump was turned off when the nasopharyngeal temperature cooled down to 18°C. Rewarming was achieved by perfusion with warmed bypass blood at a flow rate of 150 mL·kg–1·min–1. When the nasopharyngeal temperature rose to 35°C and the cardiac index stabilized, the patients were weaned from CPB. pH-stat strategy was used in the cooling phase, and alpha-stat otherwise.
The NIRS optodes (NIRO-500, Hamamatsu Photonics KK, Hamamatsu, Japan) were placed on the forehead 4 cm apart. The relative changes in the concentrations of oxygenated hemoglobin (HbO2), reduced hemoglobin (HbR), and oxidized cytochrome aa3 (Cytox) in brain tissue were measured continuously and recorded every 30 sec after induction of anesthesia. Blood samples were drawn from the jugular bulb to determine neuron-specific enolase and lactate levels at 4 time points: 2 min after CPB, at the end of cooling, at the end of rewarming, and 6 hr after bypass. Neuron-specific enolase was assayed with a commercial enzyme immunoassay kit (Cobas Core NSE EIA II; Roche, Basel, Switzerland).1 Lactate was assayed with a portable lactate analyzer (Accusport/Accutrend, Roche Diagnostics, Germany).2 A 4-channel quantitative electroencephalogram (EEG) was taken twice from each patient; preoperatively, and 6 days postoperatively. Electroencephalogram data were recorded and interpreted in a blinded fashion by a pediatric technologist board-certified in electroencephalography.
Data are presented as mean ± standard deviation. All results were analyzed using SAS version 6.12 software for Windows (SAS Institute, Cary, NC, USA). Analysis of variance and the grouped t test were used, and Pearson correlation coefficients between the NIRS data and biochemical variables were calculated. Curve fitting and curvilinear regression was used to analyze the HbO2 data. A p value of less than 0.05 was considered statistically significant.
|
RESULTS
|
|---|
The clinical and operative data are summarized in Table 1
. Among the 24 patients studied, there was no death and only one infant assigned to the DHCA group suffered a temporary increase in muscular tension 2 days after the operation in both lower extremities. No convulsion or stroke occurred. Due to the extended rewarming time, the aortic cross clamp time in the DHCA group was significantly longer than in the other 2 groups. The hematocrit and mean arterial pressure during bypass were significantly less than those pre- and post-bypass. There were no significant differences among the 3 groups.
In all 3 groups, a drop in HbO2 occurred at the start of CPB, with an increase above the baseline during further cooling (Figure 1). Beyond the cooling phase, HbO2 remained stable in the MHCPB and DHLF groups, but it dropped sharply with the onset of circulatory arrest in the DHCA group and then tapered off asymptotically. At the start of reperfusion, HbO2 recovered to the baseline value. Between the end of cooling and the start of rewarming, HbO2 values were significantly lower in the DHCA group than in the other two groups ( p < 0.05). Reduced hemoglobin in the MHCPB and DHLF groups decreased during the cooling phase (Figure 2). It increased during rewarming and reached pre-CPB levels at the end of bypass. In the DHCA group, HbR increased during the circulatory arrest period, reached a peak value before rewarming, and recovered after rewarming. Between the end of cooling and the beginning of rewarming, HbR values were significantly higher for the DHCA group than the other 2 groups. Cytox decreased consistently at the beginning of CPB in all groups, and increased after rewarming, the maximum decline was observed in the DHCA group (Figure 3).
View larger version (15K):
[in this window]
[in a new window]
|
Figure 1. Change in oxygenated hemoglobin (HbO2) in different CPB modes, calculated relative to the baseline level (in µmol·L–1 x differential path length factor). Time points: 0 = CPB onset (baseline), 1 = end of cooling, 2 = midpoint between the end of cooling and start of rewarming, 3 = start of rewarming, 4 = end of CPB, 5 = end of surgical procedure. CPB = cardiopulmonary bypass, DHCA = deep hypothermic circulatory arrest, DHLF = deep hypothermic low flow, MHCPB = moderately hypothermic CPB.
|
|
View larger version (16K):
[in this window]
[in a new window]
|
Figure 2. Change in deoxygenated hemoglobin (HbR) in different CPB modes, calculated relative to the baseline level (in µmol·L–1 x differential path length factor). Time points: 0 = CPB onset (baseline), 1 = end of cooling, 2 = midpoint between the end of cooling and start of rewarming, 3 = start of rewarming, 4 = end of CPB, 5 = end of surgical procedure. CPB = cardiopulmonary bypass, DHCA = deep hypothermic circulatory arrest, DHLF = deep hypothermic low flow, MHCPB = moderately hypothermic CPB.
|
|
View larger version (22K):
[in this window]
[in a new window]
|
Figure 3. Change in oxidized cytochrome aa3 (Cytox) in different CPB modes, calculated relative to the baseline level. Time points: 0 = CPB onset (baseline), 1 = end of cooling, 2 = midpoint between the end of cooling and start of rewarming, 3 = start of rewarming, 4 = end of CPB, 5 = end of surgical procedure. CPB = cardiopulmonary bypass, DHCA = deep hypothermic circulatory arrest, DHLF = deep hypothermic low flow, MHCPB = moderately hypothermic CPB.
|
|
Oxygenated hemoglobin decreased during cardiac arrest in the DHCA group (Figure 4). The linear decay in HbO2 can be described by a logarithmic function: HbO2 = alog(t) + b, dHbO2/dt = a/t, where t is the time after onset of DHCA, and a and b are constants. We defined the HbO2 signal as having reached the nadir value (a plateau state) when the slope of the fitted curve (the differential coefficient dHbO2/dt) exceeded –0.5. The period between reaching the nadir and reperfusion was calculated in each case and termed the "HbO2-signal nadir time". The HbO2-signal nadir time in Figure 4 was 42.26 min. The mean duration of time to the nadir in the DHCA group was 19.35 min.
View larger version (13K):
[in this window]
[in a new window]
|
Figure 4. Oxygenated hemoglobin (HbO2) decay curve during deep hypothermic circulatory arrest (DHCA). In this patient, the plateau state (when dHbO2/dt exceeded –0.5) was reached at 15.74 min, y = –7.87log(x) + 9.45, r = 0.05, p < 0.01.
|
|
Neuron-specific enolase and lactate concentrations at each time point are shown in Table 2. Cardiac surgery caused an increase of lactate at the end of cooling, a further increase toward the maximum blood levels at the end of rewarming, and then a rapid decline 6 hr postoperatively in all groups. Lactate levels at the end of rewarming were significantly higher in the DHCA group than in the MHCPB group ( p < 0.01). Lactate levels did not differ between the DHLF and MHCPB groups. In each group, the neuron-specific enolase concentrations at 2 min after CPB and at the end of cooling were within the normal range, whereas they increased at the end of rewarming, reaching the maximum value at 6 hr postoperatively. The neuron-specific enolase level at the end of rewarming was significantly higher for the DHCA group than for the MHCPB group ( p < 0.01). It did not differ between the DHLF and MHCPB groups.
No infant showed an EEG abnormality before the operation. Two patients in the DHCA group were found to have a background EEG abnormality postoperatively, with diffuse low frequency. One of them was diagnosed with a transient increase in muscular tension; the other had neither clinical symptom nor physical sign. Six months later, both patients’ EEGs had returned to normal.
The relationships between NIRS data and biochemistry data are shown in Table 3. In the DHCA group, HbO2-signal nadir time and the minimum value of HbO2 correlated closely with neuron-specific enolase and lactate values at the end of rewarming. Reduced hemoglobin and Cytox showed no correlation with the neuron-specific enolase and lactate values. Near-infrared spectrometry data for the MHCPB and DHLF groups showed no correlation with the neuron-specific enolase and lactate values. The 2 patients with definitely abnormal postoperative EEGs had a significantly longer HbO2-signal nadir time compared to those with normal EEGs. All patients whose HbO2-signal nadir time was less than 35 min were free from EEG abnormalities and had significantly higher HbO2 than patients without neuropsychological complications ( p < 0.05).
|
DISCUSSION
|
|---|
Deep hypothermia circulatory arrest may be used in 3 circumstances: surgery of the aortic arch, complex congenital heart defects, and sometimes, congenital heart disease with a low body weight, as in our study. Neuropsychological and neurological deficiencies are major causes of postoperative mortality and morbidity, especially after DHCA, hence the importance of monitoring cerebral function during the operation. Near-infrared spectrometry provides a promising method for noninvasive monitoring of cerebral oxygenation. Within the near-infrared range, HbO2, HbR, and Cytox have characteristic absorption spectra. The concentrations of HbO2, HbR and Cytox are likely to exhibit rapid changes, paralleling alterations in cerebral perfusion and oxygenation. Oxygenated hemoglobin and HbR are located in the blood vessels, and Cytox in the mitochondrial membrane. Near-infrared spectrometry may thus provide information on vascular oxygenation (changes in HbO2 and HbR) and on cellular oxygenation (changes in Cytox).3 Near-infrared spectrometry monitoring is noninvasive, continuous, and applicable during the operation. However, some methodological limitations remain. As NIRS measurements describe changes in cerebral oxygenation, it is difficult to define normal values or a limit of safety, thus the method remains controversial.4
In our study, NIRS data for the MHCPB group demonstrate a paradoxical divergence in intravascular and mitochondrial oxygenation (HbO2 and Cytox), which begins in the earliest phases of CPB and core cooling. This divergence accompanied by increasing neuron-specific enolase and lactate levels suggests that there are some disadvantages even in standard CPB. The possible reasons are decreased offloading of oxygen from hemoglobin and insufficient tissue oxygenation, and diffusion problems that limit oxygen access to the mitochondria. However, because neuron-specific enolase may be released from sources in the body other than the central nervous system, it is also possible that the significantly higher levels of neuron-specific enolase in the DHCA group might correlate with bypass-related hemolysis, coincidentally peaking at 6 hr postoperatively.
Nollert and colleagues4 suggested that a decline of Cytox indicates brain damage. However, other studies found that the reliability and reproducibility of cytochrome aa3 measurements with recent technology have been inconsistent, and the results have been variable in different preparations.5 Cytox decreased in all groups at the start of CPB and increased after rewarming, the maximum decline was observed in the DHCA group. Perhaps during circulatory arrest, there is some damage at the cellular level. However, it must be stressed that this was a relatively small study on a limited number of patients, and we plan to undertake further studies on the reliability of Cytox measurements.
Some have suggested that low-flow CPB may provide superior cerebral protection compared to circulatory arrest, but this has not been borne out by clinical studies. Our previous study indicated that the optimal perfusion flow rate for the brain during deep hypothermic CPB at 20°C is 25 mL·kg–1·min–1 or more, and the low-flow period should not go beyond 40 min. In the current study, we found no significant difference between MHCPB and DHLF. Possible reasons for this include the relatively short (12.7 min) low-flow period, the fact that the pump rate remained at 30 mL·kg–1·min–1 while rectal temperature was 20°C, and the application of pH-stat strategy during the cooling period. A tendency towards fewer postoperative neurological sequelae has been demonstrated with the use of pH-stat strategy in infants undergoing DHCA. At the onset of DHCA, the steep decline in HbO2 and Cytox suggests that hypoxia exists both in blood vessels and mitochondria, and the intravascular hypoxia exacerbates the mitochondrial hypoxia. Our study found that the association between the HbO2-signal nadir time and the biochemical values was stronger than other parameters. We therefore conclude that the HbO2-signal nadir time is a better predictor of neuropsychological outcome than HbO2, HbR, and Cytox.
The mechanism by which HbO2-signal nadir time may predict neurologic outcome is unknown.5 Other reports have confirmed that cerebral metabolism is suppressed but not eliminated during the circulatory arrest period.6 It would thus seem reasonable to speculate that the gradual decline in HbO2 during DHCA reflects ongoing oxygen metabolism in cerebral neurons and astroglia.7 The finding that HbO2 reaches a plateau suggests that the available cerebral HbO2 is depleted at this point. When all oxygen has been extracted from both arterial and venous phases within the brain, anaerobic metabolism ensues. If sufficiently prolonged, the resulting acidosis can result in neuronal injury. Thus we think the actual detrimental effects of hypothermic circulatory arrest are more related to the point at which the brain becomes severely desaturated (represented by the nadir time), than to the absolute circulatory arrest time.
At present, the safe upper limit of circulatory arrest time remains uncertain. Kin and colleagues8 reported that the period of DHCA should not be longer than 60 min for operations at 15°C. In our study, patients in whom the HbO2-signal nadir time was less than 35 min were free from neurological damage or an abnormal EEG. These patients also had lower neuron-specific enolase and lactate levels. As HbO2-signal nadir time is the difference between the duration of circulatory arrest and the time to its plateau state, it is relatively safe during DHCA to limit the HbO2-signal nadir time to be within 35 min. In our study, the time taken to reach the plateau state was 19.35 min. This result agrees with the study of Kin and colleagues.8
Our findings suggest that the HbO2-signal nadir time might be useful in predicting the need for cerebral reperfusion during circulatory arrest. However, the study should be interpreted in the light of some limitations. First, this is a relatively small study. Furthermore, we studied a very homogeneous group of children with ventricular septal defects in a single institution. As diagnosis can affect both cognitive outcome and surgical conduct, these results may not be extrapolated to a population with mixed diagnoses or age at surgery beyond infancy. However, as NIRS measurements are noninvasive and continuous, do not interfere with the operation, and provide detailed information on cerebral oxygenation, they may provide a promising method for monitoring CPB.
|
ACKNOWLEDGMENTS
|
|---|
The authors would like to thank Hai-Bo Zhang, MD, and Jing-Hao Zheng, MD, Department of Pediatric Thoracic and Cardiovascular Surgery, Shanghai Children’s Medical Center for their help during the clinical study and their intellectual contributions. We also wish to thank Zaw-Sing Su, PhD, a retired scientist from Stanford Research Institute (SRI International) of Menlo Park, California, for his careful reading and review of the manuscript.
|
REFERENCES
|
|---|
- Sterk M, Oenings A, Eymann E, Roos W. Development of a new automated enzyme immunoassay for the determination of neuron-specific enolase. Anticancer Res 1999;19:2759–62.[Medline]
- Slomovitz BM, Lavery RF, Tortella BJ, Siegel JH, Bachl BL, Ciccone A. Validation of a hand-held lactate device in determination of blood lactate in critically injured patients. Crit Care Med 1998;26:1523–8.[Medline]
- Abdul-Khaliq H, Troitzsch D, Schubert S, Wehsack A, Bottcher W, Gutsch E, et al. Cerebral oxygen monitoring during neonatal cardiopulmonary bypass and deep hypothermic circulatory arrest. Thorac Cardiovasc Surg 2002;50:77–81.[Medline]
- Nollert G, Jonas RA, Reichart B. Optimizing cerebral oxygenation during cardiac surgery: a review of experimental and clinical investigations with near infrared spectrophotometry. Thorac Cardiovasc Surg 2000;48:247–53.[Medline]
- Sakamoto T, Hatsuoka S, Stock UA, Duebener LF, Lidov HG, Holmes GL, et al. Prediction of safe duration of hypothermic circulatory arrest by near-infrared spectroscopy. J Thorac Cardiovasc Surg 2001;122:339–50.[Abstract/Free Full Text]
- McCullough JN, Zhang N, Reich DL, Juvonen TS, Klein JJ, Spielvogel D, et al. Cerebral metabolic suppression during hypothermic circulatory arrest in humans. Ann Thorac Surg 1999;67:1895–9.[Abstract/Free Full Text]
- Sakamoto T, Zurakowski D, Duebener LF, Hatsuoka S, Lidov HG, Holmes GL, 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;73:180–90.[Abstract/Free Full Text]
- Kin H, Ishibashi K, Nitatori T, Kawazoe K. Hippocampal neuronal death following deep hypothermic circulatory arrest in dogs: involvement of apoptosis. Cardiovasc Surg 1999;7:558–64.[Medline]
TOP
ABSTRACT
INTRODUCTION
