| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
*Department of Anesthesia, Cincinnati Children’s Hospital Medical Center and University of Cincinnati College of Medicine, and Institute of Pediatric Anesthesia, Cincinnati Children’s Hospital Research Foundation, Cincinnati, Ohio; and Department of Pathology, Children’s Hospital of Philadelphia and University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania
Address correspondence and reprint requests to Dr. Andreas W. Loepke, Department of Anesthesia, Cincinnati Children’s Hospital Medical Center, ML2001, 3333 Burnet Ave., Cincinnati, OH 45229. Address e-mail to andreas.loepke{at}cchmc.org.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
As a result of concerns about the neurological sequelae associated with DHCA, LF-CPB and selective cerebral perfusion have become more popular for complex infant heart surgery. However, the reduction in CPB flow during LF-CPB or selective cerebral perfusion poses the risk of incomplete ischemia to vital organs. Although not well studied, LF-CPB for neonatal heart surgery is associated with neurological impairment such as diminished cognition and motor performance (4).
Whereas the brain injury after DHCA has been well characterized, the pattern of injury after LF-CPB and selective cerebral perfusion have not been studied. Because LF-CPB improves neurological outcome compared with DHCA, the brain injury pattern of the two modalities might be different. The present study therefore aims to characterize brain injury associated with LF-CPB and to determine the factors associated with this injury in a piglet survival model.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
After induction of anesthesia with IM ketamine (33 mg/kg) and acepromazine (3.3 mg/kg), animals were endotracheally intubated and ventilated to normocapnia with an inspired oxygen concentration (Fio2) of 1.0. After IV catheter insertion, anesthesia was supplemented with fentanyl (25 µg/kg, then 10 µg · kg–1 · h–1) and droperidol (0.25 mg/kg, then 0.25 mg · kg–1 · h–1). A femoral arterial catheter was placed to monitor mean arterial blood pressure (MAP), blood gases, pH, and hemoglobin (iStat; iStat Co., East Windsor, NJ), as well as glucose concentrations (Surestep; Lifescan, Milpitas, MN). A scalp incision and a 3-mm skull hole were made over the right parietal region to accommodate an epidural thermistor and laser Doppler probe. Electrocardiogram, MAP (Gould Instrument Systems, Valley View, OH), cortical cerebral blood flow (CBF) (Laserflo; Medtronic, Minneapolis, MN), and end-expiratory CO2 (Poet IQ; CSI, Waukesha, WI), as well as brain epidural, esophageal, and rectal temperatures (models 555 and 401; Yellow Springs Instruments, Yellow Springs, OH) were measured throughout surgical preparation.
After exposure through a right-sided neck incision and after IV administration of cefazolin (25 mg/kg) and heparin (200 U/kg), the right common carotid artery and the right external jugular vein were cannulated and catheters (Biomedicus; Medtronic) were advanced to the aorta and right atrium, respectively, for CPB. Before initiation of CPB, 10 mL/kg blood was withdrawn and stored cold for postoperative transfusion. The CPB circuit used a nonpulsatile roller pump (RS 7800; Renal Systems, Minneapolis, MN) and a membrane oxygenator (Liliput; Dideco, Mirandola, Italy) primed with heparin (2000 U), fentanyl (50 µg), pancuronium (1 mg), cefazolin (25 mg/kg), furosemide (1 mg/kg), calcium chloride (500 mg), dexamethasone (30 mg/kg), sodium bicarbonate (5 mEq), and plasma-lyte A (Baxter PPI, Deerfield, IL), targeting the Hct to 25% during CPB.
After initiation of CPB at 100–150 mL · kg–1 · min–1, animals were surface- and core-cooled using pH-stat management. The arterial Pco2 during CPB cooling was increased by decreasing the oxygenator sweep flow while maintaining an arterial Po2 of more than 400 mm Hg. Arterial perfusate was kept 5°–10°C lower than all body temperatures until 22°C (brain) was reached. At this point, CPB was maintained for 150 min with the brain temperature kept constant. Three control animals underwent "reduced-flow" CPB with pump flows of 25–50 mL · kg–1 · min–1 and MAP targeted between 25 to 30 mm Hg. Twenty-two animals underwent LF-CPB with MAP targeted to 10 mm Hg with a minimal pump flow of 5 mL · kg–1 · min–1. Preliminary studies identified these LF-CPB variables to create a reproducible, survivable neurological injury. After 150 min, surface- and core-rewarming were used while the CPB pump flow was gradually increased to 100–150 mL · kg–1 · min–1 with the perfusate temperature being kept 5°–10°C higher than all body temperatures, to a maximum of 38°C. The heart was defibrillated as indicated. Mannitol (0.5 g/kg) was administered IV when the brain temperature reached 28°C. When all body temperatures were more than 34°C (after about 35 min of CPB rewarming), piglets were separated from CPB, cannulae and laser Doppler CBF probe were removed, protamine (4 mg/kg) was administered IV, and the incisions were closed.
Postoperatively, piglets were transfused with the blood collected pre-CPB, followed by an infusion of dextrose 5% in lactated Ringer’s solution at 4 mL · kg–1 · h–1. Animals’ lungs were ventilated with a Fio2 of 1.0 to a targeted arterial Pco2 of 35–45 mm Hg until return of a regular breathing pattern, airway reflexes, and purposeful movement, at which point the trachea was extubated. Brain temperature was maintained at normothermia (38°C) until tracheal extubation, at which point the temperature probe was removed and animals were returned to the pen. Postoperatively, animals were allowed to feed at will; those unable to ambulate were bottle-fed; those unable to bottle-feed were administered IV fluids.
Heart rate, MAP, temperatures, and cortical CBF were recorded every 5 min during CPB and every 15 min during reduced-flow or LF-CPB and after separation from CPB. Arterial blood gases, pH, glucose, and Hct concentrations were measured at baseline, at 15 min of CPB cooling and rewarming, every 30 min during LF-CPB, and 15 and 120 min after separation from CPB. Sodium bicarbonate was administered as needed to yield a base excess between –2 and +2.
Four of the animals underwent magnetic resonance imaging (MRI) on a Siemens Magnetom 1.5 T scanner at 2 and 9 days postoperatively. The images obtained were sagittal T1-weighted spin-echo, axial and coronal T2-weighted fast-spin-echo, axial T2-weighted gradient-echo, and axial diffusion-weighted images. The MRI scans were reviewed by a neuroradiologist unaware of the animals’ condition for pathological lesions, such as atrophy, cerebral edema, hemorrhage, periventricular leukomalacia, focal tissue loss, and infarction. The MRI scans were classified as no lesions, possible lesions, and definite lesions. After the second scan 9 days after CPB, animals undergoing MRI were euthanized and their brains were removed and processed. All other animals were killed 2 days after surgery and brains were prepared as previously described (5).
Slides were examined by a neuropathologist unaware of group assignment for neuronal cell death, hemorrhage, inflammation, and infarction and scored from 0 to 5 with a semiquantitative scoring system as follows: 0 = no damage (normal structures), 1 = rare damage (<1% neurons dead), 2 = mild damage (1%–5% neurons dead), 3 = moderate damage (6%–15% neurons dead), 4 = severe damage (16%–30% neurons dead), 5 = very severe damage (more than 30% of all neurons dead). All regions of neocortex and hippocampus were examined, as well as periventricular white matter, thalamus, basal ganglia, cerebellum, and brainstem, and cell death morphology was classified as apoptotic or necrotic.
An observer blinded to group assignment performed a daily neurobehavioral examination on the animals, assigning scores for specific neurological deficits (1). A score of 0 (no neurological deficits) to 96 (brain death) was obtained by adding the scores for deficits in level of consciousness (0–25), cranial nerve function (0–6), sensory function (0–14), gait (0–25), and behavior (0–20).
Data are presented as mean (±sd). Comparisons among groups were made with analysis of variance with Tukey’s test for multiple comparisons and with Mann-Whitney U-test for parametric and nonparametric data, respectively. Statistical significance was accepted as P < 0.05. Spearman’s rank correlation coefficients were calculated among all perioperative physiological data listed in Table 1 and neurological performance scores or histopathological damage scores for all brain regions listed in Table 2; correlations were accepted when the probability was <0.01 to adjust for multiple comparisons.
|
|
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
In the hypothermic reduced-flow CPB group, no functional disability was observed on neurological examination at 24 h or 48 h postoperatively. In the LF-CPB group, functional disability occurred in 64% of animals at 24 h postoperatively and decreased to 31% of animals at 48 h and 0% at 9 days postoperatively (Fig. 1). Disability mainly involved gait disturbance, abnormal tone, as well as decreased alertness. All animals were able to feed independently.
|
No histological damage was seen in the reduced-flow CPB group. After LF-CPB, all animals displayed histological brain damage showing a predilection for particular brain regions (Table 2). Forty-eight hours postoperatively, brain injury appeared in neocortex (100% of piglets), hippocampus (100% of piglets), basal ganglia (56% of piglets), cerebellum (44% of piglets), and white matter (44% of piglets). No neuronal damage was seen in the brainstem. Neocortex and hippocampus showed the most severe damage, followed by basal ganglia and thalamus, with minimal damage in white matter and cerebellum. After LF-CPB, damage in the neocortex appeared as small clusters interspersed throughout the gray and white matter, disproportionately affecting layers 4–6 of the gray matter (Figs. 2 and 3). Hippocampal damage affected CA1–4. Morphologically, cell death in both hippocampus and neocortex appeared necrotic and rarely apoptotic.
|
|
|
No cerebral atrophy, edema, hemorrhage, periventricular leukomalacia, focal tissue loss, or infarction was seen on the MRI scans. Lesions consistent with ischemia were detected in the anterior portion of the caudate in 2 animals, which diminished from day 2 to day 9 (Fig. 3).
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
In the present study, we used a membrane oxygenator, pH-stat strategy, nonrapid cooling and rewarming, increased Hct and arterial Po2 to simulate optimal neurological protection conditions during LF-CPB for neonatal cardiac surgery (2,9). However, the CPB pump flow was less and duration of LF-CPB was more than most clinical situations in order to incur brain damage for characterization of the injury pattern. Also, a closed-chest model sacrificing the right carotid artery and external jugular vein was used for CPB to improve survivability and minimize postoperative discomfort in an animal model. Several observations argue against this closed-chest model influencing the brain damage. In neonatal pigs, the ligation of one carotid artery has been shown to have no effect on CBF during normal or ischemic conditions (10). Moreover, no difference in histological damage between the two brain hemispheres was seen in this study. Extracorporeal perfusion in and of itself did not cause brain damage, as evidenced by the lack of neurological injury in animals in the reduced-flow hypothermic CPB group. Ketamine, an N-methyl d-aspartic acid (NMDA) blocker, was administered for anesthetic induction. It was not a factor in the present study because it is cleared within one hour in piglets.
We did not check for presence of a patent ductus arteriosus (PDA) and the experimental set-up made it impractical to ligate a PDA. However, several lines of evidence argue against the PDA influencing results of this study. First, whereas PDAs have been documented in up to 38% of newborn piglets less than 20 hours of age, they persist in <6% of piglets aged 20–48 hours (11). Thus, the prevalence of a PDA would be even less in the age group used in our study. Second, ventilation of the lungs was discontinued on initiation of CPB, limiting the hemodynamic effects of a PDA if it were present. Third, even if a PDA was present, the study’s goal to characterize the neuropathology in a low-perfusion CPB model would have been satisfied, as the laser Doppler flowmetry documented the presence of CBF at reduced levels, distinct from a DHCA model or full-flow CPB model.
The brain damage after LF-CPB was similar yet different from that reported after DHCA. After LF-CPB, the present study revealed that neuronal damage occurred most often and was most severe in neocortex and hippocampus, followed by basal ganglia, neocortical white matter, and cerebellum. In a previous study (1), we found the same regions to be vulnerable to damage after DHCA. However, whereas neuronal cell damage after DHCA appeared predominantly in the superficial gray matter (layers 2 and 3), it is located in the deeper gray matter (layers 4–6) after LF-CPB. Despite the very slow pump flow, CBF was present at approximately 10% of pre-CPB levels, as measured by laser Doppler flowmetry. Because laser Doppler interrogates approximately 1 mm3 of tissue beneath the probe, corresponding to superficial neocortex, laser Doppler may not have monitored for blood flow in the deeper layers. However, total cessation of blood flow to the deeper brain regions seems unlikely, because 150 minutes of no CBF would correspond to 150 minutes of DHCA at 22°C, which in our experience is not survivable or produces far worse brain damage than that observed in the present study (1).
In LF-CPB, brain damage appeared as selective neuronal or selective oligodendroglial death and not as infarction or hemorrhage. In our previous study, we also observed selective neuronal death after DHCA (1). After LF-CPB, selective cell death consisted of small clusters of dead cells interspersed among viable cells. Clusters varied between 1 or 2 cells to 15 or 25 cells, corresponding to areas of 0.03 to 0.8 mm3. Because cellular death evolves over time and cannot be observed histologically immediately after CPB, those animals not surviving for at least 48 hours were not included in the histological assessment. Functional recovery after LF-CPB was rapid over 24–48 hours despite the presence of neuronal cell death. Although the dead cells had been removed by nine days postoperatively and neurological function had returned to normal, the study was not designed to assess the effect of early brain cell death on later neurological function. Thus, the relationship of early neuronal cell death after neonatal CPB on neurocognitive function later in life is unknown.
In our study, functional impairment correlated best with severity of basal ganglia injury, a region associated with fine motor activity. In human survivors of neonatal cardiac surgery using LF-CPB, focal infarcts, especially in basal ganglia, periventricular leukomalacia, and subarachnoid hemorrhage have been observed by MRI one to two weeks postoperatively, and cerebral atrophy and infarcts at several months postoperatively (8). Neurobehavioral examinations reveal difficulty in fine motor skills several months postoperatively (12). These observations point to the susceptibility of the basal ganglia to LF-CPB for neonatal heart surgery. In the present study, histological brain damage was apparent in all animals, yet no infarction, hemorrhage, or focal lesions were demonstrated postoperatively on MRI, except for ischemic changes in the anterior caudate. A 1.5 T MRI is able to visualize lesions as small as 1–2 mm3 in diameter. Thus, MRI did not have the spatial resolution to visualize the selective neuronal necrosis documented on histopathology. This finding suggests that MRI underestimates the incidence of neuronal injury in human neonatal heart surgery. Cerebral edema has been observed in adults within 24 hours after cardiac surgery. However, our study did not observe cerebral edema in piglets, nor did other studies in neonates after cardiac surgery.
Despite the use of protocols to keep physiological variables at targeted values during CPB, variations in MAP, Pco2, Po2, Hct, CBF, and glucose were observed among animals. These variations reflect the result of a complex interaction between extracorporeal circulation, body core cooling, and the animals’ compensatory mechanisms. These variations, however, afforded the opportunity to examine correlations between these variables and neurological outcome. Several physiological variables were correlated with impaired neurological performance and histological damage. The variation in pH and Pco2 among animals during pH-stat CPB was noteworthy. Higher arterial Pco2 was correlated with less histological damage in neocortex, hippocampus, and white matter. Interestingly, there was no relationship between histological damage and CBF or pH. It could be speculated that during LF-CPB arterial perfusion pressure was very low, and the cerebrovasculature was maximally dilated, such that an increase in Pco2 would not have been able to further vasodilate and increase CBF (13). Correlations between intraoperative factors and neurological outcome could not be done for nine-day survival animals because no functional deficits or brain damage were observed in any of these animals.
Blood gas management with pH-stat during CPB has been shown to improve neurological outcome after DHCA and retrograde cerebral perfusion (2,14). During full-flow CPB, CBF is higher during pH-stat than 
-stat, and as a result, brain cooling is faster and more homogeneous with pH-stat management (15). It has been thought that pH-stat is protective through higher CBF and better brain cooling. However, in the present study, the lack of CBF or pH correlation with neurological outcome and the correlation of Pco2 with neurological outcome all point to CO2 protecting by a mechanism independent of CBF and arterial pH.
Vannucci et al. (16) found a protective effect of CO2 on neurological outcome in a neonatal model of normothermic hypoxic-ischemic brain injury. As Pco2 was increased from 38 to 80 mm Hg, neurological injury decreased in a stepwise manner. However, an increase in Pco2 to 100 mm Hg led to worsened neurological outcome (17). In our study, we did not find a ceiling to CO2 protection up to a Pco2 of 128 mm Hg. Vannucci et al. attributed extreme hypercapnia to produce a deleterious effect by cardiac depression, which in our study would have been negated by CPB support.
There are several metabolic mechanisms by which CO2 could confer neuroprotection. CO2 produces arterial acidosis, which will shift the oxygen-hemoglobin dissociation curve to the right, improving cellular oxygen delivery. However, because this mechanism is from blood pH, which did not correlate with better neurological outcome, it is an unlikely mechanism in the present study. In contrast to hydrogen ions, CO2 readily crosses the blood-brain barrier to produce cerebral cellular acidosis. Cellular acidosis inhibits NMDA receptors, which blocks excitotoxic neuronal death during ischemia (18). Other mechanisms for hypercarbia-induced protection may include a reduction in brain glutamate and brain tissue lactate concentrations during ischemia (19,20). Less lactate and glutamate being available also lessens several key pathways that damage the cellular machinery.
White matter injury is a common lesion in neonatal heart surgery (8). The white matter is predominantly populated by oligodendrocytes which, recent studies have shown, are damaged by different mechanisms than neurons (21). In particular, oxygen free radicals have a central role in their cell death and not glutamate-induced excitotoxicity, as in neurons (22). In the present study, higher arterial Po2 before CPB and during CPB cooling was associated with worse neuropathology in white matter. This correlation was not found in other brain regions or at any other time point during the experiment. One explanation might be that a higher Po2 supported oxygen free radicals formation during ischemia. However, no association was seen between Po2 and outcome during LF-CPB and CPB reperfusion, a time when free radical generation would be expected. Further research is needed to study white matter injury during neonatal hypothermic CPB.
Several studies advocate a Hct more than historical levels of 20%–25% to improve neurological outcome after CPB and DHCA (1,3). Interestingly, the present study did not find a correlation between Hct from 20% to 30% during CPB and neurological outcome. However, there was a strong association between higher postoperative Hct and less neurological impairment and less basal ganglia damage. After DHCA, a Hct of 30% during CPB improved neurological outcome when compared with an intraoperative Hct of 20%, even when modified ultrafiltration was used to increase the postoperative Hct to 30% (3). The present study suggests that the postoperative Hct may be important to neurological outcome.
There are fundamental differences between adults and neonates regarding the effects of hyperglycemia during ischemia on neurological outcome. Whereas hyperglycemia during ischemia has been shown to be deleterious for the adult brain, several studies have demonstrated, in neonatal animals, that hyperglycemia protects the neonatal brain (23–25). No relationship was found between high glucose levels during LF-CPB or DHCA and adverse neurological or developmental outcome in human neonates undergoing heart surgery (26). Similarly, the present study in a neonatal animal model found no correlation between hyperglycemia and adverse histological or functional outcome after LF-CPB.
In conclusion, neuronal injury after hypothermic LF-CPB in a piglet model using pH-stat strategy occurs predominantly in deep neocortex and hippocampus. MRI was unable to show the selective neuronal necrosis. Factors mitigating injury were higher arterial CO2, Hct, and blood glucose levels.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. H. Hovels-Gurich, K. Konrad, D. Skorzenski, B. Herpertz-Dahlmann, B. J. Messmer, and M.-C. Seghaye Attentional Dysfunction in Children After Corrective Cardiac Surgery in Infancy Ann. Thorac. Surg., April 1, 2007; 83(4): 1425 - 1430. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. M. Hoffman Neurologic Monitoring on Cardiopulmonary Bypass: What Are We Obligated to Do? Ann. Thorac. Surg., June 1, 2006; 81(6): S2373 - S2380. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
