| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2005;80:1955-1964
© 2005 The Society of Thoracic Surgeons
a Department of Cardiothoracic Surgery, Stanford University School of Medicine, Stanford, California
b Department of Anesthesiology, Stanford University School of Medicine, Stanford, California
* Address correspondence to Dr Amir, Department of Cardiothoracic Surgery, Stanford University School of Medicine, 300 Pasteur Dr, Stanford, CA94305 (Email: gabiamir{at}stanford.edu).
 |
Abstract |
|---|
 Top Abstract IntroductionPathophysiology of Neonatal...Human Clinical StudiesNeonatal Brain ProtectionSummaryAcknowledgmentsReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionPathophysiology of Neonatal...Human Clinical StudiesNeonatal Brain ProtectionSummaryAcknowledgmentsReferences |
|---|
Pharmacologic and surgical strategies have evolved to minimize ischemic brain injury after DHCA; nevertheless, although experimental data are abundant, controversy still exists regarding optimal protective strategies. Several contemporary issues are contemplated, including the type of blood gas management to be undertaken during cooling, how long to cool and to what temperature, the "safe" duration of DHCA, continuous regional brain perfusion (CRBP), and the use of intermittent cerebral brain perfusion during DHCA.
The purpose of this review is to summarize the experimental and clinical data available in the field of neonatal brain protection during DHCA. We will outline the pathophysiology of neonatal hypoxic brain injury associated with cardiopulmonary bypass and DHCA and will discuss the issue of selective vulnerability of the neonatal brain compared with the adult brain. Current techniques and limitations in neonatal brain protection will be described; additionally, we will discuss alternatives to DHCA. We will try to formulate, based on the available data, guidelines that might be helpful.
To that extent, an extensive literature search was conducted through "Pub-Med" using several main key word combinations, including circulatory arrest, cardiopulmonary bypass, neonate, brain injury, neurologic damage, hypothermia, hemodilution, alpha stat, pH stat, and more. The search was conducted as far back as 1965 and as recently as May 2004.
|
Pathophysiology of Neonatal Brain Injury |
|---|
|
TopAbstractIntroductionPathophysiology of Neonatal...Human Clinical StudiesNeonatal Brain ProtectionSummaryAcknowledgmentsReferences |
|---|
The second mechanism of cell death is by apoptosis, or programmed cell death, where cell death occurs despite adequate cellular energetics. Apoptosis is due to activation of specific genes, receptors, and enzymes that break down the cell in a programmed manner [8]; and it is characterized by nuclear karyorrhexis, margination of chromatin in the nucleus, but with minimal cytoplasmic and inflammatory changes. Apoptosis is mediated by a series of proteins that are sequentially activated and have a final common pathway leading to the generation of a family of cysteine proteases called caspases. The principal cysteine proteases involved in apoptosis are caspase 3 and caspase 8 [9]. Caspase 3 can be activated by two principal pathways. The extrinsic pathway is initiated by the classic inflammatory response: soluble factors such as Fas and tumor necrosis factor alpha (TNF-
) bind to cell surface receptors and activate caspase 8, which in turn activates caspase 3. Intrinsic activation of caspase 3 is initiated by the release of cytochrome c from nonlethally damaged mitochondria.
Apoptosis plays a crucial role in neuronal cell death after DHCA. Ditsworth and colleagues [10] demonstrated in a piglet model of DHCA that the apoptotic process starts within a few hours after reperfusion and continues for several days. In this study, piglets were subject to 90 minutes of DHCA at 19°C. After reperfusion and disconnection from cardiopulmonary bypass (CPB), the animals were sacrificed at several time points ranging from 4 hours to 1 week. Damaged neurons were observed as early as 8 hours after reperfusion and as long as 72 hours later. Cortical ATP levels were unchanged from control animals that underwent surgical preparation and were then sacrificed without being subjected to cardiopulmonary bypass or DHCA. In addition, caspase 3 and 8 concentration and activities were significantly higher in the DHCA group, as were cytosolic cytochrome c and Fas, which were significantly elevated 1 and 4 hours after DHCA [10].
Ischemia and hypoxia have also been linked to excessive neuronal stimulation and hyperactivity, initiating a cascade of cellular events leading to neuronal cell death [11, 12]. Excitotoxicity is a term applied to the death of cells caused by overstimulation of excitatory amino acids, mainly glutamate, and is believed to be a fundamental process involved in postischemic neuronal cell damage [6, 13]. Compromised synaptic reuptake of excitatory amino acids and membrane depolarization associated with ischemia cause a lethal flood of Ca+2 and Na+ into the neuron [6, 11]. High calcium concentrations activate nitric oxide synthetase (NOS), which leads to the production of nitric oxide (NO), a neuromodulator that in excess is thought to cause cell death [13].
The role of excitatory amino acids and NOS in the pathogenesis of ischemic brain injury after DHCA has been studied in a circulatory arrest model of mature dogs (7 to 12 months old) [14–16]. Brain regions vulnerable to injury during prolonged circulatory arrest were found to have the highest density of glutamate receptors [14]. Using intracerebral microdialysys techniques, Tseng and colleagues [17] demonstrated that DHCA increases intracerebral glutamate and aspartate concentrations along with the concentration of the coagonist glycine. Citrulline, which is produced in stoichiometric equivalent amounts with NO, was used as a marker of NO production, and was found to be elevated in parallel to the elevated concentrations of glutamate and aspartate [17]. The use of N-methyl-D-aspartate (NMDA) receptor blocker MK-801 significantly reduced citrulline production and apoptosis [14, 18]; and the use of the neuronal nitric oxide inhibitor (7-nitroindsazole) in a canine model of DHCA significantly inhibited neuronal apoptosis as seen on hematoxylin-eosin staining, TUNEL (TdT-dNTP terminal nick-end labeling), and electron microscopy [15].
Selective Vulnerability of the Developing Brain
Deep hypothermic circulatory arrest inflicts a global and diffuse ischemic insult to the brain; nevertheless, damage is most likely to occur in select groups of structures in the immature brain [19, 20]. Clinically, in adults, generalized Jacksonian type seizures follow severe hypoxic ischemic insults, whereas in neonates focal seizures are more common. Among infants aged 6 to 12 months, choreoathetosis can be a manifestation of ischemic neurologic postoperative insult, usually secondary to basal ganglia damage [21].
The topography of selective neuronal cell death associated with DHCA has been shown in an adult canine experimental model of DHCA. It closely corresponded with the distribution of excitatory amino acid receptors in the pyramidal cells of the CA-1 region of the hippocampus, the molecular layer of the dentate nucleus the entorhinal cortex, and the molecular layer of the cerebellum [14, 22]. Blocking these receptors with the specific glutamate NMDA receptor antagonist MK-801 improved the functional recovery of dogs subjected to 120 minutes of DHCA at 18°C and limited selective neuronal necrosis in the CA-1 region of the hippocampus, neocortex, basal ganglia, and cerebellum. Receptor autoradiography revealed significantly better preservation of NMDA glutamate receptor subtypes in the MK-801 treated animals [14]. Selective vulnerability of neuronal populations in neonatal piglets after DHCA has been demonstrated to mainly affect the neocortex and hipoccampus [23].
Functional imaging evidence also indicates that normothermic hypoxic ischemic injury is linked to functionally hypermetabolic areas of the brain. Positron emission scans show relatively high metabolic rates of glucose utilization in the basal ganglia, brain stem, and sensory cortex of newborn babies, with much lower rates elsewhere [24]. Blennow and colleagues [25] evaluated brain glucose metabolism in infants with normothermic hypoxic ischemic encephalopathy, and found focal elevations in selective metabolism of glucose in the basal ganglia and cerebral cortex of 5 infants who later had severe neurologic deficits. The 1 infant who had normal cerebral glucose metabolism was neurologically normal at follow up [25].
Cerebral Energetics
Metabolic and cellular energy changes have been found to be associated with low-flow cardiopulmonary bypass and DHCA [26]. Kramer and colleagues [26] evaluated cerebral ATP changes during normothermic and hypothermic circulatory arrest in mature mongrel dogs. In the normothermic circulatory arrest group, ATP levels were found to decrease to half their baseline values after 3.8 minutes (calculated half-life) of circulatory arrest. Electroencephalographioc silence occurred 20 seconds after the initiation of normothermic circulatory arrest, corresponding to a 10% reduction of ATP from baseline levels. The ATP recovery rates inversely correlated with the duration of cerebral ischemia. Deep hypothermic circulatory arrest performed at esophageal temperatures of 7.8 ± 2.5°C and at average cortical temperatures of 13 ± 2.7°C demonstrated a calculated half-life of 13.3 minutes compared with 3.8 minutes observed under normothermic conditions. The disappearance of cortical electrical activity was noted at 16.6 ± 3.5°C during cooling, coinciding with a reduction of less than 10% of control ATP concentrations.
In vivo phosphorus-31 nuclear magnetic resonance spectroscopy was used to assess the metabolic state of the brain during circulatory arrest by measuring the concentrations of high-energy phosphate compounds and intracellular pH [27]. Sheep were cooled to 15°C and were subject to either circulatory arrest or low-flow CPB. Both circulatory arrest and low-flow of 5 mL·kg-1 ·min-1 resulted in severe intracellular acidosis and depletion of high energy phosphates However, flows of 10 mL·kg-1 ·min-1 maintained both brain high-energy phosphate concentrations and intracellular pH.
|
Human Clinical Studies |
|---|
|
TopAbstractIntroductionPathophysiology of Neonatal...Human Clinical StudiesNeonatal Brain ProtectionSummaryAcknowledgmentsReferences |
|---|
|
Neonatal Brain Protection |
|---|
|
TopAbstractIntroductionPathophysiology of Neonatal...Human Clinical StudiesNeonatal Brain ProtectionSummaryAcknowledgmentsReferences |
|---|
Hypothermia and Hemodilution
The brain utilizes up to 20% of total body oxygen consumption, with 40% of its energy consumption used in the preservation of cellular integrity and 60% in the transmission of nerve impulses [28]. Electrocerebral silence and the disappearance of the somatosensory evoked response occurs at approximately nasopharyngeal temperature of 17°C [29, 30]. Thus, theoretically, lowering brain temperature would reduce much of its metabolic demands. The rate of change of a reaction for every 10°C change in temperature is referred to as Q10. Kern and colleagues [31] evaluated the rate by which cerebral O2 metabolism decreases in relation to pump flow and metabolism. They noted that, as temperatures were lowered, the cerebral oxygen extraction rate and the estimated minimal pump flow rates decreased logarithmically. The Q10 was higher for infants and children (Q10 = 3.65) than for adults (Q10 = 2.6), implying that changes in temperature have greater impact in cerebral oxygen consumption in pediatric subjects than in adults. This finding reflects the greater metabolic suppression by hypothermia in the pediatric population [32].
What temperature provides optimal brain protection during DHCA? Early experience with deep hypothermia suggested that using extremely low temperatures resulted in a dramatic increase in neurologic and pulmonary injury [33, 34]. DeLeon and colleagues [34] reported that choreoathetosis developed in 8 of 758 patients operated on using cardiopulmonary bypass and hypothermia, the incidence being significantly higher in the group of patients for whom deep hypothemia (rectal temperature lower than 25°C) was used (8 of 463 versus 0 of 295, p = 0.02) and for whom the cooling time (the time to cool the patient to a target temperature) was longer than 1 hour (7 of 243 versus 1 of 220, p = 0.05) [34]. On the other hand, recent experimental publications have demonstrated that profound hypothermia does not cause cerebral injury [35, 36] and, in marked contrast, correlated with improved neurologic outcomes. Gillinov and colleagues [36] subjected dogs to 120 minutes of DHCA either using profound hypothermia (5°C to 7°C) or deep hypothermia (18°C to 20°C). Animals were evaluated for 72 hours, then sacrificed. Animals subjected to profound hypothermia had better neurologic function and significantly less injury on histologic examination compared with animals in the deep hypothermia group [36]. Current clinical practice suggests that temperatures of 15°C and below are no worse than temperatures of 18°C to 20°C as long as the appropriate hemodilution is used [37].
Where should we monitor temperature? Recent experiments performed in our laboratory (unpublished data) suggest that the deep brain cools faster than the subcortical areas and the rest of the body, displaying a temperature gradient of 2°C to 3°C between deep brain, superficial brain, and rectal temperature. During rewarming, the deep brain temperature was significantly higher than the displayed rectal temperature, reaching temperatures as high as 38°C simultaneously with rectal temperatures of 36°C. Deep brain temperature correlated well with the pump inflow temperatures and nasopharyngeal temperature, both during cooling and rewarming. We therefore recommend that aortic inflow temperatures not exceed 37°C during rewarming, and that nasopharyngeal temperatures should be monitored closely.
Hemodilution has been used over the years in conjunction with hypothermic CPB and circulatory arrest to partially counteract some of the deleterious effects of hypothermia, such as increased blood viscosity and red blood cell rigidity. However, hemodilution reduces the oxygen-carrying capacity of blood. Clinically, a wide range of hemodilution protocols are used in conjunction with DHCA, with hematocrits ranging from 10% to 30% [38–40]. Scarce scientific data exist to support any of the currently used hemodilution protocols. Using magnetic resonance spectroscopy and near-infrared spectroscopy (NIRS), Shinoka and colleagues [41] evaluated the use of extreme (hematocrit <10%), moderate (hematocrit = 20%), and mild (hematocrit = 30%) hemodilution in conjunction with experimental DHCA. Extreme hemodilution was associated with a higher hypoxic stress during early cooling, as reflected by loss of phosphocreatine and by intracellular acidosis. The lower hematocrit group displayed a significant reduction in cytochrome aa3 levels during circulatory arrest compared with the higher hematocrit groups. In addition, neurologic scores on the first postoperative day were best preserved in the high hematocrit group (p = 0.01), although the difference diminished with time, and all the animals were neurologically comparable after 4 days [41]. Recently, a single-center randomized trial evaluated the outcomes of pediatric patients undergoing low-flow hypothermic cardiopulmonary bypass (with or without DHCA) using two hemodilution protocols (low hematocrit, 21.5% ± 2.9%; high hematocrit, 27.8% ± 3.5%). At 1 year of age, the lower hematocrit group had worse scores on the Psychomotor Developmental Index (81.9 ± 15.7 versus 89.7 ± 14.7, p = 0.08), as well as more scores that were at least 2 standard deviations below the population mean (16 of 56 [29%] versus 5 of 53 [9%], p = 0.01) [42]. Although conclusive evidence is lacking, based on the available data, it is prudent that blood hematocrit during hypothermic cardiopulmonary bypass and circulatory arrest be maintained in the 25% to 30% range.
Arterial Blood Gas Management on CPB
Two basic strategies are used in blood pH management on CPB: alpha stat and pH stat. Alpha stat management allows in vivo pH and PaCO2 to vary with temperature variations so as to keep arterial blood pH at 7.4 and PaCO2 at 40 mm Hg when measured at 37°C. This strategy is thought to optimize cellular enzymatic activity and is popularly referred to as "not-temperature corrected" [43–45]. On the other hand, in the pH stat management strategy, the goal is to maintain in vivo arterial blood pH at 7.4 and PaCO2 at 40 mm Hg regardless of temperature, by adding CO2 to the blood during cooling (temperature corrected). When sampled at 37°C, the arterial blood specimen is acidemic and hypercapnic. The pH stat is thought to improve cerebral blood flow and cerebral oxygenation and to effectively cool the brain during CPB [46–48], but at a greater risk of microembolism [49], free radical mediated injury [50], and a negative impact on brain enzymatic function [51].
The impact of arterial blood gas management during DHCA in children is still unclear. The only randomized prospective trial designed to answer this question showed no difference in early or midterm neurologic outcome between the alpha stat and pH stat groups [52, 53]. Infants less than 9 months old scheduled for open heart surgery were randomly assigned to either alpha or pH stat techniques of arterial blood gas management on CPB. Most infants did not undergo DHCA, but they all underwent deep hypothermic CPB (tympanic temperature of 18°). Although no statistically significant differences were found between the two groups in survival or other neurologic outcomes—death (2% versus 0%, p = 0.058), postoperative encephalographic seizures (9% versus 2%, p = 0.11), and clinical seizures (4% versus 2%, p = 0.44)—the encephalographic signal returned sooner among the patients assigned to the pH stat strategy. Among the homogenous D-transposition subgroup (92 patients), those assigned to the pH stat strategy tended to have a higher cardiac index despite lower inotropic support, less frequent postoperative acidosis (p = 0.02), hypotension (p = 0.05), shorter mechanical ventilation (p = 0.01), and intensive care unit stay (p = 0.01). At 1-year follow-up, the Psychomotor Developmental Index scores did not differ significantly between the alpha stat and the pH stat groups (p = 0.97). Treatment group assignment was not associated with neurologic abnormalities (p = 0.77) or EEG abnormalities (p = 0.77) at 1 year, or with parents' rating of the child's development (p = 0.99) or behavior (p = 0.27) at 2 to 4 years of age [53].
Experimental data, however, seem to suggest that pH stat arterial blood gas management provides improved brain protection during DHCA. Priestley and colleagues [54] randomly assigned 5- to 10-day-old piglets to undergo 90 minutes of DHCA using either the alpha stat or pH stat management strategy. Neurologic performance was better and functional disability was less severe in the pH stat group. In addition, histologic neuronal cell damage in the neocortex and the CA1 region of the hippocampus was less severe in the pH stat group compared with the alpha stat group [54].
Based on these observations, blood-gas management strategy using pH stat and alpha stat in succession (cooling with pH stat and switching to alpha stat before the institution of DHCA) is advocated by some [37]. Applying this combined technique to neonatal piglets resulted in improved metabolic brain suppression compared with alpha stat alone, as well as a significant enhancement in the metabolic recovery after rewarming (as reflected in cerebral metabolic rates of oxygen consumption [CMRO2], calculated from global cerebral blood flow and arterial-sagittal sinus oxygen difference) [35, 37]. From the clinician's standpoint, it should be noted that there is no difference between alpha stat and pH stat management at normothermia, and that the acid base management of the two strategies diverge as temperature moves away from normothermia. At moderate temperatures, the pH stat alpha stat issue may have little significance, and even at deep hypothermia, clinically proven benefits of one strategy over the other are difficult to come by.
A special note should be made regarding a subgroup of patients with large aortopulmonary collaterals who may have a higher incidence of neurologic abnormalities and choreoathetosis after cardiac surgery [55]. Kirshbom and colleagues [56] demonstrated in an infant piglet model that the existence of aortopulmonary collaterals caused a decrease in the rate of cooling prior to DHCA, and that even when temperatures were controlled, there was an increase in the cerebral metabolic derangement [56]. The use of pH stat strategy in the same experimental model improved cerebral metabolic recovery [51]. The use of the alpha stat strategy during cooling in patients with uncontrolled large aortopulmonary collaterals has an increased impact on the shunting of blood away from the cerebral to the pulmonic circulation owing to cerebral vasoconstriction and pulmonary vasodilatation of a large pulmonary vascular bed.
Pharmacology
Pharmacologic agents used to protect the brain during periods of ischemia exert their effect through three main mechanisms: reducing cerebral oxygen demand, increasing cerebral oxygen delivery, and arresting deleterious pathologic intracellular processes [57]. Several agents have been shown experimentally to attenuate the ischemic neurologic injury. Examples include barbiturates, volatile anesthetics, lidocaine, benzodiazepines, and calcium channel blockers [57]. The role of volatile anesthetics in neuroprotection during DHCA has been studied by Kurth and colleagues [58]. Piglets anesthesized with desflurane had better neurologic outcome, and displayed less neuronal tissue damage compared with piglets anesthesized with fentanyl alone. Attenuation of the systemic inflammatory response with either methylprednisolone [59, 60] or leukocyte filtration [61, 62] has been shown to confer neuroprotection during periods of circulatory arrest, whereas most recently a novel approach of pharmacologically induced preconditioning with diazoxide [63] has been advocated. Other agents such as the use of thromboxane A2 receptor blockers [64] and aprotinin [65] have been used experimentally to attenuate ischemic damage through periods of DHCA.
Data generated from adults cannot automatically be extrapolated to children, particularly neonates; whereas high blood glucose levels adversely affect neurologic outcomes in adults after cardiac arrests and surgery, the same does not apply to neonates [66]. Too few studies of these various pharmacologic interventions exist to allow firm conclusions regarding their efficacy, but the results suggest that further research in these areas is needed.
Intermittent Cerebral Perfusion
Cerebral energy metabolism becomes predominantly anaerobic within the first 20 minutes of deep hypothermic circulatory arrest [67]. Kimura and colleagues [67] were able to demonstrate that intermittent systemic recirculation for 10-minute periods every 20 minutes prevents cerebral anaerobic metabolism during long periods (120 minutes) of circulatory arrest. In a neonatal piglet model of DHCA, 1-minute cerebral perfusion every 15 minutes, for a 60-minute period of circulatory arrest, results in normal CMRO2 recovery; and electron microscopy evaluation was found to be identical to the one seen in control animals that did not undergo DHCA. In contrast, DHCA animals displayed impaired CMRO2 recovery, and extensive damage was demonstrated by electron microscopy characterized by perivascular intracellular and organelle edema and vascular collapse [68].
Currently, few or no clinical data exist correlating the use of intermittent cerebral perfusion during periods of circulatory arrest with improved outcomes, but based on the available experimental data, it seems reasonable to suggest that this technique should be considered whenever DHCA is chosen as the perfusion management and long periods of arrest are encountered.
Continuous Regional Brain Perfusion
For the past decade, surgical techniques have been developed to repair complex intracardiac and aortic arch lesions during cardiopulmonary bypass without the use of circulatory arrest. The choice of continuous perfusion for operations not involving the aortic arch manipulation or reconstruction simply involves standard arterial cannulation of the ascending aorta and bicaval venous cannulation with venting of the left side of the heart. When aortic arch surgery is needed, modification of systemic arterial cannulation and perfusion techniques are required. Sakamoto and associates [69] first reported selective brain perfusion during neonatal arch repair without the use of DHCA. Asou and colleagues [70] published a technique of selectively perfusing the brain using a polytetrafluoroethylene (PTFE) graft or a specially designed cannula inserted into the innominate artery during neonatal arch repair. Since then, several modifications of the original technique have been reported, all of which have a common denominator: the use of antegrade cerebral brain perfusion during arch reconstruction [71–73]. Pigula and colleagues [72] initially cannulated the main pulmonary artery and established perfusion through the patent ductus arteriosus. After creation of a proximal BT shunt, cardiopulmonary bypass and brain perfusion were achieved through the 3.5-mm shunt. Imoto and coworkers [73] achieved cerebral and systemic perfusion by cannulating both the PTFE tube graft anastomosed to the innominate artery and the descending aorta. Our current surgical practice, which has evolved since 1992 and has been used in more than 300 neonates, involves direct cannulation of the innominate artery and selective clamping of the proximal innominate, left carotid and left subclavian arteries. Continuous regional brain perfusion is achieved at flows of 20 to 30 cc·kg-1
·min-1.
Recently it has been demonstrated in an experimental model comparing DHCA and regional low-flow perfusion (RLFP) in neonatal piglets that RLFP improved neurologic outcome. Neonatal piglets subjected to 90 minutes of RLFP at a rate of 10 mL·kg-1 ·min-1 had better neurologic scores on day 1 and reduced apoptosis as seen by TUNEL compared with the DHCA group [74].
Several issues still remain unresolved when contemplating the use of regional or selective brain perfusion in complex congenital and arch repairs. First, CRBF is only required when arch repair is a component of the surgical reconstruction. Otherwise the use of standard continuous CPB is applicable in essentially all other reconstructive procedures, except in the smallest of premature infants. The use of CRBP seems intuitively rational; however, these techniques have only evolved recently. As a result, many questions remain relating to the optimal management of the standard set of variables associated with perfusion practice. Technical issues relating to the use of hypothermic antegrade cerebral perfusion, such as the perfusate temperature and the flow rate and pressure used to selectively perfuse the brain, are still unresolved. Standard methods of evaluation will likely clarify many of these issues as experience and investigations continue.
Relatively little is known about cerebral blood flow at 18°C in the neonate. Most of the data have been obtained from animal models of total body perfusion at low temperatures and extrapolated to humans. Tanaka and colleagues [75] evaluated the effect of cerebral selective perfusion pressure at 20°C on CMRO2 and cerebral blood flow. Lowering perfusion pressure gradually from 90 mm Hg to 40 mm Hg did not change cerebral blood flow or CMRO2; but when pressures were lowered below 40 mm Hg, cerebral blood flow declined abruptly, and at 30 mm Hg, CMRO2 fell significantly, suggesting that cerebral perfusion pressure should be kept within the range of cerebral autoregulation even during hypothermia [75]. In a group of children undergoing cardiac surgery, Kern and colleagues [31] demonstrated that a reduction of 45% to 70% in pump flow at 18°C to 20°C significantly reduced cerebral blood flow and CMRO2 but did not change O2 extraction, suggesting that at deep hypothermia, despite a significant reduction in pump flow rates, cerebral blood flow and cerebral oxygen supply exceed cerebral metabolic needs [31].
Using NIRS in 6 neonates undergoing RLFP, Pigula and coworkers [72] demonstrated that to maintain baseline cerebral saturation, regional perfusion had to be 20 mL·kg-1 ·min-1. Children undergoing DHCA alone showed significantly greater falls in cerebral oxygen saturations (-33.5 ± 14.6 versus -0.8 ± 5.2, p = 0.02) and change in cerebral blood volume index (-19.2 ± 14.3 versus -1.4 ± 2.7, p = 0.003) compared with neonates supported with RLFP [72].
It is difficult to conclude from the existing literature what the optimal flow for selective cerebral perfusion at low temperatures should be; perfusion pressures of 40 mm Hg and flow rates of 10 to 20 cc·kg-1 ·min-1 appear to be acceptable. These issues seem relatively easy to resolve. For example, if normal cardiac output in the neonate at normothermia is 200 cc·kg-1 ·min-1 and the brain is thought to take 20% of the normal cardiac output, then normal brain perfusion at normothermia is approximately 40 cc·kg-1 ·min-1. Using the Q10 relationship, which links metabolic rate to temperature, it is a reasonably simple task to estimate brain blood flow requirements at various degrees of hypothermia. It is interesting that these calculations are consistent with Pigula's empirically derived values above.
Another consideration is whether unilateral cerebral brain perfusion provides equal distribution of blood to both hemispheres. For that to occur, unobstructed bihemispheric blood flow through the circle of Willis must occur. Riggs and Ropp [76] studied the circle of Willis anatomy in 994 adult postmortem brain specimens and found that a perfect pattern was found only in 21% of the specimens whereas some degree of hypoplasia of the communicating arteries was found in 79%. A complete defect in the circle was rare [76]. Ye and colleagues [77] evaluated brain flow distribution using a pig model of unilateral antegrade cerebral perfusion through the right axilla. Magnetic resonance perfusion imaging showed a uniform distribution of flow in the brain, and histopathology showed normal morphology in all regions of the brain [77]. On the other hand, clinically, other investigators found differences and gradients in arterial pressures measured in head vessels on the right and left sides while unilaterally perfusing the brain during circulatory arrest [70,78], suggesting uneven distribution of blood to the hemispheres during DHCA and antegrade perfusion. Our clinical observations (unpublished) using bilateral NIRS monitoring (Invos 5100; Somanetics Corp, Lansing, Michigan) in infants shows that both frontal cortices are equally oxygenated during CRBP, suggesting that flow occurs through the circle of Willis.
There are no objective data regarding the "safe" temperature at which CRBP should be initiated. Although most surgeons use deep hypothermia, down to temperatures of 18°C [69–71], others do not. Depending on other intraoperative factors, the authors may use temperatures between 20°C and 25°C. Recently, Imoto and coworkers [79] reported that none of the 7 survivors who underwent the Norwood procedure using antegrade cerebral perfusion at moderate hypothermia (29°C to 31°C) had neurologic deficits.
|
Summary |
|---|
|
TopAbstractIntroductionPathophysiology of Neonatal...Human Clinical StudiesNeonatal Brain ProtectionSummaryAcknowledgmentsReferences |
|---|
All forms of extracorporeal circulation carry with it risks, and it remains, at the current time, the individual decision of the operating surgeon as to the specific methodology to be used during CPB. The use of DHCA does not eliminate or replace CPB; it is added to it. Therefore, the intrinsic risks of DHCA combine with those of CPB. Cardiopulmonary bypass itself involves many maneuvers: cannulations, decannulations, exposure of blood to artificial surfaces, inflammatory response, endocrine response, lack of pulsatile flow, and venting and deairing maneuvers. The use of DHCA does not eliminate or ameliorate any of these; it only reduces the length of time on CPB, which is only one of the many causative factors for brain injury related to CPB.
The available data relating to the risk of DHCA allow a limited number of definite conclusions. It is known that cessation of blood flow to the brain, no matter how cold the brain is at the time, results in almost immediate perturbation of important biochemical and bioenergetic components. It is not known with certainty what the short- and long-term functional implications of these perturbations are. There is general acknowledgment that the severity of these perturbations is directly related to the length of the ischemia time in general; however, quantitative detail is lacking. It is well known that the functional implications are dire when the perturbations are severe.
It is also unclear whether there is such a thing as a "safe period" of DHCA. However, it is known that duration of DHCA directly relates to the short- and long-term functional deficits. The 8-year follow-up data of patients with transposition of the great arteries who were randomly assigned either to DHCA or to low-flow CPB show that neurodevelopmental outcomes are not adversely affected if the duration of circulatory arrest does not exceed 41 minutes [5]. Efforts to determine a safe period are understandable; however, these efforts involve the use of many assumptions. A rationally calculated number in minutes, below which DHCA is considered safe, is misleading for at least two very important reasons. Most important, existing data only allow for a derived value that is a result of statistical modeling. Review of raw data from studies, which attempt to define a safe period, show unequivocally that brain injury can and does occur sometimes well below the modeled value, and likewise, does not always occur above the modeled value. The modeling process provides insights, but it is not valid to use it prospectively to make decisions in individual cases. The second important reason is that the endpoints used to define brain injury in neonates reflect relatively severe injury, and therefore any analysis using these endpoints will be insensitive to moderate or minor injury. One is left to conclude that brain injury related to DHCA is unpredictable on an individual basis, although the incidence of injury increases the longer the period of DHCA. Every surgeon must make the decision in individual cases whether DHCA either is required to perform the operation or is less morbid than the alternative CPB technique. Decisions will be influenced by a combination of the surgeon's own interpretation of the available, sometimes ambiguous, data on the risks of DHCA and on the surgeon's own repertoire of surgical skills related to CPB techniques and cardiac reconstruction.
Based on the available scientific data, we recommend the following in planning hypothermic cardiopulmonary bypass and DHCA. For neonates not requiring arch repair, CPB should be instituted at a rate of 200 cc·kg-1 ·min-1, cooling to moderate hypothermia (22°C to 28°C) using alpha stat strategy. After 15 minutes of cooling, flow should be reduced to 100 cc·kg-1 ·min-1 and global cerebral oxygenation should ideally be monitored by NIRS or other emerging technologies. For neonates requiring arch repair, systemic arterial cannulation should be performed in such a way to allow selective brain perfusion using the techniques previously cited. Cardiopulmonary bypass should again be instituted at 200 cc·kg-1 ·min-1. Cooling to 22°C to 25°C should be carried out over a period of 15 minutes using pH stat strategy, carefully monitored by a nasopharyngeal temperature probe. Perfusate hematocrit should be kept in a range of 25% to 30%. Once the target temperature is reached, CRBP should be instituted at a rate of 20 to 40 cc·kg-1 ·min-1, with oxygenation monitored on both hemispheres by NIRS. The head and neck arterial vessels not being perfused should be clamped to reduce steal from the brain. After the completion of the arch repair, total body perfusion should be reestablished for the remainder of the procedure. In neonates in whom DHCA is deemed necessary, CPB should be instituted at a rate of 200 cc·kg-1 ·min-1. Cooling to 18°C should be carried out over a period of 15 minutes using pH stat strategy, carefully monitored by the nasopharyngeal temperature probe. Perfusate hematocrit should be kept in a range between 25% and 30%. The use of DHCA should be minimized, but if arrest times are anticipated to extend beyond 20 minutes, intermittent cerebral perfusion should be considered. After the completion of the arch repair, total body perfusion should be reestablished, and rewarming should be carried out using alpha stat strategy, keeping inflow temperatures lower than 37°C.
Neurologic injury is the most serious and feared sequel of congenital heart surgery and DHCA. Efforts should be made to maximize cerebral protection. Future research should focus on the development of protective surgical techniques and manipulations as well as on the development of pharmacologic protective strategies based on selective agents targeted to block the effects of excitatory amino acids and apoptotic pathways.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionPathophysiology of Neonatal...Human Clinical StudiesNeonatal Brain ProtectionSummaryAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionPathophysiology of Neonatal...Human Clinical StudiesNeonatal Brain ProtectionSummaryAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Percy, S. Widman, J. A. Rizzo, M. Tranquilli, and J. A. Elefteriades Deep Hypothermic Circulatory Arrest in Patients With High Cognitive Needs: Full Preservation of Cognitive Abilities Ann. Thorac. Surg., January 1, 2009; 87(1): 117 - 123. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. C. Hirko, R. Dallasen, S. Jomura, and Y. Xu Modulation of Inflammatory Responses After Global Ischemia by Transplanted Umbilical Cord Matrix Stem Cells Stem Cells, November 1, 2008; 26(11): 2893 - 2901. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Meybohm, G. Hoffmann, J. Renner, A. Boening, E. Cavus, M. Steinfath, J. Scholz, and B. Bein Measurement of Blood Flow Index During Antegrade Selective Cerebral Perfusion with Near-Infrared Spectroscopy in Newborn Piglets Anesth. Analg., March 1, 2008; 106(3): 795 - 803. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Pretre and M. I. Turina Deep Hypothermic Circulatory Arrest Card. Surg. Adult, January 1, 2008; 3(2008): 431 - 442. [Full Text]
|
|
|
|
 |
|
 X. Pan, L. Sun, W. Ma, Y. Tang, C. Long, L. Tian, N. Liu, Z. Feng, and J. Zheng Overactivation of poly(adenosine phosphate-ribose) polymerase 1 and molecular events in neuronal injury after deep hypothermic circulatory arrest: study in a rabbit model. J. Thorac. Cardiovasc. Surg., November 1, 2007; 134(5): 1227 - 1233. [Abstract] [Full Text] [PDF] |
