Chapter 14. Deep Hypothermic Circulatory Arrest

Deep hypothermic circulatory arrest (DHCA) came about as an extension and unification of the work being done in the 1950s in cardiopulmonary bypass (Gibbon), systemic hypothermia (Bigelow), and aortic surgery (Debakey, Cooley, Crawford). Successful intracardiac surgery was based on the use of cardiopulmonary bypass (CPB); however, some procedures (eg, aortic arch surgery) could not be performed with standard CPB cannulation because of an inability to perfuse the cerebrum. Bigelow's work in the animal lab demonstrated that moderate hypothermia could be protective in cases in which CPB was halted for up to 10 minutes.1 The earliest use of DHCA in adult cardiac surgery has been accredited to Niazi and Lewis in 1958.2 Initially, DHCA was performed topically without the assistance of CPB; however, it soon became apparent that at hypothermic temperatures the heart would fibrillate or slow to a stop. Rewarming also proved to be quite difficult because of poor circulatory function during hypothermia. Many of these issues were improved on with the coordination of CPB with DHCA by the physiologist Gollan and the continued development and improvement of pump oxygenators and heat exchangers.3 Extensive work on the metabolic aspects of DHCA were studied and put into clinical practice by Griepp and others in the 1970s, leading to safe and reliable techniques that have been adopted by many into general practice.

Currently, DHCA in adult cardiac surgery is used in ascending aortic and aortic arch replacements for aneurysmal disease, dissections, and extensive aortic calcifications (porcelain aorta). Circulatory arrest has also been described for resection of complex inferior vena cava (IVC) and cardiac tumors and pulmonary thromboendarterectomies, and has been widely adopted in congenital heart surgery for the repair of complex congenital lesions.

The adult human brain is the organ most susceptible to ischemic injury. Under normal conditions, cerebral autoregulation matches cerebral blood flow (CBF) to cerebral oxygen consumption and cerebral metabolic activity (CMRO2) under a wide range of cerebral perfusion pressures.4 Yet, despite the effectiveness of cerebral autoregulation at normal physiologic perfusion pressures, the duration of cerebral ischemic tolerance at normothermia is approximately 5 minutes, making irreversible damage imminent unless adjunct perfusion methodologies or alternative strategies are employed to decrease the cerebral metabolic requirement.

The brain consumes 20% of the total body oxygen consumption, with 40% of its energy used in the preservation of cellular integrity and 60% in the transmission of nerve impulses.5 Because the brain does not have the ability to store oxygen, CMRO2 is a true index of brain metabolic activity and has been the variable of interest in studies attempting to determine the effectiveness of hypothermia in providing cerebral protection. For example, studies have demonstrated that a 10°C reduction in body temperature reduces CMRO2 by a factor of four4 (Fig. 14-1). This exponential decrease in metabolic rate differs from the linear reduction in CBF that occurs with hypothermia6; a discrepancy that results in the uncoupling between cerebral blood flow and cerebral metabolism occurring at approximately 22°C.4

Clinically, the reduction in cerebral metabolic function at hypothermic temperatures translates into periods of “safety” during circulatory arrest. Human studies have found a reasonable duration of DHCA with a low risk of neurocognitive insult to be no greater than 29 minutes at 13°C7; however, common clinical practice has pushed the limits of DHCA. It is not uncommon for DHCA to be performed at 18 to 20°C and prolonged for greater than 30 minutes; however, there continues to be controversy regarding optimal DHCA guidelines secondary to a lack of data in the current scientific literature.

During DHCA, the types of neurologic events are generally grouped into three different etiologies. These include focal cerebrovascular events, which are most often caused by thromboembolism; global cerebral necrosis, which occurs because of a failure to maintain the minimum oxygen and metabolic requirements during DHCA; and apoptosis, which occurs in specific regions of the brain and may also occur under oxygen/metabolic stress.

Focal Cerebrovascular Events

The occurrence of focal cerebrovascular events or stroke, which is not unique to DHCA but noteworthy in all instances of cardiac surgery, warrants special discussion caused by the high-risk population undergoing aortic arch surgery. The etiology of stroke in this subset of patients is not caused by a mismatch between the cerebral metabolic requirement and oxygen delivery, but rather is the consequence of an embolic event. Previous studies have found ascending aortic atherosclerosis to be an independent risk factor in the occurrence of early stroke and is likely caused by fragmentation of existing debris.8 Strategies that may prevent these intraoperative events include epiaortic scanning, which may assist in alteration of cannulation strategy.9 Furthermore, by assisting in flushing of debris from the cerebral vasculature, the adjunctive use of retrograde cerebral perfusion may be successful in reducing the incidence of focal cerebrovascular events that are caused by embolic events.

Global Cerebral Necrosis

Global cerebral necrosis occurs in those situations when the minimum oxygen and cerebral metabolic requirements are not met during DHCA. As such, the occurrence of frank cerebral necrosis is rare assuming adequate hypothermia with cerebral electrical suppression has been achieved, and the duration of circulatory arrest does not deplete the neuron's ability to maintain basic cellular processes. In instances of profound ischemia in which these requirements are not met, adenosine triphosphate (ATP) molecules are depleted and the cell reverts to anaerobic metabolism of glycolytic substrates,10 which in turn yields an inadequate amount of ATP. The depletion of ATP is the primary insult that leads to the irreversible failure of basic cellular ionic pumps and redistribution of cellular ions.10 Failure of the Na+-K+ ATPase pump results in an influx of Na+ and Cl− ions, which causes cellular swelling and neuronal depolarization,11 thereby resulting in accumulation of intracellular calcium.

The accumulation of intracellular calcium is the central and direct mediator of irreversible ischemic cellular injury.12 High intracellular concentrations of calcium causes the activation of phospholipases A and C, resulting in hydrolysis of membrane phospholipids and disruption of organelle structure and function.12 In summary, although DHCA provides adequate cerebral protection by minimizing the neuronal metabolic requirement, irreversible ischemia and frank necrosis may still occur if guidelines are not adhered to that are designed to minimize the cerebral metabolic requirement while maintaining adequate oxygen delivery.

Activation of Programmed Cerebral Death

In situations of low, but not depleted levels of ATP, the release of calcium into the cytoplasm initiates programmed cellular death, rather than uncontrolled and frank necrosis. This second mechanism of cellular death after DHCA is termed apoptosis, and occurs in the context of adequate cellular metabolism and energy stores. Apoptosis occurs after a highly specific activation of genes, receptors, and enzymes, and results in the destruction of cellular structure in a programmed and predictable manner.13 Cerebral apoptosis occurs principally via two intracellular pathways that are centered on caspases 3 and 8, a family of proteases that activate a series of biochemical cellular reactions that result in nuclear karyorrhexis and margination of the nuclear chromatin in the absence of inflammation.14 The prolonged inability of the neuron to restore calcium homeostasis and ensure the turnover of cytostructure proteins is responsible for a progressive loss of cellular function and eventual death.15 As a result, apoptosis is a progressive and active energy-dependent process characterized by the breakdown and phagocytosis of neurons.

Neuronal death in the setting of DHCA can also be accelerated secondary to excessive neuronal stimulation and hyperactivity,15 thereby causing an energetic mismatch and inducing the release of toxic neurotransmitters. Excitotoxicity is a term applied to the death of cells caused by overstimulation of excitatory amino acids (primarily glutamate), and is believed to be a fundamental process involved in postischemic neuronal cell damage.16 With partially depleted energetic stores, glutamate, the principal excitatory neurotransmitter, has been shown to have potent neurotoxic activities,17 ultimately catalyzing and accelerating neuronal cell death.

Studies have demonstrated that these apoptotic pathways, although common to all cells, occur more commonly in certain regions of the brain after DHCA.18 This concept, termed selective vulnerability, does not appear to be caused by uneven cooling19; rather, it is caused by regional cellular vulnerability.20 In the adult brain, neurons in the hippocampus, cerebellum, striatum, amygdala, lateral thalamic nucleus, and third to fifth layers of the neocortex are the primary regions in which programmed cellular death occurs.15 Cellular death via these apoptotic mechanisms begins as soon as 1 hour after reperfusion,15 and continues for several days thereafter. Clinically, patients may demonstrate neurologic deficits consistent with regions of selective apoptosis.21 These symptoms may include memory impairment, poor cognition, labile emotional state, and altered motor function. Interestingly, although clinical neurologic function continues to progressively improve throughout the postoperative period,21 ongoing cellular death may occur up to 1 week after DHCA.

Cooling

There is wide institution and surgeon variability in the methodologies of cooling during the application of DHCA. Once CPB is commenced at normothermia, the heat exchanger is used to create a temperature gradient not exceeding 7 to 10°C between pump inflow and outflow.22 The time interval required to attain equilibration of temperature between blood and tissue is directly related to the temperature gradient (between perfusate and organs), blood flow, and tissue-specific coefficients of temperature exchange. In general, standard parameters suggest that cooling should generally occupy at least 30 minutes, with some evidence suggesting that cooling times of at least 75 minutes4,23 may be necessary to ensure even and adequate cooling. A short duration of cooling results in wide variation in temperature between organs, within organs themselves, as well as a rapid increase in the rate at which core body temperature drifts upward once circulatory arrest has been achieved.

Certain factors should alter cooling strategy; occlusive vascular disease and altered vascular reactivity may reduce cerebral perfusion significantly and delay temperature equilibration. Thus, to enhance and complement deep cooling in these situations, topical cerebral cooling should be performed by placing ice packs around the head. Topical cooling assists in the maintenance of cortical and subcortical temperature while reducing heat conduction across the skull.24 Special attention should also be given to obese patients, as they often require a substantially longer period of cooling to achieve and maintain hypothermia.

Although the goal of DHCA is to obtain and maximize a period of minimal cerebral metabolism, there are little data to suggest optimal parameters for interrupting antegrade flow and commencing with circulatory arrest. Generally accepted parameters include jugular venous saturation greater than 95%, the complete absence of SSEPs and/or EEG silence, and cooling for at least 30 minutes.7 These are discussed in depth later in this chapter.

In complex aortic reconstructions requiring prolonged periods (>40 minutes) of circulatory arrest, DHCA alone may not provide adequate cerebral protection. In these situations, adjunct methodologies to DHCA, such as selective antegrade cerebral perfusion and retrograde cerebral perfusion, have been employed and have gained widespread adoption.

Selective Antegrade Cerebral Perfusion

Kazui et al. described the first use of selective antegrade cerebral perfusion in which perfusion was administered through both the innominate and left common carotid arteries.25 Since that time, selective antegrade cerebral perfusion has gained widespread popularity as a modality to provide optimal cerebral protection during complex ascending and transverse aortic reconstructions. Antegrade perfusion of the brain during DHCA is typically administered through a cannula inserted in the right axillary artery (Fig. 14-2). The axillary artery has become the artery of choice caused by its size, accessibility, and resistance to atherosclerotic disease burden. The right axillary artery is generally approached through an infraclavicular or deltopectoral groove incision. A 6-, 8-, or 10-mm Dacron graft is anastomosed in an end-to-side fashion to provide adequate flow and to ensure adequate perfusion of the right upper extremity. Direct cannulation of the axillary artery is avoided caused by the risk of inducing trauma, which may lead to dissection. The innominate artery is subsequently isolated proximal to takeoff of the right subclavian artery and perfusate temperature is usually set at 18°C with a flow between 10 and 20 mL/kg per minute and adjusted to maintain a pressure of between 40 and 50 mm Hg in the right radial artery.

The need to cannulate relatively small and often diseased arch arteries and the presence of additional cannulas in the operating field constitute the main drawbacks of the technique. Direct cannulation of the common carotid or axillary arteries can result in dissection of the arterial wall and embolism of atheromatous plaque material or air. Furthermore, arterial flow depends on proper positioning of the cannula tip within the vessel. For these reasons, some surgeons rely on a unilateral perfusion of the brain with sole cannulation and perfusion of the right subclavian artery. In these situations, the right vertebral and right common carotid artery territories are perfused in an antegrade fashion. The blood reaches the left cerebral hemisphere through the circle of Willis and, to a lesser extent, through cervicofacial connections. Therefore, it is important that the take-offs of the left common carotid and left subclavian arteries be occluded to avoid a steal of blood down these arteries. Occlusion (usually with an inflatable balloon) of the descending aorta is also a useful maneuver to improve overall body perfusion. Effective somatic perfusion (including the abdominal organs, spinal cord, and lower limb musculature) has been documented with this maneuver.26 The presence of an aberrant right subclavian artery (arteria lusoria) is a clear contraindication to the use of this perfusion method. The aberrant origin of the artery usually is identified readily by computed tomographic scanning or magnetic resonance imaging. The burst of blood from the descending aorta during opening of the aortic arch should alert the surgeon to this anatomical variation and prompt a direct cannulation of the ostium of the right and left common carotid arteries.

Sequential perfusion of the cerebral arteries provides additional safety to unilateral cerebral perfusion and avoids cannulation of small or diseased arch arteries.27 The right subclavian artery remains perfused during the whole procedure. A vascular graft is sewn immediately to a common patch of aortic wall including all the arch vessels,28 or the second branch of a multiple-arm prosthesis is anastomosed to the left common carotid artery. Perfusion then is instituted through this additional graft and enhances cerebral perfusion.

Retrograde Cerebral Perfusion

Retrograde cerebral perfusion (RCP) is performed by isolation and perfusion through the superior vena cava shortly after the interruption of antegrade arterial flow. The superior vena cava (SVC) is typically snared below the azygous vein to provide retrograde cerebral blood flow and blood flow to the anterior spinal plexus and, to a small degree, the remainder of the visceral organs (Fig. 14-3). Alternatively, the SVC is directly cannulated for bypass, the cannula snared, and arterial flow directed into the SVC from the cardiopulmonary bypass circuit after clamping the IVC cannula. Pressurization of the venous system to 20 to 25mm Hg allows for blood flow into the upper extremities, jugular veins, and spinal channels, eventually crossing the capillary bed into the arterial system and finally draining into the open aortic arch. The proposed benefits of RCP are to maintain cerebral hypothermia, flush out embolic debris and air, provide metabolic support, and allow for the removal of toxic metabolites and waste products.29

The initial report that described the use of RCP to assist in brain protection during circulatory arrest created considerable controversy.30 Investigators questioned whether perfusion of the brain via the cerebral venous system offered any metabolic delivery to the brain, whereas others felt that high venous cerebral pressure may cause edema. Based on the subsequent results of clinical and experimental studies, it is now generally accepted that RCP produces deep and homogeneous cooling of the brain hemispheres and the expulsion of solid particles or gaseous bubbles from the arch arteries.31 Although there is generally excellent backflow with RCP and an open aortic arch, occlusion of the inferior vena cava to decrease the pressure gradient between the two venous territories effectively reduces the amount of stolen blood.32 Interstitial edema is a rare but potential complication of retrograde perfusion and can lead to cerebral edema and hypertension when the perfusion pressure is set above 25 mm Hg.33

Because RCP is nonphysiologic, it is critical that the conduct of administering RCP is performed correctly. For venous-to-arterial blood flow to occur, the aortic arch must be wide open during RCP, thereby minimizing the pressure gradient between the venous and arterial vasculature. Any arterial occlusion, including severe atherosclerotic cerebrovascular disease, may impede cerebral perfusion, and may result in cerebral edema in cases of high venous pressure. Furthermore, contrary to common practice, patients should only be partially exsanguinated at the onset of circulatory arrest to assure that full venous capacitance is obtained. RCP in these situations is most likely to contribute to optimal retrograde cerebral blood flow and exceptional results when performed correctly.34

Rewarming after DHCA

Rewarming represents a critical time period during which any additional harm to neurons might induce permanent injury or death. Providing a favorable hematologic environment, ensuring optimal hemodynamic conditions, and avoiding cerebral hyperactivity should provide optimal conditions for recovery of the energy-depleted brain.14 It is crucial to begin reperfusion slowly after circulatory arrest. An initial period of “cold blood low-pressure reperfusion” washes out accumulated metabolites, buffers free radicals, and provides substrates for regeneration of high-energy molecules before the resumption of cerebral electrical activity. Furthermore, an adequate hematocrit during this reperfusion period is attractive not only because of its oxygen-carrying4 capacity, but also because of its buffer, redox, and free-radical scavenging capacity. Glycemia should be monitored closely and treated aggressively, because hyperglycemia, which is stimulated by the release of endogenous catecholamines, increases intracellular acidosis and can prevent or delay the restitution of metabolic homeostasis.35 During rewarming, cerebral vascular resistance and energetic metabolism are impaired in proportion to the severity of ischemia.36 Cerebral perfusion is reduced and glucose is derived in part from the less efficient anaerobic pathway, with oxygen coupling during oxidative phosphorylation being partially interrupted.37 This vulnerable period can last for 6 to 8 hours after initiation of reperfusion.21 During this time, an abnormally high extraction of oxygen and glucose is necessary to sustain the cerebral metabolic rate.38 Jugular venous oxygen saturation is often below 40% during this recovery period and cerebral autoregulation may not compensate for reduction in oxygen delivery, which could occur with postoperative events such as acute hypotension, hypoxemia, and anemia.

The temperature of the perfusate during the rewarming phase should be managed carefully. Detection of increased cerebral activity should prompt immediate therapeutic action, which includes deep anesthesia, appropriate sedation, and reduction of temperature. Monitoring cerebral electrical activity during the rewarming phase (and thereafter if circulatory arrest has exceeded the safe ischemic period or if signs of abnormal electrical activity are present) could help to limit the extent of secondary damage to the brain. The perfusate temperature should not exceed 37°C because a relative hypothermia may prove to be beneficial for optimal brain recovery. It is our practice to maintain perfusate temperature at 37°C for a period of at least 20 minutes when a patient has achieved normothermia to minimize a downward hypothermic trend. Electrical hyperactivity of the brain can trigger overwhelmingly destructive reactions. The disorder is not uncommon after prolonged circulatory arrest and actually is considered a sign of ischemic injury.39

Outcomes after DHCA, RCP, and SACP

Clinical neurologic deficits after DHCA encompass a variety of disorders ranging from deep coma to subtle, barely perceptible alterations in cognitive functions, to focal deficiencies and behavioral changes. In the immediate postoperative period, the return of sophisticated neurologic function is often obscured by the administration of sedative and analgesic agents. A focal deficit is caused by interruption of blood in a terminal vascular territory, usually secondary to embolism, and is often the consequence of atheromatous burden. The clinical expression, typically, is a motor-sensory deficit, aphasia, or cortical blindness. Computed tomographic scanning and magnetic resonance imaging are able to detect a sharply demarcated area of necrosis in the cerebral cortex. The prevalence of a focal deficit in clinical series ranges from 5 to 10% after elective aortic surgery with the use of deep hypothermic circulatory arrest. Age, atherosclerosis, and manipulation of aorta are more potent risk factors than the duration of circulatory arrest.24 Retrograde perfusion of the aorta during cardiopulmonary bypass (with the arterial cannula inserted into the femoral or iliac artery) also has been associated with an increased risk of focal deficit because retrograde flow can dislodge floating atheromatous plaques and thrombi loosely attached to the walls of thoracic aortic aneurysms.40

Diffuse neurologic deficits are caused by global cerebral ischemia that produces various levels of cellular dysfunction. In its mildest forms, the neurons are viable but temporarily unable to function properly. The clinical spectrum of neuropsychological disorders ranges from benign and reversible conditions such as transient confusion, stupor, delirium, and agitation to more serious and debilitating conditions such as seizures, parkinsonism, and coma. Imaging studies are often normal, although in the most severe forms, scattered areas of necrosis may appear. The wide range of incidences (from 3 to 30%) of diffuse deficits after circulatory arrest quoted in the literature reflects the subtle and often unrecognized nature of most deficits. Scrupulous postoperative evaluation of neurologic function discloses a frequency of between 10 and 20%.41,42 Age, conduct of cardiopulmonary bypass, and prolonged duration of circulatory arrest are recognized risk factors for the occurrence of diffuse deficits. Disorders that impair vascular reactivity and cerebral autoregulation, such as diabetes and hypertension, have also been associated with an increased incidence of diffuse deficits.43

With refinement in neurologic evaluation, including behavioral and cognitive testing, it appears that subtle deficits occur in a much larger proportion of patients after relatively short ischemic times. Transient neurologic dysfunction, a condition once not considered a deficit, and postoperative electroencephalographic hyperactivity appear now as definitive markers of long-lasting cerebral injury.33,41 One-fourth of patients with transient deficits perform poorly on postoperative neuropsychological testing, and the deficit, affecting mainly memory and fine motor function, persists in many of them after hospital discharge. The risk of transient neurologic deficit starts when deep hypothermic circulatory arrest exceeds 25 minutes.24 The risk initially is linearly related to the duration of circulatory arrest and rises more steeply after 50 minutes of ischemia24 (Fig. 14-4).

Based on these findings and accumulated experience, it appears that the majority of patients can support unharmed a circulatory arrest of 30 minutes at 18°C (Fig. 14-5), provided that electrocerebral silence has been obtained. No deficit, or only a transient neurologic dysfunction, is expected when the ischemic period extends to 40 minutes, provided that rewarming is performed correctly and hemodynamic stability is maintained postoperatively. With an arrest time of more than 40 minutes, neurologic deficit is prone to occur, particularly in high-risk patients, such as those presenting with diabetes, hypertension, or old age. Further cooling of the brain to 13 to 15°C reduces the risk and makes a deficit again less likely if the arrest time does not exceed 40 minutes, and renders it no more severe than a transient dysfunction if it lasts 50 minutes. In these cases, careful rewarming with close monitoring of cerebral activity, deep anesthesia, and hemodynamic stability are critical for a favorable outcome.

Outcomes after Selective Antegrade Cerebral Perfusion

Evaluating the benefit of SACP is a difficult task because most studies in the literature are limited by selection bias. The primary benefit of SACP is not to reduce the incidence of embolic events, but rather to increase the safe duration of hypothermic circulatory arrest and allow for the application of circulatory arrest under moderate temperatures.28,44 Thus, it is not surprising that the majority of clinical studies have demonstrated no reduction in the incidence of permanent stroke.27,45–49 Overall, permanent stroke incidences after SACP in the literature range from 1 to 16%,50,51 likely reflecting the diversity and complexity of the patient population in which SACP is used. Generally, however, permanent stroke after elective aortic surgery with SACP occurs in 2.5 to 5% of patients.50,52 The incidence of global neurologic dysfunction, which is caused by failure to appropriately protect the cerebrum, is reduced in patients who undergo SACP compared with DHCA alone.52 Furthermore, because SACP does not require deep hypothermia, other postoperative complications such as reintubation, renal failure, and ventilator times may also be reduced.52 There is also evidence that SACP may improve long-term survival, although the exact mechanism is not completely clear.52 Although prospective studies are lacking, it appears that SACP may be primarily advantageous in those situations in which a prolonged period of circulatory arrest is anticipated. These may include complex Type I aortic dissections and aortic arch reconstructions.

Outcomes after Retrograde Cerebral Perfusion

Recent clinical research using RCP has demonstrated encouraging results. A reduction in both mortality and neurologic damage has been documented regularly with the adjunctive use of retrograde cerebral perfusion compared with classic deep hypothermia alone.53–57 Some studies confirmed the limited capacity of retrograde perfusion to sustain cerebral metabolism.58,59 Yet, the inability of RCP to meet cerebral requirements does not indicate that it is ineffective as an adjunctive means of cerebral perfusion, because RCP has proved extremely effective at maintaining cerebral hypothermia as well as preventing intracellular acidosis, which may contribute to apoptosis and in turn result in temporary and/or permanent neurologic dysfunction.60

Most surgeons acknowledge the potential capacity of retrograde cerebral perfusion to prolong the period of safe circulatory arrest; however, it is generally not an adjunctive modality that is used for complex repairs in which a period of circulatory arrest of greater than 50 minutes is anticipated. This method should be viewed as a valuable but not an alternative adjunct to conventional methods when longer periods of circulatory arrest are anticipated. Some have found alternative uses of RCP that may prove to be beneficial. For example, at the Brigham and Women's Hospital we have recently adopted the use of DHCA with RCP for patients who undergo ascending aortic aneurysm repair to facilitate the creation of an open distal anastomosis, thereby allowing for the resection of all aneurysmal ascending aortic tissue. A recent study has demonstrated no overall difference in the incidence of neurologic events or mortality when compared with the standard aortic cross-clamp, closed distal anastomosis approach.61

Accurate temperature monitoring is vital to the prevention of cerebral injury during DHCA. There have been a number of peripheral sites used to assess body temperature, yet it has been well documented that peripheral temperature differs from core cerebral temperature62; a differential that becomes more pronounced during rapid warming or cooling. Although possible, during complex cardiovascular surgical procedures, it is simply not feasible to directly monitor cerebral temperature. As such, it becomes important to understand the limitations of temperature measurement at peripheral sites, including rectal, bladder, nasopharyngeal, and jugular venous sites. Most institutions have adopted nasopharyngeal monitoring because it appears to be the most reliable approximate of cerebral temperature. There is no single nasopharyngeal temperature that results in electrocerebral silence. In one large study, the minimal nasopharyngeal temperature necessary to obtain electrocerebral silence in all patients was 12.5°C; at the classic 18°C, 40% of patients still exhibited electrical activity.63

Accurate monitoring of temperature serves not only to assure that adequate hypothermia is achieved during circulatory arrest, but it also assists in the prevention of hyperthermia during the warming phase of CPB. Although there is a high correlation between nasopharyngeal and brain temperature during cooling, nasopharyngeal temperature underestimates brain temperature during rewarming by 2 to 3°C.64 This discrepancy is important because hyperthermia during rewarming exacerbates cerebral activity and disturbs cellular metabolism after circulatory arrest.

Arterial carbon dioxide tension is a major regulatory factor that impacts cerebral blood flow during CPB and DHCA. Two strategies for blood gas monitoring have been proposed for management during hypothermia. Alpha-stat management aims to maintain a normal pH and PaCO2 (a pH of 7.40 and a PaCO2 of 40 mm Hg) in warmed (37°C) blood. In vivo, under hypothermic conditions, when using alpha-stat management, blood becomes alkaline and hypocapnic. The second management strategy, pH-stat, aims to maintain a pH of 7.40 and PaCO2 of 40 mm Hg under hypothermic conditions. Consequently, under pH-stat management, when rewarmed to 37°C, blood becomes acidotic and hypercapnic.

Alpha-stat management preserves autoregulation of brain perfusion and optimizes cellular enzyme activity. Caused by blood alkalosis at hypothermic conditions, the curve of oxyhemoglobin dissociation is shifted to the right, corresponding to an increased affinity of oxygen for hemoglobin. With a corresponding further shift of oxyhemoglobin to the right owing to hypothermia, the availability of oxygen carried by the hemoglobin molecule becomes reduced. At deep hypothermic temperatures, oxygen dissolved in blood represents the major source of oxygen to tissues. Comparatively, the pH-stat strategy, results in a powerful and sustained dilatation of the cerebral vessels. Autoregulation of brain perfusion is lost and cerebral blood flow is increased. The time for temperature equilibration between blood and brain is shortened, resulting in quick and homogeneous cooling of the brain. Hypercapnia shifts the oxyhemoglobin dissociation curve to the left and results in an increased availability of oxygen to tissues.

There are conflicting data regarding the optimal strategy for pH management. Theoretically, both strategies are capable of maintaining energy balance and cellular hemostasis during long periods of hypothermia. Studies in the adult cardiac surgery literature have provided evidence that the alpha-stat management technique provides optimal neurological outcomes.65–67 Yet, amid animal study findings, and suspicion that infant physiology may require a different pH management strategy, retrospective68,69 and prospective studies18,70 have indicated that pH-stat management yields a decrease in patient morbidity in neonates and infants. The superiority of the pH-stat strategy has not been confirmed in adults. In fact, prospective studies in the adult cardiac surgery literature have found no difference71 and even inferior neuropsychological outcomes.65

Because the alpha-stat strategy maintains a physiologic coupling between cerebral blood flow and metabolism, it appears to result in better outcomes in adults where the risk of underperfusion or overperfusion within the brain is substantial.72 Cerebral edema which can be a consequence of cerebral overperfusion, is less likely to occur with alpha-stat management. The preservation of cerebral autoregulation may attenuate the homogeneous distribution of blood that is prone to occur in adult patients with underlying vasculopathic risk factors such as atherosclerosis, hypertension, and diabetes. Comparatively, the underlying etiology of the neonatal brain damage in patients undergoing DHCA is most often caused by hypoperfusion during periods of low systemic flow. Under pH-stat conditions cerebral oxygen supply is enhanced, thereby indicating why there may be a significant improvement in neurologic outcome compared to alpha-stat70 in this population.

The relationship between intraoperative hyperglycemia and adverse postoperative outcomes has been well established in the cardiac surgery literature.73 There are two main mechanisms of morbidity that can be the result of intraoperative hyperglycemia, especially in the context of circulatory arrest. Hyperglycemia results in the anaerobic conversion of glucose to lactate, thereby worsening intracellular acidosis and limiting cellular metabolism, which ultimately increases the likelihood of apoptosis.74 Additionally, hyperglycemia increases the release of excitatory amino acids, which can have profoundly negative effects during periods of cerebral ischemia.75 For these reasons, aggressive management of hyperglycemia with intravenous insulin administration should be pursued in patients who undergo DHCA. Although the data are sparse regarding the optimal glycemic range during DHCA, it appears that the risk of adverse neurologic events increases significantly at blood glucose levels greater than 180 mg/dL.76

Since the inception of DHCA, it has been widely accepted that a deleterious and potential fatal consequence of deep hypothermia is increased cellular rigidity and viscosity, which may increase the risk of end organ ischemic events.77 Conversely, however, hemodilution reduces the oxygen carrying capacity of the blood and in conjunction with the left-shift of the oxyhemoglobin curve, may predispose to organ damage (particularly cerebral) secondary to limited oxygen delivery. These conflicting findings have led to variability in hematocrit manipulation during DCHA with a range of post-prime CPB hematocrit values between 10 and 30%.78,79 Subsequent animal investigations have found that during DHCA, hemoglobin acts as a reservoir during the arrest period,80 with lower hematocrit levels predisposing to a higher incidence of cerebral necrosis. Previous studies have found that the optimal hematocrit that provides adequate oxygen delivery during a 60-minute course of DHCA to be approximately 30%.80

The majority of the evidence examining optimal hematocrit levels has been in the setting of CPB only, however, many of these findings have been applied to conditions of DHCA as well. For example, human studies have found that low intraoperative hematocrit levels with CPB may be independently associated with an increased risk of mortality81 with one study identifying that an intraoperative hematocrit less than 22% to be independently associated with the risk of stroke.82 Although these findings are likely generalizable to DHCA, there are other factors under manipulation during hypothermia that affect the interpretability of such data. As previously discussed, pH, temperature, and hematocrit are all intricately and independently associated with oxygen delivery during hypothermia as a result of oxyhemoglobin dissociation. Each of these parameters can encompass a wide range of values based on institutional practices, anesthesia preferences, and/or surgeon preferences. Yet, taking the interaction of these factors into account, studies have demonstrated that a hematocrit level of 30% or greater has the potential to significantly improve neurologic outcomes irrespective of other perfusion conditions.83

Electroencephalographic (EEG) monitoring during DHCA has been a useful adjunct to the assessment of cerebral function. Current EEG monitors use between 4 and 16 channels and require placement of electrodes at specific cortical sites by trained technicians. Although it has been found to be a very sensitive and clinically relevant predictor of postoperative cerebral dysfunction, EEG monitoring has poor specificity84 and often requires a dedicated technician to place electrodes and interpret tracings. As such, many centers have strayed away from continuous monitoring in lieu of other technologies that provide simpler, yet suitable alternatives to the evaluation of cerebral electrical activity and or metabolism.

Newer devices, such as the Bispectral Index (BIS) monitor, uses Fourier transformation and bispectral analysis of a one-channel processed EEG pattern to produce a single number index, which is strongly correlated with cerebral electrical activity.85 A BIS index ranges from 0 (flat line EEG) to 100 (awake state) and provides the operator with an accurate assessment of cerebral electrical activity. At the Brigham and Women's Hospital, it is currently our practice to use BIS monitoring during all circulatory arrest cases caused by its simplicity and reliability.

Somatosensory evoked potential monitoring (SSEP) is an alternative neuromonitoring technology that has also been used as a substitute to EEG. SSEP detects neural activity at the median nerves, which serve as a surrogate for cerebral activity. SSEP monitoring is generally easier to assess and interpret, because there is less electric noise interference and electrical activity is not significantly influenced by anesthetic drugs. The sensitivity and specificity of SSEP monitoring in determining the occurrence of early postoperative neurologic events approaches 100%,86 making it an extremely useful and simple modality for determining and maintaining the optimal level of hypothermia.

Jugular venous bulb saturation is viewed by some to be one of the most critical aspects of monitoring during DHCA. By assessing the SvO2, use of this modality allows the surgical team to infer the rate of cerebral oxygen uptake, and consequently ongoing cerebral metabolism. Guidelines suggest that SvO2 should exceed 95% to ensure that maximal cerebral metabolic suppression has been attained.87 Ongoing continuous oxygen extraction suggests active cerebral metabolic activity and may predispose to energy depletion and subsequent neurologic events. Studies have shown that use of a jugular venous bulb catheter is associated with lower central temperatures,88 yet their use has not been shown to be independently associated with an improvement in neurologic outcome.88 A unique advantage to the jugular venous bulb catheter is the ability to obtain a near-direct measurement of cerebral temperature, thus allowing for two modes of critical measurement, which may increase the likelihood of a safe operation.

It should be noted that the modalities that aim to assess neural cell arrest (temperature, EEG/SSEP, and jugular venous bulb saturation) each have varying levels of specificity and often produce differing results. For example, under conditions of EEG/SSEP arrest, some have found low jugular venous bulb oxygen saturation levels,89 suggesting that even in the setting of electrocerebral silence there is continuous metabolic activity. In the clinical setting, electrocerebral silence (as assessed by electroencephalography) is generally obtained at a mean nasopharyngeal temperature of 17.5°C63; however, this is not universally true, as some studies have reported continuous phasic activity at this temperature,90 which may be caused by individual patient characteristics and/or methodologies of cooling. These variations are also confounded by the fact that the brain temperature is not measured directly and can differ markedly among sites of temperature measurement. In summary, complete suppression of cerebral metabolic activity should be the goal when undergoing DHCA; however, it must often be surmised by a number of clinically relevant variables. As such, multimodality monitoring is the safest approach to assessing cerebral metabolic suppression.91

Transcranial noninvasive cerebral oximetry is a relatively new technology that uses an infrared light source contained in an adhesive patch placed on the patient's forehead. This light source transmits photons to the outer cortical layer and uses a near infrared spectroscopy to measure both arterial and venous hemoglobin (rSO2) concentration. Cerebral oximetry is primarily used through the duration of DHCA and functions to assess major changes in cerebral oxyhemoglobin concentrations that suggest an increase in cerebral metabolism that may predispose the occurrence of cerebral ischemia.

The use of rSO2 has been validated as an appropriate measure of cerebral oxygen concentration when compared with standard approaches such as SvO2. 92 Yet, because this technology is new, there are still concerns about the interpretability of the data and intervention threshold. Commonly accepted interventional thresholds include a relative desaturation of 10 scale units below the baseline reading or an absolute value less than 50.93

Barbiturates

Barbiturates have been found to decrease the energy required for synaptic transmission, thereby allowing oxygen under conditions of hypothermic circulatory arrest to be used exclusively for cellular metabolism.94 In addition to providing global cerebral protection, barbiturates appear to also be beneficial in cases of focal ischemia95 by reducing the extent and severity of cerebral infarction by mechanisms that are not completely understood. The first human study that demonstrated the neuroprotective effect of barbiturates in cardiac surgery was performed by administering thiopentone during the periods of time most likely to result in embolism; cannulating and weaning from CPB.96 There was a significant decrease in the incidence and extent of neuropsychiatric deficits in the thiopentone group. The neuroprotective benefit of barbiturates is not without risk, because the greatest benefit is appreciated at high doses, which are associated with a negative inotropic effect and may induce sedation that would prolong time to extubation. Studies using lower doses of barbiturates (thiopental at 15 to 30 mg/kg) have found that cerebral metabolism is in fact minimized,97 yet few prospective randomized clinical studies have demonstrated an appreciable effect on the reduction of neuropsychiatric events after DHCA.97 As such, widespread adoption in the use of thiopental during DHCA has not yet been appreciated.

Volatile Anesthetics

Volatile anesthetics—including halothane, isoflurane, desflurane, and sevoflurane—confer global cerebral protection at minimum alveolar concentration. Mechanistically, they provide protection via inhibition of excitotoxicity, depression of metabolic demand, and improvement of cerebral oxygenation.98 These protective effects are primarily mediated by inhibiting glutamate release, binding to N -methyl-d-aspartate and alpha-amino-3-hydroxyl-5-methyl-4-isozazole propionate receptors while upregulating GABA-aminobutyric acid receptors.99 In contrast with barbiturates, volatile anesthetics are capable of producing an isoelectric EEG with minimal effect on cardiac output and hemodynamics.100 Most of the promising data suggesting a beneficial effect of volatile anesthetics have been performed on animal models with isoflurane and have demonstrated a reduction in brain injury after permanent or transient cerebral ischemic events.101,102 However, the scarcity of clinical information makes it difficult to infer the role of these volatile anesthetics in patients who undergo DHCA.

Corticosteroids

CPB causes a systemic inflammatory response that increases capillary permeability as well as impacts vascular tone, fluid distribution, and organ function and has been implicated in postoperative cerebral dysfunction.103Many have hypothesized that this inflammatory response may exacerbate the oxidative stress that the brain experiences during periods of hypothermia and circulatory arrest, leading some to suggest the use of anti-inflammatory medications. The use of high-dose steroids has been linked to a reduction in cytokines and TNF-alpha104 and has demonstrated a survival benefit when employed during cardiac surgery and CPB likely caused by the prevention of capillary leak syndrome and decreasing cardiac output.105 Capillary leak may be particularly detrimental in the setting of DHCA since increased permeability of the blood brain barriers is associated with a higher frequency of apoptosis caused by activation of caspases.106

Corticosteroids appear to provide unique benefits when used during DHCA. In addition to their antiapoptotic properties, studies have found that steroid pretreatment is associated with a significant improvement in pulmonary function104 as well as superior cerebral oxygen metabolism and recovery of cerebral blood flow after DHCA.106 Yet, although some studies have demonstrated a neuroprotective advantage in patients who received corticosteroid pretreatment,107 others have demonstrated no appreciable clinical advantage.108 This discrepancy is likely caused by variable treatment algorithms, including dose, timing, and safe duration of DHCA making future studies necessary before widespread application is adopted.

• DHCA allows for the performance of complex ascending aortic and arch reconstructions by minimizing cerebral metabolism.
• When performed correctly, DHCA at a systemic temperature of 18°C, allows for a “safe” operative period of approximately 30 to 40 minutes.
• In cases in which a prolonged period of circulatory arrest is anticipated, selective antegrade cerebral perfusion, employed either alone or supplemented with retrograde cerebral perfusion, provides additional cerebral protection, and prolongs the “safe period” of DHCA.
• Aggressive monitoring and intraoperative management of temperature, glycemia, hematocrit, pH, and cerebral activity (via EEG, SSEPs, or transcranial Doppler monitoring) is critical to the safety of DHCA.
• Pharmacologic adjuncts such as barbiturates, corticosteroids, and the use of volatile anesthetics provide optimal cerebral protection during DHCA.