Cerebral and Spinal Cord Blood Flow

Nitric Oxide

Although it is unlikely to be directly involved in pressure autoregulation itself, NO is the subject of intense scrutiny as a mediator of vascular tone and as a neurotransmitter. ^, The interest in NO results from the identification of the multiple biologic roles it plays as a messenger molecule. Although until recently no evidence had shown it to have any biologic function at all in vertebrates, NO now appears to have at least the following major roles: (1) bactericidal and tumoricidal effects in white blood cells, (2) a neurotransmitter, and (3) a moderator/mediator of vascular tone, functioning as an “endothelium-derived relaxing factor.”

NO is synthesized from l -arginine by nitric oxide synthase (NOS). There are at least three isoforms of NOS: endothelial (eNOS), neuronal (nNOS), and inducible (iNOS). Of these, eNOS and nNOS exist in the normal brain, whereas iNOS synthesis can be induced by endotoxins and cytokines. Endogenous inhibitors of NOS, such as asymmetric dimethyl- l -arginine (ADMA), are produced during protein catabolism and may reach concentrations sufficient to inhibit NOS activity in the brain. NO action has been studied through the use of arginine analogues such as NG-nitro- l -arginine methyl ester ( l -NAME), 7-nitroindazole, and aminoguanidine, which can nonselectively or selectively block NO synthesis. NO appears to influence basal tone, as well as the endothelium-dependent response to acetylcholine in cerebral arteries and vasogenic dilation from stimulation of nonadrenergic, noncholinergic nerves. In general, topical, systemic, and intra-arterial application of NO donors increases CBF in several animal species. ^, Intra-arterial injection of the NO donor nitroprusside into angiographically normal territories in patients with cerebral arteriovenous malformations failed to augment CBF. A similar failure of intra-arterial nitroprusside was seen in healthy primates. ^, In contrast, a study in human volunteers found that systemic and intra-arterial administration of NG-monomethyl- l -arginine ( l -NMMA), a nonspecific inhibitor of eNOS, decreases CBF. ^, The latter findings suggest that NO may be involved in regulation of basal cerebrovascular tone. After synthesis, NO diffuses into the vascular myocyte and activates guanylate cyclase, forming cyclic guanosine monophosphate (cGMP). A protein kinase is stimulated by cGMP, resulting in phosphorylation of the light chain of myosin and thus vascular relaxation. NO may also act partly through calcitonin gene-related peptide (CGRP) and ATP-sensitive potassium (KATP) channels. NO partly acts also by suppressing endothelial generation of vasoconstrictors such as thromboxane A ₂ . In pathologic settings, such as vasospasm and hypoxia, Rho kinase, a serine threonine kinase, is emerging as a potent mediator of sustained vasoconstriction that in part acts through the NO pathway. Inhibition of Rho kinase increases cerebral blood flow. In middle cerebral artery occlusion models, Rho kinase inhibition improves neurologic outcome. ^, Rho kinase inhibition increases eNOS synthesis, and Rho kinase seems to negatively regulate eNOS activity. Calcium is intimately involved in vascular relaxation by NO. NO appears to be formed on demand and is not stored in vesicles—the traditional fate of neurotransmitters.

The role of NO in the vasodilitatory response due to changes in perfusion pressure or carbon dioxide (CO ₂ ) remains to be coherently defined. For example, nonspecific inhibition of NOS in primates does not affect pressure autoregulation but impairs response to CO ₂ . However, in humans, nonspecific inhibition of NOS results in a decrease in CBF but does not affect response to hypercapnia. In rodents, nonspecific inhibition of NOS impairs autoregulatory response to hypotension in basilar artery irrigation. While selective nNOS inhibition by 7-nitroindazole has no effect on baseline blood flow, 7-nitroindazole can prevent the increase in blood flow due to neural activation. In canines, 7-nitroindazole decreases collateral blood flow during middle cerebral artery occlusion.

Some investigators have reported that NO appears to play a role in dilation in response to CO ₂ . In other experiments, however, its participation in hypocapnia-induced vasoconstriction could not be demonstrated. Iadecola and Zhang proposed that NO plays either an “obligatory” or a “permissive” role in CO ₂ -induced vasodilation. Obligatory implies that NO directly mediates vasodilation through that mechanism. For example, topical application of glutamate agonists results in vasodilation that can be markedly attenuated by inhibition of NOS. Therefore, NO seems to play an obligatory role in glutamate-mediated vasodilation. Permissive implies that NO facilitates relaxation but near-complete inhibition of NOS only partly attenuates the vasodilator response. Because hypercapnic response is only partly attenuated by NOS inhibition, NO’s role is described as permissive, with other mechanisms also contributing to hypercapnic dilation. NO appears to play a much greater role in hypercapnic vasodilation in adults than in neonates. The site of action for CO ₂ -induced NO production may not be in the endothelium but, rather, in the perivascular structures, such as astrocytes.

The participation of NO in hypoxia-induced vasodilation does not appear to be physiologically important. ^{, ,} In regard to anesthetic effects on CBF, NO appears to interact with the cerebral vasodilatory effects of both halothane and isoflurane. The role of NO as a neurotransmitter undoubtedly will prove to be significant for care of the patient with neurologic disease through its interactions with anesthetic depth and cerebral ischemic states, ^, in particular the pathogenesis of vasospasm after subarachnoid hemorrhage (SAH). Inhibition of NO synthesis leads to vasoconstriction due to unopposed effects of endothelial prostanoids, such as thromboxane A ₂ and prostaglandin F _2α . Vascular abnormalities in disease states that significantly predispose the brain to damage, such as diabetes mellitus, may also be related to an NO-mediated mechanism.

Vasoactive Peptides

In the cerebral circulation, perivascular nerves contain several vasodilator peptides, including CGRP, substance P, and neurokinin A. Vasodilation with CGRP, unlike with substance P and neurokinin A, is independent of endothelin. CGRP acts by increasing intracellular cyclic adenosine monophosphate (cAMP) concentrations and partly mediates cerebral vasodilation in response to hypotension, cortical spreading depression, and cerebral ischemia. Vasodilation by NO is in part mediated by CGRP. ^, CGRP probably does not play a role in vasodilator response to hypoxia or hypercapnia. The physiologic roles of substance P and neurokinin A are not yet understood. Substance P may mediate vasodilation during pathologic derangements such as cerebral and meningeal inflammation and edema.

Potassium Channels

Of the several potassium channels in the cerebral vessels, two are of particular importance in the regulation of vascular tone: KATP channel and calcium-activated potassium (KCa) channel. A third potassium channel, pH-sensitive delayed rectifier potassium channel, may play a role in hypercapnia. Opening of potassium channels triggers potassium efflux from the vascular smooth muscle cell, hyperpolarizes the cell membrane, closes the voltage-dependent calcium channels, decreases calcium entry into the cells, and ultimately relaxes the muscles. KATP channels are opened by a decrease in intracellular pH and are inhibited by an increase in intracellular ATP concentrations and by sulfonylureas. Activation of KATP channels may partly mediate vasodilation by acetylcholine, CGRP, or norepinephrine (noradrenaline). KATP channels may play some role in vasodilation during hypotension, hypercapnia, acidosis, and hypoxia. KCa channel-mediated vasodilation is due partly to astrocyte-derived carbon monoxide, which diffuses into the smooth muscle cells. Large-conductance KCa (BKCa) channels are the most important of the several KCa channels found in the cerebral circulation. These channels can be selectively blocked by tetraethylammonium, charybdotoxin, and iberiotoxin. Inhibition of BKCa channels results in cerebral vasoconstriction in the large arteries, suggesting that BKCa channels may be involved in the regulation of basal cerebrovascular tone in these vessels. BKCa channels are activated by cGMP, cyclic adenosine monophosphate, and NO and are partly responsible for hypoxia-induced vasodilation of cerebral arteries. ^{, , ,}

Prostaglandins

Prostaglandins such as PGE ₂ and PGI ₂ are vasodilators but thromboxane A ₂ and PGF _2α are vasoconstrictors in the cerebral circulation. Synthesis of prostaglandin H ₂ from membrane phospholipids involves two critical enzymes, phospholipase and cyclooxygenase. Prostaglandin H ₂ is converted into other prostaglandins by subsequent enzymatic steps. Although cyclooxygenase can be inhibited by aspirin, naproxen, and indomethacin, only indomethacin impairs hypercapnic vasodilation in humans. ^,

Prostaglandins probably play a more significant role in the regulation of CBF in neonates than in adults. Inhibition of phospholipase by quinacrine hydrochloride abolishes the cerebrovascular response to hypercapnia, and hypoxia in newborn animals. Endothelial damage and indomethacin also abolish hypercapnia-induced vasodilation and the increase in cerebrospinal fluid (CSF) PGI ₂ concentrations. However, indomethacin-impaired CO ₂ reactivity can be restored by very low concentrations of PGE ₂ . This suggests that prostaglandins may not be direct mediators of hypercapnic vasodilation but that small amounts of prostaglandins are necessary for the CO ₂ response to hypercapnia to occur and that prostaglandins thus play a so-called permissive role.

Endothelin

Endothelin is a vasoactive peptide that is synthesized by the brain and the vascular endothelium. There are three isoforms of endothelin. The brain synthesizes endothelin-1 (ET-1) and endothelin-3 (ET-3) but not endothelin-2 (ET-2). The vascular endothelium synthesizes ET-1. The two receptors for endothelin are endothelin A (ETA) and endothelin B (ETB). Activation of ETA receptors causes vasoconstriction, and activation of ETB receptors may cause either vascular relaxation or constriction. Vascular relaxation is thought to be mediated by endothelin receptors on the endothelium, whereas constriction is probably mediated by endothelin receptors located on the smooth muscle cells. ETA receptors are probably more sensitive to ET-1 and ET-2 than to ET-3. The ETB receptor is equally sensitive to all isoforms of endothelin. ^, Endothelin most likely acts through influx of extracellular calcium, which is probably mediated by protein kinases. The vascular smooth muscle contraction caused by endothelin is sustained, suggesting that endothelin is not involved in rapid adjustment of cerebrovascular resistance (CVR). Topical applications of endothelin receptor antagonists do not alter resting CVR. Endothelin has been implicated in vascular spasm after SAH. ^, In experimental models of SAH, ETA and ETB receptor antagonists prevent evolution vasospam. Endothelin-induced vasospasm can also be reversed nonspecifically by calcium channel blockade and seems to be more responsive to intra-arterial nicardipine than to verapamil. ^, Early results of clinical trials showed that intravenous infusion of an ETA receptor antagonist, clazosentan, resulted in a decrease in the incidence of vasospasm after SAH. Intravenous clazosentan infusion reduces the severity of the established cerebral vasospasm. However, despite improvements in angiographic vasospasm with ETA receptor antagonists, more recent clinical analyses have shown no improvement in vasospasm-related cerebral infarction, new cerebral infarction, or case-fatality. Furthermore, treatment with ETA receptor antagonists was associated with higher incidence of pulmonary complications, hypotension, and anemia. Thus, enthusiasm for using these agents has waned.

Pressure Regulation

Conceptually, a convenient way to model the cerebral circulation is to envision a parallel system of rigid pipes in which Ohm’s law would apply:

where F is flow, P _i is input pressure, P _o is outflow pressure, and R is resistance.

The term P _i − P _o is usually referred to as cerebral perfusion pressure (CPP) and is calculated as MAP minus the outflow pressure. The cerebral venous system is compressible and may act as a “Starling resistor.” True CPP often is overestimated because a small gradient exists between systemic and cerebral vessels, which may be particularly important in patients with cerebral AVMs. It is useful to conceptualize pressure and resistance as independent variables in the preceding equation and flow as the dependent variable (ie, the pressure or resistance is affected by disease or treatment, and flow follows suit). For example, drugs exert effects on CBF by changing CPP and CVR (directly for vasodilators and indirectly by metabolic depressants).

Circulatory resistance can be modeled in terms of the Hagen–Poiseuille relationship (Eq. 2.2 ), as follows:

where l is length of conduit; μ is blood viscosity; and r is radius of vessel. Other symbol definitions were given previously. As is the case for Ohm’s law, when this equation is applied to an intact vascular system, a number of critical assumptions are clearly not met. The equation applies to newtonian fluids during nonturbulent flow through rigid tubes. Circulation, in contrast, is pulsatile with capacitance and the potential for turbulence. Also, a decrease in CPP can be a result of a decrease in systemic blood pressure or an increase in ICP or jugular venous pressure. Some groups have reported that the cerebral vascular bed responds in a similar way to changes in CPP, whether as a result of a decrease in the MAP or an increase in intracranial or jugular venous pressure. However, other investigators have reported that for a given change in CPP, the effect on vessel inner diameter due to an increase in ICP is different from that due to a decrease in MAP.

From a purely practical standpoint, examination of the previous relationship leaves little question as to why vessel diameter evolved into the preeminent mode of vascular regulation. Although viscosity and vessel length influence resistance in a linear manner, the fact that flow is proportional to the fourth power of the conduit’s radius makes this the most efficient means of controlling resistance.

In the normal individuals, CBF remains constant with CPP in the range of approximately 50 to 150 mmHg ( Fig. 2.4 ). As the ability of the cerebral vasculature to respond to changes in pressure is exhausted, CBF passively follows changes in CPP. At the extremes, resistance probably does not stay fixed. Vessel collapse and passive vascular dilation may actually potentiate the predicted decline or increase caused by CPP changes. Thus, resistance does not remain linearly related to pressure. Although the general concept put forth in Fig. 2.4 is important, it is only a statistical description of how the general population responds, and a value of 50 mmHg, even in a normotensive individual, does not guarantee that a particular patient’s cerebral circulation remains within the “autoregulatory plateau.” Individual responses vary widely. Ideally, at the lower limit of cerebral autoregulation, a near maximal vasodilation is thought to take place. However, evidence shows that even below the lower limit of autoregulation, pharmacologic vasodilation may be possible. ^, The relevance of the idealized cerebral autoregulation curve, in particular the lower limit of autoregulation, has been questioned by some writers.

In its simplest form, a cerebral autoregulation curve expressing CBF as a function of CPP is often represented by three straight lines. Two sloping lines intersect a horizontal line at points that represent the lower and upper limits of cerebral autoregulation. The horizontal segment represents the pressure-independent flow within the autoregulatory range, whereas the sloping lines represent pressure-dependent flow outside the range of autoregulation. In mathematical terms, an autoregulatory curve can be characterized by four principal autoregulatory parameters: lower limit of pressure autoregulation, upper limit of pressure autoregulation, slope below the lower limit of autoregulation, and slope above the lower limit of autoregulation. Using mathematical modeling, Gao and colleagues observed that the three previously described autoregulatory curves did not accurately predict the experimentally observed principal autoregulatory parameters ( Figs. 2.5 to 2.8 ). Computer modeling was most successful in predicting experimental results when the arterial resistive bed was compartmentalized into a series of four compartments on the basis of arterial/arteriolar diameter. These study findings suggest that there are multiple sites of autoregulation in the cerebral arterial resistive bed.

A time constant is associated with autoregulatory changes. Fig. 2.9A depicts the response of a simple tube (or a dysregulating vascular bed) to a step change in pressure. Because resistance does not change (assuming nonturbulent flow), flow passively follows the change in pressure. Fig. 2.9B depicts the response that is typical of a normal circulatory bed. With the step change in pressure comes an instantaneous drop in flow, but as the bed actively autoregulates and resistance decreases, flow gradually increases and returns to baseline. When the pressure is returned to normal, there is a transient period of hyperemia while the resistance is reset.

Venous Physiology

The influence of the cerebral venous system on overall autoregulation is unclear, primarily because of the difficulty of direct observation. The smooth muscle content and the innervation of the venous system are less extensive than those of the arterial system, and many believe that the venous system is a passive recipient of the “regulated” arterial inflow. Asymptomatic occlusion of cortical veins in animals can impair the local autoregulatory response to systemic hypotension. In addition, the venous system contains most of the cerebral blood volume (CBV); therefore slight changes in vessel diameter may have a profound effect on intracranial blood volume. Available evidence suggests that the venous system may be regulated more by neurogenic than by myogenic or metabolic factors.

Pulsatile Perfusion

Both a fast component and a slow component to the myogenic response to changes in perfusion pressure have been proposed. This consideration is of particular interest in the patient undergoing cardiac surgery. During cardiopulmonary bypass, the pulsatile variations in blood pressure transmitted to the cerebral vasculature appear to influence CBF, perhaps by interaction with endothelium-derived mediators of vascular tone. Although the importance of these effects has not been completely determined, the loss of pulsatility may worsen the outcome of a cerebral ischemic event. Sudden restoration of pulsatile perfusion to a previously dampened circulatory bed may be a mechanism to explain certain instances of cerebral hyperemia.

Cardiac Output

A proposed theory is that an increase in cardiac output may be responsible for improved CBF and outcome after SAH. However, there is little evidence for increased cardiac output as an operative mechanism of improving cerebral perfusion. Improvement in perfusion by volume loading is indirectly accomplished by improving blood rheology and directly accomplished by increasing systemic blood pressure and preventing occult decreases in systemic pressure. Studies examining the possible relationship between a change in cardiac output and a change in CBF have, for the most part, assessed the effect of drugs that increase cardiac output during either normotension or induced hypertension. Some investigators suggest, however, that during deliberate drug-induced hypotension, a decrease in cardiac output might be reflected by a decrease in CBF, even when blood pressure is kept above the lower autoregulatory threshold. The effects of altering cardiac output on CBF are more likely to be indirect effects on central venous pressure and large cerebral vessel tone (ie, sympathetic tone).

Rheologic Factors

Clinically, hematocrit is the main influence on blood viscosity, and, as shown in Eq. 2.2 , blood viscosity is a major determinant of vascular resistance. Muizelaar and associates have proposed that viscosity directly participates in hemodynamic autoregulation. As discussed later, viscosity may be the only determinant of CVR subject to manipulation in certain settings. An inverse relationship exists between hematocrit (Hct) and CBF. A continuing controversy concerns whether this relationship is, in fact, purely rheologic or a function of changes in oxygen delivery to the tissue.

Todd and coworkers demonstrated a significant CBF increase, from 30 ± 14 mL/100 g/min (baseline Hct = 42 ± 2%, mean ± SD) to 100 ± 20 mL/100 g/min at Hct = 12 ± 1% in normal cerebral hemispheres of rabbits. The increase in regional CBF was markedly smaller after focal cryogenic cerebral injury, suggesting that a CBF increase produced by hemodilution is an active vasodilatory process rather than a passive response to changing blood viscosity. In another animal experiment, when blood was replaced by ultrapurified polymerized bovine hemoglobin, the viscosity of which does not depend on shear rate, a fourfold increase in viscosity did not significantly affect CBF. This finding suggests that blood viscosity alone may not significantly affect CBF.

The Hagen–Poiseuille model does not accurately describe the behavior of flow at the microcirculatory level. ^, When red blood cells (RBCs) flow near vessel walls, they create shear forces, which add resistance. (The shear rate is the change in velocity moving from the wall toward the center of the vessel.) Therefore in all vessels the RBC velocity is faster in the center of the vessel and slower at the periphery. In small vessels, cells move faster than the plasma (the Fahraeus effect), thereby reducing microvascular hematocrit. This reduction in hematocrit causes a reduction in viscosity (the Fahraeus–Lindqvist effect). Another contribution of the smaller microvascular hematocrit is that as the vessels become progressively smaller, the relative size of the annular periphery (with reduced flow velocity) becomes larger.

Cerebral hematocrit in humans is approximately 75% of systemic values, but it is affected by Paco ₂ and presumably by other vasoactive influences. Relative hypercapnia reduces cerebral hematocrit, and it is presumed that the other vasodilators do as well.

Carbon Dioxide

CO ₂ is a powerful modulator of CVR. At one time, CO ₂ was thought to be the “coupler” between flow and metabolism, because an increase in metabolism generates CO ₂ and therefore releases a cerebral vasodilator into the local environment. Rapid diffusion across the blood–brain barrier (BBB) allows CO ₂ to modulate extracellular fluid pH and affect arteriolar resistance. Metabolically induced changes in pH in the systemic circulation do not have the same effect in the presence of an intact BBB, but metabolic production of H ⁺ released into the CSF or extracellular space from ischemic lactic acidosis does. The mechanism of vasodilation by CO ₂ may be different in adults and neonates ( Fig. 2.10 ). Evidence shows that NO and cyclic guanosine monophosphate pathways are probably more important in adults, whereas prostaglandins and cyclic adenosine monophosphate are more important in neonates. By active, though somewhat sluggish, exchange of HCO ₃ ⁻ the CSF eventually buffers itself against alterations in pH by CO ₂ diffusion. Although CO ₂ -induced cerebral vasoconstriction wanes over a period of 6 to 10 hours, this period can be variable in an individual patient. Also important in this regard are chronic states of either hypocapnia or hypercapnia, because sudden normalization of Paco ₂ can result in relative hypoperfusion or hyperperfusion.

At normotension, there is a nearly linear response of CBF at a Paco ₂ between 20 and 80 mmHg (CBF changes approximately 2% to 4% for each mmHg change in Paco ₂ ). The linearity of the response breaks down as Paco ₂ approaches the extremes. The values quoted for either percentage change or absolute levels in CBF change per unit CO ₂ are highly variable, depending on the methods employed and whether hemispheric or cortical flow is measured.

In general, doubling Paco ₂ from 40 to 80 mmHg doubles CBF, and halving Paco ₂ from 40 to 20 mmHg halves CBF. This highly reproducible cerebrovascular CO ₂ response is often used as a way of validating and comparing different CBF methods.

In a fashion analogous to blood pressure autoregulation, the CO ₂ response is limited by either maximal vasodilation at extreme hypercapnia or maximal vasoconstriction at extreme hypocapnia. Hypocapnia, however, may adversely affect cellular metabolism and shift the oxyhemoglobin dissociation curve to the left. Severe hypocapnia (approximately 10 mmHg) can result in anaerobic glucose metabolism and lactate production. ^, Although clinical experience clearly demonstrates impaired mentation with less severe degrees of hyperventilation, it is not clear whether this impairment represents impairment of tissue oxygenation or some effect of tissue alkalosis and transcellular ionic shifts. Clinically, inducing such extreme levels of hypocapnia is almost never necessary, and Paco ₂ levels below 25 mmHg are best avoided except in extraordinary circumstances. The routine use of profound hypocapnia in all neurosurgical settings should undergo reevaluation. ^,

Arteriolar tone, set by the systemic arterial blood pressure, modulates the effect of Paco ₂ on CBF. Moderate hypotension blunts the ability of the cerebral circulation to respond to changes in Paco ₂ , and severe hypotension abolishes it altogether ( Fig. 2.11 ). Conversely, Paco ₂ modifies pressure autoregulation, and from hypercapnia to hypocapnia there is a widening of the “autoregulatory plateau” ( Fig. 2.12 ).

There might be gender-based differences in CO ₂ reactivity due to the underlying levels of prostaglandins. For example, suppression of prostaglandin synthesis by indomethacin treatment causes a greater attenuation of CO ₂ reactivity in premenopausal women than in men. Paco ₂ responsiveness also varies by region. This difference may be due to the relative metabolic requirements present in each area, but this mechanism is not understood. Healthy female subjects demonstrated a greater increase in MCA flow velocity after 5% CO ₂ inhalation than male subjects. This finding confirms a gender-dependent response to CO ₂ in healthy subjects. Decreased CO ₂ reactivity can be a function of local decreases in CPP distal to a spastic or stenotic vessel. In addition, it may reflect deranged metabolism or structural damage in a number of disease states, including head injury, SAH, and ischemic cerebrovascular disease. In comatose patients, impaired CO ₂ reactivity suggests a poor prognostic outcome.

Oxygen

Within physiologic ranges, Pao ₂ does not affect CBF. Hypoxemia, however, is a potent stimulus for arteriolar dilation, as a result of tissue hypoxia and concomitant lactic acidosis, although the precise mechanism is unclear. Vasodilation in response to hypoxia probably involves adenosine and KATP channels. CBF begins to increase at a Pao ₂ of about 50 mmHg and roughly doubles at a Pao ₂ of 30 mmHg. States that impair CO ₂ reactivity are likely to interfere with O ₂ reactivity as well. The response of CBF to changes in both Pao ₂ and the oxygen content of blood is shown in Fig. 2.13 . Hyperoxia decreases CBF, producing a modest 10% to 15% decrease at 1 atmosphere. Hyperbaric oxygenation in humans decreases CBF, but high atmospheric pressure alone probably does not affect CBF.

Temperature

As is true for other organ systems, cerebral metabolism decreases with diminishing temperature. For each 1 °C decrease in body temperature, CMRO ₂ drops by approximately 7%. Alternatively, this relationship may be characterized by the metabolic temperature coefficient, Q10 , which is defined as the ratio of CMRO ₂ at temperature T, divided by the CMRO ₂ at a temperature that is 10 °C lower (T − 10). The value for cerebral Q10 in the physiologic range of 27° to 37 °C is between 2.0 and 3.0. Below 27 °C, however, Q10 increases to near 4.5. This finding has been explained on the basis of the neuroelectrical effects, wherein the major suppression of neuronal function occurs between 17 ° and 27 °C. Thus the lower Q10 between 27 ° and 37 °C simply reflects the decrease in the rates of biochemical reaction (basal CMRO ₂ ), and the higher Q10 between 17 ° and 27 °C is due to the additive effect of the decrease in neuronal function. ^, Because moderate hypothermia, without major suppression of neuronal functions, provides better neuroprotection than isoelectric doses of barbiturates, identifying the biochemical mechanisms that contribute to basal CMRO ₂ is important.

The regulation of CBF is known to be closely coupled to cerebral metabolism and it is not surprising that this hypothermia-induced reduction in CMRO ₂ is reflected by a parallel decrease in CBF. Some heterogeneity is found in this response, however; so CBF changes are most apparent in the cerebral and cerebellar cortex, less apparent in the thalamus, and not significant in the hypothalamus and brainstem.

Intraoperative hypothermia is most often encountered during cardiopulmonary bypass. CBF in this setting has been shown to correlate with nasopharyngeal temperature, with a maximum 55% reduction in CBF occurring, in one study, at the lowest measured temperature, 26 °C. This finding corresponds to a 56% calculated reduction in CMRO ₂ . CMRO ₂ continues to decrease with further lowering of temperature up to the point of EEG silence. In dogs, this level is reached at 18 °C. CBF during cardiopulmonary bypass with profound hypothermia (18° to 20 °C) is disproportionately maintained and is determined by arterial blood pressure and not pump flow rate. ^, However, during rewarming, CBF velocity remains lower than the pre-bypass value, probably because of hypothermia-induced changes in the cerebral vasculature. A period of cold full-flow reperfusion may improve cerebral perfusion during rewarming.

The effects of hypothermia and anesthetic drugs may be additive to the point at which EEG activity ceases. Thiopental administered during hypothermia in doses that enhance the hypothermia-induced suppression of EEG activity produces a further reduction in CMRO ₂ , which is paralleled by an additional decrease in CBF. Although similar effects on CMRO ₂ can be brought about by isoflurane, no additional drop in CBF appears to take place.

Autoregulation, as well as CO ₂ reactivity, is well preserved during cardiopulmonary bypass at moderate hypothermia. Some investigators suggest, however, that autoregulation may become impaired if the CO ₂ content of blood is allowed to rise. This effect can occur when exogenous CO ₂ is administered to provide a “normal” PaCO ₂ corrected to the patient’s actual temperature during “pH-stat” acid–base management. Recalculating the PaCO ₂ at 37 ° C for “alpha-stat” acid–base management reveals patients so treated to be markedly hypercapnic, which explains the grossly elevated values of CBF reported in some cardiopulmonary bypass studies. ^,

Pharmacology

Dose-related anesthetic or drug effects (eg, isoflurane, desflurane, and sevoflurane) can alter vasoactive responses just as blood pressure and CO ₂ do ( Fig. 2.14 ). ^, The significant vasodilatory effects of volatile anesthetic agents are apparent at minimum alveolar concentrations (MACs) exceeding 1.5. At higher MACs, volatile anesthetic agents can blunt the CO ₂ response or render CBF pressure passive. Absolute CO ₂ reactivity is preserved during intraoperative use of a narcotic, such as fentanyl or remifentanil. CO ₂ reactivity is also preserved with intravenous propofol anesthesia. Total intravenous anesthesia with propofol and remifentanil generally preserves response to CO ₂ and protects pressure autoregulation better than that with volatile anesthetic agents. Because of preserved flow-metabolism coupling, progressively increasing depth of propofol anesthesia results in a decrease in CBF. In contrast, volatile anesthetic agents in concentrations exceeding 1.5 MAC are associated with a disproportionate increase in CBF. Although intravenous anesthetic agents such as propofol seem to preserve flow-metabolism coupling better than volatile agents, the addition of nitrous oxide further impairs flow-metabolism coupling. In clinical practice, prophylactic mild hyperventilation is used to offset the vasodilatory effects of volatile anesthetics. Interestingly, intracarotid injections of intravenous anesthetic drugs in doses sufficient to cause burst suppression do not decrease blood flow, suggesting an uncoupling of blood flow and metabolism with intra-arterial injections. The apparent loss of flow-metabolism coupling with intra-arterial injections of anesthetic drugs may be due either to the biomechanical effects of the injection or to direct vascular effects.

The effect of ketamine, a N -methyl- D -aspartate (NMDA) receptor antagonist used with increasing frequency as an adjuvant intravenous anesthetic, on CBF is complex. In awake volunteers, sub-sedative doses of ketamine increased CBF and CMRO ₂ in some brain regions. In anesthetized patients, however, these effects can be reduced by coadministration of benzodiazepines and controlled ventilation. Ketamine did not abolish auto-regulation in normocapnic pigs. Dexmedetomidine, an intravenous α2 receptor agonist, reduces CBF ^, but animal studies have demonstrated it does not decrease CMRO ₂ . Dog studies suggest this may at least in part be a result of direct action on vasculature, not a result of systemic hypotension or decreased CMRO ₂ .

Vasoactive drugs may affect different aspects of autoregulatory behavior, as illustrated by evidence that nitroprusside impairs the ability of the circulation to maintain CBF when CPP is lowered but not when CPP is increased. Independent of autoregulatory impairment, anesthesia with volatile drugs appears to result in a trend for CBF to decrease over time in animal models. This does not, however, involve an effect on CSF pH. Not only do absolute flow levels decrease, but CO ₂ responsiveness changes as well. ^, This time-dependent CBF decrease has been proposed to be operative during cardiopulmonary bypass in humans.

The cause of these flow decreases (or, possibly, return to “normal”) has not been adequately explained. Evidence that flow does not decrease in other carefully controlled studies raises the question that this time effect may be a methodologic artifact. ^, In conditions of temperature flux, declines in CBF during the initial period of cardiopulmonary bypass with the skull closed probably reflect temperature equilibration in the brain. Interestingly, however, with the skull open and direct monitoring of cortical temperature, there does not appear to be a lag during cooling and rewarming during cardiopulmonary bypass.

Autonomic Nervous System

One of the most striking differences between the systemic and cerebral circulations is the relative lack of humoral and autonomic influences on normal cerebrovascular tone. The systemic circulation is regulated to a large extent by sympathetic nervous activity, but autonomic factors do not appear to control the cerebral circulation. Thus autonomic nerves are not necessary for regulatory responses, but they may modify these responses in several important ways.

The innervation of the cerebral vasculature is extensive, ^, involving serotonergic, adrenergic, and cholinergic systems of both intracranial and extracranial origin. The physiologic significance of this intricate and extensive system of innervation is not fully understood. One confounding factor in the interpretation of experimental studies is a marked interspecies difference in the CBF response to sympathetic stimulation. Thus in monkeys, acute sympathetic denervation has no effect on CBF, but acute sympathetic stimulation reduces CBF during normotension and during hypertension. In cats and dogs, by contrast, sympathetic stimulation has no effect during normotension. However, when acute hypertension is induced in cats by aortic ligation, electrical stimulation of the cervical sympathetic chain attenuates the increase in CBF and decreases disruption of the BBB.

Under normal circumstances, the presence of baseline sympathetic tone exerted on the cerebral vasculature in humans is controversial. The lack of baseline tone is supported by studies demonstrating that phentolamine-induced α-adrenergic receptor blockade does not affect CBF. In contrast, Hernandez and colleagues have demonstrated in monkeys that unilateral superior cervical ganglion excision leads to a 34% increase in CBF on the affected side, with no effect on autoregulation.

The effect of increased sympathetic tone on CBF in altered physiologic states, on the other hand, is well recognized. For example, using intense stimulation of the stellate ganglion in dogs, D’Alecy could produce a decrease in CBF greater than 60%. Thus acute sympathetic stimulation can shift the autoregulatory curve to the right. Reflex increases in sympathetic tone have been shown to attenuate the transient increases in CBF that are observed during severe hypertensive episodes. Sympathetic stimulation is also associated with a small decrease in the hyperemia seen during hypercapnia in normotensive rabbits. The cerebrovascular effects are more pronounced during bilateral sympathetic nerve stimulation. These effects are seen despite acidosis, which inhibits the release of norepinephrine. ^,

Sympathetic stimulation probably constricts the larger conductance and pial vessels, thereby interposing an additional “resistor” proximal to the arterioles. In those situations in which an increase in CBF occurs as a result of a rise in cerebral metabolic rate (ie, seizures), even bilateral activation of sympathetic nerves has no effect on CBF. In such situations, metabolic factors are the overwhelming determinants of CBF, with only a minimal contribution from the sympathetic nervous system.

At the lower limits of autoregulation, sympathetic activity modifies the autoregulatory response of CBF to a decrease in arterial blood pressure ( Fig. 2.15 ). At equivalent blood pressures, CBF is lower during hemorrhagic hypotension than during pharmacologically induced hypotension. Thus when reflex sympathetic constriction of larger cerebral arteries in response to hypotension is prevented by acute surgical sympathectomy or α-adrenergic receptor blockade, CBF is better maintained because autoregulation is preserved to a MAP that is 35% of control, in contrast to 65% of control pressure in untreated baboons. This observation explains why drug-induced hypotension during anesthesia is better tolerated than hypotension resulting from hemorrhagic shock. Although never studied, the sympathetic stimulation that occurs with severe pain may also shift the autoregulatory curve to the right.

Parasympathetic fibers surround the vessels of the circle of Willis and the cortical pial vessels. These fibers contain a wide variety of vasodilatory mediators, which include substance P, neurokinin A, and CGRP, whose mechanism of action was discussed earlier. Stimulation of parasympathetic fibers promotes a vasodilatory reaction to ischemia. Thus in rats rendered ischemic by branch occlusion of the MCA, sectioning of these nerves has been shown to lead to a greater cerebral infarction volume. Any protective effect, however, may be overshadowed by an increase in postischemic hyperemia mediated by stimulation of these same fibers. Parasympathetic fibers may also attenuate cerebral hyperemia after release of carotid arterial occlusion. Parasympathetic vasoconstrictor response is probably mediated by neuropeptide Y. Because of species differences, these results cannot reasonably be extrapolated to humans. In summary, despite extensive innervation of the intracerebral vessels, the purpose of these pathways currently remains unclear.

Local Neural-glial Regulation of Cerebral Blood Flow

There is an evolving paradigm shift in our understanding of local cerebral blood flow regulation. This new line of thought holds that input from neurons and glial cells, particularly astrocytes, regulates local blood flow directly by a “feed forward mechanism.” This new paradigm downplays the traditional “local metabolite” theory of a “negative feedback” loop consisting of increased metabolic products leading to local vasodilation. Evidence suggests that activation of neuronal NMDA receptors during glutamatergic excitatory neurotransmission leads to activation of nNOS and neuronal NO release. At least in the cerebellum, this NO appears to dilate cerebral vasculature directly. In the cortex, neuronally-generated NO vasodilates nearby vessels by inhibiting production of 20-hydroxy-eicosatetraenoic acid (20-HETE), a vasoconstrictor, in neighboring astrocytes.

Astrocytes reside between neurons and vascular smooth muscle cells with their endfoot processes wrapped around cerebral vessels, positioning them to serve as mediators of neurovascular communication. Several mechanisms have been suggested to describe how astrocytes may do this. Perhaps most notably, it has been proposed that glutamate released from nearby neurons during excitatory neurotransmission activates astrocyte metabotrophic glutaminate receptors (mGlutRs), leading to an increase in intracellular calcium concentration and activation of phospholipase A2. This leads to an increase in arachidonic acid production from membrane phospholipids and an increase in prorelaxant prostaglandins (likely PGE ₂ ) and epoxyeicosatrienoic acids (EETs), all of which are arachidonic acid metabolites. Others have proposed that elevations in astrocyte calcium concentration activates large-conductance Ca ^{2 +} -activated K ⁺ (BK) channels in astrocyte endfeet, leading to release of K ⁺ on neighboring vessels. This local, modest increase in extracellular K ⁺ is purported to hyperpolarize vascular smooth muscle cells and limit Ca ^{2 +} entry via voltage-gated Ca ^{2 +} channels. ^,

In a seeming contradiction, elevations in astrocyte calcium concentration and the resultant generation of arachidonic acid can also cause vasoconstriction. This likely results from the conversion of arachidonic acid to proncontractile 20-HETE in vascular smooth muscle itself. Whether astrocytes ultimately mediate prorelaxant or procontractile effects may depend on the existing vascular tone and local O ₂ concentration.