Effects of reduced hematocrit on erythrocyte velocity and fluorescein transit time in the cerebral microcirculation of the mouse

Control of brain capillary blood flow

While it has been widely confirmed that cerebral blood flow is closely coupled with brain metabolism, it remains a matter of controversy whether capillary flow is directly controlled to meet the energy demands of the parenchyma. Since the capillary is known to lack smooth muscle cells, it has generally been considered that capillary flow is not regulated in situ. However, we now have increasing data supporting the physiological control of capillary flow. The observation of heterogeneity in the microcirculation in vivo has suggested that intravascular factors may be involved in the flow control, including non-Newtonian rheology, red blood cell flow, leukocyte adhesion, release of vasoactive mediators, and expression of glycoproteins on the endothelial cells. Astrocytes, a key mediator of the neurovascular unit, and intrinsic innervation may also regulate capillary flow. In addition, recent findings on pericyte contractility have attracted the attention of many researchers. Finally, based on these findings, we present a new model of flow control, the proximal integration model, in which localized neural activity is detected at nearby capillaries and the vasodilation signal is transmitted proximally along the vessel. Signals are then integrated at the precapillary arterioles and other arterioles further upstream and regulate the capillary flow.

Enhancement of cerebral blood flow using systemic hypertonic saline therapy improves outcome in patients with poor-grade spontaneous subarachnoid hemorrhage

OBJECT

Systemic administration of 23.5% hypertonic saline enhances cerebral blood flow (CBF) in patients with poor-grade spontaneous subarachnoid hemorrhage (SAH). Whether the increment of change in CBF correlates with changes in autoregulation of CBF or outcome at discharge remains unknown.

METHODS

Thirty-five patients with poor-grade spontaneous SAH received 2 ml/kg 23.5% hypertonic saline intravenously, and they underwent bedside transcranial Doppler (TCD) ultrasonography and intracranial pressure (ICP) monitoring. Seventeen of them underwent Xe-enhanced computed tomography (CT) scanning for measuring CBF. Outcome was assessed using the modified Rankin Scale (mRS) at discharge from the hospital. The data were analyzed using repeated-measurement analysis of variance and Dunnett correction. A comparison was made between patients with favorable and unfavorable outcomes using multivariate logistic regression.

RESULTS

The authors observed a maximum increase in blood pressure by 10.3% (p < 0.05) and cerebral perfusion pressure (CPP) by 21.2% (p < 0.01) at 30 minutes, followed by a maximum decrease in ICP by 93.1% (p < 0.01) at 60 minutes. Changes in ICP and CPP persisted for longer than 180 and 90 minutes, respectively. The results of TCD ultrasonography showed that the baseline autoregulation was impaired on the ipsilateral side of ruptured aneurysm, and increments in flow velocities were higher and lasted longer on the contralateral side (48.75% compared with 31.96% [p = 0.045] and 180 minutes compared with 90 minutes [p < 0.05], respectively). The autoregulation was briefly impaired on the contralateral side during the infusion. A dose-dependent effect of CBF increments on favorable outcome was seen on Xe-CT scans (mRS Score 1-3, odds ratio 1.27 per 1 ml/100 g tissue x min, p = 0.045).

CONCLUSIONS

Bolus systemic hypertonic saline therapy may be used for reversal of cerebral ischemia to normal perfusion in patients with poor-grade SAH.

Higher viscosity participates in the regulation of coronary flow via nitric oxide and indomethacin-sensitive contracting factor

Few studies have reported on the association of viscosity with coronary circulation. We evaluated the change in coronary flow after dextran was added to a perfusion solution to increase viscosity in isolated rat hearts. We also measured NOx- production induced by the change in shear stress in the coronary effluent, as a marker of NO synthesis. The baseline coronary flow was not influenced by the presence of either the cyclooxygenase inhibitor indomethacin, the thromboxane A2 (TXA2)-prostaglandin H2 (PGH2) receptor antagonist ONO-3708, or the TXA2 synthase inhibitor OKY-046. After exposure to solution containing 0.5% dextran, the coronary flow first decreased and then gradually increased until 10 min. The initial decrease in coronary flow was inhibited by indomethacin, ONO-3708, and OKY-046 individually. The gradual increase was completely inhibited by the NO inhibitor L-NAME, but not by indomethacin or ONO-3708. OKY-046 partially inhibited the increase. NOx- levels in the effluent were higher after the dextran solution was administered, and the increased NOx- levels were inhibited by L-NAME. The increased NOx- levels were not inhibited by inhibitors of the cyclooxygenase pathway. It appears that a higher viscosity of perfusion solution induced a gradual increase in NO production and was associated with increased production of indomethacin-sensitive contracting factor.

Traumatic Cerebral Vascular Injury: The Effects of Concussive Brain Injury on the Cerebral Vasculature

In terms of human suffering, medical expenses, and lost productivity, head injury is one of the major health care problems in the United States, and inadequate cerebral blood flow is an important contributor to mortality and morbidity after traumatic brain injury. Despite the importance of cerebral vascular dysfunction in the pathophysiology of traumatic brain injury, the effects of trauma on the cerebral circulation have been less well studied than the effects of trauma on the brain. Recent research has led to a better understanding of the physiologic, cellular, and molecular components and causes of traumatic cerebral vascular injury. A more thorough understanding of the direct and indirect effects of trauma on the cerebral vasculature will lead to improvements in current treatments of brain trauma as well as to the development of novel and, hopefully, more effective therapeutic strategies.

Plasma viscosity and cerebral blood flow

We hypothesized that the response of cerebral blood flow (CBF) to changing viscosity would be dependent on "baseline" CBF, with a greater influence of viscosity during high-flow conditions. Plasma viscosity was adjusted to 1.0 or 3.0 cP in rats by exchange transfusion with red blood cells diluted in lactated Ringer solution or with dextran. Cortical CBF was measured by H(2) clearance. Two groups of animals remained normoxic and normocarbic and served as controls. Other groups were made anemic, hypercapnic, or hypoxic to increase CBF. Under baseline conditions before intervention, CBF did not differ between groups and averaged 49.4 +/- 10.2 ml. 100 g(-1). min(-1) (+/-SD). In control animals, changing plasma viscosity to 1. 0 or 3.0 cP resulted in CBF of 55.9 +/- 8.6 and 42.5 +/- 12.7 ml. 100 g(-1). min(-1), respectively (not significant). During hemodilution, hypercapnia, and hypoxia with a plasma viscosity of 1. 0 cP, CBF varied from 98 to 115 ml. 100 g(-1). min(-1). When plasma viscosity was 3.0 cP during hemodilution, hypercapnia, and hypoxia, CBF ranged from 56 to 58 ml. 100 g(-1). min(-1) and was significantly reduced in each case (P < 0.05). These results support the hypothesis that viscosity has a greater role in regulation of CBF when CBF is increased. In addition, because CBF more closely followed changes in plasma viscosity (rather than whole blood viscosity), we believe that plasma viscosity may be the more important factor in controlling CBF.

7-Nitroindazole impedes erythrocyte flow response to isovolemic hemodilution in the cerebral capillary circulation

The role of nitric oxide (NO) in the mechanism of hemodilution-induced cerebral hyperemia is unclear. Based on findings in hypoxemia, the authors hypothesize that NO of neuronal origin contributes to an increase in velocity of erythrocytes in the cerebral microcirculation during anemia produced by isovolemic hemodilution. The change in erythrocyte velocity in cerebrocortical capillaries was assessed by intravital fluorescence video microscopy. A closed cranial window was implanted over the frontoparietal cortex of barbiturate-anesthetized, ventilated adult rats. Erythrocytes were labeled in vitro with fluorescein isothiocyanate and infused intravenously, and their velocity in subsurface capillaries was measured by frame-to-frame image tracking. Arterial blood was withdrawn in increments of 2 mL and replaced by serum albumin; arterial blood pressure was maintained at control level with an infusion of methoxamine. Erythrocyte velocity increased progressively, reaching 215% of baseline, as arterial hematocrit was reduced from 45% to 17%. Pretreatment of a separate group of rats with 7-nitroindazole (20 mg/kg intraperitoneally), a relatively selective inhibitor of neuronal NO synthase, abolished the increase in velocity at hematocrits greater than 20%, but the maximum velocity attained at the lowest hematocrit was similar to that in the control group. The results suggest that NO from neuronal source may contribute to the increase in capillary erythrocyte flow during moderate isovolemic hemodilution.

Effect of hemodilution on RBC velocity, supply rate, and hematocrit in the cerebral capillary network

The effect of isovolemic hemodilution on the circulation of red blood cells (RBCs) in the cerebrocortical capillary network was studied by intravital videomicroscopy with use of a closed-cranial-window technique in the rat. Velocity and supply rate of RBCs were measured by tracking the movement and counting the number of fluorescently labeled cells. Arterial blood was withdrawn in increments of 2 ml and replaced by serum albumin. Arterial blood pressure was maintained constant with an infusion of methoxamine. Both velocity and supply rate of RBCs increased, by approximately equal amounts, as arterial hematocrit was reduced from 44 to 15%. The maximum increase in RBC velocity was 4.6 and in RBC supply rate was 5.2 times the baseline value. Calculated lineal density of RBC, an index of capillary hematocrit, did not change with hemodilution. The results suggest that RBC flow and oxygen supply in the cerebral capillary network are maintained during isovolemic hemodilution. The "optimal hematocrit" is as low as 15%.

Hemodilution, cerebral O2 delivery, and cerebral blood flow: a study using hyperbaric oxygenation

Hemodilution reduces blood viscosity and O2 content (CaO2) and increases cerebral blood flow (CBF). Viscosity and CaO2 may contribute to increasing CBF after hemodilution. However, because hematocrit is the major contributor to blood viscosity and CaO2, it has been difficult to assess their relative importance. By varying blood viscosity without changing CaO2, prior investigation in hemodiluted animals has suggested that both factors play roughly equal roles. To further investigate the relationship of hemodilution, blood viscosity, CaO2, and CBF, we took the opposite approach in hemodiluted animals, i.e., we varied CaO2 without changing blood viscosity. Hyperbaric O2 was used to restore CaO2 to normal after hemodilution. Pentobarbital sodium-anesthetized rats underwent isovolumic hemodilution with 6% hetastarch, and forebrain CBF was measured with [3H]nicotine. One group of animals did not undergo hemodilution and served as controls (Con). In the three experimental groups, hematocrit was reduced from 44% to 17-19%. Con and hemodiluted (HDil) groups were ventilated with 40% O2 at 101 kPa (1 atmosphere absolute), which resulted in CaO2 values of 19.7 +/- 1.3 and 8.1 +/- 0.7 (SD) ml O2/dl, respectively. A second group of hemodiluted animals (HBar) was ventilated with 100% O2 at 506 kPa (5 atmospheres absolute) in a hyperbaric chamber, which restored CaO2 to an estimated 18.5 +/- 0.5 ml O2/dl by increasing dissolved O2. A fourth group of hemodiluted animals (HCon) served as hyperbaric controls and were ventilated with 10% O2 at 506 kPa, resulting in CaO2 of 9.1 +/- 0.6 ml O2/dl. CBF was 79 +/- 19 ml. 100 g-1. min-1 in the Con group and significantly increased to 123 +/- 9 ml. 100 g-1. min-1 in the HDil group. When CaO2 was restored to baseline with dissolved O2 in the HBar group, CBF decreased to 104 +/- 20 ml. 100 g-1. min-1. When normoxia was maintained during hyperbaric exposure in the HCon group, CBF was 125 +/- 18 ml. 100 g-1. min-1, a value indistinguishable from that in normobaric HDil animals. Our data demonstrate that the reduction in CaO2 after hemodilution is responsible for 40-60% of the increase in CBF.

Effects of acute normovolemic hemodilution on T2*-weighted images of rat brain

Acute normovolemic hemodilution (HD) was induced in anesthetized rats to assess the effect of changes in hematocrit (Hct) on signal intensity in T2*-weighted magnetic resonance (MR) images. Other relevant physiological parameters were maintained invariant. Two degrees of HD were induced: mild (Hct reduced from 42.6+/-2.2% to 33.4+/-2.1%) and moderate (Hct reduced from 44.6+/-2.7% to 26.2+/-1.7%). A two-dimensional gradient-echo sequence was used to monitor signal changes with high temporal resolution before, during, and after HD protocols. The time course of signal intensity change was closely related to that of changes in Hct. Corresponding changes in R2* (deltaR2*) with respect to the pre-HD state were calculated for the brain parenchyma. Average deltaR2* values of -0.24+/-0.06 s(-1) and -0.40+/-0.07 s(-1) were obtained for the mild and moderate HD groups, respectively, during the final 2 min of MR imaging (proximal to correlative measurements of Hct). MR measured deltaR2* values were in close agreement with the expected changes in R2* predicted from theory when the measured changes in Hct were used as independent variables. These data are in good agreement with the current understanding of the effects of changes in the intravascular concentration of deoxyhemoglobin on induced magnetic susceptibility and hold promise for quantitative measurement of brain oxygenation in vivo.

Isovolemic hemodilution normalizes the prolonged passage of red cells and plasma through cerebral microvessels in the partially ischemic forebrain of rats

The objective of this study was to determine whether hemodilution could normalize the mean transit times of red blood cells (Tr) and plasma (Tp) through cerebral microvessels in a partially ischemic brain. Wistar-Kyoto (WKY) rats, aged 30-40 weeks, were divided randomly into three groups. The first group was the nonocclusion, nonhemodilution (NN) normal control group. The second group was the occlusion, nonhemodilution (ON) group, in which animals were treated with bilateral carotid artery ligation. The third group was the occlusion-hemodilution (OH) group, in which animals were treated with bilateral common carotid artery ligation and, then, isovolemic hemodilution by replacing blood with the same volume of 3% modified fluid gelatin. Local cerebral blood flow (lCBF) and microvascular volumes of red blood cells (Vr) and plasma (Vp) in 14 brain structures were measured using 14C-iodoantipyrine, iron-55 labeled red blood cells, and 14C-inulin, respectively. The amount of oxygen delivered to local brain structures (OD), cerebral microvascular blood volume (Vb), mean transit time of blood (Tb), Tr, and Tp through cerebral microvessels were calculated from the data. Two hours after carotid artery ligation, lCBF decreased by approximately 38% in forebrain structures, 22% in rostral hindbrain areas, and 8% in the caudal hindbrain (29% for all 14 structures). The decreases in ODs were parallel with those of lCBFs, at 33, 17, and 2% in the three regions, respectively (24% for all structures). In contrast, Vb increased by 68, 37, and 16% in the three regions, respectively (48% for all structures). Tr and Tp were markedly prolonged (180% for Tr and 154% for Tp) in the forebrain regions, moderately (91% for Tr and 73% for Tp) in the rostral hindbrain, and mildly (60% for Tr and 13% for Tp) in the caudal hindbrain, with a mean increase of 136% for Tr and 111% for Tp in all structures. When data in the OH and NN groups were compared, lCBF values tended to be slightly higher and Vb values were significantly higher (p < 0.05) in the OH group. ODs in the eight forebrain structures were all significantly less (p < 0.05) in the OH group than the NN group. Tr and Tp values in the forebrain were similar between the OH and the NN groups. In conclusion, occlusion of the bilateral common carotid arteries in WKY rats causes partial forebrain ischemia, in which both Tr and Tp are prolonged. These prolongations of Tr and Tp can be normalized by isovolemic hemodilution. However, the ischemic forebrain remains hypoxic after hemodilution.

Hypertonic saline does not improve cerebral oxygen delivery after head injury and mild hemorrhage in cats

OBJECTIVES

To investigate the effects of hypertonic saline for resuscitation after mild hemorrhagic hypotension combined with fluid-percussion traumatic brain injury. Specifically, the effects of hypertonic saline on intracranial pressure, cerebral blood flow (radioactive microsphere method), cerebral oxygen delivery (cerebral oxygen delivery = cerebral blood flow x arterial oxygen content), and electroencephalographic activity were studied.

DESIGN

Randomized, controlled, intervention trial.

SETTING

University laboratory.

SUBJECTS

Thirty-four mongrel cats of either sex, anesthetized with 1.0% to 1.5% isoflurane in nitrous oxide/oxygen (70:30).

INTERVENTIONS

Anesthetized (isoflurane) cats were prepared for traumatic brain injury, and then randomly assigned to the following groups: moderate traumatic brain injury only (2.7 +/- 0.2 atmospheres [atm], group 1); mild hemorrhage (18 mL/kg) only, followed immediately by resuscitation with 10% hydroxyethyl starch in 0.9% saline in a volume equal to shed blood (group 2); or both traumatic brain injury (2.7 +/- 0.1 atm) and mild hemorrhage, followed immediately by replacement of a volume equal to shed blood of 10% hydroxyethyl starch in 0.9% saline (group 3); or 3.0% saline (group 4).

MEASUREMENTS AND MAIN RESULTS

Data were collected at baseline, at the end of hemorrhage, and at 0, 60, and 120 mins after resuscitation (or at comparable time points in group 1). Intracranial pressure in group 1 was significantly increased by traumatic brain injury at the end of hemorrhage, immediately after resuscitation, and 60 mins after resuscitation (p < .02 vs. baseline). In group 2, intracranial pressure increased significantly only immediately after resuscitation (p < .0001 vs. baseline). Groups 3 and 4 exhibited higher, although statistically insignificant, intracranial pressure increases at 60 and 120 mins after resuscitation. During resuscitation, cerebral blood flow increased significantly (p < .02 vs. baseline) in the uninjured cats. In contrast, cerebral blood flow failed to increase during resuscitation in the cats subjected to traumatic brain injury before hemorrhage and resuscitation. Although cerebral oxygen delivery in group 2 decreased significantly immediately, 60 mins, and 120 mins after resuscitation (p < .001 vs. baseline) both groups 3 and 4 had significantly lower cerebral oxygen delivery at 60 and 120 mins after resuscitation (p < .01 and p < .005, respectively, vs. group 1 at 60 mins after resuscitation, and p < .01 and p < .01, respectively, vs. group 1 at 120 mins after resuscitation).

CONCLUSIONS

After a combination of hemorrhage and traumatic brain injury, neither 10% hydroxyethyl starch nor 3.0% hypertonic saline restored cerebral oxygen delivery. Although neither trauma alone nor hemorrhage alone altered electroencephalographic activity, the combination produced significant decreases in electroencephalographic activity at 60 and 120 mins after resuscitation in groups 3 and 4, suggesting that cerebral oxygen delivery is inadequately restored by either resuscitation fluid. Therefore, traumatic brain injury abolished compensatory cerebral blood flow increases to hemodilution, and neither hydroxyethyl starch nor 3.0% hypertonic saline restored cerebral blood flow, cerebral oxygen delivery, or electroencephalographic activity after hemorrhagic hypotension after traumatic brain injury.

Intravital microreflectometry of individual pial vessels and capillary region of rat

A microscopic reflectance spectrophotometer was constructed to obtain the spectra of single pial vessels and of a region containing only capillaries (capillary region). The difference in the oxygen saturation (SO2) of hemoglobin between the regional arteriole and venule [R(A - V)] and that between the regional arteriole or capillaries [R(A - C)] were calculated. The reduction of cytochrome aa3 was also estimated in the capillary region. This method was applied to the brain surface of spontaneously breathing rats subjected to hypoxic and anemic hypoxia. On decreasing the inhaled O2 from 100 to 15%, elevation of R(A - V) and R(A - C) with slight arteriolar dilatation (though statistically not significant) was observed. Below 10% O2 (especially at 4 and 3% O2), the R(A - V) and R(A - C) decreased in spite of significant arteriolar dilatation with progressive reduction of cytochrome aa3, indicating suppression of oxygen transport to mitochondria. In the case of hemodilution down to 37% hematocrit (Ht), elevation of R(A - V) and R(A - C) occurred with a slight tendency toward arteriolar dilatation. Below 32% Ht, the R(A - V) decreased but the R(A - C) remained steady, while reduction of cytochrome aa3 progressed. Altogether, the SO2 in the capillary region decreased and the reduction of cytochrome aa3 progressed with the decline of arteriolar O2 supply in both hypoxic and anemic hypoxia.

Cerebral blood flow following normovolemic hemodilution in patients with high hematocrit

The effects on cerebral hemodynamics of venisection and a 4% albumin-saline infusion were studied in six patients with high hematocrit (mean, 51.5%). Cerebral blood flow (CBF) was measured using the xenon 133 intracarotid injection method. Blood gases were measured in arterial and jugular venous blood. Rapid two-stage hemodilution, which lowered mean hematocrit by 9 and 13%, resulted in CBF increases of 19 and 23%, respectively. Jugular venous partial pressure of oxygen and oxygen delivery capacity (CBF x arterial oxygen content) did not change significantly from baseline. The cerebral metabolic rate for oxygen increased slightly following stage 1 hemodilution but returned to baseline value following stage 2. The study lends no support to the concept that patients whose hematocrit is at the high end of the normal range have generalized cerebral hypoxia.

Velocity-diameter relationships of the microcirculation

An analytical expression for the relationship between velocity (and blood flow) and diameter for vessels in the microcirculatory range up to 900 mum diameter is presented. Both the arterial and venous systems are considered. The analytical expression for these relationships has been derived from computer analysis of data obtained from individual measurements reported in the literature. Computer analysis shows that blood velocity is a nearly linear function of diameter of blood vessels for the arteriolar tree. The relationship for blood flow as a function of diameter is Q = 108d3.101 for arteries, arterioles and capillaries in the diameter range 5 to 900 mum and for the venous system in the range 5 to 115 mum is Q = 334d2.388 where Q is in mum3 per second and d is in mum. Comparison of the results of this study for the arterial system to the Hagen-Poiseuille Law reveals close agreement, if correction is made for geometric factors of blood vessels and for the variation of viscosity with diameter for blood.

Regional cerebral blood flow in the anesthetized mouse as measured by local hydrogen clearance

Platinum microelectrodes were used to measure H2 clearance in mouse brain, and the clearance curves were used to calculate regional cerebral blood flow (rCBF). The curves were usually biexponential, whether or not electrode placement was confined to the cortex. When calculated by the height over area method, rCBF in anesthetized mice averaged 37+/-14 and 49+/-15 ml/100 gm per minute in two successive groups where cortical placement had been made. After CO2 breathing, which raised PaCO2 to 77+/-18 torr, the mean rCBF of the latter group was elevated to 70+/-36 ml/100 gm per minute. Our basal rCBF values are lower than literature values for rats or mice, when compared with data obtained by other techniques. However, our data are comparable to rat rCBF data obtainted by others using H2 electrodes and are comparable also to data for whole brain CBF obtained by a variety of methods in larger anesthetized mammals. It is possible that H2 electrodes provide low values for supposedly "cortical" rCBF in the very small mouse brain, because is such brains the electrode is usually close enough to a slow clearing compartment for the electrode reading to be influenced by that compartment. At the same time one cannot rule out the possibility that other techniques when applied to small rodents may, on occasion, produce spuriously high values for CBF. Indeed, while some studies using the latter techniques do show unusually high values for cortical flow in these animals, other studies using similar methods do not.

XIII. Cerebral circulation and metabolism in stroke. Cerebral circulation and metabolism in stroke study group

An understanding of the cerebral circulation is so fundamental to comprehension of the pathogenesis of stroke that cerebral blood flow and metabolism merit review in this series of reports. The authors recognize that the research described here is very technical in nature and may appear to have little practical application to clinical medicine. Nevertheless, these matters are basic to the development of precise methods for the measurement of regional cerebral blood flow in man which could be used to monitor the therapy of stroke with greater success than is possible at present.

Constriction of pial arterioles produced by prostaglandin F2alpha

Prostaglandin F2alpha constricted pial arterioles when locally applied to the cerebral surface. Norepinephrine and serotonin each elicted similar contractile effects. The constriction produced by F2alpha in combination with either biogenic amine was greater than the constriction elicited by F2alpha or amine acting alone. The effect of one agent on the other was additive rather than potentiating. Since F2alpha norepinephrine and serotonin are all naturally occurring agents, it is possible that their combined effect is important under pathological circumstances and this combined effect should not be overlooked in the search for single spasmogens of great potency. Before ascribing a pathologically important effect of F2alpha, either alone or in combination, evidence is required showing that doses effective in experiments are similar to the concentrations occurring during disease states and/or that vessels may become hypersensitive to F2alpha during such states.

The influence of nonrespiratory alkalosis on cerebral blood flow in cats

Cerebral blood flow was measured with the 133 Xenon clearance method during short-lasting (20 minutes) and more prolonged (90 minutes) infusions of Na 2 CO 3 solutions in anesthetized cats under controlled ventilation. The infusion protocol was regulated so as to produce a given increase in the plasma [HCO 3 - ] in the first 15 minutes, followed by a constant high plasma level for the rest of the infusion period. A high Pa co 3 level was induced before and at the end of the infusion, when prolonged infusions were made. The results indicate that, in acute experiments (20 minutes), an increase in plasma [HCO 3 - ] of 14 mEq/l does not influence CBF. During more prolonged infusions (90 minutes), an increase of 12 mEq/l produces a reduction of CBF and an increase in the CSF [HCO 3 - ]. These changes are more pronounced when the increase in plasma [HCO 3 - ] is more marked (18 mEq/l).