Effects of Profound Anemia on Brain Tissue Oxygen Tension, Carbon Dioxide Tension, and pH in Rabbits

Abstract

This study sought to determine the maximum tolerable limit of anemia for the brain during halothane anesthesia. Using a multiparameter sensor, we continuously monitored brain tissue oxygen tension (PO₂), carbon dioxide tension (PCO₂), and pH during profound hemodilution and subsequent transfusion. Twelve New Zealand White rabbits were anesthetized, intubated, and mechanically ventilated at a fraction of inspired oxygen (FiO₂) of 21% to produce an arterial carbon dioxide tension (PaCO₂) of 35 to 40 mm Hg. The femoral artery was cannulated to continuously monitor arterial blood pressure and to intermittently measure arterial blood gases. The electroencephalogram (EEG) was recorded throughout the course of the study. A fiberoptic sensor was inserted into the brain for the continuous measurement of brain PO₂, PCO₂, pH, and temperature. Cerebral blood flow (CBF) was measured by the hydrogen clearance method. Severe anemia was induced by repeatedly withdrawing 50-mL aliquots of blood and infusing an equal volume of 6% hetastarch. This procedure was performed four times for each rabbit. After the forth blood draw and fluid infusion, a total of 60 mL of packed red blood cells were transfused. Upon completion of the hemodilution, the hemoglobin concentration was 2.4 ± 0.3 g/dL (mean ± SEM). Brain tissue PO₂ decreased from 27 ± 3 mm Hg to a minimum of 12 ± 2 mm Hg. Brain tissue pH also decreased from 7.22 ± 0.03 to 7.12 ± 0.05 and returned to the baseline value with transfusion. Brain PCO₂ did not change significantly during the experiment. Cerebral blood flow increased from 37 ± 3 to 66 ± 15 mL × 100g⁻¹ × min⁻¹ during hemodilution and returned to baseline after infusion of red blood cells. There was some loss of EEG amplitude and the calculated cerebral metabolic rate (CMRO₂) decreased from 4.3 ± 0.6 to 1.9 ± 0.3 mL × 100g⁻¹ × min⁻¹ at the most profound level of anemia. This is the first report of which the authors are aware of continuous monitoring of brain tissue pH, PCO₂, and PO₂ during profound hemodilution and transfusion. Hemodilution results in a decrease in brain tissue PO₂. Increases in CBF and oxygen extraction can only partially compensate for the decreased oxygen carrying capacity of the blood. Decreases in brain tissue PO₂, pH, CMRO₂, and a loss of EEG amplitude suggest that the maximum tolerable limit of hemodilution was achieved in this study.

Profound levels of anemia may occur during surgery when there is massive hemorrhage treated by volume replacement with asanguineous fluids. The accurate determination of the minimal tolerable hemoglobin concentration is important to minimize allogenic blood transfusions. Although there have been numerous studies of hemodilution, many have focused only on systemic oxygen transport, hemodynamics, or myocardial function. The effects of severe anemia on brain tissue oxygenation, pH, and metabolic rate have not been fully investigated.

During hemodilution, cerebral blood flow (CBF) and oxygen extraction are known to increase in order to compensate for the decreased oxygen carrying capacity of the blood ^(1,2). By these two mechanisms, cerebral oxygenation may be partially preserved during acute anemia. Several studies have shown that the increase in CBF associated with normovolemic hemodilution to hemoglobin levels as low as 3.0 to 4.5 g/dl allows cerebral oxygen consumption (CMRO₂) to remain unchanged ^(3,4). However, at some level of hemoglobin (Hgb), compensatory mechanisms will be exhausted and CMRO₂ will begin to decrease. This critical level of Hgb has not been well defined.

The ability to continuously measure brain tissue oxygen tension (PO₂) may provide insights into the cerebral oxygen supply and consumption during severe anemia. Recently, it has become possible to continuously measure brain tissue PO₂, carbon dioxide tension (PCO₂), pH, and temperature using a multiparameter catheter. This catheter has been introduced as a new and useful monitor of cerebral perfusion. It has been shown that changes in cerebral oxygenation because of hypoxemia or ischemia can be reliably detected with these devices ^(5–7). In patients with severe head injuries, brain tissue PO₂ and jugular venous oxygen saturation show a good correlation ⁽⁸⁾.

In this study, we continuously monitored brain tissue pH, PCO₂, and PO₂ during profound hemodilution and the subsequent transfusion of red blood cells. To our knowledge, this study is unique because simultaneous measurement of these various parameters have not been previously recorded during hemodilution at the levels achieved in this study. Using these measurements, we estimated the hemoglobin concentration below which cerebral oxygenation, metabolic rate, and physiologic function cannot be maintained.

MATERIALS AND METHODS

Animals

After approval from the Animal Care and Use Committee, 12 male New Zealand white rabbits were studied. The rabbits were 2 to 3 months old and weighed 3.8 ± 0.2 kg (mean ± SEM). The animals were fasted for 24 hours before the experiment began and were housed 1 per cage at the institutional Animal Resource Center, where they received routine veterinary care.

Anesthesia and Surgery

The animals were anesthetized in a plastic box by insufflation of 5% halothane in oxygen. Endotracheal intubation was performed with a 3-mm tube, and the lungs were mechanically ventilated with 1% halothane in air, a tidal volume of 40 to 60 mL, at a respiratory rate of 12 to 18/minute, which was adjusted to maintain the arterial carbon dioxide tension (PaCO₂) at 35 to 40 mm Hg. A catheter inserted into an ear vein administered 0.9% saline at a rate of 50 mL/h. After infiltration with 0.25% bupivacaine, the groin was incised and an arterial cannula was inserted into the femoral artery. A central venous cannula was inserted into the right atrium through the external jugular vein. Mean arterial pressure (MAP) and central venous pressure (CVP) were continuously recorded and arterial blood samples were obtained to measure PaO₂, carbon dioxide tension (PaCO₂), arterial pH (1306 pH/Blood Gas Analyzer, Instrumentation Laboratory, Lexington, MA), and Hgb concentrations (482 Co-Oximeter, Instrumentation Laboratory, Lexington, MA). The rabbit's head was secured in a stereotactic frame, with the interaural line approximately 12 cm above the midchest. The scalp was infiltrated with .25% bupivacaine, incised in the midline, and reflected laterally to expose the skull. A 2-mm burr hole was drilled 4-mm posterior and 4-mm lateral to the bregma, and after perforation of the dura with microscissors, a multiparameter catheter (Neurotrend; Diametrics Inc., St. Paul, MN) was inserted into the brain tissue. The pH, PCO₂, temperature, and PO₂ sensors were distributed along the final 2 cm of the probe, which was 0.5 mm in diameter. The void between the sensors was filled with acrylamide gel containing phenol red. Changes in hydrogen ion concentration produced color changes in phenol red, which could be detected by the pH fiber optic element. The CO₂ sensor included an ion barrier permeable to CO₂. Inside the barrier, CO₂ altered the local pH, producing a color change in phenol red, which was detected by the CO₂ fiber optic elements. The PO₂ electrode is a fluorescence sensor also using a fiberoptic element. The sensors were calibrated before insertion according to the manufacturer's instructions using the appropriate calibration gases. The catheter was then advanced 2.5 cm into the brain so that all of the sensors were in tissue. The burr hole around the catheter was sealed with bone wax. Because the catheter is flexible, the exact anatomic location of the specific sensors cannot be predicted. However, since this study involves a model of global cerebral ischemia, it is unlikely that there would be significant differences from one region to the next. In a selected number of animals, the brains were dissected at the conclusion of the study. It was found that the pH electrode and the PCO₂ electrode were positioned in the basal forebrain with the thermocouple and the PO₂ electrode in the hippocampus and the parieto-occipital cortex. A second burr hole was made over the contralateral hemisphere. Through this hole, a platinum electrode, 0.5 mm in diameter, was inserted 1.5 mm into the brain cortex to measure CBF by the hydrogen clearance method. This needle electrode has a sampling volume of approximately 1 cubic millimeter and the measured values represent an average of gray and white matter blood flows. Hydrogen was periodically added to the inspired gas mixture and CBF was calculated using an AmeflowH₂ (AMEDA AG Medizin-Technik, Switzerland) device. The frontoparietal EEG was recorded using needle electrodes in the scalp and a differential amplifier (model DP-304; Warner Instruments, Hamden, CT).

Induction of Profound Anemia

To induce profound anemia, 50 mL aliquots of arterial blood were aspirated from the arterial catheter while infusing an equal volume of warmed 6% hetastarch in 0.9% sodium chloride (Hespan; DuPoint Pharmaceuticals, Wilmington, DE) by way of the external jugular vein. Monitored variables were recorded at the end of each stage of hemodilution. This procedure was performed 4 times for each rabbit at approximately 30 minute intervals. Preliminary experiments showed that this protocol would result in a hemoglobin concentration of approximately 2g/dl. Aspirated blood was refrigerated for less than 24 hours to separate the red cells by gravity sedimentation. After the fourth blood draw and hetastarch infusion, 3 transfusions of packed red blood cells (20 mL volume each) were administered. Additional 6% hetastarch in 0.9% sodium chloride bolus (10 mL) was given if the MAP decreased to < 60 mm Hg. Small doses of sodium bicarbonate were injected as needed to ensure an arterial pH > 7.25. Twenty minutes after each step in the hemodilution and blood transfusion, arterial blood samples were obtained for assessment of Hgb concentrations, blood gases, and pH. Simultaneously, brain PO₂, PCO₂, pH, temperature, MAP, and CVP were recorded. The arterial Hgb O₂ saturation (% O₂ Hgb) was calculated from pH, PO₂ and PCO₂⁽⁹⁾. Using the calculated % O₂ Hgb, arterial oxygen content (CaO₂) was calculated as (1.34 × Hgb × % O₂ Hgb) + (0.003 × PAO₂). Similarly, the brain capillary oxygen content (CbrainO₂) was calculated from Hgb concentrations, brain pH, brain PO₂, and brain PCO₂. Cerebral oxygen consumption (CMRO₂) was calculated using the following formula: CMRO₂ = CBF × (CaO₂ - CbrainO₂). Cerebral oxygen extraction ratio (OER) was calculated as (CaO₂-CbrainO₂)/CaO₂. Cerebral blood flow was measured at baseline, during most profound anemia, and after red blood cell transfusion.

Statistical Analysis

Changes in PaCO₂, PaO₂, brain temperature, MAP, CVP, CBF, CMRO₂, brain PO₂, brain PCO₂, and brain pH were tested over time for significance using a repeated measures analysis of variance. Paired comparisons between baseline values and those at the end of hemodilution and after the final transfusion were made for variables of interest.

P values < .05 were considered statistically significant. Descriptive statistics are expressed as mean ± SEM whenever appropriate.

RESULTS

Physiologic Data and the Electroencephalogram

At the completion of hemodilution, the Hgb concentration was 2.4 ± 0.3 g/dL (Table 1). Oxygen extraction ratio increased significantly from 0.61 ± 0.06 to 0.91 ± 0.02 at end of hemodilution (D4). PaO₂ and PaCO₂ remained stable during the experiment. However, arterial pH decreased from 7.32 ± 0.02 to 7.25 ± 0.02 at D4. MAP significantly decreased from 81 ± 3 to 63 ± 1 mm Hg at D4 despite an increase in the CVP from 6 ± 0 to 9 ± 1 mm Hg. The mean excess volume of 6% hetastarch used to support the systemic blood pressure during the hemodilution averaged 17 ± 2 mL per animal.

TABLE 1: Physiological variables at baseline and during anemia

The baseline CBF was 37 ± 3 mL × 100g⁻¹ × min⁻¹, which was similar to previously reported values using the hydrogen clearance method in rabbits ⁽¹⁰⁾. There was a significant increase in CBF to 66 ± 15 mL × 100g⁻¹ × min⁻¹ (178% increase) during hemodilution and a return to baseline (35 ± 4 mL × 100g⁻¹ × min⁻¹) after infusion of red blood cells. The calculated CMRO₂ at baseline was 4.3 ± 0.6 mL × 100g⁻¹ × min⁻¹. Cerebral oxygen consumption decreased to 1.9 ± 0.3 mL × 100g⁻¹ × min⁻¹ (44% of baseline level) at D4 and returned to baseline (3.3 ± 0.5 mL × 100g⁻¹ × min⁻¹, P = .09 vs. baseline value) after blood transfusion (Fig. 1). There was some loss of EEG amplitude at the most profound level of anemia (Fig. 2).

FIG. 1.: The time course of cerebral blood flow (CBF), cerebral oxygen metabolism (CMRO₂), and cerebral oxygen transport (CTO₂). Data are expressed as mean ± SEM. *A significant difference (P < .05) versus baseline. Hemodilution = fourth blood draw; After transfusion = third red blood infusion.

FIG. 2.: Typical electroencephalogram at baseline (hemoglobin concentration: 14.5 g/dl) and during anemia (D4, hemoglobin concentration: 2.1 g/dl)

P02 Brain, PC02 Brain, and pH Brain

Mean brain baseline values were as follows: PO₂ = 27 ± 3 mm Hg, PCO₂ = 52 ± 2 mm Hg, and pH = 7.22 ± 0.03. These results were similar to values previously reported from this laboratory ⁽¹¹⁾. There was a numerical decrease in brain tissue PO₂ at each stage of hemodilution, which reached statistical significance at D4 (12 ± 2 mm Hg, 42% of baseline. After blood transfusion, brain PO₂ returned to baseline value. Brain tissue pH also decreased to 7.12 ± 0.05 at D4 and returned to baseline value after transfusion. Brain tissue PCO₂ did not change significantly during the experiment (Fig. 3).

FIG. 3.: The time course of brain tissue oxygen tension (mm Hg), brain tissue carbon dioxide tension (mm Hg), and brain tissue pH (n = 12). Data are expressed as mean ± SEM. *A significant difference (P < .05) versus baseline. D1–D4: hemodilution, I1–I3: red blood cell infusion.

DISCUSSION

Because of its high metabolic rate and low oxygen reserves, the brain requires a constant supply of oxygen. Maintenance of cerebral oxygenation may be monitored by measurement of brain tissue PO₂. Profound levels of anemia result in a decrease in blood oxygen content and may eventually decrease oxygen transport to the brain. This ultimately will result in impaired cerebral metabolism and a decrease in brain tissue PO₂. The current study demonstrated that episodes of profound anemia result in decreases in brain tissue PO₂. Cerebral oxygen consumption also decreased to approximately 50% of control levels during hemodilution, despite compensatory increases in CBF and OER. Blood transfusion reversed these changes in brain tissue PO₂, CMRO₂, CBF, and OER.

Although hemodilution is known to increase CBF, the mechanism(s) remain controversial. It is not clear whether the increase in CBF is a direct rheologic effect of a decrease in viscosity ⁽¹²⁾ or whether it may be caused by compensatory vasodilation of resistance vessels mediated by vasoactive agents such as increased levels of nitric oxide ⁽¹⁾. Other putative mediators of increased flow including extracellular hydrogen concentration have also been proposed. Whatever the mechanism, increases in CBF can only partially compensate for a decrease in oxygen content due to a fall in Hgb concentration.

A second mechanism, which maintains oxygenation during hemodilution, is an increase in the OER ⁽⁴⁾. The rightward shift of the oxygen dissociation curve secondary to a decrease in pH induces the enhanced release of oxygen ⁽²⁾. Increases in CBF and enhanced unloading of oxygen from oxyhemoglobin are two mechanisms by which cerebral oxygenation may be partially preserved during acute anemia.

Several studies have shown that normovolemic hemodilution induces an increase in CBF while CMRO₂ remains unchanged. Bauer et al. ⁽¹³⁾ reported that CBF gradually increased to 8 × the prehemodilution value when the hematocrit fell below 10% in cats anesthetized with 0.8% halothane in 70% nitrous oxide and 30% oxygen. The calculated oxygen delivery to the brain remained relatively unchanged and did not cause suppression of spontaneous EEG. A similar increase in CBF (to 5 × the level of control) was observed when Hgb concentration fell to 3 g/dL in rats anesthetized with 70% nitrous oxide and 30% oxygen ⁽³⁾. However, Maruyama et al. reported only a 2.5-fold increase in CBF during profound hemodilution ⁽¹⁴⁾. In the current study, CBF increased to 178% of control during hemodilution. Although the lowest mean arterial blood pressure (63 ± 1 mm Hg) was within what is usually considered to the autoregulatory range, failure to maintain blood pressure in this study at baseline levels may partially explain the attenuated increase in CBF. Maruyama et al. ⁽¹⁴⁾ reported that the CBF autoregulation plateau was narrowed during hemodilution. Additional fluid infusion was not effective in maintaining arterial pressure constant but did result in an increase in CVP.

The decreased CaO₂ caused by the hemodilution in the present study was partially compensated for by an increased CBF, however, oxygen transport to the brain decreased significantly and this decrease in oxygen transport resulted in a decrease in brain tissue PO₂. At each stage of hemodilution, brain tissue PO₂ decreased before reaching a nadir of 12 ± 2 mm Hg at a Hgb level of 2.4 ± 0.3 g/dL. Similar decreases were observed by Shinozuka et al. ⁽¹⁵⁾, who found that a decrease in brain tissue PO₂ occurred at a hematocrit between 30% and 35% and that a sharp decrease to about 30% of control level occurred at a hematocrit between 15% and 20% during hemodilution. Despite increases in the OER and CBF, CMRO₂ decreased by 50% from baseline at the most profound level of hemodilution and there was a loss of EEG amplitude suggesting that functional changes begin to occur in the brain between D3 (Hgb = 5.0 g/dl) and D4 (Hgb = 2.4 g/dl).

Erdmann et al. ⁽¹⁶⁾ reported that intraneuronal PO₂ levels of 4 to 8 mm Hg are normal and maintained with extracellular PO₂ values of 9 to 50 mm Hg. These data suggest that an extracellular PO₂ below 9 mm Hg may induce critical hypoxia. Hoffman et al. ⁽⁶⁾ reported a similar result in patients with clinically-defined ischemia at a mean brain tissue PO₂ of 9 ± 6 mm Hg. Work by Valadka et al. ⁽¹⁷⁾ suggests that the longer a patient experiences a brain tissue PO₂ of < 15 mm Hg, the greater the chance of death after head injury. The brain tissue PO₂ that was observed in this study (12 ± 2 mm Hg) during profound hemodilution reached or exceeded these thresholds and therefore it is not surprising that EEG changes were observed.

A variety of PO₂ sensors are currently available, although the multiparameter sensor used in this study is unique in that it also provides PCO₂, pH, and temperature data. Valadka et al. ⁽¹⁷⁾ found that the Paratrend sensor (which was designed for continuous measurement of arterial blood gases) has a tendency to overestimate the PO₂ at very low values when implanted in brain tissue. However, the Neurotrend sensor used in this study was intended for implantation into brain tissue and was calibrated using a 0% oxygen gas to improve accuracy at low brain tissue oxygen tensions.

Brain tissue PO₂, PCO₂, pH, and temperature data were continuously monitored during hemodilution and transfusion. Brain tissue PO₂ begins to decrease with even modest decrements in hemoglobin concentration. These decreases in tissue PO₂ occurred because increases in CBF and oxygen extraction could only partially compensate for decreased oxygen-carrying capacity of the blood. During profound anemia, cerebral metabolic rate was suppressed to approximately 50% of baseline at a Hgb of 2.4 g/dl and this metabolic suppression was evidenced by EEG changes and a decrease in tissue pH. These findings suggest the maximum tolerable limit of hemodilution was achieved at Hgb levels between 2.4 and 5.0 g/dl under the conditions of this study.