Cerebral arteriovenous oxygen difference as an estimate of cerebral blood flow in comatose patients

Full access

✓ The hypothesis that cerebral arteriovenous difference of oxygen content (AVDO2) can be used to predict cerebral blood flow (CBF) was tested in patients who were comatose due to head injury, subarachnoid hemorrhage, or cerebrovascular disease. In 51 patients CBF was measured daily for 3 to 5 days, and in 49 patients CBF was measured every 8 hours for 5 to 10 days after injury. In the latter group of patients, when a low CBF (≤ 0.2 ml/gm/min) or an increased level of cerebral lactate production (CMRL) (≤ −0.06 µmol/gm/min) was encountered, therapy was instituted to increase CBF, and measurements of CBF, AVDO2, and arteriovenous difference of lactate content (AVDL) were repeated. When data from all patients were analyzed, including those with cerebral ischemia and those without, AVDO2 had only a modest correlation with CBF (r = −0.24 in 578 measurements, p < 0.01). When patients with ischemia, indicated by an increased CMRL, were excluded from the analysis, CBF and AVDO2 had a much improved correlation (r = −0.74 in 313 measurements, p < 0.01). Most patients with a very low CBF would have been misclassified as having a normal or increased CBF based on the AVDO2 alone. However, when measurements of AVDO2 were supplemented with AVDL, four distinct CBF patterns could be distinguished. Patients with an ischemia/infarction pattern typically had a lactate-oxygen index (LOI = −AVDL/AVDO2) of 0.08 or greater and a variable AVDO2. The three nonischemic CBF patterns had an LOI of less than 0.08, and could be classified according to the AVDO2. Patients with a normal CBF (mean 0.42 ± 0.12 ml/gm/min) had an AVDO2 between 1.3 and 3.0 µmol/ml. A CBF pattern of hyperemia (mean 0.53 ± 0.18 ml/gm/min) was characterized by an AVDO2 of less than 1.3 µmol/ml. A compensated hypoperfusion CBF pattern (mean 0.23 ± 0.07 ml/gm/min) was identified by an AVDO2 of more than 3.0 µmol/min. These studies suggest that reliable estimates of CBF may be made from AVDO2 and AVDL measurements, which can be easily obtained in the intensive care unit.

Abstract

✓ The hypothesis that cerebral arteriovenous difference of oxygen content (AVDO2) can be used to predict cerebral blood flow (CBF) was tested in patients who were comatose due to head injury, subarachnoid hemorrhage, or cerebrovascular disease. In 51 patients CBF was measured daily for 3 to 5 days, and in 49 patients CBF was measured every 8 hours for 5 to 10 days after injury. In the latter group of patients, when a low CBF (≤ 0.2 ml/gm/min) or an increased level of cerebral lactate production (CMRL) (≤ −0.06 µmol/gm/min) was encountered, therapy was instituted to increase CBF, and measurements of CBF, AVDO2, and arteriovenous difference of lactate content (AVDL) were repeated. When data from all patients were analyzed, including those with cerebral ischemia and those without, AVDO2 had only a modest correlation with CBF (r = −0.24 in 578 measurements, p < 0.01). When patients with ischemia, indicated by an increased CMRL, were excluded from the analysis, CBF and AVDO2 had a much improved correlation (r = −0.74 in 313 measurements, p < 0.01). Most patients with a very low CBF would have been misclassified as having a normal or increased CBF based on the AVDO2 alone. However, when measurements of AVDO2 were supplemented with AVDL, four distinct CBF patterns could be distinguished. Patients with an ischemia/infarction pattern typically had a lactate-oxygen index (LOI = −AVDL/AVDO2) of 0.08 or greater and a variable AVDO2. The three nonischemic CBF patterns had an LOI of less than 0.08, and could be classified according to the AVDO2. Patients with a normal CBF (mean 0.42 ± 0.12 ml/gm/min) had an AVDO2 between 1.3 and 3.0 µmol/ml. A CBF pattern of hyperemia (mean 0.53 ± 0.18 ml/gm/min) was characterized by an AVDO2 of less than 1.3 µmol/ml. A compensated hypoperfusion CBF pattern (mean 0.23 ± 0.07 ml/gm/min) was identified by an AVDO2 of more than 3.0 µmol/min. These studies suggest that reliable estimates of CBF may be made from AVDO2 and AVDL measurements, which can be easily obtained in the intensive care unit.

In normal individuals, cerebral blood flow (CBF) is closely coupled to and regulated by the cerebral metabolic rate of oxygen (CMRO2), which is calculated, by the Fick equation, from the product of the arterial-jugular venous oxygen difference (AVDO2) and the CBF (CMRO2 = AVDO2 × CBF).11 Local CBF is increased or decreased depending on the tissue metabolic requirements. In certain altered physiological states, such as during seizures, changes in brain temperature, and anesthesia, CBF remains coupled to the CMRO2.9 If the CMRO2 is decreased, by anesthesia for example, then CBF will also decrease since requirements for metabolic substrates are less. If the CMRO2 is increased, for example by fever, then CBF will also increase. Because the ratio between CMRO2 and CBF does not change if these parameters are normally coupled, the cerebral AVDO2 remains constant.

In patients in coma due to trauma or metabolic encephalopathy, CMRO2 is typically reduced from a normal value of 1.5 µmol/gm/min to between 0.6 and 1.2 µmol/gm/min.7,10,12,16 If CBF remains coupled to CMRO2, then CBF will also be reduced. Normal coupling of CBF is retained in only 45% of comatose head-injured patients,10 however; in most of these patients, the CBF regulatory mechanisms are abnormal and, rather than being coupled to CMRO2, CBF is increased or decreased independently of the reduced CMRO2. In this situation, the ratio between CMRO2 and CBF will vary. As CBF changes occur, measurements of the reciprocal changes in AVDO2 might serve as an indicator of CBF adequacy. A normal AVDO2 would suggest that CBF is normally coupled to CMRO2, a decreased AVDO2 would indicate that CBF is excessive for cerebral metabolic requirements, and an elevated AVDO2 would indicate a decreased CBF. Figure 1 illustrates the nonlinear relationship between CBF and AVDO2 that would be expected in normal and pathophysiological conditions; a series of curves is displayed, each described by the formula: AVDO2 = CMRO2/CBF. Each individual curve defines the relationship between AVDO2 and CBF that would occur if CMRO2 were held constant at a particular value and CBF were varied. The brain extracts oxygen more completely than most tissues, and the normal cerebral AVDO2, shown by the horizontal dashed lines in Fig. 1, is 1.8 to 3.9 µmol/ml.6 If a coupled change in CMRO2 and CBF occurs, then AVDO2 remains unchanged, and the relationship between CBF and AVDO2 simply shifts to a new CMRO2 curve, as illustrated by the horizontal arrows. Assuming that CMRO2 remains constant, changes in AVDO2 (curved arrows) reflect uncoupled variations in CBF. A low AVDO2 suggests that CBF is elevated relative to cerebral metabolic requirements, while an increased AVDO2 suggests that CBF is low. This hypothesis is of particular interest, because technology has become available to monitor cerebral AVDO2 continuously.* Such continuous monitoring of CBF adequacy may allow early identification and treatment of secondary ischemic injury.

Fig. 1.
Fig. 1.

Relationship between arteriovenous oxygen difference (AVDO2) and cerebral blood flow (CBF). If a coupled change in cerebral metabolic rate of oxygen (CMRO2) and CBF occurs, then AVDO2 remains unchanged and the relationship between CBF and AVDO2 shifts to a new CMRO2 curve (horizontal arrows). If CMRO2 remains constant, then changes in AVDO2 reflect uncoupled changes in CBF (curved arrows).

Two conditions are obviously required for this hypothesis to be valid: CMRO2 must be relatively constant and in the expected range; alternatively, if CMRO2 does change, there must be a marker that CMRO2 has moved out of the expected range. Previous studies have shown that when a head injury is accompanied by cerebral infarction, the first condition may not be true.12 In such cases, CMRO2 is typically less than 0.6 µmol/gm/min in the presence of ischemic injury. However, these studies also showed that the characteristic elevation of cerebral lactate production may provide a satisfactory marker of the presence of significant cerebral ischemia, and therefore that CMRO2 may not be in the expected range.

The purpose of the present study was to assess the value of AVDO2 in predicting CBF and, in particular, in identifying patients with a low CBF.

Clinical Material and Methods
Patient Population and Management

From April 1, 1983, to March 31, 1986, 51 patients who were admitted to Ben Taub General Hospital in coma (Glasgow Coma Scale score ≤ 8) secondary to a head injury had measurements of CBF, CMRO2, and cerebral metabolic rate of lactate (CMRL) performed at least once a day for 3 to 5 days after injury. In the subsequent 22-month period, an additional 49 patients admitted in coma due to head injury, subarachnoid hemorrhage, or cerebrovascular disease had these parameters measured every 8 hours for the first 5 to 10 days after injury. The demographic characteristics of both groups of patients are shown in Table 1.

TABLE 1

Demographic characteristics of patients studied*

CharacteristicGroup 1Group 2
no. of cases5149
type of injury
 diffuse brain injury1211
 epidural hematoma33
 subdural hematoma1311
 intracerebral hematoma1712
 gunshot wound62
 subarachnoid hemorrhage08
 cerebrovascular disease02
sex
 male4230
 female919
Glasgow Coma Scale score
 3–51215
 6–83934
Glasgow Outcome Scale
 good recovery/moderate disability2014
 severe disability/vegetative1821
 dead1314

Group 1 patients were admitted between April 1, 1983, and March 31, 1986; Group 2 patients were admitted between April 1, 1986, and January 31, 1988.

All head-injured patients were treated by a standard protocol that emphasized early surgical evacuation of intracranial hematomas, controlled ventilation, and monitoring of intracranial pressure. Patients with subarachnoid hemorrhage were also treated by a standard protocol that included early clipping of aneurysms, with postoperative hypervolemic therapy guided by monitoring pulmonary wedge pressure. Routine medications included phenytoin, morphine for sedation, and antibiotics. Intracranial pressures greater than 20 mm Hg were treated with hyperventilation (pCO2 25 to 30 mm Hg), cerebrospinal fluid drainage, sedation and paralysis, mannitol, and (if necessary) barbiturates. Because of the marked effect of barbiturates on cerebral metabolism, CBF measurements obtained while patients were in barbiturate coma are not included in this analysis.

In the first group of patients, the presence of cerebral ischemia was identified retrospectively by the development of an infarction on computerized tomography as described in a previous study.12 It became evident that focal cerebral ischemia could be identified in the presence of a normal global CBF by an elevated cerebral lactate production, so in the second group of patients cerebral ischemia was defined as a CBF of 0.2 ml/gm/min or below or a CMRL of −0.06 µmol/gm/min or less. When changes characteristic of cerebral ischemia were found, treatments intended to increase CBF were undertaken. Treatments varied from patient to patient, but included allowing pCO2 to increase to 35 mm Hg, institution of hypervolemic hemodilution, and/or induction of hypertension by dopamine administration. The results of treatment were followed with repeated measurements of CBF, CMRO2, and CMRL. Treatment was considered successful if an increase in CMRO2 accompanied the improvement of CBF.

Cerebral Blood Flow

Cerebral blood flow was measured by the Kety-Schmidt technique, using N2O as the indicator.4,5,17 A total of 578 individual CBF measurements were obtained, 174 in the first group of 51 patients and 404 in the second group of 49 patients. A No. 18 Teflon catheter was inserted percutaneously into the internal jugular vein and positioned so that the tip was in the jugular bulb. The catheter was placed on the side of the most severe injury, or on the right side if the injury was diffuse. The correct position of the catheter tip was confirmed by an x-ray study. A No. 20 catheter was placed in the radial artery. Ten percent N2O was introduced into the patient's inspired gases in a stepwise fashion, and 10 timed samples of arterial and jugular venous blood were anaerobically collected during the first 15 minutes of N2O saturation. The N2O concentration was measured in the blood samples on an infrared analyzer using an extraction system modified from that described by Swedlow and Lewis.17 The CBF was calculated from curves fit to the measured N2O concentrations and integrated to 15 minutes (CBF-15) and to infinity.4 The coefficient of variation of repeated CBF measurements was 3%. The CBF values are reported uncorrected for pCO2. Mean CBF measured by this method in a normal adult population is 0.5 ml/gm/min.6

Cerebral Metabolism

Arterial and jugular venous blood samples were obtained simultaneously with measurement of CBF for determination of blood gases, oxygen saturation, hemoglobin, and lactate concentration. The blood gases were measured on a Corning 165/2 or 170 blood gas analyzer and the hemoglobin and oxygen saturation on an IL-282 co-oximeter. Whole-blood lactate concentrations were measured by an enzymatic method or with a lactate analyzer.§

The CMRO2 and CMRL values were calculated by multiplying the CBF-15 by the AVDO2 and the arterial-jugular venous difference of lactate (AVDL), respectively. As a measure of the ratio of the amount of glucose metabolized anaerobically to the amount metabolized aerobically, the lactate-oxygen index (LOI) was calculated by the formula, LOI = -AVDL/AVDO2. Although it does not accurately reflect the stoichiometry of glucose metabolism, the LOI is a simple calculation that does not require measurement of CBF or arterio-venous difference of glucose. The LOI value is normally less than 0.03. If CBF is decreased but cerebral oxygen consumption is maintained by increased extraction of oxygen, both AVDO2 and -AVDL will be increased and the LOI ratio will be unchanged. However, if increased oxygen extraction cannot compensate for the reduced CBF, then cerebral oxygen consumption will decrease, cerebral lactate production will increase, and the ratio LOI will be increased. In a previously reported series of patients, a LOI of 0.08 or more accurately predicted increased cerebral lactate production.12,13

By the conventions used in this study, CMRO2 is a positive number, since there is always a net consumption of oxygen by the brain. In a normal adult, mean CMRO2 is 1.5 µmol/gm/min.6 The CMRL can be a positive or negative number, depending on whether there is a net uptake or excretion of lactate by the brain. Normally, there is a small but measurable cerebral lactate production (mean CMRL −0.02 µmol/gm/min). Elevated production of lactate by the brain is indicated by a more negative CMRL and by an increased LOI.

Statistical Analysis

All summary data are expressed as the mean ± standard deviation. For the 578 measurements of CBF, the relationship between AVDO2 and CBF was analyzed by nonlinear regression analysis. For the patients who were identified as having developed cerebral ischemia, the CBF, CMRO2, and CMRL measured before and after treatment of the ischemia were compared by a paired t-test. A p value of < 0.05 was considered significant.

Results

When all 578 of the CBF measurements obtained in the 100 patients were examined, the expected nonlinear relationship between CBF and AVDO2 was found to be present (Fig. 2), but with a correlation coefficient of only −0.24 (p < 0.01). Most (72%) of the CMRO2 values fell between 0.6 and 1.2 µmol/gm/min, which was the expected range for patients with severe head injury. However, 112 (25%) of the CMRO2 values were less than 0.6 µmol/gm/min. Only 12 (3%) of the values were over 1.2 µmol/gm/min. The patients with a very low CBF (≤ 0.2 ml/gm/min) could not be identified by an elevated AVDO2 since their values ranged from 0.45 to 5.44 µmol/ml. Most patients with a CBF of 0.2 ml/gm/min or less would have been misclassified as having a normal or increased CBF if only the AVDO2 had been known.

Fig. 2.
Fig. 2.

Correlation between arteriovenous oxygen difference (AVDO2) and cerebral blood flow (CBF) in 100 comatose patients with and without cerebral ischemia (r = −0.24 in 578 measurements, p < 0.01).

When CBF measurements that were accompanied by an LOI of 0.08 or greater, indicating elevated cerebral lactate production, were excluded from the analysis, the correlation between CBF and AVDO2 was significantly improved (r = −0.74 in 313 measurements, p < 0.01). As shown in Fig. 3, most of the patients with very low CMRO2 values were eliminated by identifying elevated cerebral lactate production. In addition, the remaining patients without cerebral ischemia could be divided into three CBF categories by the level of their AVDO2. All of the patients with a CBF of 0.2 ml/gm/min or less had an AVDO2 over 3.0 µmol/ml. All of the patients with a very elevated CBF (> 0.8 ml/gm/min) had an AVDO2 of less than 1.3 µmol/ml.

Fig. 3.
Fig. 3.

Correlation between arteriovenous oxygen difference (AVDO2) and cerebral blood flow (CBF) in 55 comatose patients without cerebral ischemia (r = −0.74 in 313 measurements, p < 0.01).

From these data, a classification of CBF abnormalities was developed based on measurements of AVDO2 and AVDL. The mean values for CBF and CMRO2 in each CBF category are listed in Table 2. Patients with the ischemia/infarction CBF pattern, identifiable by an LOI of 0.08 or more and a variable AVDO2, had a mean CBF of 0.388 ± 0.200 ml/gm/min and a mean CMRO2 of 0.49 ± 0.30 µmol/gm/min. Patients without ischemia, identifiable by an LOI of less than 0.08, had a mean CMRO2 of 0.82 ± 0.21 µmol/gm/min regardless of the level of the CBF. Patients with an AVDO2 of less than 1.3 µmol/ml had an elevated or hyperemic CBF (mean 0.529 ± 0.181 ml/gm/min). Patients with an AVDO2 between 1.3 and 3.0 µmol/ml had a normal CBF (mean 0.416 ± 0.124 ml/gm/min). Patients with an AVDO2 of more than 3.0 µmol/ml had a low CBF (mean 0.234 ± 0.069 ml/gm/min). Because this last category of patients did not have elevated cerebral lactate production or a CMRO2 of less than 0.6 µmol/gm/min, this CBF pattern was called “compensated hypoperfusion” rather than ischemia.

TABLE 2

Classification of CBF abnormalities from AVDO2 and AVDL*

ClassificationAVDO2 (µmol/ml)LOICBF (ml/gm/min)CMRO2 (µmol/gm/min)
nonischemic patterns< 0.08
 hyperemia< 1.30.529 ± 0.1810.87 ± 0.08
 normal CBF1.3–3.00.416 ± 0.1240.82 ± 0.22
 compensated hypoperfusion> 3.00.234 ± 0.0690.84 ± 0.18
ischemia/infarctionvariable≥ 0.080.338 ± 0.2000.49 ± 0.30

CBF = cerebral blood flow; AVDO2 = arteriovenous oxygen difference; AVDL = arterial-jugular venous difference of lactate; LOI = lactate-oxygen index; CMRO2 = cerebral metabolic rate of oxygen.

Because of the shape of the CMRO2 curves, AVDO2 was more sensitive for identifying patients with a low CBF. Small changes in CBF were associated with large changes in AVDO2 when CBF was low. There was considerable overlap in AVDO2 values between patients with a normal CBF and those with hyperemia. Nevertheless, increases in CBF in individual patients could be distinguished by changes in AVDO2. An example of findings in a patient with a temporal lobe contusion who developed a transient period of hyperemia is shown in Fig. 4. During the time that the CBF was elevated, the CMRO2 was unchanged, and the CBF abnormality was evident from a decrease in AVDO2.

Fig. 4.
Fig. 4.

Findings in a patient with a transient episode of hyperemia. Left: Graphs showing changes in cerebral blood flow (CBF) and arteriovenous oxygen difference (AVDO2) over time after injury. On Day 2, CBF increased with no change in the cerebral metabolic oxygen rate (CMRO2). Right: The relationship between AVDO2 and CBF is shown for each CBF measurement. All of the points fall close to the 0.9-µmol/gm/min CMRO2 curve. The increase in CBF on Day 2 is easily identifiable by the decrease in AVDO2.

Eleven (22%) of the 49 patients who were evaluated for ischemia prospectively were found to have the characteristic changes of compensated hypoperfusion: decreased CBF and AVDO2 greater than 3.0 µmol/ml, but a normal LOI. Their CBF averaged 0.221 ± 0.038 ml/gm/min. One of the patients who had intracranial hypertension being managed with mechanical hyper-ventilation was treated by permitting the pCO2 to increase to 35 mm Hg. The other patients received hypervolemic hemodilution, and one was treated, in addition, with dopamine-induced hypertension. With treatment, all of the patients had an increase in CBF to a mean value of 0.336 ± 0.050 ml/gm/min. In all cases, AVDO2 decreased into the normal range as CBF was increased by the therapy, from 3.92 ± 0.84 to 2.61 ± 0.39 µmol/ml (Fig. 5). The CMRO2 was unchanged by the CBF treatment, confirming the conclusion based on the normal cerebral lactate production that these patients were not truly ischemic, but had been able to compensate for the low CBF by increased extraction of oxygen.

Fig. 5.
Fig. 5.

Summary of treatment of compensated hypoperfusion in 11 patients. Left: Graphs showing the mean values for cerebral blood flow (CBF) and cerebral metabolic oxygen rate (CMRO2) before and after treatment. The CBF increased with treatment, while CMRO2 remained unchanged. Right: The relationship between CBF and the arteriovenous oxygen difference (AVDO2) is shown before and after treatment for the individual patients. The changes in CBF did not alter cerebral metabolism and were accompanied by reciprocal changes in AVDO2.

Four (7%) of the 49 patients studied prospectively for ischemia developed an ischemia/infarction CBF pattern after having an initially normal CBF. In one of the patients, the ischemia was due to severe intracranial hypertension that developed 24 hours after evacuation of a traumatic intracerebral hematoma. A second patient developed ischemia due to vasospasm 72 hours after a subarachnoid hemorrhage. The other two patients developed cerebral ischemia 2 to 3 days after closed head injury. Regardless of the etiology, all four patients changed from a pattern of normal CBF, with a normal AVDO2 and a normal cerebral lactate production, to a pattern of infarction (Fig. 6) between the routine CBF measurements (approximately 8 hours). The CBF dropped from 0.342 ± 0.035 to 0.214 ± 0.042 ml/gm/min, CMRO2 decreased from 0.76 ± 0.09 to 0.18 ± 0.13 µmol/gm/min, and cerebral lactate production increased from −0.020 ± 0.017 to −0.042 ± 0.007 µmol/gm/min. The AVDO2 decreased from 2.23 ± 0.36 to 0.81 ± 0.54 µmol/ml, which, if interpreted alone, might suggest that CBF had increased. However, the LOI dramatically increased, from 0.03 ± 0.02 to 0.36 ± 0.29, indicating the presence of ischemia/infarction, and that AVDO2 would not accurately reflect CBF. In each case, a lateral skull x-ray film documented that the venous catheter remained in the jugular bulb; however, it cannot be ruled out that, with the drop in CBF, increased extracerebral contamination of the jugular venous blood might have accounted for the increase in venous oxygen saturation.

Fig. 6.
Fig. 6.

Findings during the development of cerebral ischemia in four patients who initially had a normal cerebral blood flow (CBF). Left: Graphs showing the mean values for CBF and cerebral metabolic oxygen rate (CMRO2). Over a period of 8 hours, CBF and CMRO2 decreased, while cerebral lactate production increased from −0.020 to −0.042 µmol/gm/min. Right: The relationship between CBF and the arteriovenous oxygen difference (AVDO2) is shown for the individual patients. As the CBF decreased, the AVDO2 also decreased.

Ten (20%) of the 49 patients evaluated for ischemia prospectively had an ischemic pattern on the initial CBF measurement or developed cerebral ischemia during their subsequent hospital course. In two patients in whom the cerebral ischemia was due to intractable intracranial hypertension, it was not possible to follow treatment of the ischemia with CBF measurements, because they died despite aggressive treatment. In the remaining eight patients, nine separate episodes of treatment of cerebral ischemia were documented with CBF measurements before and after treatment. The treatments that were examined included increasing pCO2 to 35 mm Hg (three patients), hypervolemic hemodilution (five patients), and dopamine-induced hypertension (one patient). In all nine cases, CBF increased with treatment from 0.163 ± 0.038 to 0.303 ± 0.107 ml/gm/min. With five of the episodes of ischemia, an increase in CMRO2 accompanied the increase in CBF, from a mean value of 0.26 ± 0.17 to 0.68 ± 0.14 µmol/gm/min (Fig. 7 upper). In four patients, CMRO2 remained unchanged, although CBF was increased by the therapy (Fig. 7 lower). The change in AVDO2 was quite variable, and it was not possible to determine from the changes in AVDO2 whether CMRO2 was improved or not. However, the LOI decreased in the five episodes of cerebral ischemia where treatment resulted in an increase in CMRO2, and, in contrast, increased in the four patients in whom CMRO2 remained unchanged. Both of the patients whose CMRO2 did not change with treatment eventually died of their neurological injury. In the four patients in whom the improvement in CMRO2 was sustained for the remainder of the CBF monitoring, the LOI decreased to less than 0.08, clearly into the nonischemic range, while in the patient in whom the improvements in cerebral metabolism were transient despite continued treatment, the LOI decreased only from 0.45 to 0.13.

Fig. 7.
Fig. 7.

Treatment of cerebral ischemia in nine patients. When cerebral blood flow (CBF) increased, cerebral metabolic oxygen rate (CMRO2) increased in five patients (upper) and remained unchanged in four (lower). Left: Graphs showing the mean CBF and CMRO2 values before and after treatment. When an increase in CBF improved CMRO2, cerebral lactate production decreased (upper pair). When the ischemic changes were irreversible, CMRO2 and cerebral metabolic lactate rate were unchanged by treatment (lower pair). Right: The relationship between CBF and the arteriovenous oxygen difference (AVDO2) is shown for the individual patients. The changes in AVDO2 as CBF increased were variable and not predictive of whether or not CMRO2 was improved.

Discussion

From the data presented in this study, the model shown in Fig. 8 was developed. The model is based on the hypothesis, which was confirmed in this study, that in the absence of ischemia, CMRO2 is relatively constant in the comatose patient, typically ranging from 0.6 to 1.2 µmol/gm/min, and CBF may vary independently of CMRO2. In these patients without ischemia, AVDO2 and CBF have the relationship represented by the constant CMRO2 curves. If CBF is in excess of cerebral metabolic requirements, AVDO2 will be decreased. If CBF is decreased, AVDO2 will proportionally increase as the brain compensates for the decreased flow by extracting a greater amount of oxygen. As long as increased extraction of oxygen completely compensates for the decreased blood flow, CMRO2 remains unchanged. However, a point will be reached at which further decreases in CBF cannot be compensated for by increased oxygen extraction and ischemia follows, manifested by a fall in CMRO2 and an increase in cerebral lactate production. Initially these metabolic changes of ischemia may be reversible. However, as time passes, irreversible ischemic injury or infarction may develop. The time required for the changes to become irreversible depends on the severity of the reduction in CBF. In experimental studies, a CBF of 0.18 ml/gm/min is tolerated for several hours before changes become irreversible, while a CBF of less than 0.10 ml/gm/min produces infarction in minutes.2,3 As the tissue dies, CMRO2 falls, although cerebral lactate production may remain elevated. Increases in CBF after the tissue has become infarcted will not result in a significant increase in CMRO2, but instead will be expressed as a decrease in AVDO2.

Fig. 8.
Fig. 8.

Model diagramming the relationship between cerebral blood flow (CBF) and cerebral metabolism in comatose patients. In the absence of cerebral ischemia, the arteriovenous oxygen difference (AVDO2) and CBF have the relationship illustrated by the solid curve, with cerebral metabolic rate of oxygen (CMRO2) averaging 0.9 µmol/gm/min. In the presence of cerebral ischemia/infarction (open arrows), AVDO2 and CBF have an unpredictable relationship.

In this series, it was uncommon to find a cerebral metabolic pattern that has been considered representative of early cerebral ischemia.18 This pattern, characterized by a low CBF, an elevated AVDO2, and an increased cerebral lactate production, with an increase in CMRO2 and a decrease in cerebral lactate production on treatment of the low CBF, occurred on only one occasion in the 100 patients studied. Much more common was a pattern of low CBF, an increased AVDO2, and a normal CMRL, which we have called “compensated hypoperfusion” or a pattern of low CBF, normal or decreased AVDO2, and an increased cerebral lactate production, which we have called “ischemia/infarction.”

The compensated hypoperfusion pattern may be the most important to recognize clinically. The increased AVDO2 suggests that the brain is compensating for the decreased blood flow by extracting oxygen more completely, but because anaerobic metabolism is not increased and because CMRO2 does not change as CBF is increased, the low blood flow does not appear to have altered overall cerebral energy metabolism. Although the patients are not truly ischemic at the moment of the CBF measurement, they have exhausted the brain's compensatory mechanism for maintaining cerebral metabolism in the presence of decreased oxygen availability. In such cases, small decreases in either CBF or arterial oxygen content could produce significant cerebral ischemia.

The ischemia/infarction pattern is also important to identify, because sometimes this ischemia pattern is reversible. Because AVDO2 is low, this pattern is often assumed to indicate that brain tissue is irreversibly injured and that the CBF is adequate for the markedly reduced metabolic requirements. Although this is often the case, the presence of reversible ischemic changes are best determined by a trial of increasing CBF. The reason for the low AVDO2 in patients with reversible ischemia is not entirely clear. It may be that the blood flow is not homogeneous and that areas of very low flow and markedly increased oxygen extraction are mixed with areas of normal oxygen extraction. Alternatively, it may be that, with very low CBF, venous blood in the jugular bulb is contaminated with more extracerebral blood than when blood flow is normal. Studies conducted in normal adults have shown that extracerebral contamination is only 2% to 3%;15 however, no studies have documented that this remains the case when blood flow to the brain is markedly reduced.

Several factors could potentially affect the ability of the LOI to identify patients with ischemic injury. Other causes for cerebral lactic acidosis, such as ventriculitis, could cause an increase in the LOI in the absence of ischemia. Since the samples for AVDO2 and AVDL are obtained from only one internal jugular vein, the possibility exists of overlooking an infarction in brain tissue draining by the opposite jugular vein. Studies of oxygen and glucose concentrations performed on samples drawn simultaneously from both internal jugular veins have demonstrated small differences in normal adults, while marked differences can occur in patients with unilateral lesions.1 Since the movement of lactate across the blood-brain barrier is dependent upon the relative concentrations of lactate in the brain tissue and in blood, an increase in cerebral lactate production could be obscured in the presence of a systemic lactic acidosis.8,14 As long as arterial lactate concentration is normal, increases in cerebral lactate production should be reflected by a net efflux of lactate. In a previous study, an increased cerebral lactate production was present in patients that developed a cerebral infarction, even when arterial lactate concentration was increased.12

Although the underlying physiology of the injury is more easily understood if CBF measurements are available, reliable deductions about CBF can be made from measurements of AVDO2 and AVDL, which can be easily obtained in the intensive care unit. If the LOI is less than 0.08, cerebral ischemia is probably not present and CBF can be predicted reliably from the AVDO2. This permits identification of those patients with compensated hypoperfusion who may benefit from increasing CBF and of those patients with severe hyperemia. When the LOI is 0.08 or greater, ischemia or infarction is present, and the exact level of the CBF cannot be reliably predicted from the AVDO2. Actual measurement of CBF in this circumstance would be necessary. However, if treatment of the ischemia is successful in increasing CMRO2, the LOI will decrease, so that some clinically useful information can still be obtained without measuring CBF.

For future studies, continuous monitoring of jugular venous oxygen saturation, with periodic measurements of AVDO2 and AVDL, may be a valuable adjunct to intracranial pressure monitoring in comatose patients. A drop in jugular venous oxygen saturation or an increase in the LOI should trigger a search for causes of decreased oxygen delivery, such as a decrease in arterial pO2, hemoglobin, or CBF. Early identification and treatment of ischemic episodes may prevent secondary injury and improve overall outcome from neurological injury.

References

  • 1.

    Gibbs ELLennox WGGibbs FA: Bilateral internal jugular blood. Comparison of A-V differences, oxygen-dextrose ratios and respiratory quotients. Am J Psychiatry 102:1841901945/1946Am J Psychiatry 102:

  • 2.

    Hossman KASchuier FJ: Experimental brain infarcts in cats. 1. Pathophysiological observations. Stroke 11:5835921980Stroke 11:

  • 3.

    Jones THMorawetz RBCrowell RMet al: Thresholds of focal cerebral ischemia in awake monkeys. J Neurosurg 54:7737821981J Neurosurg 54:

  • 4.

    Kety SS: The theory and applications of the exchange of inert gas at the lungs and tissues. Pharmacol Rev 3:1411951Kety SS: The theory and applications of the exchange of inert gas at the lungs and tissues. Pharmacol Rev 3:

  • 5.

    Kety SSSchmidt CF: The determination of cerebral blood flow in man by the use of nitrous oxide in low concentrations. Am J Physiol 143:53661945Am J Physiol 143:

  • 6.

    Kety SSSchmidt CF: The nitrous oxide method for the quantitative determination of cerebral blood flow in man: theory, procedure, and normal values. J Clin Invest 27:4764831948J Clin Invest 27:

  • 7.

    Lassen NA: Cerebral blood flow and oxygen consumption in man. Physiol Rev 39:1832381959Lassen NA: Cerebral blood flow and oxygen consumption in man. Physiol Rev 39:

  • 8.

    Nemoto EMHoff JTSeveringhaus JW: Lactate uptake and metabolism by brain during hyperlactatemia and hypoglycemia. Stroke 5:48531974Stroke 5:

  • 9.

    Nilsson BRehncrona SSiesjo BK: Coupling of cerebral metabolism and blood flow in epileptic seizures, hypoxia and hypoglycaemiaPurves M (ed): Amsterdam: Elsevier1978199214

  • 10.

    Obrist WDLangfitt TWJaggi JLet al: Cerebral blood flow and metabolism in comatose patients with acute head injury. Relationship to intracranial hypertension. J Neurosurg 61:2412531984J Neurosurg 61:

  • 11.

    Raichle MEGrubb RL JrGado MHet al: Correlation between regional cerebral blood flow and oxidative metabolism. In vivo studies in man. Arch Neurol 33:5235261976In vivo studies in man. Arch Neurol 33:

  • 12.

    Robertson CSGrossman RGGoodman JCet al: The predictive value of cerebral anaerobic metabolism with cerebral infarction after head injury. J Neurosurg 67:3613681987J Neurosurg 67:

  • 13.

    Robertson CSNarayan RKGrossman RGet al: Monitoring for cerebral ischemia after severe head injuryMinor J (ed): Neurotrauma. Stoneham, Mass: Butterworths (In press)Neurotrauma.

  • 14.

    Rowe GGMaxwell GMCastillo CAet al: A study in man of cerebral blood flow and cerebral glucose, lactate and pyruvate metabolism before and after eating. J Clin Invest 38:215421581959J Clin Invest 38:

  • 15.

    Shenkin HAHarmel MHKety SS: Dynamic anatomy of the cerebral circulation. Arch Neurol Psychiatry 60:2402521948Arch Neurol Psychiatry 60:

  • 16.

    Sokoloff L: Neurophysiology and neurochemistry of coma. Exp Biol Med 4:15331971Sokoloff L: Neurophysiology and neurochemistry of coma. Exp Biol Med 4:

  • 17.

    Swedlow DBLewis LE: Measurement of cerebral blood flow in children. Anesthesiology 53 (Suppl):S1601980Anesthesiology 53 (Suppl):

  • 18.

    Wise RJSBernardi SFrackowiak RSJet al: Serial observations on the pathophysiology of acute stroke. The transition from ischemia to infarction as reflected in regional oxygen extraction. Brain 106:1972221983Brain 106:

Monitoring system manufactured by Oximetrix, Inc., Mountain View, California.

Infrared N2O analyzer manufactured by Vital Signs, Inc., East Rutherford, New Jersey.

Blood gas analyzers manufactured by Ciba Corning Diagnostics Corp., Medfield, Massachusetts; co-oximeter manufactured by Instrumentation Laboratories, Lexington, Massachusetts.

YSI-23L lactate analyzer manufactured by Yellow Springs Instruments, Yellow Springs, Ohio.

Article Information

Address reprint requests to: Claudia S. Robertson, M.D., Department of Neurosurgery, Baylor College of Medicine, One Baylor Plaza, Houston, Texas 77030.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Relationship between arteriovenous oxygen difference (AVDO2) and cerebral blood flow (CBF). If a coupled change in cerebral metabolic rate of oxygen (CMRO2) and CBF occurs, then AVDO2 remains unchanged and the relationship between CBF and AVDO2 shifts to a new CMRO2 curve (horizontal arrows). If CMRO2 remains constant, then changes in AVDO2 reflect uncoupled changes in CBF (curved arrows).

  • View in gallery

    Correlation between arteriovenous oxygen difference (AVDO2) and cerebral blood flow (CBF) in 100 comatose patients with and without cerebral ischemia (r = −0.24 in 578 measurements, p < 0.01).

  • View in gallery

    Correlation between arteriovenous oxygen difference (AVDO2) and cerebral blood flow (CBF) in 55 comatose patients without cerebral ischemia (r = −0.74 in 313 measurements, p < 0.01).

  • View in gallery

    Findings in a patient with a transient episode of hyperemia. Left: Graphs showing changes in cerebral blood flow (CBF) and arteriovenous oxygen difference (AVDO2) over time after injury. On Day 2, CBF increased with no change in the cerebral metabolic oxygen rate (CMRO2). Right: The relationship between AVDO2 and CBF is shown for each CBF measurement. All of the points fall close to the 0.9-µmol/gm/min CMRO2 curve. The increase in CBF on Day 2 is easily identifiable by the decrease in AVDO2.

  • View in gallery

    Summary of treatment of compensated hypoperfusion in 11 patients. Left: Graphs showing the mean values for cerebral blood flow (CBF) and cerebral metabolic oxygen rate (CMRO2) before and after treatment. The CBF increased with treatment, while CMRO2 remained unchanged. Right: The relationship between CBF and the arteriovenous oxygen difference (AVDO2) is shown before and after treatment for the individual patients. The changes in CBF did not alter cerebral metabolism and were accompanied by reciprocal changes in AVDO2.

  • View in gallery

    Findings during the development of cerebral ischemia in four patients who initially had a normal cerebral blood flow (CBF). Left: Graphs showing the mean values for CBF and cerebral metabolic oxygen rate (CMRO2). Over a period of 8 hours, CBF and CMRO2 decreased, while cerebral lactate production increased from −0.020 to −0.042 µmol/gm/min. Right: The relationship between CBF and the arteriovenous oxygen difference (AVDO2) is shown for the individual patients. As the CBF decreased, the AVDO2 also decreased.

  • View in gallery

    Treatment of cerebral ischemia in nine patients. When cerebral blood flow (CBF) increased, cerebral metabolic oxygen rate (CMRO2) increased in five patients (upper) and remained unchanged in four (lower). Left: Graphs showing the mean CBF and CMRO2 values before and after treatment. When an increase in CBF improved CMRO2, cerebral lactate production decreased (upper pair). When the ischemic changes were irreversible, CMRO2 and cerebral metabolic lactate rate were unchanged by treatment (lower pair). Right: The relationship between CBF and the arteriovenous oxygen difference (AVDO2) is shown for the individual patients. The changes in AVDO2 as CBF increased were variable and not predictive of whether or not CMRO2 was improved.

  • View in gallery

    Model diagramming the relationship between cerebral blood flow (CBF) and cerebral metabolism in comatose patients. In the absence of cerebral ischemia, the arteriovenous oxygen difference (AVDO2) and CBF have the relationship illustrated by the solid curve, with cerebral metabolic rate of oxygen (CMRO2) averaging 0.9 µmol/gm/min. In the presence of cerebral ischemia/infarction (open arrows), AVDO2 and CBF have an unpredictable relationship.

References

1.

Gibbs ELLennox WGGibbs FA: Bilateral internal jugular blood. Comparison of A-V differences, oxygen-dextrose ratios and respiratory quotients. Am J Psychiatry 102:1841901945/1946Am J Psychiatry 102:

2.

Hossman KASchuier FJ: Experimental brain infarcts in cats. 1. Pathophysiological observations. Stroke 11:5835921980Stroke 11:

3.

Jones THMorawetz RBCrowell RMet al: Thresholds of focal cerebral ischemia in awake monkeys. J Neurosurg 54:7737821981J Neurosurg 54:

4.

Kety SS: The theory and applications of the exchange of inert gas at the lungs and tissues. Pharmacol Rev 3:1411951Kety SS: The theory and applications of the exchange of inert gas at the lungs and tissues. Pharmacol Rev 3:

5.

Kety SSSchmidt CF: The determination of cerebral blood flow in man by the use of nitrous oxide in low concentrations. Am J Physiol 143:53661945Am J Physiol 143:

6.

Kety SSSchmidt CF: The nitrous oxide method for the quantitative determination of cerebral blood flow in man: theory, procedure, and normal values. J Clin Invest 27:4764831948J Clin Invest 27:

7.

Lassen NA: Cerebral blood flow and oxygen consumption in man. Physiol Rev 39:1832381959Lassen NA: Cerebral blood flow and oxygen consumption in man. Physiol Rev 39:

8.

Nemoto EMHoff JTSeveringhaus JW: Lactate uptake and metabolism by brain during hyperlactatemia and hypoglycemia. Stroke 5:48531974Stroke 5:

9.

Nilsson BRehncrona SSiesjo BK: Coupling of cerebral metabolism and blood flow in epileptic seizures, hypoxia and hypoglycaemiaPurves M (ed): Amsterdam: Elsevier1978199214

10.

Obrist WDLangfitt TWJaggi JLet al: Cerebral blood flow and metabolism in comatose patients with acute head injury. Relationship to intracranial hypertension. J Neurosurg 61:2412531984J Neurosurg 61:

11.

Raichle MEGrubb RL JrGado MHet al: Correlation between regional cerebral blood flow and oxidative metabolism. In vivo studies in man. Arch Neurol 33:5235261976In vivo studies in man. Arch Neurol 33:

12.

Robertson CSGrossman RGGoodman JCet al: The predictive value of cerebral anaerobic metabolism with cerebral infarction after head injury. J Neurosurg 67:3613681987J Neurosurg 67:

13.

Robertson CSNarayan RKGrossman RGet al: Monitoring for cerebral ischemia after severe head injuryMinor J (ed): Neurotrauma. Stoneham, Mass: Butterworths (In press)Neurotrauma.

14.

Rowe GGMaxwell GMCastillo CAet al: A study in man of cerebral blood flow and cerebral glucose, lactate and pyruvate metabolism before and after eating. J Clin Invest 38:215421581959J Clin Invest 38:

15.

Shenkin HAHarmel MHKety SS: Dynamic anatomy of the cerebral circulation. Arch Neurol Psychiatry 60:2402521948Arch Neurol Psychiatry 60:

16.

Sokoloff L: Neurophysiology and neurochemistry of coma. Exp Biol Med 4:15331971Sokoloff L: Neurophysiology and neurochemistry of coma. Exp Biol Med 4:

17.

Swedlow DBLewis LE: Measurement of cerebral blood flow in children. Anesthesiology 53 (Suppl):S1601980Anesthesiology 53 (Suppl):

18.

Wise RJSBernardi SFrackowiak RSJet al: Serial observations on the pathophysiology of acute stroke. The transition from ischemia to infarction as reflected in regional oxygen extraction. Brain 106:1972221983Brain 106:

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 2 2 2
Full Text Views 148 148 106
PDF Downloads 44 44 34
EPUB Downloads 0 0 0

PubMed

Google Scholar