Correlation between Cerebral Blood Flow,Substrate Delivery,and Metabolism in Head Injury: A Combined Microdialysis and Triple Oxygen Positron Emission Tomography Study

Abstract

Microdialysis continuously monitors the chemistry of a small focal volume of the cerebral extracellular space. Conversely, positron emission tomography (PET) establishes metabolism of the whole brain, but only for the duration of the scan. The objective of this study was to apply both techniques to head-injured patients simultaneously to assess the relation between microdialysis (glucose, lactate, lactate/pyruvate [L/P] ratio, and glutamate) and PET (cerebral blood flow [CBF], cerebral blood volume, oxygen extraction fraction (OEF), and cerebral metabolic rate of oxygen) parameters. Microdialysis catheters were inserted into the frontal cerebral cortex and adipose tissue of the anterior abdominal wall of 17 severely head-injured patients. Microdialysis was performed during PET scans, with regions of interest defined by the location of the microdialysis catheter membrane. An intervention (hyperventilation) was performed in 13 patients. The results showed that combining PET and microdialysis to monitor metabolism in ventilated patients is feasible and safe, although logistically complex. There was a significant relation between the L/P ratio and the OEF (Spearman r = 0.69, P = 0.002). There was no significant relation between CBF and the microdialysis parameters. Moderate short-term hyperventilation appeared to be tolerated in terms of brain chemistry, although no areas were sampled by microdialysis where the OEF exceeded 70%. Hyperventilation causing a reduction of the arterial carbon dioxide tension by 0.9 kPa resulted in a significant elevation of the OEF, in association with a reduction in glucose, but no significant elevation in the L/P ratio or glutamate.

Keywords

Microdialysis Positron emission tomography Cerebral blood flow Cerebral metabolism Head injury

Recent advances in the management of patients with acute brain injury caused by trauma and subarachnoid hemorrhage have included monitoring of cerebral metabolism using microdialysis, and functional imaging of the brain using positron emission tomography (PET). Microdialysis is currently under investigation as a monitor of substrate (glucose) supply and metabolite (lactate and pyruvate) production in patients with head injury, subarachnoid hemorrhage, and epilepsy (Whittle, 1990; Ungerstedt, 1991; Editorial, 1992; Persson and Hillered, 1992; Vespa et al., 1998; Bullock et al., 1998; Goodman et al., 1999; Hutchinson et al., 1999a). The technique also enables the extracellular concentration of amino acids and drugs to be determined. The introduction of bedside analyzers has facilitated the application of microdialysis, enabling the metabolic state of the brain to be followed real-time in intensive care or in the operating theatre. PET scanning provides a comprehensive assessment of cerebral metabolism by providing quantitative data for regional cerebral blood flow (CBF), cerebral blood volume (CBV), oxygen extraction fraction (OEF), and the cerebral metabolic rate of oxygen (CMRO₂) (Frackowiak et al. 1980a,c; Baron et al., 1982; Lebrun-Grandie et al., 1983; Sette et al., 1989; Heiss, 1998).

Both microdialysis and PET scanning have limitations. One of the major considerations when interpreting microdialysis results is appreciation of the focal nature of the technique, where monitoring of only a small volume of brain tissue is possible. Consequently, catheter location (e.g., an area of diffuse injury compared to pericontusional tissue) is of prime importance in the interpretation of results. However, microdialysis does provide data continuously over long periods of hours and days. In contrast, PET scanning provides data for the whole brain with the ability to quantify global, regional, or local regions of interest, but only for short periods. By combining the focal but continuous technique of microdialysis with the global but intermittent “snap-shot” images derived from PET, a comprehensive picture of cerebral metabolism can potentially be achieved.

The objectives of this study were to (1) investigate the feasibility of performing microdialysis with sampling during PET scanning; (2) define the relation between microdialysis (glucose, lactate, lactate/pyruvate [L/P] ratio and glutamate) and PET (CBF, CBV, OEF, CMRO₂) parameters; (3) to cross-correlate ischemic thresholds for these parameters; and (4) to determine the effects of hyperventilation on PET and microdialysis parameters.

MATERIALS AND METHODS

Patient selection

The study was approved by the Local Cambridge Ethics Committee and UK Radiation Protection Committee. Consent was obtained from the next of kin. Patients older than age 16 years with severe head injury requiring sedation, ventilation, and intracranial pressure monitoring were eligible for the study. Exclusion criteria were deranged coagulation parameters and/or low platelet count preventing the placement of a microdialysis catheter and the presence of certain other injuries. These included unstable spinal injuries contraindicating transfer to the PET scan suite and thoracic complications requiring high (>60%) percentage of inspired oxygen (Fio₂), preventing sufficient delivery of the isotopes to the brain.

Patients were managed according to standard protocols (Menon, 1999) including sedation with propofol and fentanyl, and paralysis with atracurium. Monitored parameters included electrocardiogram, invasive blood pressure, peripheral oxygen saturation, continuous arterial blood gas analysis (Paratrend 7 sensor, Diametrics, High Wycombe, U.K.), end-tidal carbon dioxide, jugular venous oxygen saturation (Sjvo₂) and intracranial pressure. Comprehensive neurointensive-care monitoring and treatment was maintained during the PET scans in the adjacent Wolfson Brain Imaging Center. With regard to hyperventilation in patients with raised intracranial pressure, our policy is to maintain the arterial partial pressure of carbon dioxide in the region of 4.5 kPa. For the patients in this study, the baseline partial pressure of carbon dioxide ranged from 4.22 kPa to 5.52 kPa.

To compare the PET data from the head injury patients with the uninjured brain, 10 unsedated volunteers were studied with the triple oxygen technique.

Microdialysis technique

Microdialysis catheters (CMA70 10-mm membrane, CMA, Stockholm, Sweden) were inserted into the frontal cerebral parenchyma, in conjunction with an intracranial pressure sensor (Codman, Raynham, MA, U.S.A.) using a specially designed triple-lumen cranial access device (Hutchinson et al., 2000c). This device allows a secure fixation of the catheter. The catheters were inserted to a depth of 30 mm, that is, monitoring predominantly white matter (depth below pia 20 to 30 mm with a 10-mm microdialysis membrane). The cranial access device is inserted via an 8-mm skull hole allowing direct visual and opening of the dura before insertion of the device. The catheters were perfused with Ringer's solution (K⁺ 4 mmol/L, Na⁺ 147 mmol/L, Ca²⁺ 2 mmol/L, Cl- 155 mmol/L) at a rate of 0.3 μL/min using the CMA106 pump. Collecting vials were changed every 20 minutes, thus providing a sample volume of 6 μL sufficient for analysis of four metabolites (glucose, glutamate, lactate, and pyruvate) using the CMA600 bedside microdialysis analyser.

This instrument allows the online analysis of ultrasmall volumes of microdialysis samples on the Intensive Care Unit, with each assay requiring less than 1 μL per analyte and taking 1 to 1.5 minutes. The principle is based on the reaction of the microdialysis sample, with enzyme reagents and the measurement of the product using a single-beam filter photometer with mercury lamp light source. The measurement is performed in a 10-mm (2 μL) capillary flow cell at wavelengths 365 and 546 nm, with the rate of formation of the dye proportional to the substance concentration. The analyzer is controlled by a standard personal computer and calibrates automatically every 6 hours. We have previously validated the analysis of microdialysis samples for glutamate using the CMA600 by comparing measurements obtained using the gold standard of high-performance liquid chromatography (Hutchinson et al., 2000d).

The delay from microdialysis catheter tip to collecting vial is 17 minutes, and this was accounted for when analyzing the results. In addition to the cerebral microdialysis catheter, a subcutaneous microdialysis catheter (CMA60 30-mm membrane) was placed into the adipose tissue of the anterior abdominal wall to monitor peripheral chemistry.

PET technique

The study was performed in the Wolfson Brain Imaging Center adjacent to the neuro-intensive care unit and allows safe imaging of ventilated patients. The PET scans were performed on a GE Advance scanner (General Electric Medical Systems, Milwaukee, WI, U.S.A.) with a 15.3-cm axial field of view. Emission data were acquired after, sequentially, a steady-state infusion of H₂¹⁵O (800 MBq for two frames, each acquired during 5 minutes in 3D mode), a 60-second inhalation of C¹⁵O (300 MBq) with data acquired during the subsequent 5-minute period in 3D mode, and a steady-state inhalation of O (7,200 MBq, with data acquired in two 5-minute frames) in 2D mode. The total radiation burden from each set of measurements was 4.18 mSv EDE. Regional CBF, CBV, OEF, and CMRO₂ were calculated using simultaneous measurements of arterial tracer activity and standard modeling techniques (Frackowiak et al., 1980b).

During the baseline imaging, the Fio₂ and rate and tidal volume of ventilation were unchanged to maintain a constant level of arterial oxygenation, arterial CO₂ pressure, and cerebral perfusion pressure. After completion of the baseline scan, the minute ventilatory volume was increased, aiming to reduce the arterial CO₂ pressure by 1 kPa, providing that the Sjvo₂ did not decrease to less than 50%. When the arterial CO₂ pressure was stable, the scans were repeated in the reverse order.

Data analysis

After PET, a computed tomography (CT) scan was acquired to determine the location of the monitoring probes, which were not visible on the PET images. The CT scan was coregistered to the least noisy PET image, the summed H₂¹⁵O frame, using the mutual information algorithm developed by Studholme et al., (1996, 1997). This effectively coregistered the CT scan to all the PET images, because patient movement is negligible for these anesthetized and paralyzed patients. A 20-mm diameter region of interest (ROI) was defined on the coregistered CT scan around the probe location (Fig. 1) on two contiguous slices (each 4.5 mm) to represent the volume of brain on PET directly relevant to the volume of brain sampled by the microdialysis catheter (10-mm membrane length). The size of the ROI was determined by the desire to keep the region as focal as possible while limiting parameter variance. By superimposing this ROI over the PET emission scans, tissue concentration values were derived. These values, together with the blood radioactivity concentrations and other blood factors, were used to produce ROI values for CBF, CBV, OEF and CMRO₂ using established kinetic models (Frackowiak et al., 1980c; Alpert et al., 1988). The physiologic data (heart rate, Fio₂, Sao₂, Sjvo₂, blood gases, and intracranial pressure) and microdialysis (cerebral and peripheral glucose, lactate, pyruvate, and glutamate) concentrations were recorded on a spreadsheet and combined with the PET-derived data. Results are expressed as mean values (± SDs), with correlations between parameters determined using Spearman correlation coefficients with Bonferroni correction.

Fig. 1.

Positron emission tomography coregistration for identification of microdialysis catheter location. The region of interest is shown by the circle.

RESULTS

Demography

Seventeen patients (11 male and 6 female) with severe head injury were studied with microdialysis and PET. The mean (± SDs) age of the patients was 34 (± 15) years (range: 16 to 66 years). The demographic data from the patients are shown in Table 1 and physiologic data are shown in Table 2. Ten volunteers (8 male and 2 female) with mean (± SD) age 45 (± 9) years (range: 29 to 59 years) were studied with PET to provide control data.

Table 1.

Demographic data for each patient at time of positron emission tomography scan, location of microdialysis catheter relative to lesion, time between injury and catheter insertion, time between catheter insertion and scan, and total duration of microdialysis monitoring

Patient no.	Sex	Age	Lesion	ICP	Catheter in region of contusion	Time injury to μd (h)	Time μd to PET (h)	Duration μd (h)
1	M	45	L temporal contusion	21	N	25	38	47
2	F	20	Diffuse swelling	14	N	42	17	140
3	F	40	L acute subdural	20	N	12	4	170
4	M	29	L parietal contusion	17	N	23	44	73
5	F	51	L and R frontal and temporal contusions	6	Y	23	66	118
6	F	26	L acute subdural	6	N	19	23	97
7	M	17	L frontal contusions	25	Y	19	48	221
8	M	31	L frontal contusions	21	N	22	144	242
9	M	24	L and R frontal contusions	30	N	103	19	134
10	M	35	L frontal contusions	7	N	32	13	45
11	M	25	L temporal and parietal contusions	22	Y	87	15	110
12	M	17	L parietal contusions	15	N	37	24	142
13	F	49	Diffuse injury	14	N	29	43	70
14	M	16	L extradural	6	N	20	2	47
15	F	66	L and R frontal and temporal contusions	9	N	35	28	74
16	M	29	L temporal extradural	16	N	12	96	162
17	M	52	R tSAH	11	N	14	2	67

ICP, intracranial pressure; μd, microdialysis; PET, positron emission tomography; L, left; R, right.

Table 2.

Physiologic baseline data (mean and SD) for all patients

	Mean	SD
Sao₂ (%)	99	1.2
HR	77	17
MAP (mm Hg)	95	9.6
ICP (mm Hg)	15	7.3
Sjvo₂ (%)	71	9
Pao₂ (kPa)	16.8	4.6
Paco₂ (kPa)	4.85	0.34
pHa	7.43	0.051

Sao₂, peripheral oxygen saturation; HR, heart rate; MAP, mean arterial blood pressure; ICP, intracranial pressure; Sjvo₂, jugular venous oxygen saturation; Pao₂, arterial partial pressure of oxygen; Paco₂, arterial partial pressure of carbon dioxide; pHa, arterial pH.

Microdialysis results

During PET scanning

The results for the cerebral and peripheral catheters showed low cerebral glucose concentrations (mean [± SD] of 1.64 ± 1.04 mmol/L) compared with peripheral glucose concentrations (mean 5.23 ± 2.69 mmol/L), high cerebral lactate concentrations (mean 4.0 ± 1.92 mmol/L) compared with peripheral lactate concentrations (mean 2.52 ± 1.79 mmol/L), high cerebral L/P ratios (mean 26.3 ± 8.1) compared with peripheral L/P ratios (mean 16.5 ± 8.0), and a wide variation in cerebral (mean 13.4 ± 12.8 mmol/L) and peripheral (mean 22.4 ± 33.5 mmol/L) glutamate concentrations.

Comparison of peripheral and cerebral microdialysis parameters revealed a good correlation between peripheral glucose and cerebral glucose (r = 0.69, P = 0.004), but no relation between peripheral lactate and cerebral lactate (r = 0.06, P = 0.84), peripheral pyruvate and cerebral pyruvate (r = −0.14, P = 0.59) and peripheral L/P ratio and cerebral L/P ratio (r = 0.047, P = 0.86).

Comparison of cerebral microdialysis substrates and metabolites showed a good correlation between cerebral lactate and cerebral pyruvate (r = 0.74, P = 0.001), a trend between cerebral glutamate and cerebral L/P ratio (r = 0.42, P = 0.092), and a weak relation between cerebral glutamate and cerebral lactate (r = 0.32, P = 0.22) that did not achieve significance. There was no relation between the other parameters.

Other microdialysis data

The overall microdialysis data monitoring compared with the microdialysis recordings determined during the PET scans are shown in Table 3. Adverse clinical events, notably secondary insults such as hypoxia and intracranial hypertension, were associated with derangements in the microdialysis measurements, with the demonstration of periods of improving (increase in glucose, decrease in L/P ratio, and increase in glutamate) and deteriorating metabolism (decrease in glucose, increase in the L/P ratio and decrease in glutamate).

Table 3.

Microdialysis data for each patient showing overall data for duration of monitoring (mean, min, max) compared with baseline data obtained during positron emission tomography scans

Patient No.	Overall data						PET data
	Glc		L/P		Glt		Glc		L/P		Glt
	m	r	m	r	m	r	m	r	m	r	m	r
	1	2.04	0.11 2.99	29.4	10.9 45.9	15.5	3.85 74.8	1.52	1.61 1.65	25.3	15.7 25.8	8.05	7.60 8.61
	2	2.39	0.48 4.59	22.3	14.3 34.3	5.18	2.4 23.7	2.86	2.46 3.10	19.1	18.6 19.5	9.23	7.10 9.0
3	1.77	0.44 3.60	20.0	9.6 31.8	4.19	1.67 7.5	1.74	1.74 1.75	33.6	33.2 34.0	6.6	5.6 7.4
4	0.89	0.42 1.74	28.7	23.3 40.0	2.62	1.04 8.26	0.62	0.60 0.63	30.3	30.0 30.9	2.07	1.74 2.26
5	*	*	24.6	16.3 33.6	31.1	7.7 85.7	*	*	26.6	25.3 29.1	34.9	32.5 37.4
6	2.75	1.74 4.48	22.0	16.6 33.6	9.45	2.12 26.8	3.47	3.43 3.50	21.7	19.8 25.2	17.1	11.1 19.9
7	1.13	0.28 2.29	20.5	14.7 31.7	2.48	1.0 11.5	1.24	1.19 1.30	16.8	16.3 17.4	1.8	1.48 2.43
8	1.27	0.18 4.2	26.5	20.3 46.8	12.4	1.72 33.7	0.21	0.20 0.22	37.8	36.8 38.3	12.7	12.2 13.5
9	0.91	0.45 1.43	22.7	12.1 34.1	3.67	1.97 17.6	1.19	1.18 1.19	15.3	12.8 17.8	7.34	3.38 11.3
10	3.24	2.82 3.8	22.5	20.0 24.9	11.4	3.35 14.0	3.19	3.13 3.24	22.0	21.6 22.2	11.9	11.3 12.2
11	1.78	0.69 3.60	28.0	21.5 43.2	2.07	1.03 37.2	0.91	0.74 1.00	24.3	23.9 24.7	1.73	1.68 1.76
12	1.14	0.34 2.18	25.8	21.1 38.2	5.20	1.32 35.9	1.00	0.98 1.01	29.6	28.1 32.3	8.70	7.74 9.24
13	1.80	0.54 2.82	23.6	17.7 66.2	10.9	5.61 90.3	1.50	1.36 1.70	20.7	20.5 20.9	8.19	8.04 8.45
14	1.49	0.31 2.64	19.0	14.7 26.9	6.48	1.60 96.1	0.81	0.31 1.12	32.4	18.8 52.9	46.1	17.5 96.1
15	1.95	0.19 4.73	18.2	14.0 30.1	8.25	1.87 66.7	1.11	0.96 1.2	18.1	17.1 18.6	4.60	4.19 4.91
16	1.55	0.19 2.60	29.9	16.5 70.0	13.1	1.74 33.0	1.32	1.31 1.34	28.8	28.4 29.3	12.1	10.9 13.0
17	2.16	0.11 4.5	20.3	14.9 85.4	5.31	1.0 50.3	3.51	0.50 5.87	46.1	22.0 85.4	34.3	10.5 50.3

PET, positron emission tomography; Glc, glucose; L/P lactate/pyruvate ratio; Glt, glutamate; m, mean; r, range; * not measured.

PET results

The PET results for the patients showed a mean (± SD) CBF of 27 ± 7.5 mL · 100 g⁻¹ · min⁻¹, CBV of 0.03 ± 0.01, OEF of 0.41 ± 0.10, and CMRO₂ of 68 ± 16 μmol · 100 g⁻¹ · min⁻¹. Only 2 patients had CBF less than 18 mL · 100 g⁻¹ · min⁻¹, and no patients had OEF greater than 70%. Using equivalent frontal regions of interest in the controls showed CBF values (mean ± SD) of 27 ± 3.4 mL · 100 g⁻¹ · min⁻¹, CBV of 0.03 ± 0.01, OEF of 0.42 ± 0.03, and CMRO₂ of 92 ± 10.5 μmol · 100 g⁻¹ · min.

PET-microdialysis correlation

The baseline PET and microdialysis data were compared to determine whether there was an association between image derived blood flow and metabolic data, and the extracellular chemistry (Table 4). There was a significant relation between the OEF and L/P ratio (r = 0.69, P = 0.002, Bonferroni correction) (Fig. 2). There were no significant relations between CBF, CBV, and CMRO₂, and the microdialysis parameters.

Fig. 2.

Relation between oxygen extraction fraction (OEF) and cerebral microdialysis parameters in 17 patients. There was a significant correlation between the lactate/pyruvate (L/P) ratio and OEF. Note that an OEF of 0.4 corresponded to a L/P ratio of 26, 25 being regarded as the upper limit of normal. The r value is the Spearman coefficient.

Table 4.

Correlation between positron emission tomography and microdialysis parameters cerebral blood flow, substrate delivery, and metabolism

	Glucose	Lactate	Pyruvate	L/P ratio	Glutamate
CBF
r	−0.15	−0.24	−0.15	−0.29	−0.43
P	0.58	0.36	0.56	0.25	0.08
CBV
r	−0.29	0.40	0.55	−0.04	−0.15
P	0.28	0.11	0.02	0.87	0.57
OEF
r	−0.09	0.29	−0.14	0.69	0.31
P	0.75	0.26	0.60	0.002	0.23
CMRO₂
r	−0.15	−0.03	−0.32	0.23	−0.05
P	0.58	0.91	0.21	0.37	0.84

The r value is the Spearman coefficient. The OEF-L/P ratio remains significant after Bonferroni correction for multiple comparisons.

L/P ratio, lactate/pyruvate ratio; CBF, cerebral blood flow; OEF, oxygen extraction fraction; CMRO₂, cerebral metabolic rate of oxygen.

Hyperventilation

Hyperventilation was performed in 13 patients and was tolerated in all cases (Sjvo₂ did not decrease to less than 50%). The significant relation between the OEF and L/P ratio observed during baseline was also seen after hyperventilation but shifted to the right (Fig. 3 regression line), indicating that hyperventilation significantly increases the OEF but does not increase the L/P ratio. The effects of hyperventilation comparing paired values for each parameter before and after hyperventilation (paired t-test) are shown in Table 5. As anticipated, hyperventilation induced a significant reduction in Paco₂ and Sjvo₂. There was a small increase in mean arterial blood pressure and reduction in intracranial pressure. There was a significant increase in OEF and reduction in cerebral glucose, but this did not translate into increases in the L/P ratio or release of glutamate. An example of the effect of hyperventilation is shown in Fig. 4.

Fig. 3.

Relation between oxygen extraction fraction (OEF) and lactate/pyruvate (L/P) ratio at baseline and after hyperventilation in 13 patients. The r value is the Spearman coefficient.

Fig. 4.

Patient 3. Effect of hyperventilation on microdialysis and positron emission tomography parameters showing increase in oxygen extraction fraction (OEF; %), reduction in jugular venous oxygen saturation (%), reduction in glucose (mmol/L), and no change in the lactate/pyruvate ratio or glutamate (umol/L). Each integer on the horizontal axis (1–7) represents a 20-minute epoch. Hyperventilation was performed at epoch 2. of lactate, suggesting that pyruvate is metabolized to lactate and that anaerobic metabolism is occurring in these patients despite the OEF being within the normal range. All patients with high glutamate levels had high L/P ratios, although the converse was not true.

Table 5.

Effect of hyperventilation (n = 13); results expressed as mean ± SD

Variable	Baseline	Hyperventilation	P
MAP (mm Hg)	96 ± 9.9	100 ± 9.8	0.001
ICP (mm Hg)	17 ± 7.5	14 ± 6.2	0.040
Paco₂ (kPa)	4.92 ± 0.31	4.08 ± 0.47	0.001
Pao₂ (kPa)	16.3 ± 4.44	15.9 ± 4.72	0.593
pHa	7.43 ± 0.053	7.49 ± 0.074	0.001
Sjvo₂ (%)	69 ± 10	57 ± 12	0.001
CBF (ml · 100 g⁻¹ · min⁻¹)	27 ± 7.8	23 ± 11	0.108
OEF	0.40 ± 0.10	0.50 ± 0.11	0.002
CBV	0.026 ± 0.009	0.032 ± 0.011	0.089
CMRO₂ (μmol · 100 g⁻¹ · min⁻¹)	68 ± 17	86 ± 41	0.146
Glucose (mmol/L)
Peripheral	4.81 ± 2.40	5.12 ± 1.98	0.409
Cerebral	1.63 ± 1.03	1.49 ± 1.01	0.015
Pyruvate (μmol/L)
Peripheral	142 ± 78.2	152 ± 67.5	0.419
Cerebral	160 ± 49.7	179 ± 72.6	0.24
Lactate (mmol/L)
Peripheral	2.61 ± 1.92	2.70 ± 1.70	0.303
Cerebral	3.99 ± 1.94	4.66 ± 2.16	0.167
L/P ratio
Peripheral	17.1 ± 8.58	17.6 ± 9.25	0.501
Cerebral	24.9 ± 6.61	25.8 ± 5.01	0.325
Glutamate (μmol/L)
Peripheral	11.0 ± 10.3	11.5 ± 10.1	0.729
Cerebral	10.0 ± 8.69	9.79 ± 7.85	0.835

MAP, mean arterial blood pressure; ICP, intracranial pressure; Paco₂, arterial partial pressure of carbon dioxide; Pao₂, arterial partial pressure of oxygen; pHa, arterial pH; Sjvo₂, jugular venous oxygen saturation; CBF, cerebral blood flow; OEF, oxygen extraction fraction; CBV, cerebral blood volume; CMRO₂, cerebral metabolic rate of oxygen; L/P, lactate/pyruvate.

DISCUSSION

The objective of this study was to investigate the relation between fundamental metabolic parameters (CBF, OEF, glucose, L/P ratio, and glutamate) that relate to the pathophysiology of head injury. The results confirm that combining these techniques is feasible, although logistically complex (Enblad et al., 1996). Continuation of intensive care during PET scanning is possible using the same principles of ventilatory management and hemodynamic monitoring as in the intensive care unit (Menon et al., 1997). The application of small, portable microdialysis pumps and mobile bedside chemical analyzers has enabled microdialysis to continue when patients are transferred from the intensive care unit to the operating theater or scanning suite.

Microdialysis analysis and control data

Microdialysis was performed during the PET imaging using the same perfusion rate (0.3 μL/min) as on the intensive care unit. This enabled the microdialysis recovery rate (i.e., the microdialysis concentration as a percentage of the true extracellular concentration) to remain constant and relatively high. We have previously shown that the recovery rate is approximately 70% for glucose, pyruvate, lactate, and glutamate using this flow rate with 10-mm-length membrane microdialysis catheters (Hutchinson et al., 2000d). The L/P ratio is independent of in vivo recovery (Persson and Hillered, 1996). However, slow flow rates have the disadvantage in that a relatively small volume of dialysate (6 μL at 0.3 μL/min) is obtained during 20-minute sampling. This is sufficient for analysis using the CMA600 analyzer but does not produce enough volume for delayed off-line analysis of other chemicals—for example, other amino acids and urea—which may be used as an internal reference for monitoring of in vivo probe performance (Ronne-Engstrom et al., 2001). Increasing the sampling time to increase the dialysate volume would have resulted in decreased time resolution. Many studies with microdialysis monitoring on intensive care units have used a 1-hour sampling time for logistical reasons. Although overall this is of sufficient resolution to capture events, there is a concern that rapid and brief insults may be missed. To investigate this, we have sampled patients during aneurysm surgery with 10-minute vial changes demonstrating transient biochemical changes (Hutchinson et al., 1999a, 2000b). Other groups are developing microdialysis analysis systems without vials allowing continuous monitoring of glucose and lactate concentrations (Strong et al., 2001).

Concerns regarding damage to the brain caused by microdialysis catheters have been addressed in several studies. Early animal studies demonstrated uncoupling of regional CBF from local brain glucose metabolism for the first 24 hours after probe insertion (Benveniste et al., 1987; Benveniste, 1989) and reticular fiber deposition, and gliosis after several days, which may affect the performance of the catheter (Benveniste and Diemer, 1988). However, this was probably a consequence of the damage caused by the relatively large and nonsterile devices available at the time. Other studies showed that recovery rates of substances stabilize after 30 minutes, with histology demonstrating only occasional microhemorrhages (Benveniste and Diemer, 1988) and the blood brain-barrier remaining intact (Major et al., 1990). In humans, our own data and those from the literature (Hillered et al., 1990; Meyerson et al., 1990) indicate a time from catheter implantation to baseline to be between 20 and 60 minutes, with a tendency for glutamate to take longer to reach baseline than glucose, pyruvate, and lactate. The sterile CMA 70 catheter designed for clinical use has been evaluated by Whittle et al. in sheep brains revealing minimal or no disturbance to the cerebral parenchyma as a result of microdialysis catheter implantation (Whittle et al., 1998). We have examined the brain of a patient at postmortem who died of cardiac complications after a subarachnoid hemorrhage and found no evidence of injury induced by the microdialysis catheter compared with the tract created by an external ventricular drain (Hutchinson et al., 1999b).

One of the other potential problems concerning microdialysis is determining whether the catheter is monitoring gray or white matter. This is influenced by the depth of the catheter (constant across patients using the cranial access device) and the relation of the catheter to sulci. In the present study, we are monitoring predominantly from white matter. Data from the literature indicate that the response to ischemia in gray and white matter are different (Hillered et al., 1990; Dohmen et al., 2001). The development of the CMA gold-tipped microdialysis catheter will assist in determining the exact location of the microdialysis catheter on the CT scan.

The cerebral microdialysis results are similar to those of previous studies in head injury (Persson and Hillered, 1992; Zauner et al., 1997; Goodman et al., 1999; Hutchinson et al., 2000d) demonstrating low cerebral glucose and high cerebral lactate relative to peripheral chemistry and controls (frontal cortex catheters in patients with posterior fossa lesions) (Reinstrup et al., 2000), and a wide variation in glutamate levels (Table 6). The findings are in keeping with the high metabolic activity of the brain of head-injured patients.

Table 6.

Comparison of microdialysis data during positron emission tomography in head-injured patients with control data taken from normal human brain, anesthetized and awake* and putatively normal values†

	Head-injured patients during PET scan	Control data		Normal levels
		Anesthetized	Awake
		Membrane length (mm)	10		10	10	10
Flow rate (μL/min)	0.3	1	0.3	0.3
Glucose (mmol/L)	1.6 ± 1.0	1.2 ± 0.6	0.9 ± 0.6	1.46 ± 0.77
Lactate (mmol/L)	4.0 ± 1.9	1.2 ± 0.6	1.4 ± 0.9	2.56 ± 0.57
Pyruvate (μmol/L)	160 ± 50	70 ± 24	103 ± 50	132 ± 24.6
L/P ratio	26.38 ± 8.1	22 ± 6	21 ± 6	24.1 ± 4.2
Glutamate (μmol/L)	10 ± 8.7	17 ± 12	7 ± 5	2.04 ± 0.6

Methodology studies have shown the relative recovery (i.e., microdialysis concentration as a percentage of the true extracellular concentration) to be approximately 70% at 1.0 μL/min and 30% at 0.3 μL/min (Hutchinson et al., 2000d).

Data from *Reinstrup et al. (2000) and from †Ungerstedt and Nordstrom (personal communication).

PET, position emission tomography; L/P, lactate/pyruvate.

In the patients with head injury, there was no relation between peripheral and cerebral chemistry, with the exception of glucose. The results suggest that cerebral extracellular glucose levels are a reflection of systemic levels and the metabolic rate of glucose. Cerebral glucose did not correlate with cerebral lactate and pyruvate. Although microdialysis measures the concentration and not production of substances, the data do indicate that high levels of pyruvate were associated with high levels of lactate, suggesting that pyruvate is metabolized to lactate and that anaerobic metabolism is occurring in these patients despite the OEF being within the normal range. All patients with high glutamate levels had high L/P ratios, although the converse was not true.

Although the main objective of this study was to compare microdialysis with PET measurements during the PET scans, this period of microdialysis monitoring represented only a small proportion of the total microdialysis monitoring for these patients. Comparison of the overall microdialysis data with that during the PET scan demonstrates that wide variations are seen in the concentration of substrates and metabolites. To proceed from the neuro-intensive care unit to PET, patients had to be stable in terms of physiologic parameters, and this is reflected in the microdialysis data. We have demonstrated a relation between microdialysis results and outcome in patients with head injury, demonstrating the importance of the L/P ratio and wide variation in glutamate levels (Hutchinson et al., 2000a). Our studies and those from other centers are currently assessing the ability of microdialysis to assist in the management of individual patients; in particular, the effects of interventions on chemistry are being addressed.

PET control data and effect of sedation

The PET control data (nonsedated) obtained from the volunteer studies using equivalent frontal regions of interest revealed values are similar to those from the head-injury patients. Normal OEFs in our laboratory agree with those of other authors (Leenders et al., 1990) and are defined as approximately 0.4. In a PET-microdialysis primate middle cerebral artery occlusion study from Uppsala, the baseline OEF was determined as 0.48 ± 0.11 (penumbra), severe ischemia and reperfusion (0.55 ± 0.066), and severe ischemia with no reperfusion (0.44 ± 0.033) (Enblad et al., 2001). The effect of sedation on microdialysis and PET results is difficult to determine. The Lund study (Reinstrup et al., 2000) of control microdialysis values indicated that overall there were no significant differences between anesthetized and awake patients with the exception of glutamate (Table 6). In terms of blood flow and metabolic data, propofol has been shown to induce metabolic depression and reduction in CBF (Alkire et al., 1995; Menon, 2001).

Microdialysis-PET correlation

There has been a previous clinical study from Uppsala combining PET with microdialysis (Enblad et al., 1996) showing that the presence of regional ischemia within the region of the microdialysis probe is associated with increased levels of energy-related metabolites and excitatory amino acids derived from microdialysis. The same authors proceeded to examine the effects of middle cerebral artery occlusion and reperfusion on microdialysis and PET parameters in primates (Enblad et al., 2001). This study demonstrated that in catheter regions with severe ischemia (defined as CMRO₂ less than 60% that of the contralateral hemisphere), high and broad peaks of L/P ratio, hypoxanthine, and glutamate occurred that persisted in cases of both reperfusion and no reperfusion. In penumbral regions, there were only transient increases during middle cerebral artery occlusion that returned to baseline. We have performed a study comparing PET with a multiparameter sensor (Neurotrend) demonstrating no correlation in absolute values between brain tissue oxygen and PET-derived oxygen concentrations but a good correlation between changes in values induced by hyperventilation (Gupta et al., 2000).

We have attempted to further explore these relations, to determine the cross-correlation of ischemic thresholds and to interpret the results in terms of the pathophysiology of the injured brain. The comparison between PET and microdialysis parameters was made using a 20-mm ROI (by 9-mm thickness). Sampling from microdialysis catheters is within the order of a few millimeters from the membrane. A 20-mm ROI was selected as the optimum balance between restricting the ROI to tissue surrounding the probe while achieving optimal image statistics. As part of a study comparing PET with brain tissue oxygen sensors (Gupta et al., 2002) and in the present study, we attempted to apply ROIs of 10 mm, but because of increased scatter and consequently decreased signal-to-noise ratios, CBF, OEF, CBV, and CMRO₂ could not be quantified with accuracy for these small regions.

There was no significant correlation between CBF and glucose, lactate, the L/P ratio, or glutamate. However, the majority (15/17) of patients had CBF values more than the threshold for ischemia (18 mL · 100 mg⁻¹ · min⁻¹) (Yonas et al., 1989), and significant correlations may therefore be seen in patients with critical ischemia.

There was a strong correlation between the OEF and L/P ratio but not with lactate, thus supporting the concept of measuring the L/P ratio in these patients in addition to lactate per se. An OEF of 0.4 corresponded to an L/P ratio of 26, 25 being regarded as the upper limit of normal (Persson et al., 1996; Reinstrup et al., 2000). Previous studies (Enblad et al., 1996) have indicated that in ischemia, increases in the OEF are related to increases in the L/P ratio. Our data indicate that OEF values in the “nonischemic” range (i.e., less than 70%) are also associated with increases in the L/P ratio, which as an indicator of increased glycolysis with anaerobic metabolism suggests that the metabolic needs of the brain are not entirely covered by oxidative phosphorylation.

There was no significant relation between glutamate and OEF, with the majority of patients having normal glutamate despite an elevated OEF. However, in no patients did the OEF in the relevant ROI approach 75%, which is conventionally accepted as indicating ischemic tissue that is at risk of infarction (Sutton et al., 1990; Powers, 1992; Yundt and Diringer, 1997). CMRO₂ demonstrated poor correlation with the microdialysis parameters.

In 1994, Pellerin and Magistretti reported that glutamate drives glycolysis in astrocytes, resulting in increased extracellular lactate, which can be used as a substrate for neurons in preference to glucose (Pellerin and Magistretti, 1994). The results of this study (from predominantly nonischemic frontal lobe), demonstrating high metabolic activity of the brain with low cerebral glucose and high cerebral lactate compared with the peripheral adipose tissue, no correlation between cerebral lactate and glucose, correlation between cerebral lactate and pyruvate, and the weak relation between high glutamate and high lactate, are compatible with their conclusion.

Effect of hyperventilation

Moderate short-term hyperventilation (reduction of arterial carbon dioxide from 4.92 ± 0.31 to 4.08 ± 0.47 kPa, increase in arterial pH 7.43 ± 0.053 to 7.49 ± 0.074) with preservation of the jugular venous oxygen saturation more than 50%, produced reduction in CBF and increase in OEF, but this appeared to be tolerated in the ROI. Although there was a reduction in the glucose supply, this was not sufficient to produce significant elevation of the L/P ratio or release of glutamate, and overall produced no deleterious effects on cerebral chemistry. However, no areas were sampled by microdialysis where the OEF exceeded 70%, and hyperventilation was only performed for the short-term duration of the scan. Longer periods and more vigorous hyperventilation may result in adverse chemical changes.

CONCLUSIONS

In conclusion, this study has demonstrated that combining microdialysis with PET in ventilated patients is feasible and safe. Relating chemistry to PET-derived metabolic data revealed a strong relation between the L/P ratio and OEF. The data showed that despite L/P ratios of greater than 25 (indicative of anaerobic metabolism), the OEF remained within the normal range. CBF did not significantly correlate with the microdialysis parameters. Moderate hyperventilation produced a reduction in cerebral glucose, but this did not translate into elevation of the L/P ratio or glutamate levels in patients in whom the Sjvo₂ was maintained at more than 50% and with OEF less than 70%.

Further studies are indicated to determine how representative the values recorded from microdialysis catheters are in terms of larger regions (e.g., whole gyri and lobes). Targeting catheters using stereotaxy to abnormal regions defined by functional imaging will enable the pathophysiology of, for example, the progression of contusions to be evaluated.

Footnotes

Acknowledgements

The authors acknowledge the support of members of the Wolfson Brain Imaging Team (S. Downey, J. Jenkins, I. Kendall, L. Maskell, J. Mathews, H. Richards, and C. Turner) and collaboration of Professors U. Ungerstedt and R. Bullock. They thank Dr. R. Tasker for helpful discussion regarding the hyperventilation data and Dr. D. B. A. Hutchinson for supplying the cranial access device. Patents are pending on the design of the device, which is the subject of commercial development by Dr. Hutchinson and the authors (under the company name Technicam).

References

Alkire

Haier

Barker

(1995) Cerebral metabolism during propofol anaesthesia in humans studied with positron emission tomography. Anesthesiology 82:393–403

Alpert

Buxton

Correia

Katz

Ackerman

(1988) Measurement of end-capillary pO2 with positron emission tomography. J Cereb Blood Flow Metab 8:403–410

Baron

Lebrun-Grandie

Collard

Crouzel

Mestelan

Bousser

(1982) Noninvasive measurement of blood flow, oxygen consumption, and glucose utilization in the same brain regions in man by positron emission tomography: concise communication. J Nucl Med 23:391–399

Benveniste

(1989) Brain microdialysis. J Neurochem 52:1667–1679

Benveniste

Diemer

(1988) Early postischaemic 45Ca accumulation in rat dentate hilus. J Cereb Blood Flow Metab 8:713–719

Benveniste

Drejer

Schousboe

Diemer

(1987) Regional cerebral glucose phosphorylation and blood flow after insertion of a microdialysis fiber through the dorsal hippocampus in the rat. J Neurochem 49:729–734

Bullock

Zauner

Woodward

Myseros

Choi

Ward

Marmarou

Young

(1998) Factors affecting excitatory amino acid release following severe human head injury. J Neurosurg 89:507–518

Dohmen

Kumura

Schneweis

Grond

Heiss

W-D

Graf

(2001) Importance of the probe position in animal and human brain microdialysis: differences between gray matter and white matter. J Cereb Blood Flow Metab 21(suppl 1):S176

Editorial (1992) Microdialysis. Lancet 339:1326–1327

10.

Enblad

Valtysson

Andersson

Lilja

Valind

Antoni

Langstrom

Hillered

Persson

(1996) Simultaneous intracerebral microdialysis and positron emission tomography in the detection of ischaemia in patients with subarachnoid haemorrhage. J Cereb Blood Flow Metab 16:637–644

11.

Enblad

Frykholm

Valtysson

Silander

Andersson

Fasth

K-J

Watanabe

Langstrom

Hillered

Persson

(2001) Middle cerebral artery occlusion and reperfusion in primates monitored by microdialysis and sequential positron emission tomography. Stroke 32:1574–1580

12.

Frackowiak

Jones

Lenzi

Heather

(1980a) Regional cerebral oxygen utilization and blood flow in normal man using oxygen-15 and positron emission tomography. Acta Neurol Scand 62:336–344

13.

Frackowiak

Lenzi

Jones

Heather

(1980b) Quantitative measurement of regional cerebral blood flow and oxygen metabolism in man using 15O and positron emission tomography: theory, procedure and normal values. J Comput Assist Tomogr 4:727–736

14.

Frackowiak

Lenzi

Jones

Heather

(1980c) Quantitative measurement of regional cerebral blood flow and oxygen metabolism in man using 15O and positron emission tomography: theory, procedure, and normal values. J Comput Assist Tomogr 4:727–736

15.

Goodman

Valadka

Gopinath

Uzura

Robertson

(1999) Extracellular lactate and glucose alterations in the brain after head injury measured by microdialysis. Crit Care Med 27:1965–1973

16.

Gupta AK, Hutchinson PJ, Fryer T, Wolfson Brain Imaging Team, Menon DK (2002) Measurement of brain tissue oxygenation performed using positron emission tomography scanning to validate a novel monitoring method. J Neurosurg 96:263–268

17.

Heiss

(1998) Application of positron emission tomography to the study of cerebral ischaemia. In: Cerebrovascular disease: pathophysiology, diagnosis and management ( Ginsberg

Bogousslavsky

, eds), Oxford: Blackwell Science, pp 761–772

18.

Hillered

Persson

Ponten

Ungerstedt

(1990) Neurometabolic monitoring of the ischaemic human brain using microdialysis. Acta Neurochir (Wien) 102:91–97

19.

Hutchinson

Al-Rawi

O'Connell

Gupta

Maskell

Hutchinson

DBA

Pickard

Kirkpatrick

(1999a) Monitoring of brain metabolism during aneurysm surgery using microdialysis and brain multiparameter sensors. Neurol Res 21:352–358

20.

Hutchinson

O'Connell

Al-Rawi

Maskell

Gupta

Hutchinson

Pickard

Kirkpatrick

Wharton

(1999b) Neuropathological findings after microdialysis catheter implantation [Letter]. Neuroreport 10:1

21.

Hutchinson

Al-Rawi

O'Connell

Gupta

Maskell

Hutchinson

Pickard

Kirkpatrick

(2000a) On-line monitoring of substrate delivery and brain metabolism in head injury. Acta Neurochir Suppl 76:431–435

22.

Hutchinson

Al-Rawi

O'Connell

Gupta

Pickard

Kirkpatrick

(2000b) Biochemical changes related to hypoxia during cerebral aneurysm surgery: combined microdialysis and tissue oxygen monitoring: case report. Neurosurgery 46:201–206

23.

Hutchinson

Barr

Burgess

Kirkpatrick

Pickard

(2000c) A new cranial access device for cerebral monitoring. Br J Neurosurg 14:46–48

24.

Hutchinson

O'Connell

Al-Rawi

Maskell

Kett-White

Gupta

Richards

Hutchinson

Kirkpatrick

Pickard

(2000d) Clinical cerebral microdialysis—a methodological study. J Neurosurg 93:37–43

25.

Lebrun-Grandie

Baron

Soussaline

Loch'h

Sastre

Bousser

(1983) Coupling between regional blood flow and oxygen utilization in the normal human brain. A study with positron tomography and oxygen 15. Arch Neurol 40:230–236

26.

Leenders

Perani

Lammertsma

Heather

Buckingham

Healy

Gibbs

Wise

Hatazawa

Herold

Beaney

Brooks

Spinks

Rhodes

Frackowiak

Jones

(1990) Cerebral blood flow, blood volume and oxygen utilisation. Normal values and effect of age. Brain 113:27–47

27.

Major

Shdanova

Duffek

Nagy

(1990) Continuous monitoring of blood-brain barrier opening to 51Cr-EDTA by microdialysis following probe injury. Acta Neurochir Suppl (Wien) 51:46–48

28.

Menon

(1999) Cerebral protection in severe brain injury: physiological determinants of outcome and their optimisation. Br Med Bull 55:226–258

29.

Menon

(2001) Mapping the anatomy of unconsciousness- imaging anaesthetic action in the brain. Br J Anaesth 86:607–610

30.

Menon

Minhas

Herrod

Clark

Downey

Kemp

Czosnyka

Pickard

(1997) Heterogeneity of cerebrovascular reactivity in acute brain injury measured by 15-O positron emission tomography. J Cereb Blood Flow Metab 17(suppl 1):S78

31.

Meyerson

Linderoth

Karlsson

Ungerstedt

(1990) Microdialysis in the human brain: extracellular measurements in the thalamus of parkinsonian patients. Life Sci 46:301–308

32.

Pellerin

Magistretti

(1994) Glutamate uptake into astrocytes stimulates aerobic glycolysis: a mechanism coupling neuronal activity to glucose utilization. Proc Natl Acad Sci USA 91:10625–10629

33.

Persson

Hillered

(1992) Chemical monitoring of neurosurgical intensive care patients using intracerebral microdialysis. J Neurosurg 76:72–80

34.

Persson

Hillered

(1996) Intracerebral microdialysis [Letter]. J Neurosurg 85:984–985

35.

Persson

Valtysson

Enblad

Warme

Cesarini

Lewen

Hillered

(1996) Neurochemical monitoring using intracerebral microdialysis in patients with subarachnoid hemorrhage. J Neurosurg 84:606–616

36.

Powers

(1992) Hemodynamics and metabolism in ischemic cerebrovascular disease. Neurol Clin 10:31–48

37.

Reinstrup

Stahl

Mellergard

Uski

Ungerstedt

Nordstrom

(2000) Intracerebral microdialysis in clinical practice: baseline values for chemical markers during wakefulness, anesthesia and neurosurgery. Neurosurgery 47:701–709

38.

Ronne-Engstrom

Cesarini

Enblad

Hesselager

Marklund

Nilsson

Salci

Persson

Hillered

(2001) Intracerebral microdialysis in neurointensive care: the use of urea as an endogenous reference compound. J Neurosurg 94:397–402

39.

Sette

Baron

Mazoyer

Levasseur

Pappata

Crouzel

(1989) Local brain haemodynamics and oxygen metabolism in cerebrovascular disease: positron emission tomography. Brain 112:931–951

40.

Strong

Wong

Jones

Parkin

Boutelle

(2001) Detection and analysis of peri-infarct glucose and lactate transients with rapid-sampling microdialysis. J Cereb Blood Flow Metab 21(suppl 1):S86

41.

Studholme

Hill

Hawkes

(1996) Automated 3-D registration of MR and CT images of the head. Med Image Anal 1:163–175

42.

Studholme

Hill

Hawkes

(1997) Automated three-dimensional registration of magnetic resonance and positron emission tomography brain images by multiresolution optimization of voxel similarity measures. Med Phys 24:25–35

43.

Sutton

McLaughlin

Dante

Kotapka

Sinwell

Mills

(1990) Cerebral venous oxygen content as a measure of brain energy metabolism with increased intracranial pressure and hyperventilation. J Neurosurg 73:927–932

44.

Ungerstedt

(1991) Microdialysis—principles and applications for studies in animals and man. J Intern Med 230:365–373

45.

Vespa

Prins

Ronne-Engstrom

Caron

Shalmon

Hovda

Martin

Becker

(1998) Increase in extracellular glutamate caused by reduced cerebral perfusion pressure and seizures after traumatic brain injury: a microdialysis study. J Neurosurg 89:971–982

46.

Whittle

(1990) Intracerebral microdialysis: a new method in applied clinical neuroscience research. Br J Neurosurg 4:459–462

47.

Whittle

Glasby

Lammie

Ball

Ungerstedt

(1998) Neuropathological findings after intracerebral implantation of microdialysis catheters. Neuroreport 9:2821–2825

48.

Yonas

Sekhar

Johnson

Gur

(1989) Determination of irreversible ischaemia by xenon-enhanced computed tomographic monitoring of cerebral blood flow in patients with symptomatic vasospasm. Neurosurgery 24:368–372

49.

Yundt

Diringer

(1997) The use of hyperventilation and its impact on cerebral ischaemia in the treatment of traumatic brain injury. Crit Care Clin 13:163–184

50.

Zauner

Doppenberg

Woodward

Choi

Young

Bullock

(1997) Continuous monitoring of cerebral substrate delivery and clearance: initial experience in 24 patients with severe acute brain injuries. Neurosurgery 41:1082–1093