The importance of cerebral blood flow to maintain neuronal viability is well known . In brain tissue this can be defined as maintaining tissue oxygenation and a metabolic environment for neuronal function [2-4]. A multiparameter sensor (Paratrend 7 Registered Trademark; Biomedical Sensors, High Wycombe, UK) is available which can simultaneously measure oxygen pressure (PO2), carbon dioxide pressure (PCO2), temperature, and pH using a combined electrode-fiber-optic system. This sensor has been validated for continuous intraarterial monitoring of blood gases  but the principles of the sensor should be valid for tissue measurement.
Jugular bulb oxygen content measures have been used to evaluate brain oxygenation [6-8], However, this technique may not provide an indication of regional ischemia. An alternative is to measure PO2 directly in brain tissue [9-11]. During ischemia, PCO2 increases due to decreased clearance and pH decreases [2,3,12]. We hypothesized that patients with clinical evidence of ischemia would have changes in brain tissue PO2, PCO2, and pH consistent with those seen during acute ischemia. Therefore, we evaluated baseline levels of these variables, in patients with and without a compromised cerebral circulation.
These studies were approved by our institutional review board for clinical research, and informed consent was received. Fourteen patients undergoing craniotomy for cerebrovascular procedures were studied Table 1. Patients with suspected ischemia were identified before surgery, based on the presence of hemodynamically associated neurologic deficits. A regional cerebral perfusion deficit was verified by single photon emission computed tomography in patients with suspected ischemia. An acetazolamide challenge was also given to these patients to confirm the loss of cerebrovascular reactivity within this region. Cerebral angiograms were performed in all patients and hypoperfusion was confirmed in ischemic patients by decreased filling of ischemic tissue. These procedures were used to identify the tissue at risk for sensor placement in patients with suspected ischemia.
At the end of the study, the presurgical patient records of each of the 14 patients were examined by an investigator who was blinded to data collection in order to separate patients into two groups, those with and without a compromised cerebral circulation. A determination of a compromised cerebral circulation was based on the neurological diagnosis of transient ischemic episodes or stroke and confirmation of regional tissue hypoperfusion by single photon emission computed tomography and/or cerebral angiography.
The sensor used in this study has been designed for intravascular blood gas monitoring. It is supplied as a sterile, disposable device comprised of two modified optical fibers for the measurements of PCO2 and pH, a miniaturized Clark electrode for PO2 measurement and a thermocouple for determining temperature. The void between the sensors is filled with acrylamide gel containing phenol red. Changes in hydrogen ion concentration produce color changes in phenol red, which are detected by the pH fiber optic elements. The CO2 sensor includes an ion impermeable barrier which excludes the movement of hydrogen ions but allows the movement of CO2. Within the CO2 sensor, CO2 alters the local pH, producing a color change in phenol red which is detected by the CO2 fiberoptic elements. All measured gas and pH variables are corrected for temperature. The diameter of the probe is 0.5 mm. The four sensing elements are located at different intervals along the final 4 cm of the probe Figure 1. Approximately 4 cm must be inserted for all of the elements to be in the tissue but the sensors are no further apart than 3 cm. The outer surface of the sensor has a covalently bonded heparin coating.
The sensor is packaged with a tonometer containing buffer solution which serves as a calibrating medium. The sensor is calibrated with three precision gases supplied with the monitor before insertion into the patient. These gases are: 1 = 2% CO2, 15% O2, balance N2; 2 = 5% CO2, 15% O2, balance N2; 3 = 10% CO2, 15% O2, balance N2. Each gas is bubbled into the calibrating solution for 10 min. The oxygen calibration curve (0-120 mm Hg) is constructed using an electric zero and the 15% O2 gas assuming linear properties of the electrode. The CO2 and pH calibration curves are constructed within the range of 10-80 mm Hg using the three CO2 gas concentrations: 2%, 14 mm Hg (pH 7.83); 5%, 36 mm Hg (pH 7.43); and 10%, 71 mm Hg (pH 7.13). The range and 95% confidence limits for each sensor have been determined in in vitro testing: O2, range 0-120 mm Hg, 95% confidence limits +/- 1 mm Hg; CO2, range 10-80 mm Hg, 95% confidence limits +/- 3 mm Hg; pH, range 6.80-7.80, 95% confidence limits +/- 0.03. The 0%-90% response time for each of the electrodes is: O2 70 s, CO2 143 s, pH 78 s . Bicarbonate is calculated using pH and PCO2 from the Henderson-Hasselbalch equation .
For insertion of the sensor, the outer, nonsterile introducer system is covered with a sterile sheath. The sterile sensor is then extended from the end of the system and visually inserted into cortex tissue of interest during the surgery. The probe is inserted 4 cm to ensure that all sensing elements are in brain tissue. An RS-232 connector on the back of the Paratrend 7 Registered Trademark monitor supplied PO2, PCO2, pH, bicarbonate, and temperature measures for continuous data collection.
Patients were anesthetized with 10-15 micro gram/kg fentanyl and 3-5 mg/kg thiopental, paralyzed with 100 micro gram/kg vecuronium, intubated, and ventilated with 0.5%-1.0% end-tidal isoflurane. A radial artery catheter was inserted to measure arterial pressure (Marquette, Milwaukee, WI). The fraction of inspired oxygen was adjusted between 0.4 and 0.6 to maintain PaO2 > 100 mm Hg. An attempt was made to maintain PaCO2 between 31 and 36 mm Hg before the craniotomy, to decrease brain mass. This was not possible in Patient 10 due to pulmonary dysfunction. It was decided to maintain Patient 1 at a higher PaCO2 during surgery because of the presence of severe cerebral occlusive disease.
After the craniotomy, the dura was opened and the Paratrend 7 Registered Trademark sensor inserted 4 cm into the cortex within the distribution of the compromised circulation. In nonischemic patients, the sensor was inserted into tissue of interest for the surgical procedure. After a 30-min equilibration period, an arterial blood gas sample was taken and baseline measures of end-tidal CO2 and anesthetic gases, mean arterial pressure, and brain tissue PO2, PCO2, pH, and temperature were recorded for 10 min. This measurement was made between 2 and 3 h after the start of the surgery. All measures are reported at a temperature of 37 degrees C. All data were collected by computer using Labview Registered Trademark (National Instruments, Dallas, TX) every 10 s.
Groups were compared statistically using t-tests with Bonferroni correction for multiple testing. In all patients, baseline brain tissue PCO2 and pH were plotted as a function of PO2. A Pearson product moment correlation was performed between tissue measures of tissue PCO2 and hydrogen ion concentration.
Based on patient records, 8 of the 14 patients were identified with a compromised cerebral circulation at the time of surgery. Two of these patients, scheduled for aneurysm repair, had subarachnoid hemorrhage before surgery and Patient 5 had cerebral vasospasm as determined by angiography. Extracerebral to intracerebral, bypass was performed in four patients with cerebral occlusive disease. According to patient records, Patient 7 had a normal cerebrovascular circulation at the time of surgery. However, an extracerebral to intracerebral bypass was performed in anticipation of carotid sacrifice during resection of a skull base meningioma.
The patients with clinically defined cerebrovascular deficits had significantly lower tissue PO2, higher tissue PCO2, lower tissue pH, and tissue bicarbonate compared to noncompromised patients Table 1. There was no difference in MAP, blood gases, or brain temperature between the two groups at the time of testing.
Tissue PCO2 and pH were plotted as a function of PO2 in all patients Figure 2. In noncompromised patients, PCO2 remained in a range of 40-60 mm Hg and pH remained above 7.0 over a range of PO2 of 20-60 mm Hg. In compromised patients, PO2 was decreased below 20 mm Hg. Coincidentally, PCO (2) was increased above 60 mm Hg and pH was decreased below 7.0.
When tissue hydrogen ion concentration (H+) was plotted as a function of tissue PCO2 for each patient, there was a trend for increasing H+ as PCO2 increased. However, the data for Patient 5, who was diagnosed with cerebral vasospasm, was clearly an outlier from this trend. When the data from Patient 5 was excluded there was a significant correlation between H+ and PCO2 (r = 0.92, P < 0.001; Figure 3).
The sensors were well tolerated during their insertion into brain tissue and no bleeding or inflammation were observed after their removal.
These results show that patients with a compromised cerebral circulation have significantly lower tissue PO2, increased PCO2 and decreased pH compared to noncompromised controls. These differences in tissue gases and pH are consistent with decreased oxygen delivery and CO2 clearance. When evaluating PCO2 and pH as a function of PO2, we observed that PCO2 increased above 60 mm Hg and pH decreased to less than 7.0 when PO2 was below 20 mm Hg. This supports the conclusion that brain tissue PO2, PCO2, and pH are related to tissue perfusion and oxygen delivery and that they have potential use to indicate the status of tissue at risk for ischemia.
Previous studies have measured brain tissue PO2 in noncompromised brain tissue. A brain interstitial PO2 of 28 +/- 7 mm Hg has been reported in dogs with a PaO2 of 112 +/- 6 mm Hg . In humans, Assad et al.  used a polarographic multiwire surface electrode to measure PO2 in normal and tumor tissue. In three patients they reported average values of 33-36 mm Hg in normal tissue with a PaO2 of 170 mm Hg. Meixenberger et al.  measured interstitial brain PO2 using a Clark electrode and reported a tissue of PO2 of 48 +/- 13 mm Hg when PaO2 was 170 +/- 28 mm Hg and PaCO2 was 30 +/- 3 mm Hg. Our study agrees with these earlier reports; six patients with a noncompromised brain circulation had a tissue PO2 of 37 +/- 12 mm Hg.
Thews  theoretically calculated hypoxic thresholds in brain tissue. He determined that a tissue PO2 of 17 mm Hg would be the lowest normal level and that 4-11 mm Hg would be the critical range for hypoxic injury. This is consistent with our finding that patients with clinically defined ischemia had a baseline PO (2) of 9 +/- 6 mm Hg. The ischemic patients also showed significant changes in tissue PCO2 and pH. This supports the suggestion that an ischemic threshold occurred which can be defined by simultaneous changes in PO2, PCO2, and pH.
Tissue CO2 easily diffuses across the blood-brain barrier and tissue PCO2 may approximate cerebral venous PCO2 [14,15]. That would agree with our normal tissue values of 49 +/- 5 mm Hg. During ischemia, tissue perfusion and clearance of CO2 may decrease, increasing tissue PCO2. A tissue PCO (2) > 60 mm Hg was seen only in cerebrovascularly compromised patients. The increased tissue PCO2 observed in patients with a compromised cerebrovasculature is consistent with ischemic tissue PCO2  and with changes in tissue PCO2 during temporary ischemia .
Our values of tissue pH in noncompromised patients agree with other authors [2,4]. Ischemic changes in brain tissue pH have also been reported. During three minutes of complete ischemia, extracellular pH measured on the cortical surface of the cat decreased 0.5 units . Estimates in intracellular pH in rats show that 10 minutes of ischemia decrease pH from 7.04 to 6.48 . This is consistent with our measures of tissue pH in patients with chronic cerebral ischemia.
There was a trend for increasing H+ as PCO2 increased. The correlation for this trend was r = 0.92, P < 0.001 when Patient 5 was excluded from the calculation. This relationship between PCO2 and pH is consistent with previous reports . However, increases in brain PCO2 are associated with increases bicarbonate concentration in brain tissue and cerebrospinal fluid [16-18]. In contrast, in this study, patients with ischemia had higher tissue PCO2 and lower bicarbonate. The depletion of tissue bicarbonate suggests the presence of metabolic acidosis in ischemic patients. This was especially severe in Patient 5, who had cerebral vasospasm. These results suggest that measurement of tissue PCO2 and pH can better define the state of ischemia by evaluating the presence of metabolic acidosis.
It is a concern that the sensing elements of this probe are separated by up to 3 cm in brain tissue. This suggests that different brain regions may be measured during monitoring of PO2, PCO2, and pH. On the cortical surface, brain PO2 varies markedly over a distance of a few millimeters . However, regional PO2 within brain tissue may be less variable [9-11]. The interelectrode distance between the cathode and anode and the need for oxygen to diffuse through the acrylamide gel of the Paratrend 7 Registered Trademark oxygen sensor would expand the tissue region from which oxygen is sampled. CO2 is highly diffusable in brain tissue and local measurements may be influenced by diffusion from surrounding tissue. Further, clinical verification of a compromised circulation in this study probably indicates that a significant portion of brain tissue is affected. This may explain why measures of PO2, PCO2, and pH were internally consistent for validating ischemia within each patient.
Since this sensor is not designed for tissue measurement, it should be considered whether these measurements are valid. The Clark electrode has been used for measurement of oxygen in animal and human brain tissue previously [9-11] and was calibrated for the measurements made here. One CO2 measurement and two pH measurements lie outside the calibration range constructed for the CO2 and pH sensors, respectively. The confidence limits determined for CO2 and pH would not be valid for these three measures. However, in vivo studies in pigs have shown that the accuracy of blood CO2 and pH measures is not markedly altered when measurements are made beyond the normal calibration range . This suggests that, although the measurements of CO2 and pH outside the calibration range may be less reliable, they are still valid.
Finally, there is the question of whether this sensor is compatible with brain tissue. The sensor is constructed of physiologically inert polyethylene with a covalently bonded surface of heparin. The special bonding process, which does not allow diffusion of heparin from the surface, is a critical design to allow the sensor to remain in arterial blood for up to a week without clotting. We observed no adverse tissue reactions or bleeding when this sensor was inserted into brain tissue. This was facilitated by visual insertion of the sensor and the small size of the sensor (0.5 mm in diameter). We conclude the sensor is physiologically inert, consistent with its design for arterial blood.
In summary, these results show that patients with a compromised cerebral circulation have decreased brain tissue PO2, increased PCO2, and decreased pH compared to noncompromised controls. Evaluation of PCO2 and pH in relation to PO2 suggests that critical ischemia may occur when tissue PO2 decreases below 20 mm Hg. This is consistent with theoretical analyses. Simultaneous measures of brain tissue PO2, PCO2, and pH may improve the evaluation of ischemia in neurosurgical patients.
1. Siesjo BK. Brain energy metabolism. New York: John Wiley, 1973.
2. Astrup J, Heuser D, Lassen NA, et al. Evidence against H+
as the main factors in the regulation of cerebral blood flow during epileptic discharges, acute hypoxemia, amphetamine intoxication, and hypoglycemia. A microelectrode study. In: Betz E, ed. Ionic actions on vascular smooth muscle Berlin: Springer-Verlag, 1972;110-5.
3. Ljunggren B, Schutz H, Siesjo BK. Changes in energy state and acid-base parameters of the rat brain during complete compression ischemia. Brain Res 1974;73:277-89.
4. Waddell WJ, Bates RG. Intracellular pH. Physiol Rev 1969;49:285-329.
5. Venkatesh B, Brock TH, Hendry SP. A multiparameter sensor for continuous intra-arterial blood gas monitoring: a prospective evaluation. Crit Care Med 1994;22:588-93.
6. Nakjima T, Ohsumi H, Kuro M. Accuracy of continuous jugular bulb venous oximetry during cardiopulmonary bypass. Anesth Analg 1993;77:1111-5.
7. Robertson CS, Narayan RK, Gokaslan ZL, et al. Cerebral arteriovenous oxygen difference as an estimate of cerebral blood flow in comatose patients. J Neurosurg 1989;70:222-30.
8. Matta BF, Lam AM, Mayberg TS, et al. A critique of the intraoperative use of jugular venous catheters during neurosurgical procedures. Anesth Analg 1994;79:45-50.
9. Mass AIR, Fleckenstein W, de Jong A, van Santbrink H. Monitoring cerebral oxygenation: experimental studies and preliminary clinical results of continuous monitoring of cerebrospinal fluid and brain tissue oxygen tension. Acta Neurochir 1993;59:50-7.
10. Assad F, Scultheiss R, Leniger-Follert E, Wullenweber R. Measurement of local oxygen partial pressure (PO2
) of the brain cortex in cases of tumors. Adv Neurosurg 1984;12:263-6.
11. Meixenberger J, Dings J, Kuhnigk H, Roosen K. Studies of tissue PO2
in normal and pathological human brain cortex. Acta Neurochir (Wein) 1993;59:58-63.
12. Hoffman WE, Charbel FT, Edelman G, et al. Brain tissue oxygen pressure, carbon dioxide pressure and pH during ischemia. Neurol Res. In press.
13. Thews G. Implications to physiology and pathology of oxygen diffusion at the capillary level. In: Schade JP, McMenemey WH, eds. Scientific vulnerability of the brain in hypoxaemia. Oxford: Blackwell, 1979;27-35.
14. Ponten U, Siesjo BK. Gradients of CO2
in the brain. Acta Physiol Scand 1966;67:129-40.
15. Friis ML, Paulson OB, Hertz MM. Carbon dioxide permeability of the blood-brain barrier in man. Microvasc Res 1980;20:71-80.
16. Kazemi H, Shannon DC, Carvallogh E. Brain CO2
buffering capacity in respiratory acidosis and aldalosis. J Appl Physiol 1967;22:241-6.
17. Siesjo BK. The regulation of cerebrospinal fluid pH. Kidney Int 1972;1:360-74.
18. Hoffman WE, Charbel FT, Edelman G, Ausman JI. Brain tissue acid-base response to hypercapnia in neurosurgical patients. Neurol Res 1995;17:417-20.
19. Leniger-Follert E, Lubbers DW, Wrabetz W. Regulation of local tissue PO2
of the brain cortex at different arterial O2
pressures. Pflugers Arch 1975;359:81-95.
20. Clutton-Brock TH, Fink S, Luthra AJ, Hendry SP. The evaluation of a new intravascular blood gas monitoring system in the pig. J Clin Monit 1994;10:387-91.