British Journal of Anaesthesia 1997; 78: 169–171 Brain tissue oxygenation in patients with cerebral occlusive disease and arteriovenous malformations W. E. HOFFMAN, F. T. CHARBEL, G. EDELMAN AND J. I. AUSMAN Summary Patients and methods It is not clear if ventilation with oxygen increases The studies were approved by the University of brain tissue oxygen pressure (PO2) during Illinois Institutional Review Board for Clinical ischaemia. We have measured brain tissue PO2 Research and informed consent was obtained. carbon dioxide pressure (PCO2) and pH during Patients in group 1 (n:9) served as controls for the baseline anaesthesia and oxygen ventilation in study. None of these patients bled or had clinical non-ischaemic control patients (n:9), patients signs of ischaemia before surgery. Neurosurgery was with cerebral occlusive disease (n:11) and performed in these patients for clipping of cerebral patients with arteriovenous malformations (AVM, aneurysms. Patients in group 2 (n:11) had evidence n:12). The same anaesthetic treatment was given of cerebral ischaemia, as determined by neurological to all groups and anaesthesia was constant during examination or the presence of transient ischaemic the study. Arterial pressure, brain temperature and episodes. Decreased regional cerebral perfusion was arterial blood-gas tensions were similar between confirmed in these patients by single photon emis- groups. Under baseline conditions, brain tissue PO2 sion computed tomography (SPECT) or cerebral was mean 4.2 (SD 1.4) kPa in the controls and was angiography. Six of these patients were undergoing 70% lower in patients with ischaemia and AVM. extracerebral to intracerebral vascular bypass, two Patients with occlusive disease also had elevated patients cerebral aneurysm surgery with confirmed tissue PCO2 and acidosis. During oxygen ventila- cerebral vasospasm and three patients had a cerebral tion, PO2 increased to 7.5 (2.9) kPa in controls and embolism. Patients in group 3 (n:12) were under- this was 50% greater than the increase in the going neurosurgery for AVM resection. Nine of ischaemia and AVM patients. The results showed these patients had previous neuroradiological AVM that baseline tissue oxygenation and increases in embolization procedures performed before AVM PO2 during hyperoxia were attenuated in patients resection. with ischaemia or AVM. (Br. J. Anaesth. 1997; 78: All patients were anaesthetized with thiopentone 169–171) 3–5 mg kg91 and fentanyl 10–15 g kg91. Tracheal intubation was facilitated with vecuronium 0.1 mg Key words kg91 and the lungs ventilated with 0.5–1.5% isoflu- Brain, ischaemia. Brain, carbon dioxide. Complications, rane and oxygen in room air (inspired oxygen frac- arteriovenous malformation. Carbon dioxide, partial tion:0.4). Oesophageal temperature was measured pressure. and allowed to decrease to approximately 34 C. Arterial carbon dioxide tension ( P aCO 2 ) was adjusted to 3.9–4.6 kPa. Monitoring included mean radial Under normal conditions, brain tissue PO2 is arterial pressure (MAP) measured by a Marquette reported to be in the range 3.3–5.2 kPa in animals Electronics Tramscope (Milwaukee, WI, USA); and patients.1–3 Tissue oxygenation may be end-tidal isoflurane and PCO2 were measured with a inhibited by cerebral occlusive disease or by arterial Datex Ultima (Helsinki, Finland). venous malformations (AVM), which attenuate After craniotomy, a PO2, PCO2, pH and tempera- cerebral perfusion pressure by producing a low ture sensor (Paratrend, Biomedical Sensors, resistance shunt for blood flow.4–6 It has been Malvern, PA) was inserted into cortex tissue reported that the upper limit of tissue oxygenation is (diameter:0.5 mm). The sensor is a sterile, dispos- controlled in normal brain tissue during oxygen able device comprised of two modified optical fibres ventilation, but that these regulating mechanisms are missing after head injury.3 Little is known of WILLIAM E. HOFFMAN*, PHD, GUY EDELMAN, MD (Department of tissue changes in PO2 during hyperoxia in tissue Anesthesiology); FADY T. CHARBEL, MD, JAMES I. AUSMAN, MD, affected by occlusive disease or AVM. The purpose PHD (Department of Neurosurgery); University of Illinois at of this study was to evaluate the effect of oxygen Chicago, Chicago, IL, USA. Accepted for publication: September 29, 1996. ventilation in patients with cerebral occlusive *Address for correspondence: Anesthesiology Department, disease and AVM compared with non-ischaemic University of Illinois Hospital, 1740 W Taylor, Suite 3200, controls. Chicago, IL 60612, USA. 170 British Journal of Anaesthesia Table 1 Mean arterial pressure (MAP), arterial oxygen and carbon dioxide partial pressures and pH, tissue carbon dioxide partial pressure, pH and brain temperature during 40% (baseline) and 100% oxygen ventilation (oxygen) in controls, and in those with cerebral occlusive disease (COD) and arteriovenous malformations (AVM) (mean SD)). PaO2 :arterial PO2, PaCO2 :arterial PCO2, pHa:arterial pH, PtCO2 :tissue PCO2, pHt:tissue pH. *P:0.05 compared with baseline, †P:0.05 compared with control MAP PaO2 PaCO2 PtCO2 Brain temp Group Treatment n (mm Hg) (kPa) (kPa) pHa (kPa) pHt (ºC) Control Baseline 9 86 (15) 24.9 (7.3) 4.3 (0.4) 7.47 (0.04) 6.4 (0.8) 7.14 (0.15) 35.6 (1.3) Oxygen 87 (15) 54.0 (7.6)* 4.3 (0.6) 7.43 (0.07) 6.3 (0.8) 7.18 (0.10) 35.6 (1.3) COD Baseline 11 88 (12) 29.2 (4.1) 4.5 (0.4) 7.42 (0.06) 9.2 (4.9)† 6.85 (0.25)† 35.0 (0.6) Oxygen 88 (12) 58.7 (3.3)* 4.8 (0.3) 7.38 (0.03) 9.1 (3.7)† 6.88 (0.25)† 34.9 (0.6) AVM Baseline 12 80 (8) 23.3 (4.7) 4.3 (0.7) 7.43 (0.07) 6.7 (1.2) 7.09 (0.20) 35.2 (0.9) Oxygen 81 (8) 52.3 (8.4)* 4.3 (0.5) 7.44 (0.05) 6.5 (1.2) 7.17 (0.17) 35.2 (0.9) for measurement of PCO2 and pH, a miniaturized Clark electrode for PO2 measurement and a thermo- couple for measurement of temperature. The sensor was calibrated with three precision gases supplied with the monitor before insertion into the patient. The gases are: 1:2% carbon dioxide, 15% oxygen, balance nitrogen; 2:5% carbon dioxide, 15% oxygen, balance nitrogen; 3:10% carbon dioxide, 15% oxygen, balance nitrogen. The calibration range and 95% confidence limits for each sensor have been determined in in vitro testing: oxygen (range 0–15.7 kPa, 95% confidence limits ;0.1 kPa); carbon dioxide (range 1.3–10.5 kPa, 95% con- fidence limits ;0.4 kPa); pH (range 6.80–7.80, 95% confidence limits ;0.03). The 0% to 90% response time for each sensor is: oxygen:70 s, carbon dioxide:143 s, pH:78 s.7 All monitored variables were collected by computer using Labview (National Figure 1 Changes in PO2, PCO2 and pH during oxygen Instruments, Austin, TX) every 10 s. ventilation in a non-ischaemic patient. The scale for PO2 and When the sensor was inserted, baseline tissue gas PCO2 (kPa) is shown on the left and that for pH on the right. tensions, pH and temperature, MAP and blood-gas Oxygen ventilation produced an increase in PO2 with no change in PCO2 or pH. tensions were measured after a 30-min equilibration period. Inspired oxygen fraction was increased from 0.4 to 1.0 for 10 min and then returned to baseline levels. Samples for arterial blood-gas measurements were obtained under baseline conditions and during oxygen ventilation. Data are reported as mean (SD). Differences in mean values between the three groups were analysed by analysis of variance with post hoc testing if a signif- icant F value was found. Differences between base- line and oxygen administration were analysed using paired t tests. If data distribution failed the underly- ing assumptions for normality and equal variance, non-parametric analyses using Kruskal–Wallis or Wilcoxon tests were performed. Results Under baseline conditions, the physiological vari- ables in table 1 were similar in control, ischaemic Figure 2 Tissue PO2 during baseline measurements (Base) and oxygen ventilation (O2). Baseline PO2 was greater in controls and AVM patients. With administration of 100% compared with those with cerebral occlusive disease and AVM. oxygen in the inspired gases, PaO increased in all 2 The increase in PO2 during hypoxia was attenuated in cerebral three groups to a similar degree. The response to occlusive disease and AVM compared with controls. *P:0.05 oxygen ventilation in a normal patient is shown in compared with baseline; †P:0.05 compared with control. figure 1. Tissue PO2 increased in these patients during the increase in PaO but there was no 2 change in tissue PCO2 or pH. Figure 2 show mean attenuated in ischaemic and AVM patients tissue PO2 data for all patients. Under baseline during hyperoxia. Tissue PCO2 increased and conditions, tissue PO2 decreased in ischaemic and pH decreased in ischaemic patients during the AVM patients compared with controls, and the baseline period but neither variable changed increase in PO2 during oxygen administration was during oxygen ventilation (table 1). Brain tissue oxygen 171 and AVM were different but that inadequate Discussion tissue oxygenation was a consistent problem in In non-ischaemic patients, we found a baseline both groups. tissue PO2 of 4.2 kPa. This is consistent with pre- vious reports.1–3 8 During hyperoxia, tissue PO2 increased to 7.5 kPa in non-ischaemic patients. This References agrees with the conclusions of Meixenberger and 1. Hoffman WE, Charbel FT, Edelman G. Brain tissue oxygen, colleagues3 that local mechanisms regulate the upper carbon dioxide, and pH in neurosurgical patients at risk for limit of tissue oxygenation. Tissue oxygenation ischemia. Anesthesia and Analgesia 1996; 82: 582–586. decreased significantly in patients with cerebral 2. Mass AIR, Fleckenstein W, de Jong A, van Santbrink H. occlusive disease and AVM under baseline condi- Monitoring cerebral oxygenation: experimental studies and preliminary clinical results of continuous monitoring of tions and during oxygen ventilation. This indicates cerebrospinal fluid and brain tissue oxygen tension. Acta that mechanisms which promote tissue oxygenation Neurochirurgica 1993; 59: 50–57. are attenuated in occlusive disease and AVM. 3. Meixenberger J, Dings J, Kuhnigk H, Roosen K. Studies of Previous studies have suggested that cerebral tissue PO2 in normal and pathological human brain cortex. occlusive disease and AVM produce different types Acta Neurochirurgica 1993; 59: 58–635. 4. Batjer HH, Devous MD sr, Meyer YJ, Purdy PD, Samson of ischaemia. Acute and chronic brain ischaemia DS. Cerebrovascular hemodynamics in arteriovenous during brain artery occlusion is associated with a malformation complicated by normal perfusion pressure decrease in PO2, an increase in PCO2 and acidosis.1 8–11 breakthrough. Neurosurgery 1988; 22: 503–509. This is accompanied by loss of cerebrovascular 5. Young WL, Prohovnik I, Ornstein E, Ostapkovich N, Sisti MB, Solomon RA, Stein BM. The effect of arteriovenous reactivity to increases in PaCO2 .12 In contrast, with an malformation resection on cerebrovascular reactivity to AVM, cerebral tissue perfusion pressure and blood carbon dioxide. Neurosurgery 1990; 27: 257–267. flow are decreased because of shunting of arterial 6. Young WL, Kader A, Prohovnik I, Ornstein E, Fleischer LH, blood flow away from normal tissue.4–6 13 Although Ostapkovich N, Jackson LD, Stein BM. Pressure autoregula- AVM have been described as producing vascular tion is intact after arteriovenous malformation resection. Neurosurgery 1993; 32: 491–496. paralysis and ischaemia,14 studies have shown that 7. Venkatesh B, Brock TH, Hendry SP. A multiparameter carbon dioxide reactivity and pressure autoregula- sensor for continuous intra-arterial blood gas monitoring: a tion are intact.5 6 In spite of the reported differences prospective evaluation. Critical Care Medicine 1994; 22: between these two types of ischaemia, we found that 588–593. 8. Van Santbrink H, Maas AIR, Avezaat CJJ. Continuous the tissue oxygenation response to oxygen ventila- monitoring of partial pressure of brain tissue oxygen in tion was attenuated in both groups. This suggests patients with severe head injury. Neurosurgery 1996; 38: that inadequate brain tissue oxygen delivery is a con- 21–31. sistent problem in patients with cerebral occlusive 9. Hoffman WE, Charbel FT, Edelman G, Ausman JI. Brain disease and AVM. tissue acid–base response to hypercapnia in neurosurgical patients. Neurological Research 1995; 17: 417–420. In this study, patients with occlusive disease had 10. Hoffman WE, Charbel FT, Edelman G, Hannigan K, significantly increased tissue PCO2 and decreased Ausman JI. Brain tissue oxygen pressure, carbon dioxide pH. This supports previous reports that carbon pressure and pH during ischaemia. Neurological Research dioxide clearance is attenuated and metabolic 1996; 18: 54–56. 11. Siesjo BK, Kajallquist A. A new theory for the regulation of acidosis may be present in these regions.9 11 15 16 In the extracellular pH in the brain. Scandinavian Journal of contrast, in tissue adjacent to an AVM, PCO2 and pH Clinical Laboratory Investigation 1969; 24: 1–9. did not differ from control values, possibly because 12. Dettmers C, Young A, Rommel T, Hartmann A, Weingart decreased tissue metabolism would lower carbon O, Baron JC. CO2 reactivity in the ischaemic core, dioxide production and normalize pH in hypoxic penumbra, and normal tissue 6 hours after acute MCA occlusion in primates. Acta Neurochirurgica (Wien) 1993; tissue.17 One concern is that the lack of change in 125: 150–155. brain tissue PCO2 and pH results from a lack of sensi- 13. Norbash AM, Marks MP, Lane B. Correlation of pressure tivity of the sensors rather than absence of a tissue measurements with angiographic characteristics predisposing change. However, in previous studies we have shown to hemorrhage and steal in cerebral arteriovenous malforma- tions. American Journal of Neuroradiology 1994; 15: 809–813. that tissue PCO2 and pH change in patients during 14. Spetzler RF, Wilson CB, Weinstein P, Mehdorn M, hypercapnic challenge.9 This supports the fact that Townsend J, Telles D. Normal perfusion pressure break- the lack of changes in tissue PCO2 and pH during through theory. Clinical Neurosurgery 1978; 25: 651–672. increases in oxygen ventilation were accurate. 15. Bradley RD, Semple SJG. A comparison of certain acid–base In summary, these results showed that brain tissue characteristics of arterial blood, jugular venous blood and cerebrospinal fluid in man, and the effect on them of some PO2 was lower in patients with cerebral occlusive acute and chronic acid base disturbances. Journal of disease and AVM compared with non-ischaemic Physiology (London) 1962; 160: 381–391. patients. During oxygen ventilation, tissue PO2 16. Mines AH, Morrill CG, Sorenson SC. The effect of isocarbic ; increased (4.6 kPa) in control patients; this was metabolic acidosis in blood on [H ] and [HCO39] in CSF with deductions about the regulation of an active transport of significantly greater than the response in patients H;/HCO39 between blood and CSF. Acta Physiologica with occlusive disease and AVM. Simultaneous Scandinavica 1971; 81: 234–245. measurements of tissue PCO2 and pH indicated that 17. Fink GR. Effects of cerebral angiomas on perifocal and the metabolic characteristics of occlusive disease remote tissue. Stroke 1992; 23: 1099–1105.