Hypothermia has been used for cerebral protection during cardiopulmonary bypass (CPB) , but this protective effect may not be present at critical periods . During the rewarming phase of hypothermic CPB, cerebral venous oxyhemoglobin desaturation to less than 50% (as measured by jugular venous oximetry [SjO2]) may occur [3-5].
We hypothesized that pharmacological intervention to reduce cerebral metabolic rate (CMRO2) during the rewarming phase of hypothermic CPB might prevent or ameliorate the decrease in cerebral venous oxyhemoglobin saturation. Thiopental has previously been used in doses that produce electroencephalographic (EEG) burst suppression as a cerebral protectant during cardiac surgery . However, it results in a prolonged requirement for mechanical ventilation and an increased use of inotropic drugs. Propofol also produces burst suppression on the EEG in appropriate doses  and reduces the CMRO2 during normothermia  without affecting flow-metabolism coupling . Moreover, because it is more rapidly eliminated than thiopental, adverse effects of propofol on consciousness and the cardiovascular system would be anticipated to be present for a shorter duration. Therefore, propofol may be preferable to thiopentone for reduction of CMRO2 during CPB.
Reduction in SjO2 has frequently been used as an aerobic estimate of cerebral hypoperfusion during cardiac surgery and has been related to brain injury in the form of cognitive deficits . In contrast, anaerobic estimates such as cerebral lactate production (AVDL) and its combination with aerobic estimates, in the form of lactate-oxygen index (LOI), have been infrequently explored in cardiac surgery. Because AVDL and LOI are valuable in predicting adverse cerebral outcome after neurotrauma, these estimates are worthy of study during cardiac surgery. Unlike reductions in SjO2, the relationship of AVDL and LOI to brain injury associated with cardiac surgery is unknown.
For the above reasons, the primary aim of this study was to investigate the effect of propofol on SjO2 during the rewarming stage of hypothermic CPB. Subsidiary aims of this study were to explore the influence of propofol on levels of AVDL and LOI during hypothermic CPB.
The study received local ethics committee approval and was undertaken in the cardiac surgery unit at the Royal Infirmary of Edinburgh. Thirty patients, aged 46-75 yr, scheduled for elective coronary artery bypass grafting, who gave informed consent were recruited. The study had an open, randomized, prospective, and parallel group design. Patients who had epilepsy, cerebral vascular accident, transient ischemic attack, diabetes mellitus, treatment with barbiturates or sedative agents, a history of alcohol abuse, or allergy to propofol were excluded.
Patients were premedicated with lorazepam 1-4 mg orally the night before surgery and temazepam 20-30 mg orally 1 h before the induction of anesthesia. Venous and arterial cannulae were inserted during preoxygenation. Anesthesia was induced with midazolam up to 0.2 mg/kg, fentanyl 5 [micro sign]g/kg, and etomidate as required to abolish the eyelash reflex. Neuromuscular blockade was obtained using pancuronium 0.15 mg/kg IV, and the trachea was intubated. Ventilation was controlled mechanically, before and after CPB, to obtain normocapnia (5.2 kPa). Nasopharyngeal temperature (NPT) was measured. Anesthesia was maintained with IV infusions of fentanyl 0.1 [micro sign]g [center dot] kg-1 [center dot] min-1 and midazolam 0.5 [micro sign]g [center dot] kg-1 [center dot] min-1 supplemented with isoflurane as required to control increases in mean arterial pressure (MAP) before CPB. Processed EEG was monitored throughout surgery using a cerebral function analyzing monitor (Cerebrotrac; SRD Medical Limited, Shorashim, Israel) with five electrodes over a bilateral fronto-centro-parietal distribution. Impedance was less than 5000 ohms on each electrode with a sensitivity range of 1-200 [micro sign]V.
A central venous catheter and/or a pulmonary artery catheter was inserted via the right internal jugular vein. A spectrophotometric catheter (Opticath; Abbott Laboratories, Maidenhead, Berkshire, UK) was inserted into the jugular vein using a previously described technique , and the catheter tip position was confirmed radiographically. The catheter was connected to an Oximetrix 3 measurement system (Abbott Laboratories) and calibrated in vivo against jugular venous samples analyzed by using a Corning 270 Co-oximeter (Ciba-Corning Diagnostics Ltd., Halstead, Essex, UK).
Before aortic cannulation, heparin was administered IV to obtain an activated clotting time greater than 450 s. A standard CPB circuit was used, primed with 2 L of lactated Ringer's solution and sodium bicarbonate 50 mM, incorporating a membrane oxygenator (I-3500-2A; AVECOR Cardiovascular Inc., Plymouth, MN) and a nonpulsatile roller pump (Stockert Instruments, Munich, Germany) with flow rates maintained at 2.4 L [center dot] min-1 [center dot] m-2. Acid-base management followed alpha-stat principles. Moderate hypothermia (NPT 28[degree sign]C) was induced in all patients. During rewarming, the temperature of the water from the heat exchanger was no greater than 10[degree sign]C warmer than NPT, and never more than 42[degree sign]C. Whenever the MAP decreased below 50 mm Hg, 2-mg bolus doses of methoxamine were given IV until MAP returned to that level. Similarly, bolus doses of phentolamine 1 mg IV were administered when the MAP exceeded 90 mm Hg, until it decreased. Hypovolemia was corrected by administering Hartmann's solution (hemoglobin >8 g/100 mL) or concentrated red cells (hemoglobin <8 g/100 mL).
After initiation of CPB, patients in the propofol group were given bolus doses of propofol at 30 mg/min until a burst suppression ratio of 80% was obtained on the Cerebrotrac monitor. An IV infusion of propofol was commenced to maintain a burst suppression ratio of 80% throughout CPB, until rewarming had produced a NPT of 36[degree sign]C, at which time the infusion was discontinued. The Oximetrix monitor and Kontron Colormon patient monitor (Kontron Instruments Ltd., Watford, UK) were interfaced to a portable computer to allow data logging of physiological parameters.
Arterial and jugular venous blood samples were taken for analysis of oxygen and lactate content, together with an assay of propofol concentrations, 20 min after surgery had started (baseline), at the start of rewarming, when NPT = 32[degree sign]C, when NPT = 36[degree sign]C, and immediately before weaning from CPB (propofol assay only). Glucose and lactate analyses were performed using a YSI 2300 Stat G/L analyzer (YSI Inc., Yellow Springs, OH). Arterial blood gas tension analysis was undertaken using a Radiometer ABL4 (Radiometer Ltd., Copenhagen, Denmark). Propofol assays were performed by the Department of Anesthesia, University of Glasgow, using high-performance liquid chromatography. The LOI was calculated as the ratio of the arterial-jugular venous difference in lactate to the arterial-jugular venous difference in oxygen content.
Based on the variance of previous measurements of Sjo2 , the population of this study was calculated to give an 80% power of detecting a 20% difference between groups with a 5% probability of alpha-type error. Data analysis incorporated analysis of variance for repeated measures, Student's t-test, and the Kruskal-Wallis test, using SPSS-PC (SPSS Inc., Chicago, IL) and EPI-INFO (Centers for Disease Control, Atlanta, GA). Because the duration and magnitude of desaturation may be important in terms of cerebral hypoperfusion, the area of the SjO2 <50% curve was collapsed into a summary measure by summing the degree of desaturation below 50% for each monitored minute to produce an integrated total desaturation for each patient. Data are presented as the mean +/- SD unless stated otherwise. A 5% level of significance was used.
There were no significant differences in demographic factors between the groups, nor were there any significant differences in perioperative variables (Table 1). The rate of rewarming was 0.28[degree sign]C +/- 0.06[degree sign]C per minute.
SjO2 values measured using the bench oximeter were below 50% in only 11 of 90 blood samples obtained during the rewarming period. There were no significant differences in SjO2 between the groups at any of the predefined temperature points (Table 2). The propofol group had five desaturations below 50%, with an average desaturation to 47.3% +/- 1.8% compared with six desaturations with an average desaturation to 46% +/- 1.6% in the control group (P = 0.71). The mean change (over rewarming) in SjO2 for the propofol group was -12.2% +/- 7.3%, and for the control group, the mean change was -12.3% +/- 6.0%.
Using fiberoptic oximetry, 122 SjO2 desaturations below 50% were observed, with an average of 5 per patient. There were no significant differences between the groups in the frequency or duration of the desaturations (Table 3). The mean total duration of desaturations per patient was 18 +/- 22.6 min, and the mean duration of a single desaturation was 4 min (range 1-26 min). There was no significant difference in the integrated total desaturations (despite transformation of the data). Desaturations lasted 5 min or longer in 27 observations, thereby excluding artifact , and no significant difference was found between groups using this criterion (P = 0.4). There were no significant differences in the mean minimal observed values of SjO2.
Most SjO2 reductions (n = 115) were associated contemporaneously with decreases in MAP. There were 245 reductions in MAP below 50 mm Hg, 47 during the rewarming period (mean 8.4 episodes per patient). There were no significant differences in the mean minimal MAP value between the groups. Propofol administration was associated with an increased incidence of reductions in MAP below 50 mm Hg (P = 0.023) and an increased requirement for vasoconstrictive drugs (P = 0.025). The odds ratio for the occurrence of MAP reduction <50 mm Hg if receiving propofol was 4.63 (P = 0.009, Mantel-Haenszel).
The AVDL across the cerebral circulation became negative in 59 of 120 blood samples. There was no significant difference between comparison groups when examining overall means of AVDLs. However, there was a significant difference between groups both at baseline (P = 0.02) and at the 28[degree sign]C point (P = 0.001), at which times the mean AVDLs were more negative in the propofol than in the control group (Table 2). The difference increased from baseline to 28[degree sign]C and decreased thereafter. LOI was significantly larger in the propofol than the control group. This was most marked at the 28[degree sign]C point (P = 0.001), and the difference decreased to statistical insignificance by 36[degree sign]C.
An average of 648 +/- 120.2 mg of propofol was infused over the period of CPB. Plasma propofol assays during rewarming produced 2.82 +/- 0.92 mg/L. The 32[degree sign]C assay of propofol produced a mean concentration of 3.17 +/- 0.63 mg/L, and the 36[degree sign]C assay had a mean concentration of 3.24 +/- 0.81 mg/L.
Despite using a population size similar to that of other studies reporting reductions in SjO2 [4,5,11], we observed neither the same magnitude nor duration of desaturations. Methodological differences may account for these differences. The size of the temperature differential between the heat exchanger and the venous blood returning to the CPB circuit affects the rate of rewarming. Because the arterial blood from the CPB circuit rewarms the brain faster than the body as a whole, this may result in cerebral hyperthermia . Although the NPT-heat exchanger temperature differential was limited to 10[degree sign]C in this study, it may be that other studies used different thermal gradients. However, it is not possible to be conclusive, because most other studies have not clearly defined the thermal gradients used for rewarming. Differences in CPB flow rate may be another explanation. Although variations in flow rate have little effect on cerebral blood flow during stable hypothermic CPB, this may not be the case during rewarming. Most studies examining SjO2 have used pump flow rates of 2.2-2.4 L [center dot] min-1 [center dot] m-2. However, it is not unusual to vary the flow rate to manipulate MAP, e.g., reducing flow rate to control hypertension. This is unlikely in our study because protocol compliance by our perfusionists was confirmed by monitoring the flow rate.
Regardless of the differences in the duration and magnitude of SjO2 reduction compared with other studies, propofol did not reduce the incidence of cerebral venous desaturation in our study during the rewarming phase of hypothermic CPB. In contrast to our findings, Newman and colleagues  found that propofol reduced CBF, with a consequent reduction in CMRO2. They also found that propofol had no effect on cerebral oxygen extraction while the supply/demand ratio remained constant. Because of the reduction in CBF, they postulated that propofol may reduce the embolic load to the brain and therefore have a cerebral-protective effect. One explanation of the failure of propofol to affect SjO2 in our study could be that an insufficient dose of propofol was used. Indeed, the average concentration of propofol required for burst suppression was less than that reported by Newman and colleagues . However, our propofol levels were obtained from direct plasma assay, rather than from the computed pharmacokinetic estimates reported by Newman et al. . Our use of isoflurane to control hypertension before CPB may have had a residual effect during CPB, possibly reducing the amount of propofol required to produce EEG burst suppression. However, it seems unlikely that there would be sufficient isoflurane remaining at rewarming to have an important effect. Moreover, any effect on our results would have been controlled, as isoflurane was used in both groups. The failure to demonstrate an additive protection of propofol during rewarming is consistent with previous suggestions of a limited effect of pharmacological protection compared with the effect of hypothermia . This is supported by Koorn and colleagues , who report a lack of additive effect between pharmacological drugs and hypothermia in a rat model of ischemic brain injury. This suggests an "either/or" effect of the two modes of cerebral protection, with hypothermia being the more effective.
To be truly effective, pharmacological drugs must not merely reduce aerobic metabolism, they must also reduce anaerobic metabolism. The temperature coefficient (Q10) values for common biological enzyme systems converge around a limited range of values , and hypothermia tends to exert a similar effect on both anaerobic and aerobic systems. Whether this parity of action is also true for pharmacological suppression is unknown. Lactate is normally produced in a very small quantity by the brain, generating a negligible negative AVDL . There was a chance difference between groups at baseline measurement, which may be caused by individual variations in cerebral perfusion after the induction of anesthesia. This effect would most likely be removed with the start of CPB and a normalization of perfusion pressure, together with inclusion of lactate in the CPB prime. In fact, the difference in AVDL in our study became more pronounced over the period of CPB, with the propofol group having more cerebral lactate production. Our findings are paralleled by the effect of propofol on systemic metabolism during hypothermic CPB, during which it significantly reduces systemic oxygen uptake but does not reduce lactate production, which tends to increase . On the basis of our data, propofol would seem not to depress anaerobic pathways of cerebral metabolism during rewarming from hypothermic CPB.
LOI has been used as a measure of the ratio of aerobic to anaerobic metabolism within the brain . In neurotrauma, values of less than 0.03 are accepted as normal, whereas values of 0.08 or greater are deemed to be pathological . We observed levels of LOI beyond both these thresholds, and they were significantly associated with the use of propofol (P < 0.001). The relationship between LOI and propofol is confounded by the associated increased incidence of hypotension, and a causal relationship cannot be concluded. Our strategy to prevent hypotension using bolus doses of methoxamine failed in both groups, probably because methoxamine was only administered once the MAP decreased below 50 mm Hg, and, if the dose was insufficient, it would have been several minutes before an additional dose was given. Although the validity of the LOI thresholds quoted for predicting cerebral tissue infarction have been proven in neurotrauma, these findings may not be applicable to patients undergoing cardiac surgery. However, any change toward anaerobic metabolism indicates a deleterious effect, and our findings suggest that the relationship between LOI and brain injury associated with cardiac surgery merits further investigation.
We thank Abbott Laboratories for their contribution of Opticath catheters.
1. Royston D. Interventions to reduce cerebral injury during cardiac surgery: the effect of physical and pharmacological agents. Perfusion 1989;4:153-61.
2. Newman MF, Murkin JM, Roach G, et al. Cerebral physiologic effects of burst suppression doses of propofol during nonpulsatile cardiopulmonary bypass. Anesth Analg 1995;81:452-7.
3. Croughwell ND, Frasco P, Blumenthal JA, et al. Warming during cardiopulmonary bypass is associated with jugular bulb desaturation. Ann Thorac Surg 1992;53:827-32.
4. Andrews PJD, Colquhoun AD. Detection of cerebral hypoperfusion during cardiopulmonary bypass. Anaesthesia 1994;49:949-53.
5. Nakajima T, Kuro M, Hayashi Y, et al. Clinical evaluation of cerebral oxygen balance during cardiopulmonary bypass: online continuous monitoring of jugular venous oxyhemoglobin saturation. Anesth Analg 1992;74:630-5.
6. Nussmeier NA, Arlund C, Slogoff S. Neuropsychiatric complications after cardiopulmonary bypass: cerebral protection by a barbiturate. Anesthesiology 1986;64:165-70.
7. Weir DL, Laycock GJA, Alston RP. EEG burst suppression induced by propofol during hypothermic cardiopulmonary bypass. J Cardiothorac Anesth 1990;4:106.
8. Fox J, Gelb AW, Enns J, et al. The responsiveness of cerebral blood flow to changes in arterial carbon dioxide is maintained during propofol-nitrous oxide anesthesia in humans. Anesthesiology 1992;77:453-6.
9. Croughwell ND, Newman MF, Blumenthal JA, et al. Jugular bulb saturation and cognitive dysfunction after cardiopulmonary bypass. Ann Thorac Surg 1994;58:1702-8.
10. Souter MJ, Andrews PJD. Validation of the Edslab dual lumen oximetry catheter for continuous monitoring of jugular bulb oxygen saturation after severe head injury. Br J Anaesth 1996;76:744-6.
11. Cook DJ, Oliver WC Jr, Orszulak TA, Daly RC. A prospective, randomized comparison of cerebral venous oxygen saturation during normothermic and hypothermic cardiopulmonary bypass. J Thorac Cardiovasc Surg 1994;107:1020-9.
12. Drummond JC. Brain protection during anesthesia: a reader's guide. Anesthesiology 1993;79:877-80.
13. Michenfelder JD. Anesthesia and the brain: clinical, functional, metabolic and vascular correlates. New York: Churchill Livingstone, 1988.
14. Koorn R, Lesser J, Darmon P, et al. Cerebral protective effects of normal saline, thiopental and etomidate during either normothermia or hypothermia in Wistar-Kyoto rats [abstract]. Anesth Analg 1995;80:S244.
15. Harper JA. Review of physiological chemistry. Los Altos, CA: Lange Medical Publishers, 1973.
16. Robertson CS, Grossman RG, Goodman JC, Narayan RK. The predictive value of cerebral anaerobic metabolism with cerebral infarction after head injury. J Neurosurg 1987;67:361-8.
17. Laycock GJA, Alston RP. Propofol and hypothermic cardiopulmonary bypass: vasodilation and enhanced metabolic suppression. Anaesthesia 1992;47:382-7.
18. 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.