Cardiopulmonary Bypass Temperature, Hematocrit, and Cerebral Oxygen Delivery in Humans

Transcription

1 Cardiopulmonary Bypass Temperature, Hematocrit, and Cerebral Oxygen Delivery in Humans David J. Cook, MD, William C. Oliver, Jr, MD, Thomas A. Orszulak, MD, Richard C. Daly, MD, and Rex D. Bryce, MD Department of Anesthesiology and Section of Cardiothoracic Surgery, Department of Surgery, Mayo Clinic and Foundation, Rochester, Minnesota Background. The neurologic effects of warm heart operations is a subject of popular interest. The purpose of this study was to examine the adequacy of cerebral oxygenation during normothermic cardiopulmonary bypass and better define the relationship between hematocrit, temperature, and brain oxygen delivery. Methods. Cerebral blood flow, metabolic rate, and oxygen delivery were measured in 6 patients randomized to normothermic (37 C) or hypothermic (27 C) cardiopulmonary bypass. The nitrous oxide saturation technique of Kety and Schmidt was used for cerebral blood flow determinations. Both temperature groups underwent moderate (31%) hemodilution. Results. During normothermic cardiopulmonary bypass, cerebral blood flow increased secondary to hemodilution and decreased cerebral vascular resistance; a nor- mal matching of oxygen demand and delivery was maintained. During hypothermic bypass, hemodilution and hypothermia had essentially equal, opposing effects on cerebral vascular resistance and blood flow. With hypothermia, brain oxygen demand and delivery were both reduced but not closely coupled. Conclusions. From the standpoint of global cerebral perfusion and oxygenation, our data support the practice of "warm" heart operations. It clarifies the marked influence of hematocrit on cerebral blood flow and delineates the interaction of temperature and hematocrit on cerebral oxygen delivery. It also suggests that additional investigation to better define "temperature-appropriate" hemodilution during cardiopulmonary bypass is indicated. (Ann Thorac Surg 1995;6:1671-7) he practice of cardiopulmonary bypass (CPB) in T North America has been undergoing change during the past 3 years. Traditionally, CPB has been performed with induced hypothermia (25 to 32 C) because reductions in temperature lower tissue metabolic rate and have myocardial and cerebral protective effects [1]. Recently, many institutions have shifted their CPB practice to higher temperatures because warm cardioplegia with systemic normothermia improves myocardial function after CPB [2]. Although myocardial performance may be enhanced, there are significant concerns about the impact of normothermic CPB on the brain. A recent report [3] suggested that warm CPB may be associated with a higher incidence of neurologic injury than hypothermic CPB. There are also physiologic data that suggest normothermic bypass may be a cerebral stress [4, 5]. Cerebral venous oxygen saturation generally reflects the balance between cerebral oxygen delivery (CDO2) and consumption. Decreases in cerebral venous oxygen saturation have been documented during rewarming from hypothermic CPB [4, 6] and occurs with high frequency during early normothermic CPB [4]. An explanation of these results could be that the higher cerebral oxygen demand of warm bypass may not be coupled to reciprocal increases in CDO 2. Accepted for publication July 13, Address reprint requests to Dr Cook, Department of Anesthesiology, Mayo Clinic, 2 First St SW, Rochester, MN Although measurements of cerebral blood flow (CBF) or cerebral metabolic rate for oxygen (CMRO2) cannot confirm or refute the safety of a temperature management strategy during cardiac operations, the measurement of CBF and CMRO 2 is fundamental to demonstrate that this change in cardiac surgical practice has at least a prima facie physiologic acceptability with respect to the cerebral circulation. In this study, we measure CBF, CMRO2, and CDO 2 in patients undergoing hypothermic and normothermic bypass, our goal being to assess the adequacy of cerebral oxygenation during normothermic CPB and better understand the relationship of temperature, CDO2, and hematocrit during CPB. Material and Methods After Institutional Review Board approval and written informed consent, 6 patients undergoing elective firsttime coronary artery bypass grafting were studied. Patients were prospectively randomized (3 per group) to undergo CPB with either systemic hypothermia (27 C) or systemic normothermia (37 C). Patients with clinical or laboratory evidence of cerebrovascular disease, allergy to radiographic contrast, or insulin-dependent diabetes mellitus were excluded from the study. Anesthesia consisted of fentanyl and midazolam (loading dose, fentanyl 3 t~g/kg and midazolam 1 /zg/kg, followed by an infusion of fentanyl and midazolam,.3 and.4 t~g " kg 1 min 1, respectively). Infusion rate 1995 by The Society of Thoracic Surgeons /95/$9.5 SSDI (95)648-5

2 1672 COOK ET AL Ann Thorac Surg TEMPERATURE, HEMATOCRIT, AND BRAIN OXYGENATION 1995;6: was reduced by 5% with the onset of CPB. Inspired oxygen fraction was maintained between.4 and.7. Arterial, pulmonary artery, and right atrial blood pressures, heart rate, end-tidal carbon dioxide concentrations, and nasopharyngeal temperature were continuously measured. A catheter was placed percutaneously in the right jugular bulb for sampling of cerebral venous blood. Catheter position was confirmed by fluoroscopy in all patients. During CPB, a nonpulsatile pump flow of 2.2 to 2.4 L min 1. m 2 was maintained. Arterial carbon dioxide tension (PaCO2) was adjusted to normocapnic levels (35 to 4 mm Hg) without temperature correction (c~-stat regulation). The bypass pump was primed to maintain a hematocrit of 23% or greater during CPB and a membrane oxygenator was used. Sodium nitroprusside or phenylephrine infusions were used during CPB to maintain a mean arterial blood pressure of 5 to 7 mm Hg. To maintain nasopharyngeal temperature at 27, 32, or 37 C, the perfusate was set at 25 to 26, 32 to 33, and 37 to 39 C, respectively. Cerebral blood flow and the arteriovenous oxygen difference (AVDO 2) were measured during four periods: (1) between sternotomy and aortic cannulation; (2) 3 minutes after the onset of CPB (37 C) in normothermic patients, and when a stable nasopharyngeal temperature of 27 C was achieved in hypothermic patients; (3) 6 minutes after the onset of CPB in normothermic patients, and at 32 C during rewarming in hypothermic patients; and (4) 3 minutes after weaning from CPB. Mean arterial blood pressure, pump flow rate during CPB, and nasopharyngeal temperature were stable for 5 minutes before measurements were made. Measurement of CBF and AVDO 2 allowed subsequent calculation of CMRO 2, CDO2, and the CDO 2 to CMRO~ ratio. Cerebral vascular resistance was calculated as mean arterial blood pressure divided by CBF. Cerebral blood flow was measured using the nitrous oxide (N2) washin technique of Kety and Schmidt [7] according to previously established methods [8]. Ten percent N2 was introduced into the ventilator or oxygenator fresh gas flow with an air-oxygen mixture. Ten paired arterial and jugular bulb venous samples for N2 measurement were taken over the 15-minute saturation period [8]. During CPB, arterial blood was drawn from a shunt taken off the arterial inflow line of the CPB machine rather than the radial artery. The CBF was calculated from arterial and venous saturation curves fit to the measured N2 concentrations and integrated to infinity as follows [7, 8]: 1 ;~ V(t) CBF - - ml- 1g 1. min 1, (1) (a - v)dt f where ;~ is the brain blood solubility coefficient for N2; V(t) is the venous N2 reading at saturation; and f (a - v)dt is the area circumscribed by the difference in the arterial and venous N2 concentration curves. All CBF values were normalized to a PaCO 2 of 37 mm Hg by correcting measured values by 3% for each mm Hg difference in PaCO 2 from 37 mm Hg [9, 1]. This was done so CBF comparisons could be made at equivalent PaCO 2 values both within and between groups. Cerebral Metabolism Measurement The CMRO 2 was determined from the product of the AVDO 2 and the CBF, using the equation: (CBF AVDO2) ml 2 1 g ~. min 1 (2) CMRO2 = 1 Arterial and jugular bulb blood gas tensions and saturations were determined (IL-BGE Analyzer, IL Co-Oximeter; Instrumentation Laboratories Inc, Boston, MA) during each CBF measurement period. Blood gas tensions were measured at 37 C and the arterial and venous blood oxygen tensions were subsequently backcorrected to a blood temperature of 27 C or 32 C when the patient was hypothermic. The formula of Severinghaus was used [8, 11]. The arterial and venous oxygen content (CxO2), CDO 2, and CMRO 2 at 27 C and 32 C were calculated after this temperature correction. Arterial or venous oxygen content (CxO2): CxO 2 ~ 1.34 Hgb(SxO2 +.3 (PxO2). (3) Cerebral oxygen delivery (CDO2): CDO2 ~ CBF (CaO2) ml 2 1 g 1. min-1. (4) Arteriovenous oxygen content difference (AVDO2): AVDO2 - (CaO2 - CvO2) ml dl 1 (5) For both temperature groups, within-group comparisons were performed using repeated-measures analysis of variance followed by Dunnett's test, when indicated. Between-group comparisons for each of the four study periods was performed using a Wilcoxon rank sum test. All data are expressed as mean _+ standard deviation and p value less than.5 was considered significant. Complete sets of measurements for all four study periods were obtained in 3 of 3 normothermic patients. Cerebral blood flow and derived measurements were obtainable in 29 of 3 hypothermic patients during periods III and IV. Results Patient groups did not differ with respect to age, aortic cross-clamp time, or total CPB time. Before CBP, the two groups did not differ with respect to hemodynamic or blood gas values (Table 1), or with regard to any cerebral physiologic measurement (Figs 1-3, Table 2) (p >.5 by rank-sum test). In both groups during CPB mean arterial blood pressure, hemoglobin, and arterial oxygen tension decreased, whereas PaCO 2 increased relative to the control period before CPB. Hemoglobin decreased approximately 31% with the onset of CPB in both groups. In the hypothermic group, nasopharyngeal temperature decreased during bypass, and mean temperature in the normothermic

3 Ann Thorac Surg COOK ET AL ;6: TEMPERATURE, HEMATOCRIT, AND BRAIN OXYGENATION Table 1. Physiologic Variables Describing Both Patient Groups During the Four Study Periods Prebypass Study (Control) Variables Period I Temperature ( C) Mean arterial blood pressure (mm Iqg) Hemoglobin (g/dl) CaO 2 (ml/dl) Cardiac index/pump flow (L. rain i. m z) Glucose (mg/dl) PaO 2 (mm Hg) PaCO2 (mm Hg) Hypothermic Hypothermic Hypothermic HypothermJc Normothermlc Normotherm~c 35.5 ± ±.6 83 ± ± ± ± ± ±.9 13 ± ± ± ± 15 33±3 33±4 Cardiopulmonary Bypass II 11[ Post-CPB 3 min IV a'b 32. ±.2 *~'b a'b 37.2 ±.8 '~ 37.4 ±.6 "~ ± 9 "~ '~ 71 ± 1 '~ "~ "~ a "~ '~ *~ 8.1 ± 1. "~ '~ 8.3 ± 1.2 "~ 11.1 ± 1.3 ~ " "~ ~ " 11.6 ± 1.4 "~ b ± ± ± ± 39 '~ 155 ± 38 ~'b '~ 136 ± ~ 53 b 222 ± "~ a'b 38 ± 4" ~ 39 ± 3" ~ *' Identifies change from before cardiopulmonary bypass with groups, repeated measures analysis of variance followed by" Dunnett's test (p <.5). b Identifies between-group differences by Wilcoxon rank sum test (p <.5). Values are mean ± standard deviation; n - 3 except periods Ill and IV in hypothermic group, where n 29. CaO 2 arterial oxygen content; PaO 2 (PaCO2) arterial oxygen (carbon dioxide) tension. group temperature increased. Because the PaCO 2 varied within and between the groups (Table 1), all reported CBF values are normalized to a PaCO 2 of 37 mm Hg so direct comparisons could be made. In hypothermic (27 C) patients, CBF and cerebral vascular resistance did not differ from the value before CPB. Although, as body temperature declined with hypother- mia, CMI~O 2 decreased (Figs 1, 2). During CPB with hemodilution at 27 C, arterial oxygen content and CDO 2 ~6 were significantly decreased relative to the period before CPB (Fig 3). However, the ratio of CDO 2 to CMRO 2 was increased (Fig 4). During period III (32 C rewarming in the hypothermic group), CBF remained unchanged from control, whereas the CMRO 2 at 32 C remained lower than the value before CPB (Figs 1, 2). Cerebral vascular resistance decreased with rewarming (Table 2). Cerebral oxygen delivery remained less at 32 C than before CPB (Fig 3) and the CDO 2 to CMRO 2 ratio at 32 C remained higher than the control value (Table 2). Thirty minutes after CPB (period IV), the hypothermic group demonstrated large increases in CBF relative to control or bypass values and large reductions in cerebral *e- 45 3O U. 15 en Pre-CPB CPB II CPB III Post-CPB I~ Hypothermic ].E E 4 ~" 3 O 2 i Fig 1. Cerebral blood flow (CBF) (ml loo g i. rain ~) during the four study periods. Hypothermic group (~), normothermic,croup (I); n 3 in all periods except periods III and IV qf hypothermic group, where n = 29. *ldent(fies change from before cardiopulmonary bypass (CPB) within groups by repeated-measures analysis of variance followed by Dunnett's test (p ~:.5). "Identifies ber~eengroup di~:erences by Wilcoxon rank sum test (p <.1). Values are mean ~ standard deviation. Pre-CPB CPB II CPB III Post-CPB.yo,he,m.c..o mo,hermio l Fig 2. Cerebral metabolic rate for oxygen (CMRO2) (ml 2 ' 1 g J min ~) during the four study periods. Hypothermic group (E2), nonnothermic group (m). Statistical analysis was as in Fig 1. (CPB - cardiopulmonary bypass.)

4 1674 COOK ET AL Ann Thorac Surg TEMPERATURE, HEMATOCRIT, AND BRAIN OXYGENATION 1995;6: E E E 1:1 6 4 o,, # #,,'" "t ~.oo, f= [Z 8 6 e,l 4 o 2 2 II III IV Study Period Fig 3. Cerebral oxygen delivery (CDO 2) during the roar study periods. Hypothermic group (dashed line), normothermic group (solid line). Statistical analysis was as in Fig CMRO2 (ml. 1 g-1. rain-l) Fig 4. The cerebral oxygen delive~ to cerebral metabolic rate of oxygen ratio (CDO2:CMRO2) plotted against cerebral metabolic rate for oxygen is 238 points. All values for both groups, all study periods are depicted. The cerebral metabolic rate for oxygen is in ml. loo g i. min vascular resistance (Fig 1, Table 2). This occurred although the CMRO2 after CPB was not different from the CMRO 2 before CPB (Fig 2). The CDO 2 in the period after CPB also did not differ from the control value (Fig 3), nor did the CDO 2 to CMRO 2 ratio (Table 2). In the normothermic group at 3 minutes of bypass (period II), CBF increased relative to the period before CPB as cerebral vascular resistance decreased (Table 2). The CMRO2 was unchanged (Figs 1 and 2). As in the cold group, arterial oxygen content decreased with bypass hemodilution, but with normothermia CDO 2 was unchanged from control (Figs 3, 4). The ratio of CDO 2 to metabolic rate was also unchanged (Table 2). During period III in the normothermic group (6 minutes on CPB), the CBF remained higher, and cerebral vascular resistance lower than control whereas the CMRO 2 continued unchanged (Figs 1, 2), the CDO 2 and the CDO 2 to CMRO 2 ratio were also unchanged (Figs 3, 4; Table 2). After CPB, CBF and cerebral vascular resistance in the warm group remained greater than control, whereas the CMRO 2 remained unchanged (Figs 1, 2). Likewise, the CDO 2 and the CDO 2 to CMRO 2 ratio remained unchanged from before CPB (Table 2). When the two groups were compared, they did not differ before bypass with respect to any cerebral physiologic value. During the first CPB period (period II), the "warm" and "cold" bypass groups differed with regard to temperature and all cerebral physiologic variables, although mean arterial blood pressure, PaCO2, and hemoglobin did not differ (Table 1). Cerebral blood flow, metabolic rate, and oxygen delivery were all higher in the normothermic group (Figs 1-3) and cerebral vascular Table 2. Cerebral Physiologic Values Describing Both Patient Groups During the Four Study Periods Prebypass Cardiopulmonary Bypass Study (Control) Variables Period I li IIl CBF(mL.1g '-min ') CMRO 2 (ml O 2 1 g 1. rain i) CVR (MABP/CBF) CDO 2 (ml 2" 1 g ' rain ') CDO 2 to CMRO 2 ratio Post-CPB 3 min IV Hypothermic " ~ "~ " Hypothermic ±.3 a'b 1.6 _+.6 a'b = Hypothermic b ~'b 1.4 _+.5 '~ ±.4 "~ ~ 1.4 ±.5 ~ Hypothermic a'b a'b 6.2 _ ± 2.1 Hypothermic ±.8 *~'b ~ 2.6 ± ± ± ± " Identifies change from before cardiopulmonarv bypass with groups, repeated measures analysis of variance followed by Dunnett's test (p <,5). b Identifies between-group differences b~' Wilcoxon rank sum test (p <:.5). Values are mean + standard deviation; n 3 except periods 111 and IV in hypothermic group, where n 29. CBF : cerebral blood flow; CDO, cerebral oxygen delivery; CMRO2 cerebral metabolic rate for oxygen; CVR - cerebral vascular resistance,

5 Ann Thorac Surg COOK ET AL ;6: TEMPERATURE, HEMATOCRIT, AND BRAIN OXYGENATION resistance was lower. The ratio of CDO 2 to demand was significantly higher in the cold group (Table 2). All cerebral physiologic measurements between groups differed at a level of p ~.1 by Wilcoxon rank-sum test. During the second CPB period (period III), the normothermic and hypothermic groups differed with respect to temperature and each of the cerebral physiologic variables while the hemodynamic values, hemoglobin, and PaCO 2 again did not differ between groups (Table 1). The CBF, CMRO2, and CDO 2 were all higher in the warm group, whereas cerebral vascular resistance was lower (Figs 1-3, Table 2) (p ~.5 by Wilcoxon rank-sum test). During this period, the CDO2 to CMRO 2 ratio did not differ between groups (Table 2). After CPB, the hypothermic and normothermic groups differed only with respect to two variables. Thirty minutes after CPB the normothermic group had a higher nasopharyngeal temperature than the hypothermic group, and the hypothermic group had a higher arterial glucose concentration (Table 1). Formal neurologic testing was not part of our study design; however, no patient in either temperature group demonstrated a focal neurologic deficit postoperatively. There was no perioperative morbidity. Comment There are three major findings of this study. First, global cerebral oxygenation is well maintained during warm bypass. During normothermia, an increased CBF maintains CDO 2 at the level before CPB despite bypass hemodilution and a close coupling of oxygen demand and delivery exists. Second, during hypothermic bypass, the coupling of CMRO 2 and CBF or CDO 2 is not maintained. Because of hemodilution, CBF is unchanged during hypothermia in spite of the large decrease in metabolic rate. Cerebral oxygen delivery is decreased during hypothermia; however, CMRO2 undergoes a proportionately greater decrease than does CDO2, and the ratio of CDO 2 to CMRO 2 rises significantly (Fig 4). Third, after CPB both groups demonstrate high CBF and low cerebral vascular resistance, although CMRO 2 is unchanged. This is presumably a function of persistent hemodilution. Cerebral oxygen delivery after CPB remains at control levels in the warm CPB group and returns to control levels in the cold CPB group. After CPB, cerebral oxygen supply and demand are closely matched. If global oxygen delivery is adequate during warm bypass, we are left to explain reports of low cerebral venous oxygen saturation during rewarming from hypothermia [4, 6] and during the early phase of normothermic CPB [6]. With rewarming, low cerebral venous oxygen saturation values may result from brain hyperthermia [12]. Measurements of brain oxygenation during stable temperatures may not be predictive of physiology when the brain is hyperthermic or when its temperature is actively changing, therefore, reports of low cerebral venous oxygen saturation values during rewarming are not inconsistent with our results. Similarly, initiation of normothermic CPB may be associated with an oxygenation stress [4, 5]. The transition to CPB is associated with hemodilution, a fall in mean arterial blood pressure, and with normothermic CPB, a rise in temperature. Aortic manipulation in this period probably also results in a cerebral embolic load. These acute changes may account for the low cerebral venous oxygen saturation values that we reported previously during the initial phase of warm CPB [4]. In this investigation, cerebral physiologic measurements were made after this transitional period of instability. One might speculate that the higher CBF associated with warm bypass may predispose the cerebral circulation to a higher embolic load and a worsened neurophysiologic status when ischemia does occur. However, with normothermic CPB and hemodilution, blood flow to all organs is increased so it is unclear if a greater proportion of emboli will in fact be delivered to the central nervous system. This has yet to be documented. In addition, in this and a previous study [4], our temperature groups did not differ in neurologic outcome and larger prospective studies on neurologic outcome and bypass temperature have generated conflicting results [3, 13]. On the basis of presented data, we can only conclude that global cerebral oxygenation is well maintained with warm CPB. This study differs from previous investigations on the relationship between temperature and cerebral physiology during human CPB in several ways. First, our CBF method, the Kety-Schmidt technique, differs from the xenon-133 washout technique that is typically used in human CPB studies [14]. The CBF and CMRO2 values we report using this technique are significantly higher than those reported with xenon-133 washout during bypass in humans [14] and our results more closely approximate predicted values [15, 16] and those obtained with a variety of techniques in humans [17, 18] and animal models [19, 2]. Second, our CBF results under hypothermic conditions differ from results obtained during hypothermia under nonbypass conditions [21]. We document an appropriate temperature-dependent decrease in CMRO 2 but this was not associated with a decrease in CBF. Under nonbypass conditions, other investigators have demonstrated that a decrease in CMRO 2 with hypothermia is associated with a decrease in CBF [21]. However, those non-cpb studies differ from our own because of the absence of hemodilution. In this report, the hematocrit decreased approximately 31% with the onset of CPB (Table 1). Hemodilution increases CBF [22] secondary to a decrease in cerebral vascular resistance. Hemodilution and reduced temperature had offsetting effects on cerebral vascular resistance during cold bypass, therefore no change in CBF was seen. This effect of hemodilution on cerebral vascular resistance and CBF is evident in the normothermic group where hemodilution without temperature change occurs. In the normothermic group, a 3% decrease in hematocrit was associated with a 48% decrease in cerebral vascular resistance and a 43% increase in CBF. Therefore, if no hemodilution occurred during hypothermic bypass, our data suggest that hypothermia would have resulted in

6 1676 COOK ET AL Ann Thorac Surg TEMPERATURE, HEMATOCR1T, AND BRAIN OXYGENATION 1995;6: approximately a 5% decrease in CBF at 27 C. This is consistent with investigations under hypothermic conditions without hemodilution [21]. The results of three other CPB investigations are supportive of this interpretation. Hindman and colleagues [19] used microspheres to determine CBF in rabbits during hypothermic (27 C) CPB with moderate hemodilution and documented CBF values similar to ours. Schwartz and associates [23] obtained similar results using internal carotid xenon-133 injection in baboons during hypothermic (32 C) CPB, and Stephan and colleagues [24] in a clinical study using the Kety-Schmidt technique during hypothermic (26 C) CPB with moderate hemodilution obtained results nearly identical to our own. This report also differs from many other cerebral physiologic studies because CDO 2 is emphasized together with changes in CBF and CMRO 2. Typically, the CBF to CMRO 2 ratio is reported, with deviations from established values of 15 to 2 [14], being suggestive of pathophysiology. However, in the context of CPB, this relationship can be misleading because CPB hemodilution has a profound effect on CBF and is not reflected adequately in the CBF to CMRO 2 ratio [251. Hemodilution and decreasing cerebral vascular resistance explains the maintenance of CBF during cold bypass as well as the increase in CBF that occurs during warm bypass and in the period after CPB. Each of these results would be unexpected if one looked only at the predicted relationship between CMRO 2 and CBF. Because bypass results in relatively dramatic changes in hematocrit, the importance of CDO 2, not simply CBF determinations, becomes evident. However, this is not sufficient to say that the brain primarily regulates CDO 2 rather than CBF. During warm bypass, CBF rises with hemodilution and CDO 2 is maintained at control levels, although during hypothermia neither CBF nor CDO 2 appear to be coupled to brain oxygen demand. In Figure 4, the CDO 2 to CMRO 2 ratio is plotted against CMRO 2 for the 238 measurements obtained in this study. As CMRO 2 is decreased with hypothermia, the CDO 2 to CMRO 2 ratio rises in an approximately logarithmic fashion. Both flow and oxygen delivery increase relative to CMRO 2 as temperature is reduced. These data argue against either a normal coupling of CDO 2 and CMRO 2 (or CBF and CMRO2) during hypothermia. Similar conclusions can be drawn from the results of Michenfelder and Milde [21 who documented increasingly high CBF values relative to metabolic rate at equivalent and more extreme levels of CPB hypothermia in dogs. It is unclear if this is a result of a disturbance in a cerebral regulatory process or a function of changing biophysical characteristics of blood at reduced temperature. Regardless of the mechanism, these results provide potential insights into management of hemodilution during hypothermic CPB. During hypothermia, CDO 2 greatly exceeds oxygen demand, even with moderate hemodilution. Our data suggest that hemoglobin concentrations approaching 5 g/dl should be well tolerated at 27 C. A narrower margin for hematocrit will exist at normothermia as the compensatory increase in flow that can occur with hemodilution is limited. Therefore, hematocrit may need to be actively manipulated as temperature changes with cooling and rewarming. Further definition of "temperature-appropriate hemodilution" during CPB could have a significant impact on our practice. This report supports the safety of normothermic bypass with respect to global cerebral oxygen supply and demand. During "warm" bypass, oxygen demand and delivery are closely coupled. It also demonstrates a significant excess of CBF and oxygen delivery relative to demand during hypothermia and indicates that flowmetabolism coupling is not well maintained under this condition. The results also suggest that greater degrees of hemodilution should be well tolerated during hypothermic CPB and that hyperemia after CPB is probably an expected consequence of hemodilution. Finally, the relevance of hematocrit and CDO 2 calculations to studies of this type is emphasized. This study was supported by the American Heart Association-- Minnesota Affiliate, Bentley Division of Baxter Healthcare Corporation, and Mayo Foundation. References 1. Kirklin JW, Barratt-Boyes BG. Cardiac surgery, 2nd ed. New York: Churchill Livingstone, 1992: Christakis GT, Koch JP, Deemar KA, et al. A randomized study of the systemic effects of warm heart surgery. Ann Thorac Surg 1992;54: Martin TD, Craver JM, Gott JP, et al. Prospective, randomized trial of retrograde warm blood cardioplegia: myocardial benefit and neurologic threat. Ann Thorac Surg 1994;57: 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;17: Cook DJ, Sharbrough FW, Oliver WC Jr, Orszulak TA, Daly RC. Electroencephalographic (EEG) changes during cardiopulmonary bypass (CPB) with systemic normothermia [Abstract]. Anesthesiology 1994;81:A Croughwell ND, Frasco P, Blumenthal JA, Leone BJ, White WD, Reves JG. Warming during cardiopulmonary bypass is associated with jugular bulb desaturation. Ann Thorac Surg 1992;53: Kety SS, Schmidt CF. The determination of cerebral blood flow in man by the use of nitrous oxide in low concentrations. Am J Physiol 1945;143: Cook DJ, Anderson RE, Michenfelder JD, et al. Cerebral blood flow during cardiac operations: comparison of Kety- Schmidt and xenon-133 clearance methods. Ann Thorac Surg 1995;59: Forster A, Juge O, Morel D. Effects of midazolam on cerebral hemodynamics and cerebral vasomotor responsiveness to carbon dioxide. J Cereb Blood Flow Metab 1983;3: Kety SS, Schmidt CF. The effects of altered arterial tensions of carbon dioxide and oxygen on cerebral blood flow and cerebral oxygen consumption of normal young men. J Clin Invest 1948;27: Severinghaus JW. Simple, accurate equations for human blood O~ dissociation computations. J App1 Physiol 1979;46:

7 Ann Thorac Surg COOK ET AL ;6: TEMPERATURE, HEMATOCRIT, AND BRAIN OXYGENATION 12. Cook DJ, Orszulak TA, Daly RC, Buda DA. Cerebral hyperthermia during cardiopulmonary bypass in adults. J Thorac Cardiovasc Surg (in press) 13. Randomized trial of normothermic versus hypothermic coronary bypass surgery. The Warm Heart Investigators. Lancet 1994;343: Schell RM, Kern FH, Greeley WJ, et al. Cerebral blood flow and metabolism during cardiopulmonary bypass. Anesth Analg 1993;76: Prough DS, Rogers AT. What are the normal levels of cerebral blood flow and cerebral oxygen consumption during cardiopulmonary bypass in humans? Anesth Analg 1993;76: Lassen NA. Normal average value of cerebral blood flow in younger adults is 5 ml/1 g/rain. J Cereb Blood Flow Metab 1985;5: Ingvar DH, Cronqvist S, Ekberg R, Risberg J, Hoedt- Rasmussen K. Normal values of regional cerebral blood flow in man, including flow and weight estimates of gray and white matter. Acta Neurol Scand 1965;41: Powers WJ, Grubb RL Jr, Darriet D, Raichle ME. Cerebral blood flow and cerebral metabolic rate of oxygen requirements for cerebral function and viability in humans. J Cereb Blood Flow Metab 1985;5: Hindman BJ, Dexter F, Cutkomp J, Smith T, Tinker JH. Hypothermic acid-base management does not affect cerebral metabolic rate for oxygen at 27 C. Anesthesiology 1993;79: Michenfelder JD, Milde JH. The relationship among canine brain temperature, metabolism, and function during hypothermia. Anesthesiology 1991;75: Busija DW, Leffier CW. Hypothermia reduces cerebral metabolic rate and cerebral blood flow in newborn pigs. Am J Physiol 1987;253:H Hudak ML, Koehler RC, Rosenberg AA, Traystman J, Jones MD Jr. Effect of hematocrit on cerebral blood flow. Am J Physiol 1986;251:H Schwartz AE, Kaplon KJ, Young WL, Sistino JJ, Kwiatkowski P, Michler RE. Cerebral blood flow during low-flow hypothermic cardiopulmonary bypass in baboons. Anesthesiology 1994;81: Stephan H, Weyland A, Kazmaier S, Henze T, Menck S, Sonntag H. Acid-base management during hypothermic cardiopulmonary bypass does not affect cerebral metabolism but does affect blood flow and neurological outcome. Br J Anaesth 1992;69: Cook DJ. The CBF-CMRO2 ratio is a misleading concept during cardiopulmonary bypass (CPB) in adults [Abstract]. Anesth Analg 1995;8:SCA69. INVITED COMMENTARY This carefully performed study is from an experienced group in the field of the brain and cardiac surgery. It is important to not over-interpret the results of this excellent study. This study related temperature, cerebral blood flow, and oxygen consumption. A much broader issue we all face as cardiac surgeons is brain protection during operation. If one accepts the well-documented fact that there is cerebral ischemia produced (mainly from microemboli) during all cardiac operations involving cardiopulmonary bypass, then the question that needs to be answered is what is the most appropriate temperature for cardiopulmonary bypass. This article in no way answers this question. It is important for the reader to carefully read the conclusion that "From the standpoint of global cere- bral perfusion and oxygenation...", normothermic cardiopulmonary bypass was not deleterious. Because there was no careful neurologic examination of these patients and no detailed psychometric testing in this very small group of patients, no conclusion can be made as to the clinical usefulness of normothermic bypass. One should keep in mind the deleterious neurologic results that were found in association with normothermic bypass in the author's reference number three. Julie A. Swain, MD 638 8th Ave Suite 237 Kenosha, WI 53143