Clinical Study of Normothermic Cardiopulmonary Bypass in 100 Patients With Coronary Artery Disease

Transcription

1 Clinical Study of Normothermic Cardiopulmonary Bypass in 100 Patients With Coronary Artery Disease Ulrik Hvass, MD, and Jean-Pol Depoix, MD Departments of Cardiovascular Surgery and Anesthesiology, H6pital Bichat, Paris, France During normothermic cardiopulmonary bypass (CPB), the body temperature is maintained at 37 C. Since 1987, it has been our standard practice to use normothermic CPB in our patients undergoing a cardiac operation, and our experience now consists of more than 3,000 consecutive patients. Myocardial protection is achieved through the combination of cold intermittent antegrade blood cardioplegia, no topical cooling, and a terminal "hot shot" of blood cardioplegia. We disagree with the stance of the Toronto group that normothermic CPB requires the administration of large volumes of cardioplegic and crystalloid solutions and the frequent use of phenylephrine hydrochloride to ensure a low systemic vascular resistance. To establish a routine technique of cold heartwarm body bypass, we conducted a clinical study in 100 consecutive patients with coronary artery disease. We found that the total cardioplegia volume needed in our patients was 1, ml, versus 4,700 -" 1,900 ml in the Toronto study, and an additional crystalloid volume loading of 400 ± 141 ml during CPB was needed in 26% of our patients, versus a total volume of 3,650 ± 800 ml in the Toronto series. Phenylephrine (250/~g) was used in 16% of our patients, versus 88% of the patients in the Toronto study (mean dose, 1.3 mg). During normothermic CPB, the mean radial arterial pressure was 57.3 ± 9.4 mm Hg. In our experience, normothermic CPB in combination with cold intermittent cardioplegia and a terminal hot shot has the following advantages: it facilitates the conduct of surgical procedures because it produces a clear operative field, it is easy for the perfusionist to manage, and it is associated with no adverse side effects. (Ann Thorac Surg 1995;59:46-51) here is growing interest in the performance of warm T heart surgical procedures since the introduction of the normothermic technique by Salerno [1] and Lichtenstein [2] and their colleagues. Acceptance of this technique among surgeons has grown despite insufficient data concerning the systemic effects of warm heart procedures. Continuing debate now seems to be even more focused on the accepted standard of the cardiopulmonary bypass (CPB) technique itself: should CPB also be normothermic, should the temperature just be allowed to drift to 33 C during the procedure, or should it be deliberately kept hypothermic? The innocuousness of normothermic CPB has been questioned. The work of Christakis and associates [3] pointed up the importance of fluid overload and the significantly higher phenylephrine hydrochloride requirements during and after CPB, with the same requirements during the first 3 postoperative hours. Other complaints voiced by discussants of their paper include the greater risks of pancreatitis, acute tubular necrosis, jaundice, and multiorgan failure. Stone heart and myocardial necrosis in hypertrophied hearts have also been reported (Fifth European Meeting of the Society of Cardiopulmonary Bypass. Aries, France, June 1993). Our experience with normothermic CPB in combination with cold heart myocardial preservation in more Accepted for publication June 10, Address reprint requests to Dr Hvass, Chirurgie Cardiaque, H6pital Bichat, 46 rue Henri Huchard, Paris 75877, France. than 3,000 patients has given us a different insight into the complications attributed to normothermic CPB. The intention of this study was to compare our prospective data with those data reported by Christakis and associates [3]. Material and Methods A clinical study of the systemic effects of normothermic CPB was started in January 1993, and the patient population consists of 100 consecutive patients undergoing coronary artery bypass grafting. Patients undergoing any additional cardiac or noncardiac procedures were excluded. Patients with chronic renal failure being treated with dialysis were also excluded. Data Collection Data regarding volume requirements were collected in the operating room during and after bypass. Hemodynamic variables (pressures, calculated systemic vascular resistance (SVR = [mean arterial pressure - right atrial pressure/cardiac output] 80), flow, and venous saturation) were determined and recorded at 10-minute intervals throughout the procedure. Results were expressed as the mean z the standard deviation. Data were also collected in the intensive care unit regarding the initial SVR, hematocrit value, renal function, the maximum myocardial bound creatine kinase (from blood sampled every 4 hours for 24 hours), the appearance of new Q waves, and the duration of assisted ventilation by The Society of Thoracic Surgeons [95/$ (94)00611-A

2 Ann Thorac Surg HVASS AND DEPOIX ;59:46-51 NORMOTHERMIC BYPASS Anesthesia and Cardiopulmonary Bypass Technique Anesthesia was induced with high doses of fentanyl (100 ~g/kg) and pancuronium bromide (0.1 mg/kg). At the onset of bypass, midazolam (1 to 5 rag) and fentanyl were administered to promote and maintain anesthesia. No gas anesthetic agents were used. Monitoring with a Swan-Ganz thermodilution catheter and a radial artery pressure catheter was performed in all patients. The central core temperature was monitored continuously with an endourinary bladder catheter (Hi-Lo Temp Foley Catheter; Mallinckrodt, St. Louis, MO). Patients were continuously warmed to maintain a central body temperature of 37 ± 0.5 C. Ascending aortic and two-stage single venous cannulation techniques were employed. Only membrane oxygenators were used. Moderate hemodilution was accomplished with a pump prime (total volume, 1,400 ml) consisting of the following: 600 ml of Ringer's lactate, 500 ml of a colloid solution (Plasmion), 300 ml of bicarbonate, and 6,000 IU of heparin. A pump pressure of 35 to 90 mm Hg was considered adequate with a flow rate of 2.4 L omin a.m-2. When the pump pressure decreased below 35 mm Hg, a bolus of phenylephrine (250 /xg) was administered. When the pump pressure exceeded 90 mm Hg, a bolus of nicardipine (1 rag) was administered. After initiating CPB and cross-clamping the aorta, all patients received cold antegrade blood cardioplegia; this was administered every 30 minutes, or earlier in the event of ventricular activity. After discontinuation of CPB, all patients received blood from the Cell Saver III (Haemonetics, Braintree, MA), which recovered shed blood from the operating field and from the residual volume left in the oxygenator and circuit at the end of the procedure. Cardioplegia Techniques The blood cardioplegic solution was composed of equal volumes of St. Thomas' solution (magnesium, 16 retool/l; sodium, 147 mmol/l; potassium, 20 mmol/l; calcium, 2 mmol/l; and chloride, 203 mmol/l, prepared by Aguettant Laboratory, Lyon, France) and oxygenated blood obtained from the CPB circuit after the onset of bypass. Cardioplegia was delivered through a coil cooled in ice (CCAS-R; Bentley Laboratories, Irvine, CA). The temperature at the pump was 4.9 ± 0.96 C and the temperature at delivery was 6.5 ± 0.6 C. The mean values of the various characteristics of the cardioplegia at delivery were as follows: hematocrit, 11% ± 0.02%; potassium, 19.6 ± 0.8 mrnol/l; oxygen tension, kpa; and ph, 7.42 ± All proximal anastomoses were performed before aortic cross-clamping was done using a partial occluding clamp. After aortic cross-clamping, ml of cold cardioplegic solution was infused through the ascending aorta. The infusion pressure and myocardial temperatures were not measured. No topical cooling was used. Cold blood cardioplegia (300 ml) was administered every 30 minutes, or earlier if myocardial activity occurred. Before the release of the aortic cross-clamp, 1 L of warm antegrade blood cardioplegia was infused at a rate of 200 ml per minute for 5 minutes. The composition of this warm blood shot was the same as that of the cold blood cardioplegia. Results Patients Patients were predominantly male (84%) and the mean age was 62 ± 10.6 years. The distribution of the coronary lesions was as follows: left main coronary artery, 21%; three-vessel disease, 46%; two-vessel disease, 23%; and one-vessel disease, 10%. Thirty-five percent of the patient had suffered a myocardial infarction. The left internal mammary artery was used in 98% of the patients. Two arterial grafts (mammary arteries) were used in 31% of the patients; three-artery anastomosis was performed sequentially in 7%. Veins were used as the only grafts in 2% of the patients. Six percent of the procedures were redo operations. Spontaneous defibrillation occurred as a rule (94%) after the aortic cross-clamp was removed despite episodes of recurrent electrical activity during cross-clamping in 52% of the patients. One patient was reoperated on for the control of bleeding. There were no instances of mediastinitis, sternal instability, or postoperative phrenic palsy. Volume Requirements During and After Cardiopulmonary Bypass The patients received ml of cold blood cardioplegia and 1,000 ml (delivered over 5 minutes at a rate of 200 ml/min) of warm blood cardioplegia before aortic unclamping (which corresponds to a crystalloid loading dose of 1 L). No additional volumes were added during CPB in 66% of the patients. In 18%, 250 ± 353 ml of homologous banked blood was added to the pump. In 26% of the patients, an additional 400 ± 141 ml of crystalloid was needed. After discontinuation of CPB, all the patients received 758 _+ 158 ml of blood from the Cell Saver, and this blood had a hematocrit of 56% + 8%. After discontinuation of bypass, all patients received a drip infusion of 300 to 500 ml of colloid. In 5% of the patients, 250 ml of banked blood was added. Hemodynamic Variables PERFUSION PRESSURES. The range of perfusion pressures is depicted in Figure 1 and summarized in Table 1. The mean highest pressure recorded was 77 _+ 12 rnm Hg and the mean lowest pressure was 44.6 _+ 8.6 mm Hg. In 61% of the patients, at some time during CPB the perfusion pressure was less than 50 mm Hg. The mean overall perfusion pressure was mm Hg. FLOW RATES. The range of flow rates is shown in Figure 2 and summarized in Table 1. The mean highest was 5.15 ± 0.64 L/rain and the mean lowest was L/min. The mean overall pump flow rate was 4.55 ± 0.61 L/rain and closely approximates the mean calculated flow rate of L/rain noted for the entire group of patients. SYSTEMIC VASCULAR RESISTANCE. The range of SVRs is shown in Figure 3 and given in Table 1. The mean highest

3 48 HVASS AND DEPOIX Ann Thorac Surg NORMOTHERMIC BYPASS 1995;59:46-51 == E E G) i == _-Ji 20 0 P i i q Fig 1. Perfusion pressure, showing highest and lowest values. SVR was 1, dynes, s 1. cm-5. The mean lowest SVR was dynes.s 1. cm-5. The mean SVR throughout CPB was 1, dynes, s -~ cm 5. Upon arrival in the intensive care unit, the mean SVR was 1, dynes, s -1. cm -5. Biology The mean hematocrit value was 23.5% _+ 3.6% after the first administration of cold blood cardioplegia and 25 % + 3.7% at the end of CPB. The rise in the lower values (15% to 21%) reflects the addition of blood products: 18% of the patients received a mean of ml of banked blood. At their arrival in the intensive care unit, the patients' mean hematocrit value was 32.7 _+ 4.1%. The increase in the hematocrit values was due to the transfusion of blood from the Cell Saver. Only 1 patient received furosemide in the operating room. The mean potassium level at the end of CPB was mmol/l. There was a high percentage of patients who experienced spontaneous defibrillation after the warm blood shot (94%), even though 52% had exhibited intermittent electrical activity during complete aortic cross-clamping. The biologic data are summarized in Table 2. Medications With regard to the use of vasopressors during CPB, a 250-/~g bolus of phenylephrine chloride was adminis- Table 1. Hemodynamics During Cardiopulmonary Bypass Highest Lowest Hemodynamic Variable Value Value Perfusion pressure (rnrn Hg) Mean + SD 72.5 _ _ Range Flow rates (L/rnin) Mean + SD 5.5 _ _ Range Systemic vascular resistance (dyne. s -1. crn -s) Mean _+ SD 2,160 _+ 2,036 1,033 +_ 876 Range 720-3, ,653 Venous saturation (%) Mean _+ SD Range SD = standard deviation. I I I J I O0 Fig 2. Pump flow, showing highest and lowest values. tered in 16% of the patients and 2% of the patients required a second dose to obtain pump pressures greater than 35 mm Hg. No patients received adrenaline or any other vasopressor during bypass. With regard to the use of vasodilators during CPB, 20% of the patients received nicardipine (0.5 to 2 rag) to lower the pump pressure during bypass when it exceeded 90 mm Hg. Six percent of the patients received both hypotensive and hypertensive drugs. In 67% of the patients, a dopamine drip delivered at a rate of 5/~g. kg -1. min -1 was started at the discontinuation of bypass. This is the protocol we follow to achieve adequate sinus rhythm, and is more efficient than ventricular pacing. Indicators of Adequate Perfusion Adequate perfusion was evaluated in the first 20 patients. The mean arteriovenous oxygen difference was 2.54 _ and the mean lactate level was mmol/l at the beginning of CPB and mmol/l during CPB. The higher initial lactate values are attributed to the 600 ml of Ringer's lactate crystalloid solution contained in the priming volumes. Acidosis was not encountered, with a ph of 7.45 _ recorded on the arterial catheter. The venous saturation during CPB was 75% _+ 4.8% at the lowest flow rates and 78% +_ 5% at the highest flow rates. Saturations were continuously monitored on the venous catheter. When the venous saturation tended to be less than 70%, an increase in the flow rate would restore the saturation to a normal level. E o 2ooo * 1500 T ~ g O Fig 3. Systemic vascular resistanc~ showing highest and lowest values.

4 Ann Thorac Surg HVASS AND DEPOIX ;59:46-51 NORMOTHERMIC BYPASS Table 2. Biologic Data Variable Mean ~ SD Range Hematocrit (%) After first cardioplegia _ After CPB 24.5 _ In ICU 29 _ Potassium (meq/l) Before CPB After CPB 4.65 _ In ICU CPB = cardiopulmonary bypass; ICU = intensive care unit ~ " i Postoperative Findings Upon arrival in the intensive care unit, the patients' SVR averaged a mean of 1,411 _ 433 dynes.s 1.cm 5. The hematocrit was 32.7% ± 4.1%. Renal function was not altered (Fig 4): the creatinine values in 98% of our patients varied between less than 30% to greater than 10% of the preoperative values. Two patients had a decrease in renal function: 1 patient had a high preoperative creatinine level (>15 mg/dl) and I patient suffered a perioperative myocardial infarction necessitating intraaortic balloon counterpulsation. Assisted ventilation was maintained for less than 24 hours in 86% of the patients. The myocardial bound creatine kinase titer (Fig 5) during the patients' stay in the intensive care unit was usually less than 20 units. The maximum titers in 92% of the patients were less than 60 units (this being considered the borderline value between infarction and noninfarction at our institution). Three patients showed new Q waves: the first was an acute case involving a recently occluded left anterior descending coronary artery; the second patient had undergone a redo procedure because of incomplete revascularization; and the third had undergone an extensive right coronary artery endarterectomy. Two patients died, one from pulmonary infection and the other, despite a thorough postmortem examination, showed no evidence of lung disease, hepatitis, or pancreatitis. Neurologic complications were unusual: 1 patient /' ~ - PREOP [] 0 [~--+ I I I! I I ~ I I -I I Fig 4. Renal function. plasma creatlnlne mgldl [] [ I -t I -- I I Fig 5. Maximum titers of myocardial bound creatinine kinase (CK-MB). suffered a stroke due to a cerebral atheroma embolism that migrated from the ascending aorta, with multiple defects seen on a computed tomographic scan of the brain. The 5 patients with a tight asymptomatic carotid stenosis remained asymptomatic. Comment Effects of Normothermic Cardiopulmonary Bypass SYSTEMIC VASCULAR RESISTANCE. Although sometimes considered a meaningless measurement in the setting of CPB, SVR is a well-studied variable. General variations in SVR, as well as the influence of temperature and the priming solution hematocrit value on SVR, have been reported [4]. The usual response of SVR to hypothermic CPB is an initial decrease followed by a progressive rise. This elevation persists after the discontinuation of CPB. The initial drop in SVR may be related to dilution. The elevation in the SVR may be induced by an increase in the plasma epinephrine levels that is triggered by hypothermia [5]. Normothermic CPB is expected to bring about a decline in the SVR. Although normothermic CPB is considered responsible for causing low systemic vascular resistance and this seems to be a major concern, the only precise data concerning SVR reported by Christakis and colleagues [3] were obtained after the discontinuation of CPB, with no significant difference observed between the SVR after hypothermic CPB (865 ± 326 dynes-s 1. cm-s) and that after normothermic CPB ( dynes, s -1 cm 5). In our study, we calculated a mean SVR during bypass of 1, dynes- s -1- cm -5 with a mean pump flow rate of 4.55 ± 0.61 L/rain that closely paralleled the calculated mean flow rate of 4.47 ± 0.49 L/rain, resulting in a mean perfusion pressure of 57 ± 9.4 mm Hg measured on a radial artery catheter. In the study conducted by Christakis' group, a low SVR was found to result in the infusion of significantly greater volumes of crystalloid solution (3.650 ± 800 ml under normothermic conditions versus 3,100 ± 700 ml under hypothermic conditions). The large volume overload is due

5 50 HVASS AND DEPOIX Ann Thorac Surg NORMOTHERMIC BYPASS 1995;59:46-51 to the increased volume of continuous blood cardioplegia delivered and the additional crystalloid solution added to the bypass circuit because of the low SVR. The total cardioplegia volume delivered in their normothermic group (4, ,900 ml) was higher than that delivered to their hypothermic group (2,600 _+ 1,900 ml). The total crystalloid volume ( ml) and total cardioplegia volume (1,946 _+ 257 ml) administered to our patients during normothermic bypass are considerably less. The low SVR observed in Christakis and associates' study was responsible for the large doses of phenylephrine hydrochloride required during and after CPB. They used phenylephrine hydrochloride to keep the systemic pressure above 50 mm Hg. The cumulative dose of phenylephrine administered during normothermic CPB in their patients averaged 1.3 mg. They state that 88% of their patients on normothermic CPB required phenylephrine. An infusion drip for more than 20 minutes was necessary in 5%. The requirements for phenylephrine remained the same for the first 3 hours postoperatively. In their hypothermic group, 64% of the patients received phenylephrine, whereas, in our study, only 18% of the patients received phenylephrine during normothermic CPB. This was always administered in a bolus of 250 /zg, and a second bolus was needed in 2% of the patients. No permanent infusion drip was necessary. Nicardipine (0.5 to 2 mg) was administered in 20% of the patients to lower the SVR during normothermic bypass. More patients in our study needed hypotensive drugs to lower perfusion pressure than they needed hypertensive drugs to raise the pressure. Although the perfusion pressure in 61% of our patients declined at some time during CPB to below 50 mm Hg, we did not think this was a mandatory indication for phenylephrine. It is generally agreed that a minimal perfusion pressure of 35 to 40 mm Hg is necessary to maintain capillary patency [4] and that neurologic impairment is precipitated by low flow [6] and not by low pressure. INDICATORS OF ADEQUATE PERFUSION. Poor peripheral perfusion may be revealed by a wide arteriovenous oxygen difference, a low venous oxygen saturation, increased lactate production, and acidosis. Poor perfusion with a high venous oxygen saturation is found in the setting of septic shock. Normal or high venous oxygen saturation such as that encountered in our study indicates correct peripheral perfusion. Regional perfusion may still be questioned, but we did not encounter increased lactate production. The initial higher-than-normal values were related to the Ringer's lactate in the priming solution. There was no alteration in renal function. This has also been reported by other teams of investigators [1, 7]. If one accepts impairment of renal function as being significant when the postoperative creatinine level exceeds the preoperative values by 50% [8], then only 2 of our patients met this criteria. One had a high preoperative creatinine level (>15 mg/dl) and I suffered a perioperative myocardial infarction necessitating intraaortic balloon counterpulsafion. NEUROLOG1C INJURY. Controversy exists as to whether normothermic CPB has an effect on the neuropsychologic outcome in patients undergoing heart procedures [8]. Higher stroke rates have also been reported in this context [9]. Only 1 of our patients had a stroke, and this was related to an atheromatous embolism, with multiple bilateral defects seen on the computed tomographic scan of the brain. This neurologic episode was not related to the bypass temperature. General Considerations Why did we switch to using normothermic CPB at a time when hypothermia was the standard? Our department of cardiac surgery opened in In the first few years, we used hypothermic CPB with cold crystalloid cardioplegia and topical cooling with crushed ice. We began using cold blood cardioplegia in 1985, but this did not confer any obvious advantages. We began to administer a terminal warm shot of blood cardioplegia in 1986 in patients undergoing heart transplantation. The results in these patients were so convincing that the warm shot was soon used routinely for all open heart operations. The use of crushed ice was abandoned because of the high prevalence of phrenic palsy. The incidence of phrenic palsy of 5% when crushed ice was used for topical cooling declined to 1% with the use of a phrenic protecting pad [10]. The incidence then declined to zero when all forms of topical cooling were abandoned. The hearts were proving to rewarm more without topical cooling (septal myocardial temperature was monitored until 1987), but no disadvantages appeared. We then questioned the necessity of using hypothermic CPB as a source of collateral cooling. As a result, normothermic bypass has been our standard practice since March In conclusion, normothermic CPB ensures adequate perfusion, as demonstrated by normal renal function, no increased lactate production, normal ph, and the high venous saturation during bypass. Normothermic bypass does not precipitate volume overload and high doses of adrenergic drugs are not needed during and after CPB. We have not encountered the general systemic complications reported for warm heart surgical procedures. Finally, normothermic bypass at 37 C with cold intermittent cardioplegia and a terminal hot shot confers the following advantages: it facilitates a clear operative field, there is a high rate of spontaneous defibrillation, it is easy for perfusionists to manage, and, in our experience, it has produced no adverse effects. References 1. Salerno TA, Houk JP, Borrozo CAM, et al. Retrograde continuous warm blood cardioplegia. A new concept in myocardial protection. Ann Thorac Surg 1991;51: Lichtenstein SV, Asche KA, E1 Dalati H, Cusimano RJ, Panos A, S utsky AS. Warm heart surgery. J Thorac Cardiovasc Surg 1991;101: Christakis GT, Koch JP, Deemar KA, et al. A randomized study of the systemic effects of warm heart surgery. Ann Thorac Surg 1992;54: Robicsek F, Masters TN, Niedluchowski W, Yearger JC, Duncan GD. Vasomotor activity during cardiopulmonary bypass. In: Utley JR, ed. Pathophysiology and techniques of cardiopulmonary bypass (Vol 2, Cardiothoracic surgery series).

6 Ann Thorac Surg HVASS AND DEPOIX ;59:46-51 NORMOTHERMIC BYPASS 5. Lehot JJ, Villard J, Piriz H. Hemodynamic and hormonal responses to hypothermic and normothermic cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1992;6: Brusino FG, Reves JG, Prough DS, Stump DA. Cerebral blood flow during cardiopulmonary bypass in a patient with occlusive cerebrovascular disease. J Cardiothorac Anesth 1989;3: Corwin HL, Sprague SM, DeLaria GA, Norusis MJ. Acute renal failure associated with cardiac operations. A case control study. J Thorac Cardiovasc Surg 1989;98: Mora CT, Henson MB, Weintraub WS, et al. Neuropsycho- logical functions following cardiopulmonary bypass. Does warm bypass adversely affect central nervous system outcome? [Abstract]. In: Proceedings of the Eighth Annual Meeting of the European Association of Cardiothoracic Anesthesiologists, Zurich, April Nussmeier N, Fish KJ. Neuropsychological dysfunction after cardiopulmonary bypass: a comparison of two institutions. J Cardiothorac Vasc Anesth 1991;5: Debrux JL, Popoff G, Hvass U, et al. Paralysies phreniques apr6s chirurgie cardiaque: r61e du froid, int6r6t d'un nouveau prot6ge phr6nique isolant. Ann Chir 1991;45: INVITED COMMENTARY Doctors Hvass and Depoix have produced an excellent documentation of their experience with normothermic cardiopulmonary bypass. They have demonstrated a parsimonious use of vasoconstrictors during normothermic cardiopulmonary bypass and have attributed this, in part, to the use of small amounts of cold blood cardioplegia. Upon superficial inspection, it would appear that their experience with normothermic cardiopulmonary bypass differs significantly from ours. However, 61% of the patients in their study (at some point in time, or times, during cardiopulrnonary bypass) were perfused with systemic pressures less than 50 mm Hg. Hvass and Depoix were comfortable with perfusion pressures of 35 to 40 mm Hg, as long as the perfusion flow rates were at predicted levels (2.4 L. rain 1.m 2). If those patients with perfusion pressures less than 50 mm Hg had received infusions of vasoconstrictors (as is the protocol at our institution), the results of this study would have been almost identical to those we have reported previously. Previous studies have demonstrated an excessive incidence of neurologic events and poor cerebral perfusion at systemic pressures below 50 mm Hg. The fact that 61% of patients in this study tolerated normothermic perfusion pressures of less than 50 mm Hg, and presumably in the 35 to 40 mm Hg range, without any neurologic deficits arising, raises important questions about the conduct of normothermic cardiopulmonary bypass and acceptable perfusion pressures. First, can a low-pressure perfusion under normothermic conditions and normal flow rates produce end-organ dysfunction? Second, does the duration of low perfusion pressures influence outcome? Third, what is the lowest perfusion pressure (at normal flow rates) that one should allow without infusing vasoconstrictors? Fourth, how long would a surgeon allow perfusion pressures to remain at 35 mm Hg with flows of 2.4 L. rnin -1 m-2? And fifth, does the fact that patients or elderly or diabetic or the presence of high-grade carotid, mesenteric, or renal stenoses (which patients with coronary artery disease are more likely to possess) influence a surgeon's criteria for lower limits of perfusion pressure and the duration of low perfusion pressures? Another point this study raises is the definition of systemic vascular resistance on cardiopulmonary bypass. Is systemic vascular resistance on bypass a valid and meaningful measurement? The equation for calculating systemic vascular resistance is based on certain theoretical assumptions that are not valid during bypass. The formula used assumes pulsatile flow, a constant blood viscosity, and a direct and proportional relationship between the mean arterial pressure and the central venous pressure. Cardiopulmonary bypass flow is usually nonpulsatile and blood viscosity changes during bypass, depending on temperature, the hematocrit, the plasma protein concentration, and the quantity of fluid infusions. The proportionality relationship between the mean arterial pressure and the central venous pressure is valid only if central venous pressure is kept constant. However, both the mean arterial pressure and central venous pressure vary during cardiopulmonary bypass. Furthermore, measurement of systemic vascular resistance on cardiopulmonary bypass is arbitrary and artificial. Systemic vascular resistance can be increased or decreased by cardiac and anesthetic medications that affect vasomotor tone. It can also be altered by increasing or decreasing flow rates. The collapse of capillaries during low-flow cardiopulrnonary bypass (as may be necessary at times) may result in an artificially increased systemic vascular resistance. If systemic vascular resistance is measured immediately after an increase or decrease in flow rates, and before equilibration and the venous capacitance response takes place, it may be arbitrarily high or low. Alterations in venous capacitance brought about by variations in the size and positioning of venous cannulas, kinks in tubing, "leaving blood behind," and infusions of nitroglycerin may also artificially alter systemic vascular resistance. Although the mechanics of cardiopulmonary bypass have been well established and understood for many years, the recent interest in normothermic cardiopulmonary bypass may necessitate a review of our current understanding and assumptions about the technique. George T. Christakis, MD Department of Surgery, University of Toronto Sunnybrook Health Science Centre 2075 Bayview Ave, Suite H-406 Toronto, ON M4N 3M5 Canada