Effect of temperature during cardiopulmonary bypass on gastric mucosal perfusion

Transcription

1 British Journal of Anaesthesia 1997; 78: Effect of temperature during cardiopulmonary bypass on gastric mucosal perfusion N. D. CROUGHWELL, M. F. NEWMAN, E. LOWRY, R. D. DAVIS JR, K. P. LANDOLFO, W. D. WHITE, J. L. KIRCHNER AND M. G. MYTHEN Summary The purpose of our study was to prospectively study the splanchnic response to hypothermic and tepid cardiopulmonary bypass (CPB) using alphastat management of arterial blood-gas tensions. Twenty-four patients for elective CABG surgery were allocated randomly to tepid (35 36 C) or hypothermic (30 C) bypass groups. Measurements were made at four times: (1) baseline, (2) stable during CPB (inflow temperature nasopharyngeal temperature) 30 C for hypothermic patients, bypass 20 min for tepid patients, (3) 10 min before the end of bypass, (4) after bypass, skin closure. Both groups demonstrated a significant reduction in gastric intramucosal ph (ph im ) from time 1 to time 4 and there was no difference in the incidence of a low ph im between the tepid and cold groups (4/12 vs 3/12; ns) at time 4. ph im was significantly lower in the tepid group at time 3 (P 0.03) but this discrepancy may have been because of an artefactually high ph im in the cold group. There was a significantly higher incidence of postoperative non-cardiac complications in patients who had a low ph im at time 4 (P ) Therefore, we conclude that although the temperature during CPB had a transient effect on ph im it is unlikely to be a major determinant in the pathogenesis of gut mucosal hypoperfusion after bypass. (Br. J. Anaesth. 1997; 78: ) Key words Gastrointestinal tract, ph. Gastrointestinal tract, mucosal perfusion. Heart, cardiopulmonary bypass. Surgery, cardiovascular. Temperature, effect. Despite advances in surgical technique, cardiopulmonary bypass (CPB) and anaesthetic management, non-cardiac complications occur frequently after surgery that requires CPB. 1 Global measures of oxygen consumption and delivery are not reliable predictors of regional organ dysfunction. Indirect measurement of gastric intramucosal ph (ph im ) is being used as a monitor of a tissue bed that is extremely sensitive to reduced peripheral perfusion and may be important in the pathogenesis of multiple organ dysfunction syndrome (MODS). An abnormally low ph im is common after CPB and a sensitive indicator of a poor outcome. The pathogenesis of the observed decrease in ph im remains poorly denned. Hypothermia during cardiopulmonary bypass is thought to provide protection from ischaemic injury during CPB. The use of hypothermia during CPB has fallen out of favour in many institutions which has raised concerns regarding the potential for increased ischaemic injury in poorly perfused organs (e.g. the gut) during warm CPB. As a result, the purpose of our study was to prospectively compare the effect of hypothermic with tepid CPB on gastric mucosal ph. Patients and methods After obtaining local Ethics Committee approval and informed consent, we studied 24 patients undergoing elective CABG surgery; patients were allocated randomly to a tepid (35 36 C) or cold (30 C) bypass group. Premedication comprised oral diazepam 0.1 mg kg 1, oral methadone hydrochloride 0.1 mg kg 1 and oral ranitidine 150 mg, 90 min before induction of anaesthesia. Anaesthesia was induced with midazolam g kg 1 and fentanyl citrate 5 10 g kg 1, and maintained with a continuous infusion of midazolam 0.5 g kg 1 min 1 and fentanyl 0.05 g kg 1 min 1 throughout the study. Pancuronium was given as needed to maintain complete neuromuscular block. Immediately after induction of anaesthesia a tonometer (Tonometrics Division Instrumentarium Corp., Helsinki, Finland) was inserted into the stomach and its position confirmed by injection of air down the nasogastric tube while auscultating over the epigastrium. The tonometer was primed for each measurement with 2.5 ml of 0.9% saline. The PCO 2 of the lumen of the stomach was measured indirectly by sampling the saline in the gas-permeable silicone NARDA D. CROUGHWELL, CRNA, MARK F. NEWMAN, MD, EDWARD LOWRY, BA, R. DUANE DAVIS JR, MD, KEVIN P. LANDOLFO, MD, WILLIAM D. WHITE, MPH, JERRY L. KIRCHNER, BS, MICHAEL G. MYTHEN*, MB, BS, FRCA, MD, Department of Anesthesiology and Surgery, Duke University Medical Center, Durham, NC 27710, USA. Accepted for publication: September 13, *Address correspondence: Department of Anesthesiology, Box 3094, Duke University Medical Center, Durham, NC 27710, USA.

2 Cardiac surgery and gastric mucosal perfusion 35 balloon on the end of the tonometer after an equilibration period of more than 30 min and analysing the aspirate in a blood-gas analyser. At the same time an arterial sample was obtained to measure arterial bicarbonate concentration. ph im was calculated by inserting the values of gastric tonometer PCO 2 and arterial bicarbonate into a modified Henderson- Hasselbalch equation (eqn (1)), as described and validated previously. 2 3 ph i =6.1+log 10(arterial [HCO 3 ] / PCO 2 (tonometer) K) where K time-dependent equilibration constant provided by the manufacturer. Measurements were made at four times: (1) baseline (before aortic cannulation); (2) stable during CPB (inflow temperature nasopharyngeal temperature) 30 C for cold patients, bypass 20 min for tepid patients; (3) before the end of bypass; and (4) after bypass, skin closure. The perfusion apparatus consisted of a Cobe CML oxygenator (Cobe Laboratories, Lakewood, CO, USA) and a Sarns 7000 max pump (Sarns Inc, Ann Arbor, MI, USA). Non-pulsatile perfusion of 2.4 litre min 1 m 2 was maintained throughout the study. The pump was primed with a colloid and crystalloid solution designed to achieve a packed cell volume (PCV) of 0.18 or higher during extracorporeal circulation. Packed red blood cells were added when necessary to achieve the desired PCV. All patients underwent CPB with an ascending aortic cannula, an arterial filter and single stage venous cannula included in the system. Blood-gas and tonometer saline analyses were made with a BGE IL1400 blood-gas electrolyte analyser (Instrumentation Laboratories, Lexington, Table 1 Patient data (mean (SD or range) or number) Hypothermic group Tepid group n Age (yr) 58.5 (33 77) 66.4 (56 81) Weight (kg) 99.2 (37.9) 87.5 (14.9) Sex (M/F) 7/5 10/2 Cross-clamp time (min) 55 (20) 37 (12) CPB time (min) 118 (40) 83 (33) MA, USA). Oxygen saturation and haemoglobin measurements were made on an IL482 co-oximeter (Instrumentation Laboratories, Lexington, MA, USA). Arterial carbon dioxide tension was maintained throughout CPB at kpa (uncorrected for body temperature) and P a O at kpa. 2 Data management and analysis were performed using the SAS statistical system, version 6.08 for Windows on a mhz Insight pc. All data were examined for normality of distribution, and group comparisons were made using parametric t tests or non-parametric Wilcoxon rank sum tests as appropriate; descriptive data are presented as mean (SD). For comparisons of patient data, Wilcoxon s rank sum test of group equality was used. Cardiovascular variables were analysed by overall repeated measures analysis of variance followed by pairwise comparisons with baseline within groups (paired t test) and also between groups (Wilcoxon rank sum test). P 0.05 was considered significant, without adjustment for multiple comparisons in the follow-up descriptive comparisons. Results There was no difference between the hypothermic (n 12) and tepid (n 12) groups in patient characteristics, duration of bypass or aortic cross-clamp time (table 1). In the primary repeated measures analysis, there was no overall significant difference between the two groups in any of the cardiorespiratory variables, haemoglobin or ph im (table 2, fig. 1). Our sample size of 24 (12 per group) gave 80% power to distinguish a significant difference in ph im as small as or larger between the groups. In the follow-up descriptive comparisons there was a significant difference between the two groups in mean arterial pressure (MAP) at time 1 and nasopharyngeal temperature at times 2 and 3. Both groups demonstrated a significant increase in tonometer PCO 2 and a decrease in ph im from baseline to time 4. The increase in tonometer PCO 2 was independent of arterial PCO 2 which was unchanged in both groups. ph im was significantly lower in the tepid group at time 3 (P 0.03). The cold group had Table 2 Physiological variables. Times: (1) baseline; (2) stable (inflow temperature nasopharyngeal temperature) 30 C for cold patients, bypass 20 min for tepid patients; (3) before the end of bypass; (4) after bypass skin closure. phim was calculated by substituting gastric PCO 2 and arterial bicarbonate measured from arterial blood-gas analysis into a modified Henderson Hasselbalch equation (see patients and methods). Hg haemoglobin content, MAP mean arterial pressure, Pa CO arterial partial pressure of carbon dioxide, pha arterial ph, 2 HCO3 arterial bicarbonate concentration, BE arterial base excess, CO 2 gap arterial carbon dioxide splanchnic carbon dioxide TONCO2 tonometer PCO 2. Data mean (SD). *P 0.05 between groups, P 0.05 compared with baseline Time 1 Time 2 Time 3 Time 4 Group Cold Tepid Cold Tepid Cold Tepid Cold Tepid Temp. ( C) 35.5 (1.3) 35.4 (1.0) 30 (0.4) 35.5 (0.5)* 36.3 (0.3) 35.8 (0.8)* 35.8 (0.4) 36.1 (0.6) ph im 7.42 (0.06) 7.41 (0.06) 7.42 (0.05) 7.38 (0.04) 7.43 (0.05) 7.37 (0.06)* 7.34 (0.06) 7.35 (0.05) CO 2 gap (kpa) 0.16 (0.47) 0.13 (0.97) 0.17 (0.40) 0.37 (0.52) 0.25 (0.48) 0.44 (0.61)* 0.62 (0.52) 0.76 (0.75) Hg (g dl 1 ) 13.0 (2.2) 11.8 (2.5) 8.7 (2.0) 8.8. (1.3) 8.6 (1.3) 8.5 (1.2) 10 (1.7) 9.1 (2.7) MAP (mm Hg) 85 (10) 73 (11)* 71 (17) 68 (11) 65 (18) 69 (12) 76 (9) 72 (14) ph a 7.42 (0.04) 7.41 (0.04) 7.41 (0.03) 7.39 (0.04) 7.39 (0.07) 7.38 (0.05) 7.37 (0.05) 7.39 (0.04) P (kpa) 32.4 (23.5) 22.8 (8.1) 27.6 (11.5) 27.7 (10.5) 24.7 (8.3) 29.7 (11.9) 18.4 (8.0) 24.4 (13.9) a O2 a CO 2 P (kpa) 4.9 (0.53) 5.1 (0.93) 4.8 (0.53) 5.2 (0.53) 5.1 (0.53) 5.1 (0.53) 4.9 (0.53) 4.9 (0.53) HCO 3 (mmol litre 1 ) 24 (2) 24 (3) 23 (2) 24 (2) 23 (2) 23 (2) 22 (1) 23 (2) BE (mmol litre 1 ) 0.9 (2.2) 0.02 (1.7) 0.6 (2.3) 0.4 (1.9) 1.8 (1.9) 0.8 (2.2) 2.7 (1.5) 1.5 (1.7) TONCO 2 (kpa) 5.2 (0.67) 5.2 (0.67) 4.9 (0.53) 5.6 (0.67)* 4.8 (0.53) 5.5 (0.67)* 5.6 (0.67) 5.7 (0.80)

3 36 British Journal of Anaesthesia Of the seven patients who had a low ph im at time 4, four developed postoperative non-cardiac complications that resulted in an increased (greater than 7 days) duration of stay in hospital (table 3). One patient in the group that had a normal ph im at the end of surgery had an increased length of hospital stay. Logistic regression analysis, testing association between ph im (as a continuous measure) and risk of postoperative complications, demonstrated a significant relationship (P ) between ph im at time 4 and prolonged hospital stay (fig. 2). Figure 1 Intramucosal ph (phi) at each of the measurement times in the cold ( ) and tepid ( ) CPB groups. ph im was calculated by substituting gastric PCO 2 and arterial bicarbonate measured from arterial blood-gas analysis into a modified Henderson Hasselbalch equation (see patients and methods). Times: (1) baseline; (2) stable (inflow temperature nasopharyngeal temperature) 30 C for cold patients, bypass 20 min for tepid patients; (3) before the end of bypass; (4) after bypass, skin closure. *P 0.05 between groups, P 0.05 compared with baseline. Table 3 Postoperative complications. Time 4 after bypass, skin closure, LOS length of stay, RML right middle lobe ph im (time 4) LOS Complication Post-pericardiotomy syndrome Stroke Wound infection, RML collapse, pneumonia Wound cellulitis rt leg Wound infection, debridement Figure 2 Duration of hospital stay related to carbon dioxide (CO 2) gap and ph im. CO 2 gap is the gap between gastric luminal PCO2 and arterial PCO 2. Gastric hypercapnia relative to arterial PCO2 is indicative of gut mucosal hypoperfusion. phim is calculated by substituting gastric PCO2 and arterial bicarbonate measured from arterial blood-gas analysis into a modified Henderson Hasselbalch equation (see patients and methods). Logistic regression analysis examining the association between ph im (as a continuous measure) and risk of postoperative complications demonstrated a significant relationship (P ) between ph im at time 4 and prolonged hospital stay. a significant increase in base deficit compared with baseline at times 3 and 4. In both groups there was a significant decrease in haemoglobin concentration over time. Seven patients had an abnormally low ph im ( 7.32) by time 4. There was no difference in the incidence of a low ph im between the two groups (four in the cold group compared with three in the tepid group). Discussion This is the first prospective randomized study comparing gastric mucosal perfusion using tepid and cold bypass. The results indicate that the reduction in gastric mucosal perfusion after cardiac surgery was independent of the temperature during CPB. The warm group had a significantly lower ph im at time 3 during CPB. This was not evident after rewarming. The results of this study are consistent with previous studies that reported that decreases in gastric mucosal ph im are common after cardiac surgery and associated with prolonged hospital stay Kutilla, Niinikoski and Haglund examined the relationship between gastric ph im and peripheral tissue perfusion in patients after admission to the intensive therapy unit after hypothermic CPB. 8 They found that ph im decreased steadily over 4 h as the patient s core temperature increased to normal. From 4 to 8 h after operation, ph im increased again to normal. Over the same 8-h period they recorded a steady improvement in peripheral perfusion, as judged by subcutaneous and transcutaneous tissue oxygen tension, laser Doppler skin red cell flux and finger tip temperature, all measured on the same arm. They found no change in central haemodynamic state (including cardiac output). They attributed the changes to redistribution of blood flow after hypothermic CPB. Landow and colleagues studied patients during CPB to determine the relationship between gastric ph im and several other more invasively derived indices of splanchnic perfusion. 9 They found that mean ph im increased during the period of hypothermic CPB and then became abnormal only after rewarming. Gastric ph im correlated with hepatic venous lactate concentration and hepatic venous oxygen, suggesting supplydependent oxygen deficit across the splanchnic bed. They proposed that with rewarming after hypothermic cardiopulmonary bypass there is regional oxygen consumption/delivery mismatch. This could have occurred in our patients and this postulate is consistent with our own ph im data in the hypothermic group. In a canine model, Ohri and colleagues demonstrated that during the rewarming phase of CPB there was jejunal mucosal hypoperfusion in spite of an increase in local blood flow. 10 They proposed a shunting phenomenon where blood flowed freely in the serosal but not mucosal layers. All of these studies suggest that temperature change is an important factor in the pathogenesis of gut mucosal hypoperfusion after CPB. Our results do not support this hypothesis because after cardiopulmonary

4 Cardiac surgery and gastric mucosal perfusion 37 bypass there was a significant decline compared with baseline but no difference between the two temperature groups. Previous studies have suggested that the decrease in gastric ph to abnormal values is in proportion to the duration of CPB Suggested causes of impaired gut perfusion on CPB include occult hypovolaemia, ingestion of environmental endotoxin and release of endogenous vasoconstrictors during CPB During CPB large quantities of angiotension II are released. 14 In animal models, angiotension II was found to be a potent selective splanchnic vasoconstrictor. 15 Cardiopulmonary bypass also causes activation of platelets, neutrophils and inflammatory pathways. 16 Resulting microaggregate formation may cause microvascular occlusion and local reduction in oxygen delivery. 17 We have demonstrated previously that pre-emptive plasma volume expansion with a starch solution significantly reduces (but does not eliminate) gut mucosal hypoperfusion after bypass. 7 We have also noted that higher concentrations of endogenous endotoxin antibodies confer some protection against activation of inflammatory pathways and gut mucosal hypoperfusion. 18 Other investigators have tried to improve gut mucosal perfusion in the postbypass period using cardioactive drugs such as dopexamine and dobutamine. 19 Despite increasing both whole body and splanchnic oxygen delivery, these drugs failed to improve gastric mucosal perfusion, as judged by the persistence of a low ph im. One suggestion was that persistence of a gastric mural acidosis in the face of increased oxygen delivery was because of serosal shunting of blood in response to rewarming after cold CPB. We now believe that this is unlikely and think that microvascular occlusion secondary to formation of microaggregates or emboli is more consistent with the currently available data, as suggested in a study of retinal microembolism. 17 The changes in ph im noted during hypothermic CPB in this and other studies may be artefactual as no allowance was made in the tonometric technique for the effects of temperature on carbon dioxide production, solubility and diffusion through tissues and the tonometer balloon. The tonometric technique is dependent on both aerobic and anaerobic production of carbon dioxide. The main reason for deliberately cooling patients during CPB is to reduce tissue metabolic rate and thus aerobic carbon dioxide production. Although this has not been studied systematically it seems reasonable to assume that background carbon dioxide production is reduced proportionally to temperature reduction and this calculated ph im is increased artefactually, independent of gut perfusion. In addition, reducing temperature increases tissue solubility of carbon dioxide and decreases diffusion through both tissues and the tonometer balloon. The manufacturers state that the device is intended for determination of intramucosal ph of the gastric mucosal tissue at 37 C. All of these factors would tend to give a falsely higher ph im, irrespective of perfusion of the gastric mucosa. The possible artefactual effects of hypothermia on the tonometric readings have been poorly validated and are worthy of further investigation. This is a potential limitation of the interpretation of time 2 and 3 measurements in this study; however it should not affect conclusions drawn of the effect from beginning to end of CPB measurements. There has been much recent debate on the validity of the calculation of ph im from gastric luminal PCO 2 and systemic arterial bicarbonate Calculated ph im can be influenced by systemic acid-base disturbances independent of splanchnic perfusion. It has been suggested that a more specific measure of gut perfusion is the gap between gastric luminal PCO 2 and arterial PCO 2. Gastric hypercapnia relative to arterial PCO 2 is indicative of gut mucosal hypoperfusion. In this study the significance of the results comparing the two groups was unaltered using calculated ph im or arterial to gastric PCO 2 gap. However, one patient who had a poor outcome had an increased carbon dioxide gap but a normal ph im while another had a low ph im but a normal gap (fig. 2). Again, the significance of the results was unaltered but underlines the current suggestion of changing from calculation of ph im to reporting carbon dioxide gap. Calculated ph im is a sensitive indicator of tissue hypoperfusion whereas the carbon dioxide gap is a more specific indicator of gastric hypoperfusion. The approach to arterial blood-gas management during hypothermic CPB remains controversial. In studies comparing ph and alpha-stat management, ph-stat management produced greater cognitive dysfunction after bypass, especially in patients requiring longer bypass times However, there are no human data comparing the two techniques in gut perfusion. It is the practice in our institution to use alpha-stat management of arterial blood-gas tensions. The complications reported (table 3) occurred in patients in whom a low ph im developed. These data are consistent with the hypothesis that gut mucosal hypoperfusion may result in translocation of endotoxin, additional activation of inflammatory pathways and subsequent end-organ dysfunction. We conclude that splanchnic hypoperfusion is common after CPB and associated with prolonged hospital stay. There was a lower ph im during warm CPB; this may have been because of an artefactually high ph im in the cold group. After CPB, using alphastat management of arterial blood-gas tensions, there was a significant reduction in ph im in both groups and the incidence of a low ph im and major complications were the same in both groups. Therefore, we conclude that temperature during CPB had no significant effect on the degree of gut mucosal hypoperfusion after CPB. Acknowledgement This research was supported in part by NIH grant RO1- AG References 1. Weintraub WS, Jones EL, Craver J, Guyton R, Cohen C. Determinants of prolonged length of hospital stay after coronary bypass surgery. Circulation 1989; 80: Fiddian-Green RG, Baker S. Predictive value of the stomach

5 38 British Journal of Anaesthesia wall ph for complications after cardiac operations: comparison with other monitoring. Critical Care Medicine 1987; 15: Antonsson JB, Boyle CC, Kruithoff KL, Wang HL, Sacristan E, Rothschild HR, Fink MP. Validity of tonometric measures of gut intramural ph during endotoxemia and mesenteric occlusion in pigs. American Journal of Physiology 1990; 259: G519 G Mythen MG, Purdy G, Mackie IJ, McNally T, Webb AR, Machin SJ. Post-operative multiple organ dysfunction syndrome associated with gut mucosal hypoperfusion, increased neutrophil degranulation and Cl-esterase inhibitor depletion. British Journal of Anaesthesia 1993; 71: Mythen MG, Webb AR. Intra-operative gut mucosal hypoperfusion is associated with increased post-operative complications and cost. Intensive Care Medicine 1994; 20: Mythen MG, Webb AR. The role of gut mucosal hypoperfusion in the pathogenesis of post-operative organ dysfunction. Intensive Care Medicine 1994; 20: Mythen MG, Webb AR. Pre-operative plasma volume expansion reduces the incidence of gut mucosal hypoperfusion during cardiac surgery. Archives of Surgery 1995; 130: Kuttila K, Niinikoski J, Haglund U. Visceral and peripheral tissue perfusion after cardiac surgery. Scandinavian Journal of Thoracic and Cardiovascular Surgery 1991; 25: Landow L, Phillips DA, Heard SO, Prevost D, Vanderslam TJ, Fink MP. Gastric tonometry and venous oximetry in cardiac surgery patients. Critical Care Medicine 1991; 19: Ohri SK, Becket J, Brannan J, Keogh BE, Taylor KM. Effects of cardiopulmonary bypass on gut blood flow, oxygen utilization, and intramucosal ph. Annals of Thoracic Surgery 1994; 57: Anderson LW, Landow L, Baek L, Jansen E, Baker S. Association between gastric intramucosal ph and splanchnic endotoxin, antibody to endotoxin, and tumour necrosis factor alpha concentrations in patients undergoing cardiopulmonary bypass. Critical Care Medicine 1993; 21: O Dwyer ST, Michie HR, Ziegler TR, Revhaug A, Smith RG, Wilmore DW. A single dose of endotoxin increases intestinal permeability in healthy humans. Archives of Surgery 1988; 123: Fiddian-Green RG. Gut mucosal ischemia during cardiac surgery. Seminars in Thoracic and Cardiovascular Surgery 1990; 2: Taylor K M, Pulsatile perfusion. In: Taylor K M, ed. Cardiopulmonary Bypass: Principles and Management. Baltimore: Williams and Wilkins, 1986; Porter JM, Sussman MS, Bulkley GB. Splanchnic vasospasm in circulatory shock. In: Marston A, Bulkley GB, Fiddian- Green RG, Haglund QH, eds. Splanchnic Ischaemia and Multiple Organ Failure. London: Edward Arnold, 1989; Royston DJ, Fleming J, Desai B, Taylor KM. Neutrophil activation and pulmonary injury during open heart surgery. Journal of Thoracic and Cardiovascular Surgery 1986; 91: Blauth C, Kohner EM, Arnold J, Taylor KM. Retinal microembolism during cardiopulmonary bypass demonstrated by fluorescine angiography. Lancet 1986; II: Mythen MG, Barclay GR, Purdy G, Hamilton-Davies C, Mackie IJ, Webb AR, Machin SJ. The role of endotoxin immunity, neutrophil degranulation and contact activation in the pathogenesis of post-operative organ dysfunction. Blood, Coagulation and Fibrinolysis 1993; 4: Uusaro A, Ruokonen E, Takala J. Gastric mucosal ph does not reflect changes in splanchnic blood flow after cardiac surgery. British Journal of Anaesthesia 1995; 74: Boyd O, Mackay CJ, Lamb G, Bland JM, Grounds RM, Bennett ED. Comparison of clinical information gained from routine blood-gas analysis and from gastric tonometry for intramural ph. Lancet 1993; 34: Schlichtig R, Bowles SA. Distinguishing between aerobic and anerobic appearance of dissolved CO 2 in intestine during low flow. Journal of Applied Physiology 1994; 76: Stephen H, Weyland A, Kazmaier S. Acid-base management during hypothermic cardiopulmonary bypass does not affect cerebral metabolism but does affect blood flow and neurological outcome. Annals of Thoracic Surgery 1992; 59: Murkin JM, Martzke JS, Buchan AM, Bentley C, Wong CJ. A randomized study of the influence of perfusion technique and ph management strategy in 316 patients undergoing artery bypass surgery. I. Mortality and cardiovascular morbidity. Journal of Thoracic and Cardiovascular Surgery 1995; 110: