Despite improvements in anesthetic techniques and cardiopulmonary bypass (CPB) equipment, postoperative pulmonary dysfunction remains a frequent complication after CPB (1). Exposure of blood to the artificial surfaces of the extracorporeal circuit not only activates platelets and leukocytes, but also initiates inflammatory, coagulation, fibrinolytic, and kallikrein cascades during CPB. These responses may result in clinical effects such as bleeding diatheses and compromised organs such as the lungs (2). Pulmonary endothelial dysfunction is provoked by complement activation during CPB (3).
The use of heparin-coated circuits has a protective effect on pulmonary function in CPB patients (4). However, this beneficial effect was demonstrated at the end of surgery and did not influence intubation time or length of stay in the intensive care unit (ICU). During the intra- and post-CPB period, coated circuits have demonstrated improved biocompatibility by reduction of inflammatory mediators from the complement cascade, granulocytes (5), and cytokine release (6).
Other mediators involved in the inflammatory response are phospholipases A2 (PLA2s) (7), which are produced during and after CPB (8). PLA2s are enzymes that specifically hydrolyze the fatty acid side chain ester bond at the sn-2 position of phospholipids. They are classified according to their amino acid sequences into two main groups—the intracellular and the secretory PLA2s—and several subtypes (7). PLA2 has proinflammatory properties through regulating the release of platelet-activating factor (PAF) and arachidonic acid (AA). The transformed products of AA can modulate various neutrophil functions, increase vasodilation, and stimulate platelet aggregation (9).
Type II secretory PLA2 (sPLA2-II) can be produced and released by a number of cells, including the alveolar macrophages, which are in contact with the pulmonary surfactant (10). Moreover, sPLA2-II is proposed to play a key role in inflammatory lung damage (11).
It would therefore be interesting to know the nature of the relationships between sPLA2, platelets, leukocytes, and pulmonary indices after CPB procedures. Further, it would be of practical value to know whether heparin coating reduces sPLA2 release and subsequent pulmonary injury.
Patients with depressed left ventricular function produce proinflammatory cytokines (12). These patients are more likely to receive large doses of inotropic drugs and vasoconstrictors, which may indirectly lead to an increase in the inflammatory response.
We studied the effect of heparin coating on lung pulmonary indices, platelet number and activation (β-thromboglobulin [β-TG]), leukocyte number and activation (elastase-α1-antitrypsin complex), complement activation (C3b/c), and sPLA2 during and after CPB. Additionally, blood loss, blood use, mortality, and ICU stay were evaluated as clinical outcomes.
After approval from the medical ethics committee and written, informed consent, 51 patients undergoing elective coronary artery bypass grafting were admitted to the study. Inclusion criteria were patients scheduled for elective first-time coronary artery bypass grafting, aged 45 to 70 yr, and with a left ventricular ejection fraction more than 40%. Exclusion criteria were patients with known previous pulmonary dysfunction, preoperative pulmonary therapy, preoperative immunosuppressive therapy, preoperative use of a nonsteroidal antiinflammatory drug, intra-aortic balloon support, aneurysmectomy requirement, insulin-dependent diabetes mellitus, or a plasma creatinine level more than 150 μmol/L. The patients were randomly allocated in a double-blinded fashion into groups to be perfused with either a heparin-coated or an uncoated extracorporeal circuit. Physicians involved in postoperative patient care were also blinded for randomization. As a precaution against interference, preoperative left ventricular ejection fraction and pulmonary function tests were performed according to hospital protocol.
On the day of surgery, patients received their usual early-morning dose of antianginal medication and 5 mg of lorazepam; no diuretics were given. Anesthesia was induced by using IV sufentanil 3–7 μg/kg, pancuronium bromide 0.1 mg/kg, and midazolam 0.1 mg/kg, and general anesthesia was maintained by a continuous infusion of midazolam 0.1 mg · kg− 1 · h−1 and sufentanil 1 mg · kg−1 · h−1. After endotracheal intubation, patients were mechanically ventilated at a tidal volume of 8 mL/kg with an inspiratory mixture of 50% nitrous oxide and 50% oxygen at a frequency of 14–18 breaths/min and 5 cm H2O of positive end-expiratory pressure (PEEP) to achieve normocapnia. After the induction of anesthesia, patients received dexamethasone 1 mg/kg and 1500 mg of cefuroxime.
Radial artery and thermodilution pulmonary artery catheters were inserted for blood sampling and hemodynamic monitoring. During CPB, a continuous positive airway pressure of 5 cm H2O with a fraction of inspired oxygen (FIO2) of 0.21 was maintained in the lungs. If hemodynamic instability occurred, fluid replacement and inotropic support with dopamine were the first steps taken to stabilize the patient before and after CPB. Separation from CPB took place with dopamine 2 μg · kg−1 · min−1 and nitroglycerin 0.5 μg · kg−1 · min−1.
The components of the extracorporeal circuit were the same for all patients and included a roller pump (SIII; Stockert GmbH, Munich, Germany), a hollow-fiber oxygenator (Quadrox; Jostra Medizintechnik GmbH, Hirrlingen, Germany), a soft-shell venous reservoir, a cardiotomy reservoir, a tubing system (Jostra Medizintechnik GmbH), and an arterial line filter (Auto Vent-SV, 40-μm polyester screen; Pall Biomedical Ltd., Portsmouth, UK). The entire circuit in the coated group except for the cannulas was treated with Bioline® coating (Jostra Medizintechnik GmbH). This heparin coating uses both covalent and ionic bonding to attach heparin in a polypeptide hydrogel matrix to the polymer surface. The heparin used in this coating is Liquemin (Hoffman-LaRoche, Switzerland). The prime consisted of 1400 mL:1000 mL of modified fluid gelatin, a plasma volume substitute, 50 mL of lactated Ringer's solution, 200 mL of aprotinin (Trasylol®, 2 × 106 kallikrein-inhibiting units; Bayer AG, Leverkusen, Germany), 100 mL of mannitol, and 50 mL of sodium bicarbonate (8.4%) containing 1500 mg of cefuroxime and 5000 IU of bovine heparin.
A standard cannulation technique was used to perform CPB in which cannulas were placed in the ascending aorta and right atrium (two-stage venous cannula). CPB was initiated when, after systemic heparinization (300 IU/kg), the celite activated clotting time was more than 480 s. During CPB, the target hematocrit (HCT) was 23%, and, if possible, autologous blood was removed via the central venous catheter and collected in a blood transfer bag before the onset of CPB. The amount of predonated autologous blood was replaced by the same amount of modified fluid gelatin. The nonpulsatile flow rate was maintained at 1.9–2.4 L · min− 1 · m− 2 during moderate hypothermia (28°C–30°C). All patients received antegrade high-potassium cold crystalloid cardioplegia (800–1000 mL; 4°C) for myocardial protection. Suction blood was collected in the cardiotomy reservoir and continuously returned to the patient's circulation.
After termination of CPB, heparin was neutralized with an equal dose of protamine sulfate (3 mg/kg) at an infusion rate of 100 mL/h (1 mL per 1000 IU of heparin). Autologous blood and pump residual volume were reinfused to the patient as the first choice of fluid replacement.
In the ICU, patients were treated according to standard protocol. During ventilation at normocapnia, the arterial oxygen tension was maintained at more than 80 mm Hg with continuous positive pressure ventilation and a PEEP of 5 cm H2O. After fulfilling the following criteria, patients were tracheally extubated: awake, cooperative, a respiratory frequency of 10–20 breaths/min (spontaneous), arterial oxygen tension more than 80 mm Hg, FIO2 ≤0.40, and a positive pressure support ≤5 cm H2O.
After removal of the chest tube and without the need for continuous monitoring and IV inotropes, the patients were discharged from the ICU. Hemodynamic pressures were measured according to protocol, and pulmonary vascular resistance index (PVRI), pulmonary shunt fraction (PS), and PaO2/FIO2 ratio were calculated with the following formulas:
EQUATION EQUATION EQUATION EQUATION
Assuming that at the end capillary there is a balance between the alveolar and capillary pressure,EQUATION
Blood gas measurements were performed by collecting blood samples from the arterial and venous sides of the patient and the extracorporeal circuit and processing them in an external blood gas monitor (Rapidlab 865; Chiron Diagnostics Corp., East Walpole, MA). From each sample, the following arterial and venous values were recorded: partial oxygen pressure (PaO2 and PvO2), partial carbon dioxide pressure (PaCO2 and PvCO2), saturation (SaO2 and SvO2), aHb, vHb, aHCT, and vHCT. Each time arterial and venous blood samples were obtained, the FIO2, gas flow, nasopharyngeal temperature, and blood flow were recorded.
Blood was collected in EDTA (0.01 M) to assess platelet and leukocyte counts. Cell counts were performed by using the impedance method on a Sysmex SE9000 (Sysmex, Kobe, Japan).
Samples for cell counts were taken at the following times: before CPB (before the induction of anesthesia), on arrival in ICU, 2 h in the ICU, 8 h in the ICU, and postoperative day (POD) 1. The samples for biochemical measurements were drawn from the arterial catheter before CPB (before the induction of anesthesia), l5 min after starting CPB, at cessation of CPB, 30 min after the start of protamine infusion, on POD 1, and on POD 2.
All biochemical measurements were corrected for HCT during CPB, at cessation of CPB, and 30 min after the start of protamine infusion. Blood was collected in tubes with a medium containing EDTA and soybean trypsin inhibitor that blocks nonspecific activation.
After the platelet-poor plasma was prepared for analysis, it was immediately stored at −80°C. Complement C3b/c (13), elastase-α1-antitrypsin complex (14), and sPLA2 (15) were determined as previously described. In addition to the measurements of biocompatibility, β-TG was determined according to the manufacturer's instructions (Assarachrom, Asnieressur-Seines, France) at Hemoprobe BV.
Data were gathered on a case report form with anonymous coding of each patient, stored in a database, and analyzed with standard computer software (SPSS 7.5; SPSS Inc., Chicago, IL). Data are presented as means with SD or SEM or as medians with interquartile ranges (25th–75th percentile). Analysis of variance was used. If data were not normally distributed, a log transformation was performed.
According to the nature of the data, we used Student's t-test, the Mann-Whitney U-test, or Fisher's exact test, and within-group differences were analyzed with the paired Student's t-test or Wilcoxon's matched pairs test. A two-sided P < 0.05 was considered to be statistically significant. Pearson or Spearman rank correlation tests were used to assess the degree of association between variables.
Demographic and surgical data are shown in Table 1. No adverse events occurred. No differences, as shown in Table 2, were found in cardiac output or blood loss (use of blood bank products, Hb, or HCT) after CPB.
PS increased significantly in both groups at 2 h in the ICU as compared with baseline (P < 0.05), and then it returned almost to baseline values at 8 h in the ICU (Table 3). Moreover, the PS in the uncoated group increased to significantly higher values as compared with the coated group after CPB (at 30 min after protamine infusion) and at 2 h in the ICU (P < 0.05).
The PaO2/FIO2 ratio in both groups decreased from baseline values to significantly lower levels at 8 h in the ICU (P < 0.05;Table 3). The PaO2/FIO2 ratio in the coated group was significantly increased (P < 0.05) compared with the uncoated group post-CPB and in the ICU.
PVRI (Table 3) decreased significantly in both groups after CPB compared with baseline values (P < 0.05). At arrival in the ICU, values started to increase again to significantly larger values compared with baseline (P < 0.05). After CPB and at all ICU measuring points, the PVRI was significantly higher in the coated than in the uncoated group.
Platelet counts (Fig. 1A) were significantly reduced in both groups compared with baseline, but this reduction was more pronounced in the uncoated group, which resulted in significant differences between the coated and uncoated groups in the ICU and in the postoperative period (P < 0.05).
β-TG (Fig. 1B) showed a significant increase in both groups (coated group, P < 0.01; uncoated group, P < 0.001). No differences were found between groups in β-TG at cessation of CPB.
The number of leukocytes (Fig. 2A) in both groups increased significantly in the ICU (P < 0.05). At arrival and at 2 h in the ICU, the leukocyte counts in the coated group were significantly less compared with the uncoated group. In both groups there was a significant increase in elastase-α1-antitrypsin complex (Fig. 2B) from baseline levels to cessation of CPB, and the elastase concentration was significantly larger in the uncoated than in the coated group (P < 0.001).
The complement activation product C3b/c (Fig. 3) demonstrated a significant increase from baseline levels to 30 min after protamine in both groups (P < 0.001). At 30 min after protamine, the C3b/c concentrations were significantly larger in the uncoated group than in the coated group (P < 0.05).
The sPLA2 levels increased significantly in the uncoated group during CPB (Fig. 4) and on POD 2. At 15 min on bypass, the end of bypass, and POD 2, this resulted in significant differences between the coated and uncoated groups (P < 0.005, P < 0.005, and P < 0.05, respectively). The sPLA2 concentrations at the end of CPB and after protamine correlated significantly with elastase (r = 0.583 and r = 0.527, respectively; both P < 0.01) and C3b/c concentrations (r = 0.536 and r = 0.542, respectively; both P < 0.01) after protamine. Additionally, a significant correlation was found between PVRI after CPB and sPLA2 levels after protamine (r = 0.505; P < 0.01).
This study demonstrated that the use of heparin-coated extracorporeal circuits resulted in a diminished activation of cellular blood elements and humoral mediators, and this could be beneficial for pulmonary function as compared with uncoated circuits. The improved pulmonary indices as reflected by PVRI, PS, and the PaO2/FIO2 ratio in the coated group are comparable to the findings of an animal study (16). In particular, PS appears to be a determinant of pulmonary dysfunction after CPB (3,17). The results of a lower PaO2/FIO2 ratio and a higher PS in the uncoated group suggest that gas exchange in the coated group was less impaired. However, the results of both groups lie within normally expected ranges. Probably, therefore, no significant difference was found in the duration of mechanical ventilation in the ICU between groups, although this duration was on average slightly longer in the uncoated group.
PLA2 regulates PAF and AA release and is therefore a precursor of the generation of the inflammatory lipid mediators leukotrienes, prostaglandins, and thromboxane (18). The generation and activation of sPLA2 is probably related to the use of heparin and to complement activation. In accordance with others (8,19), we found that anticoagulation with heparin enhanced the sPLA2 level, which originates mainly from the visceral organs (19). After protamine administration, the level of sPLA2 decreased in the uncoated group. However, in the coated group, sPLA2 levels were significantly lower during CPB, whereas after protamine administration, sPLA2 levels were enhanced. This may be partly explained by the binding of the PLA2-II fraction to immobilized heparin while protamine stimulates PLA2 through another mechanism (20).
Activated complement factors may also induce PLA2 release and increase activity. This is due to an increased release of PLA2 from activated leukocytes and has been shown in a model of glomerular epithelial cell injury, which appeared to be due to increased cPLA2 expression and activation after sublytic C5b-9 exposure (21). It is important to note that this effect of C5b-9 lasts for more than 24 hours. The marked difference in complement activation in this study between the coated and uncoated groups could therefore be considered as an indirect cause of reduced PLA2 release. Moreover, heparin coating is probably capable of binding PLA2. Thus, heparin coating reduces the inflammatory reaction in two ways: through binding of PLA2 and through reduced complement activation. The correlation of PLA2 and elastase and their reduction during use of heparin-coated circuits are indicative of common mediators. A positive feedback mechanism also has to be considered in which PLA2 induces thromboxane A2 and PAF.
The lower platelet count and increased β-TG release, as seen in the uncoated group, could be explained by the activity of these components. Not only does the active precursor of thromboxane A2 activate platelet aggregation, but thromboxane A2 is also the major source of pulmonary vasoconstriction, as demonstrated by an increase in PVRI (22).
The lungs have been shown to contain a severe aggregation of intravascular neutrophils that is mediated by the production of thromboxane A2 (23). Neutrophil adherence to vascular endothelium is stimulated by complement split products such as C5a. Complement activation also stimulates neutrophil degranulation to release the serine protease elastase (2). Elastase mediated by activated and accumulated neutrophils in the lungs may induce lung injury (24). In the lungs, α1-antitrypsin can form complexes with elastase and neutralize its effects. The results demonstrated significantly lower levels of elastase and elastase-α1-antitrypsin complexes in the coated group, supporting the smaller number of leukocytes in this group. This reduced leukocyte count after CPB in the coated group may have been beneficial, because returning leukocyte-depleted blood showed postoperative improvement of pulmonary function compared with blood that contained leukocytes (25).
Next to the indirect effect of sPLA2 on polymorphonuclear neutrophils and platelets, sPLA2 may also directly affect lung surfactant (11). Alveolar macrophages are phagocytic cells that may release several inflammatory mediators, but they are also the major cell source for sPLA2 (10). These macrophages are in contact with pulmonary surfactant. Pulmonary surfactant contains phospholipids, which are susceptible to hydrolysis by sPLA2 (10). As demonstrated by the relationship between surfactant degradation and sPLA2-II levels, this also suggests that alveolar surfactant is the primary target of sPLA2-II (11). Transformation of pulmonary surfactant is manifested in alveolar collapse and impaired gas exchange (10). Impaired pulmonary function appeared to correlate with the release of sPLA2 (11). Although the inflammatory reaction was reduced with the use of this type of heparin-coated circuit, no differences in clinical outcome—such as blood loss, blood use, time on the ventilator, ICU stay, hospital stay, or mortality—were found.
One of the limitations of this study is the preoperative use of dexamethasone and aprotinin in the pump prime, which is standard therapy in our center. Dexamethasone inhibits leukocyte activity and can reduce hemodynamic instability after CPB (26). Moreover, a reduction of pulmonary inflammation improves pulmonary function. However, in a post hoc study, the observed differences were no longer present by three hours in the ICU, and no difference in the duration of ICU stay could be demonstrated (27). Aprotinin improves hemostasis, but the use of small-dose (2 million IU) aprotinin in the prime has been shown to not reduce the inflammatory reaction (28).
Another limitation was that patients who were smokers were included in the study. Smokers are more at risk for developing a postoperative pulmonary complication (29). Moreover, alveolar macrophage function may be decreased as a result of smoking (30). Although more smokers were included in the heparin-coated group than in the uncoated group, no preoperative differences between the groups were found in pulmonary function.
The groups consisted of elective low-risk cardiac surgical patients, and this could also be a limitation. The inclusion of higher-risk patients may capture the protective effect from heparin-coated circuits, leading to larger differences. This study may also have been limited by its sample size. However, studies with large sample sizes that have been performed on the clinical benefits of heparin coating have failed to reveal clinical benefits (31).
In conclusion, the findings of our study suggest that sPLA2 may play a role in determining the degree of pulmonary injury and may therefore be a mediator of postoperative lung dysfunction. Heparin-coated circuits reduced sPLA2 concentration during CPB and showed improved pulmonary indices and less injury to blood elements and humoral mediators, suggesting that the lungs may be better preserved during CPB with the use of such circuits.
We thank A. Eerenberg and G. van Mierlo for performing the assays.
1. Ng CS, Wan S, Yim AP, Arifi AA. Pulmonary dysfunction after cardiac surgery. Chest 2002;121:1269–77.
2. Butler J, Rocker GM, Westaby S. Inflammatory response to cardiopulmonary bypass. Ann Thorac Surg 1993;55:552–9.
3. Park KW, Tofukuji M, Metais C, et al. Attenuation of endothelium-dependent dilatation of pig pulmonary arterioles after cardiopulmonary bypass is prevented by monoclonal antibody to complement C5a. Anesth Analg 1999;89:42–8.
4. Ranucci M, Cirri S, Conti D, et al. Beneficial effects of Duraflo II heparin-coated circuits on postperfusion lung dysfunction. Ann Thorac Surg 1996;61:76–81.
5. Harig F, Feyrer R, Mahmoud FO, et al. Reducing the post-pump syndrome by using heparin-coated circuits, steroids, or aprotinin. Thorac Cardiovasc Surg 1999;47:111–8.
6. Shore-Lesserson L. Pro: heparin bonded circuits represent a desirable option for cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1998;12:705–9.
7. Dennis EA. Diversity of group types, regulation, and function of phospholipase A2. J Biol Chem 1994;269:13057–60.
8. Nakamura H, Kim DK, Philbin DM, et al. Heparin enhanced plasma phospholipase A2 activity and prostacyclin synthesis in patients undergoing cardiac surgery. J Clin Invest 1995;95:1062–70.
9. Petrak RA, Balk RA, Bone RC. Prostaglandins, cyclo-oxygenase inhibitors, and thromboxane synthetase inhibitors in the pathogenesis of multiple systems organ failure. Crit Care Clin 1989;5:303–14.
10. Touqui L, Arbibe L. A role for phospholipase A2 in ARDS pathogenesis. Mol Med Today 1999;5:244–9.
11. Arbibe L, Koumanov K, Vial D, et al. Generation of lysophospholipids from surfactant in acute lung injury is mediated by type-II phospholipase A2 and inhibited by a direct surfactant protein A-phospholipase A2 protein interaction. J Clin Invest 1998;102:1152–60.
12. Torre-Amione G, Kapadia S, Lee J, et al. Tumor necrosis factor-α and tumor necrosis factor receptors in the failing human heart. Circulation 1996;93:704–11.
13. Wolbink GJ, Brouwer MC, Buysmann S, et al. CRP-mediated activation of complement in vivo: assessment by measuring circulating complement-C-reactive protein complexes. J Immunol 1996;157:473–9.
14. Nuijens JH, Abbink JJ, Wachtfogel YT, et al. Plasma elastase alpha 1-antitrypsin and lactoferrin in sepsis: evidence for neutrophils as mediators in fatal sepsis. J Lab Clin Med 1992;119:159–68.
15. Wolbink GJ, Schalkwijk C, Baars JW, et al. Therapy with interleukin-2 induces the systemic release of phospholipase-A2. Cancer Immunol Immunother 1995;41:287–92.
16. Redmond JM, Gillinov AM, Stuart RS, et al. Heparin-coated bypass circuits reduce pulmonary injury. Ann Thorac Surg 1993;56:474–9.
17. Magnusson L, Zemgulis V, Wicky S, et al. Atelectasis is a major cause of hypoxia and shunt after cardiopulmonary bypass: an experimental study. Anesthesiology 1997;87:1153–63.
18. Murakami M, Nakatani Y, Atsumi G, et al. Regulatory functions of phospholipase A2. Crit Rev Immunol 1997;17:225–83.
19. Kern H, Johnen W, Braun J, et al. Heparin induces release of phospholipase A2 into the splanchnic circulation. Anesth Analg 2000;91:528–32.
20. Davies GC, Sobel M, Saltzman EW. Elevated plasma fibrinopeptide A and thromboxane B2 levels during cardiopulmonary bypass. Circulation 1980;61:808–14.
21. Cybulsky AV, Takano T, Papillon J, McTavish AJ. Complement-induced phospholipase A2 in experimental membranous nephropathy. Kidney Int 2000;57:1052–62.
22. Friedman M, Selke FW, Wang SY, et al. Parameters of pulmonary injury after total or partial cardiopulmonary bypass. Circulation 1994;90(5 Pt 2):II262–8.
23. Cave AC, Manche A, Derias NW, Hearse DJ. Thromboxane A2 mediates pulmonary hypertension after cardiopulmonary bypass in the rabbit. J Thorac Cardiovasc Surg 1993;106:959–67.
24. Anderson BO, Brown JM, Bensard DD, et al. Reversible lung neutrophil accumulation can cause lung injury by elastase-mediated mechanisms. Surgery 1990;108:267–8.
25. Gu YJ, de Vries AJ, Boonstra PW, van Oeveren W. Leukocyte depletion results in improved lung function and reduced inflammatory response after cardiac surgery. J Thorac Cardiovasc Surg 1996;112:494–500.
26. Jansen NJ, van Oeveren W, van den Broek L, et al. Inhibition by dexamethasone of the reperfusion phenomena in cardiopulmonary bypass. J Thorac Cardiovasc Surg 1991;102:515–25.
27. Yared JP, Starr NJ, Torres FK, et al. Effects of single dose, postinduction dexamethasone on recovery after cardiac surgery. Ann Thorac Surg 2000;69:142–4.
28. Ashraf S, Tian Y, Cowan D, et al. “Low-dose” aprotinin modifies hemostasis but not proinflammatory cytokine release. Ann Thorac Surg 1997;63:68–73.
29. Bluman LG, Mosca L, Newman N, Simon DG. Preoperative smoking habits and post-operative pulmonary complications. Chest 1998;113:883–9.
30. Kotani N, Hashimoto H, Sessler D, et al. Smoking decreases alveolar macrophage function during anesthesia and surgery. Anesthesiology 2000;92:1268–77.
© 2004 International Anesthesia Research Society
31. Wildevuur ChRH, Jansen PGM, Bezemer PD, et al. Clinical evaluation of Duraflo®
II heparin treated extracorporeal circulation circuits (2nd version): the European Working Group on heparin coated extracorporeal circulation circuits. Eur J Cardiothorac Surg 1997;11:616–23.