- Email: [email protected]
Inflammatory Mediators in Adults Undergoing Cardiopulmonary Bypass: Comparison of Centrifugal and Roller Pumps
Inflammatory Mediators in Adults Undergoing Cardiopulmonary Bypass: Comparison of Centrifugal and Roller Pumps Saeed Ashraf, FRCS(C/Th), John Butler, FRCS(I), Yi Tian, MCh, Dahlia Cowan, BSc, Simon Lintin, FRCA, Nigel R. Saunders, FRCS, Kevin G. Watterson, FRACS, and Paul G. Martin, PhD Cardiothoracic Department, Killingbeck Hospital, Leeds, United Kingdom
Background. The nonocclusive centrifugal pump is used for cardiopulmonary bypass (CPB) and mechanical cardiac assistance. This study examined its impact on proinflammatory cytokine release. Methods. Forty-one patients undergoing elective coronary artery bypass grafting were randomized prospectively to either a standard roller pump group (n 5 21) or a centrifugal vortex pump group (n 5 20) for CPB. The two groups were well matched in age, sex, severity of disease, and duration of CPB and aortic cross-clamping. Plasma levels of the cytokines tumor necrosis factor-a, interleukin-1b, interleukin-6, and interleukin-8, as well as terminal complement, neutrophil counts, and leukocyte elastase, were analyzed before, during, and after CPB. Results. In both groups, traces of tumor necrosis factor-a were observed infrequently and interleukin-1b was not detected. Plasma levels of interleukin-6 and interleukin-8 increased during and after CPB, reaching a peak at 2 hours after protamine administration in both groups
before returning toward baseline at 24 hours. The release of interleukin-6 was significantly greater in the centrifugal group (p < 0.05), whereas the interleukin-8 concentration did not differ between the groups throughout the study period. Levels of terminal complement increased in both groups perioperatively, reaching a peak 30 minutes after protamine administration, whereas neutrophil counts and elastase peaked 2 hours after protamine administration. Plasma terminal complement, neutrophil counts, and elastase release were significantly higher in the centrifugal group (p < 0.05). Peak terminal complement correlated (r 5 0.64, p < 0.01) with peak elastase in the centrifugal group only. Conclusions. This study confirms the proinflammatory nature of CPB in adults and demonstrates that use of the centrifugal pump induces a greater systemic inflammatory response than use of the standard roller pump. (Ann Thorac Surg 1998;65:480 – 4) © 1998 by The Society of Thoracic Surgeons
T
(IL-1b) in response to CPB have been inconsistent [3, 6 –10]. Plasma levels of the cytokines interleukin-6 (IL-6) and interleukin-8 (IL-8) are elevated after CPB and may be associated with post-CPB cardiac dysfunction [5]. It has been proposed that if cytokine release could be reduced during CPB, some of the hazardous clinical effects of CPB might be avoided. Both roller and centrifugal pumps are used in extracorporeal circuits. The intrinsic safety features of centrifugal pumps contribute to their widespread use for long-term extracorporeal support as well as for routine cardiac operations [11, 12]. Studies have shown beneficial effects of the centrifugal pump over the roller pump in terms of reduced hemolysis, reduced neutrophil and complement activation, and improved platelet preservation [11–14]. The aim of this prospective, randomized study was to evaluate the systemic inflammatory profiles associated with the use of the standard roller pump and the centrifugal vortex pump in adult patients undergoing elective coronary artery bypass grafting.
he systemic inflammatory response that occurs to a varying extent in all patients after cardiopulmonary bypass (CPB) continues to be a significant cause of morbidity and occasional mortality. Postoperative bleeding, infection, systemic organ dysfunction, and acute respiratory distress syndrome are complications characteristic of CPB with the occurrence of a “wholebody inflammatory reaction” [1, 2]. Several studies have shown that this involves complement activation, neutrophil degranulation, and free radical production, together with cytokine release, leading to an increase in capillary permeability, the accumulation of interstitial fluid, and organ dysfunction [3–5]. Proinflammatory cytokines are potent intercellular signaling molecules and, particularly when produced in excess, potential mediators of vascular injury and organ dysfunction [2–5]. Reports on the release of tumor necrosis factor-a (TNF-a) and interleukin-1b
Accepted for publication Aug 20, 1997. Address reprint requests to Dr Ashraf, Cardiothoracic Surgery, Leeds General Infirmary, Great George St, Leeds, UK, LS1 3UQ.
© 1998 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
0003-4975/98/$19.00 PII S0003-4975(97)01349-0
Ann Thorac Surg 1998;65:480 – 4
Material and Methods After ethical committee approval and informed consent were obtained, 41 patients undergoing coronary artery bypass grafting were randomized to either a standard roller pump group (n 5 21) or a nonocclusive centrifugal pump group (n 5 20) for CPB. All patients had effortinduced angina pectoris refractory to maximal antianginal therapy, multivessel coronary artery disease (.70% vessel occlusion), and an ejection fraction of greater than 0.40. Exclusion criteria were unstable angina, myocardial infarction within the previous 3 months, reoperation, diabetes mellitus, liver or kidney failure, chronic obstructive airways disease, and oral anticoagulant or immunosuppressant therapy. The techniques of anesthesia and CPB were standardized. After premedication, anesthesia was induced with fentanyl (30 mg/kg, given intravenously) and muscle relaxation was achieved with pancuronium bromide (0.1 to 0.2 mg/kg, given intravenously). Anesthesia was supported by the inhalation of 1% isoflurane. The extracorporeal circuit consisted of either a Stockert roller pump (Stockert Instrumente, Munich, Germany) or a centrifugal vortex pump (Medtronic Biomedicus Inc, Eden Prairie, MN), a hollow-fiber membrane oxygenator (D703A; Dideco), polyvinylchloride tubing, a two-stage venous cannula, and a venous reservoir. The only difference in the entire perfusion circuit was the arterial pump. In both groups, the cardiotomy reservoir was integral to the venous reservoir and the cardiotomy suction was provided by three roller pumps. Patients were heparinized just before the institution of CPB with 300 IU/kg, with additional doses given as necessary to maintain the activated clotting time at greater than 480 seconds. Nonpulsatile extracorporeal circulation was initiated at flows of 2.4 to 2.6 L z m22 z min21. Moderate systemic hypothermia 28° to 30°C by nasopharyngeal probe) was used uniformly. After aortic cross-clamping, cardiac arrest was achieved by the antegrade infusion of 1 L of cold-blood cardioplegic solution supplemented by topical saline slush. All distal anastomoses were performed, the cross-clamp was removed, and the proximal anastomoses to the aorta were completed during the rewarming period. Extracorporeal circulation was terminated at a nasopharyngeal temperature of 37°C. Heparin was neutralized after CPB with protamine sulfate (1 mg/100 IU of heparin). The operations were performed by two surgeons using a similar operative technique and the same perfusion protocol.
Biochemical Measurements Venous blood samples (20 mL) were drawn and taken into sodium citrate (0.32% wt/vol, 1:9 parts blood) immediately after the induction of anesthesia, 5 minutes after the onset of CPB, at the end of CPB, and 30 minutes, 2 hours, and 24 hours after the administration of protamine sulfate. Specimens were centrifuged (10 minutes, 3,000 rpm, 4°C) immediately to obtain plasma, which was stored at 280°C before batch analysis. Enzyme-linked immunosorbent assay techniques were used to measure
ASHRAF ET AL BYPASS PUMPS AND INFLAMMATORY MEDIATORS
481
Table 1. Clinical Characteristics of the 41 Patients Who Underwent Coronary Artery Bypass Graftinga Study Group Characteristic No. of patients Sex ratio (male:female) Age (y) CPB duration (min) Cross-clamp duration (min) No. of distal grafts
Centrifugal Pump
Roller Pump
20 16:4 61 (37–72) 108 (51–155) 54 (34 – 87) 4 (2– 4)
21 17:4 67 (51–79) 102 (46 –143) 45 (24 – 80) 4 (2–5)
a
Data are presented as medians, with ranges in parentheses, except as noted; p 5 not significant. CPB 5 cardiopulmonary bypass.
each of the cytokines (R & D Systems, Minneapolis, MN) and terminal complement (C5b-9; Quidel, San Diego, CA). Plasma neutrophil elastase concentrations were determined using an autoanalyzer technique (E Merck Diagnostica, Darmstadt, Germany). The limit of sensitivity of each assay undertaken was as follows: TNF-a 5 5 pg/mL, IL-1b and IL-6 5 3 pg/mL, IL-8 5 20 pg/mL, C5b-9 5 16 ng/mL, and elastase 5 20 ng/mL. Neutrophil counts were measured at each sampling point. Results were not adjusted for hemodilution. Blood loss and blood transfusion were recorded until 24 hours after the operation. Packed red blood cells were infused when the hematocrit level was less than 30%. Cardiac output was measured by the thermodilution technique through a Swan-Ganz catheter (Baxter Edwards, Irvine, CA) from the mean of three readings. Cardiac index was calculated and reported in liters per square meter. Systemic and pulmonary vascular resistance were calculated by conventional formulas and were not indexed.
Statistical Analysis Results are expressed as the median, with the range given in parentheses. Area under the curve analysis [15] assessed the overall difference between the groups over the course of the study. Nonparametric tests were used to assess at which time points differences between the groups were most marked. Spearman’s rank correlation coefficient was used to assess association (Statistica software). In all cases, p less than 0.05 was considered statistically significant.
Results The clinical characteristics of the 41 patients are summarized in Table 1. There were no statistically significant demographic differences observed between the centrifugal and roller pump groups in terms of age, CPB and cross-clamp times, or number of grafts. No patient required exploration for postoperative bleeding and there were no operative deaths or adverse complications. All patients were discharged from the intensive care unit on the first postoperative day. The total 24-hour postoperative blood loss into the chest drain was similar in both groups: 727 mL (range,
482
ASHRAF ET AL BYPASS PUMPS AND INFLAMMATORY MEDIATORS
Fig 1. Changes in the circulating median concentration of interleukin-6 (IL-6) with time in patients who underwent cardiopulmonary bypass (CPB) with a roller (squares) or a centrifugal (triangles) pump. Error bars relate to the 25th to 75th interquartile range. (post-p 5 after protamine administration.)
Ann Thorac Surg 1998;65:480 – 4
Fig 2. Changes in the circulating median concentration of interleukin-8 (IL-8) with time in patients who underwent cardiopulmonary bypass (CPB) with a roller (squares) or a centrifugal (triangles) pump. Error bars relate to the 25th to 75th interquartile range. (NS 5 not significant; post-p 5 after protamine administration.)
C5b-9 Complex 135 to 1,900 mL) in the centrifugal group versus 722 mL (range, 345 to 1,140 mL) in the roller group. No statistically significant difference was found between the two groups in terms of transfusion requirements: 380 mL (range, 0 to 820 mL) in the centrifugal group versus 410 mL (range, 0 to 1,120 mL) in the roller group.
Tumor Necrosis Factor-a and Interleukin-1b Plasma levels of TNF-a did not alter significantly with CPB in either group. Low levels of TNF-a were detected in only 3 patients in the centrifugal group and 5 patients in the roller group at variable time points. Interleukin-1b was not detectable at all in either group.
Interleukin-6 The initial baseline level of IL-6 was below the sensitivity threshold of the assay in both groups but rose progressively during CPB (Fig 1). The level peaked 2 hours after protamine administration at 341 pg/mL (range, 109 to 509 pg/mL) in the centrifugal group and 260 pg/mL (range, 103 to 522 pg/mL) in the roller group, then returned toward baseline in most of the patients after 24 hours. Interleukin-6 release was significantly higher in the centrifugal group than in the roller group (p , 0.05). Moreover, at the end of CPB, IL-6 levels were significantly higher (p 5 0.02) in the centrifugal group compared with the roller group.
In both groups, C5b-9 levels increased significantly over baseline during CPB, but they were significantly higher in the centrifugal group than in the roller group during and after CPB (p , 0.05). Moreover, C5b-9 levels in the centrifugal group were significantly higher than those in the roller group both at the end of CPB (p 5 0.03) and 30 minutes after protamine administration (p 5 0.02). The levels peaked at 765 ng/mL (range, 195 to 1,202 ng/mL) in the centrifugal group and 509 ng/mL (range, 92 to 863 ng/mL) in the roller group before returning to preoperative levels after 24 hours (Fig 3).
Neutrophil Counts The initial neutrophil counts did not differ significantly between the two groups. After initial CPB hemodilution, there was a progressive rise in the neutrophil count initiated during rewarming. The neutrophil count reached a significantly higher level in the centrifugal group (8.15 3 109 [range, 4.0 to 14.89 3 109]) than in the roller group (6.2 3 109 [range, 2.4 to 11.6 3 109]) (p 5 0.03). Moreover, the total neutrophil count during and after
Interleukin-8 Changes in plasma concentrations of IL-8 with time are shown in Figure 2. The peak level of IL-8 occurred 2 hours after protamine administration at 143 pg/mL (range, 42 to 538 pg/mL) in the centrifugal group and 150 pg/mL (range, 23 to 987 pg/mL) in the roller group (p 5 not significant), then fell toward the preoperative level in both groups after 24 hours. There was no difference in the release of IL-8 between the two groups (p 5 not significant).
Fig 3. Changes in the circulating median concentration of terminal complement (C5b-9) with time in patients who underwent cardiopulmonary bypass (CPB) with a roller (squares) or a centrifugal (triangles) pump. Error bars relate to the 25th to 75th interquartile range. (post-p 5 after protamine administration.)
Ann Thorac Surg 1998;65:480 – 4
ASHRAF ET AL BYPASS PUMPS AND INFLAMMATORY MEDIATORS
483
Comment
Fig 4. Changes in the circulating concentration of elastase with time in patients who underwent cardiopulmonary bypass (CPB) with a roller (squares) or a centrifugal (triangles) pump. Error bars relate to the 25th to 75th interquartile range. (post-p 5 after protamine administration.)
CPB was significantly higher in the centrifugal group than in the roller group (p , 0.05)
Plasma Leukocyte Elastase In both groups, elastase levels increased during and after CPB, peaking 2 hours after protamine administration at 202 ng/mL (range, 96 to 560 ng/mL) in the centrifugal group and 183 ng/mL (range, 67 to 532 ng/mL) in the roller group (Fig 4). Plasma leukocyte elastase release during and after CPB was significantly higher in the centrifugal group than in the roller group (p , 0.05). Moreover, in the centrifugal group, the elastase level was significantly higher than in the roller group both at the end of CPB (p 5 0.03) and 30 minutes after protamine administration (p 5 0.02). In the centrifugal group, peak levels of elastase and C5b-9 were correlated significantly with each other (r 5 0.64, p , 0.01) (Fig 5). There was no correlation between elastase and IL-8 levels in either group, and there was no correlation between IL-6 levels and hemodynamic parameters such as systemic vascular resistance and cardiac index.
Fig 5. Correlation of circulating peak levels of terminal complement (C5b-9) and elastase in the centrifugal group.
This study confirms the proinflammatory response to CPB in adults. The plasma markers IL-6, C5b-9, and leukocyte elastase are more pronounced immediately after CPB with the use of centrifugal compared with roller pumps. Plasma TNF-a and IL-1b have been detected in some studies [6 – 8], but not in others [4, 9, 10]. Our study did not detect circulating IL-1b and did not find significant levels of TNF-a in either group. The short half-life of TNF-a and IL-1b, the presence of soluble receptors, and differences in the methods of cytokine measurement used may explain the discrepancies between different studies. Further, failure to detect plasma levels of these cytokines does not exclude a local production and paracrine function, although it makes it less likely that they play an important role in systemic inflammatory responses to CPB. Interleukin-6 is a key mediator in the acute-phase response to injury or infection, inducing the synthesis of hepatic acute-phase proteins [2, 16, 17]. It is thought to be a useful marker of the degree of tissue injury. Elevation of plasma IL-6 levels occurs both after cardiac operations with CPB [16] and after major noncardiac operations [17–19]. In the present study, IL-6 release was significantly higher in the centrifugal group compared with the roller group. This implies the induction of a greater acute-phase response and suggests a greater degree of tissue trauma with the centrifugal pump. The role of neutrophil activation in lung and myocardial injury after CPB has been well documented. This led us to measure IL-8, a potent neutrophil chemotactic and activating factor [3, 5, 20]. Unlike IL-6, there was no significant difference in IL-8 release between the centrifugal and roller groups. A possible explanation for this disparity in the release of these two cytokines is that IL-8 release occurs mainly in situations of ischemiareperfusion injury [4, 20, 21], whereas IL-6 release reflects the body’s response to any kind of acute insult. Given comparable CPB durations, the production of IL-8 after reperfusion would be expected to be similar in the two groups. Activation of the complement cascade, predominantly through the alternative pathway, occurs during CPB [1, 2] and results in the formation of the terminal complement complex C5b-9. Additional activation of the complement cascade by heparin-protamine complexes is reflected by further increments in C5b-9 levels after protamine infusion. In the present study, concentrations of C5b-9 were higher at the end of CPB and 30 minutes after protamine administration in patients perfused with centrifugal rather than standard roller pumps. Complement fractions activate neutrophils with release of their proteolytic enzymes, which can contribute to the adverse effects of extracorporeal circulation. To monitor neutrophil activation, leukocyte elastase was measured in plasma, where it is present in an inactive form complexed with a1-antiprotease [2, 4, 22]. The release of elastase was significantly higher in the centrifugal group compared with the roller group, and elastase
484
ASHRAF ET AL BYPASS PUMPS AND INFLAMMATORY MEDIATORS
levels were more pronounced immediately after CPB. Further, peak levels of elastase correlated with peak levels of C5b-9 in the centrifugal group. This is consistent with complement stimulation of neutrophils leading to elastase release. If this occurs at the endothelium (ie, that of the reperfused lung) as a result of IL-8 chemotaxis, it could contribute to endothelial injury and capillary leak after CPB [3–5]. Although we observed no statistically significant correlation between elastase levels and IL-8 levels, an association has been found by others [3, 4]. Clearly, the complex interactions of these molecules and cells at the endothelial level cannot be reflected accurately in statistical correlations of plasma levels. In contrast to our results, another clinical study [12] with a similar duration of CPB demonstrated less complement activation with the centrifugal pump, although in a much smaller number of patients. An in vitro study [13] showed less C5b-9 release with the centrifugal pump during extracorporeal circulation that exceeded 8 hours. It is possible that a roller pump induces less of an inflammatory response over a short duration of extracorporeal circulation, whereas a centrifugal pump is less traumatic over a longer duration. There were no differences in clinical outcomes between the two study groups. This finding does not preclude the potential clinical relevance of our biochemical findings. The magnitude of complement activation and cytokine release during and after CPB has been reported to be correlated significantly with the postoperative clinical outcome [5, 23, 24]. In summary, the greater production of IL-6, C5b-9, and elastase with centrifugal pumps suggests that these pumps are less “biocompatible” for short-term extracorporeal circulation. We thank Mr Martin Ward for excellent technical assistance.
References 1. Kirklin JK, Westaby S, Blackstone EH, Kirklin JW, Chenoweth DE, Pacifico AD. Complement and the damaging effects of cardiopulmonary bypass. J Thorac Cardiovasc Surg 1983;86:845–57. 2. Butler J, Rocker GM, Westaby S. Inflammatory response to cardiopulmonary bypass. Ann Thorac Surg 1993;55:552–9. 3. Finn A, Naik S, Klein N, Levinsky RJ, Strobel S, Elliott M. Interleukin-8 release and neutrophil degranulation after pediatric cardiopulmonary bypass. J Thorac Cardiovasc Surg 1993;105:234– 41. 4. George JF. Cytokines and mechanisms of capillary leakage after cardiopulmonary bypass. J Thorac Cardiovasc Surg 1993;106:566–7. 5. Hennein HA, Ebba H, Rodriguez JL, et al. Relationship of the proinflammatory cytokines to myocardial ischemia and dysfunction after uncomplicated coronary revascularization. J Thorac Cardiovasc Surg 1994;108:626–35. 6. Haeffner-Cavaillon N, Roussellier N, Ponzio O, et al. Induction of interleukin-1 production in patients undergoing cardiopulmonary bypass. J Thorac Cardiovasc Surg 1989;98: 1100– 6.
Ann Thorac Surg 1998;65:480 – 4
7. Jansen NJG, van Oeveren W, Gu YJ, von Vliet MH, Eijsman LL, Wildevuur CRH. Endotoxin release and tumor necrosis factor formation during cardiopulmonary bypass. Ann Thorac Surg 1992;54:744– 8. 8. Tonz M, Mihaljevic T, von Segesser LK, et al. Normothermia versus hypothermia during cardiopulmonary bypass: a randomized, controlled trial. Ann Thorac Surg 1995;59:137– 43. 9. Butler J, Pillai R, Rocker GM, Westaby S, Parker D, Shale DJ. Effect of cardiopulmonary bypass on systemic release of neutrophil elastase and tumor necrosis factor. J Thorac Cardiovasc Surg 1993;105:25–30. 10. Frering B, Philip I, Dehoux M, Rolland C, Langlois JM, Desmonts JM. Circulating cytokines in patients undergoing normothermic cardiopulmonary bypass. J Thorac Cardiovasc Surg 1994;108:636– 41. 11. Hoerr HR, Kraemer MF, Williams JL, et al. In vitro comparison of the blood handling by the constrained vortex and twin roller blood pumps. J Extracorpor Technol 1987;19: 316–20. 12. Weeldon DR, Bethune DW, Gill RD. Vortex pumping for routine cardiac surgery: a comparative study. Perfusion 1990; 5:135– 43. 13. Moen O, Fosse E, Braten J, et al. Roller and centrifugal pumps compared in vitro with regard to haemolysis, granulocyte and complement activation. Perfusion 1994;9:109–17. 14. Nishinaka T, Nishida H, Endo M, Koyanagi H. Less platelet damage in the curved vane centrifugal pump: a comparative study with the roller pump in open heart surgery. Artif Organs 1994;18:687–90. 15. Matthews JNS, Altman DG, Campbell MJ, Royston P. Analysis of serial measurements in medical research. Br Med J 1990;300:230–5. 16. Butler J, Chang GL, Baigrie RJ, Pillai R, Westaby S, Rocker GM. Cytokine responses to cardiopulmonary bypass with membrane and bubble oxygenation. Ann Thorac Surg 1992; 53:833– 8. 17. Nijsten MW, DeGroot ER, Tenduis HJ. Response of serum interleukin-6 in patients undergoing elective surgery of varying severity. Clin Sci (Colch) 1990;79:1611–5. 18. Cruickshank AM, Fraser WD, Burns HJ, Van Damme J, Shenkin A. Response of serum interleukin-6 in patients undergoing elective surgery of varying severity. Clin Sci (Colch) 1990;79:161–5. 19. Ohzato H, Yoshizaki K, Nishimoto N, et al. Interleukin-6 as a new indicator of inflammatory status: detection of serum levels of interleukin-6 and C-reactive protein after surgery. Surgery 1992;111:201–9. 20. Herbert CA, Baker JB. Interleukin-8: a review. Cancer Invest 1993;11:743–50. 21. Kawamura T, Wakusawa R, Okada K, Inada S. Elevation of cytokines during open heart surgery with cardiopulmonary bypass: participation of interleukin-8 and 6 in reperfusion injury. Can J Anaesth 1993;40:1016–21. 22. Faymonville ME, Pincemail J, Duchateau J, et al. Myeloperoxidase and elastase as markers of leukocyte activation during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1991;102:309–17. 23. Seghaye MC, Duchateau J, Grabitz RG, et al. Complement activation in infants and children: relation to postoperative multiple system organ failure. J Thorac Cardiovasc Surg 1993;106:978– 87. 24. Jansen PGM, Velthuis H, Huybregts RAJM, et al. Reduced complement activation and improved postoperative performance after cardiopulmonary bypass with heparin coated circuits. J Thorac Cardiovasc Surg 1995;110:829–34.
Background. The nonocclusive centrifugal pump is used for cardiopulmonary bypass (CPB) and mechanical cardiac assistance. This study examined its impact on proinflammatory cytokine release. Methods. Forty-one patients undergoing elective coronary artery bypass grafting were randomized prospectively to either a standard roller pump group (n 5 21) or a centrifugal vortex pump group (n 5 20) for CPB. The two groups were well matched in age, sex, severity of disease, and duration of CPB and aortic cross-clamping. Plasma levels of the cytokines tumor necrosis factor-a, interleukin-1b, interleukin-6, and interleukin-8, as well as terminal complement, neutrophil counts, and leukocyte elastase, were analyzed before, during, and after CPB. Results. In both groups, traces of tumor necrosis factor-a were observed infrequently and interleukin-1b was not detected. Plasma levels of interleukin-6 and interleukin-8 increased during and after CPB, reaching a peak at 2 hours after protamine administration in both groups
before returning toward baseline at 24 hours. The release of interleukin-6 was significantly greater in the centrifugal group (p < 0.05), whereas the interleukin-8 concentration did not differ between the groups throughout the study period. Levels of terminal complement increased in both groups perioperatively, reaching a peak 30 minutes after protamine administration, whereas neutrophil counts and elastase peaked 2 hours after protamine administration. Plasma terminal complement, neutrophil counts, and elastase release were significantly higher in the centrifugal group (p < 0.05). Peak terminal complement correlated (r 5 0.64, p < 0.01) with peak elastase in the centrifugal group only. Conclusions. This study confirms the proinflammatory nature of CPB in adults and demonstrates that use of the centrifugal pump induces a greater systemic inflammatory response than use of the standard roller pump. (Ann Thorac Surg 1998;65:480 – 4) © 1998 by The Society of Thoracic Surgeons
T
(IL-1b) in response to CPB have been inconsistent [3, 6 –10]. Plasma levels of the cytokines interleukin-6 (IL-6) and interleukin-8 (IL-8) are elevated after CPB and may be associated with post-CPB cardiac dysfunction [5]. It has been proposed that if cytokine release could be reduced during CPB, some of the hazardous clinical effects of CPB might be avoided. Both roller and centrifugal pumps are used in extracorporeal circuits. The intrinsic safety features of centrifugal pumps contribute to their widespread use for long-term extracorporeal support as well as for routine cardiac operations [11, 12]. Studies have shown beneficial effects of the centrifugal pump over the roller pump in terms of reduced hemolysis, reduced neutrophil and complement activation, and improved platelet preservation [11–14]. The aim of this prospective, randomized study was to evaluate the systemic inflammatory profiles associated with the use of the standard roller pump and the centrifugal vortex pump in adult patients undergoing elective coronary artery bypass grafting.
he systemic inflammatory response that occurs to a varying extent in all patients after cardiopulmonary bypass (CPB) continues to be a significant cause of morbidity and occasional mortality. Postoperative bleeding, infection, systemic organ dysfunction, and acute respiratory distress syndrome are complications characteristic of CPB with the occurrence of a “wholebody inflammatory reaction” [1, 2]. Several studies have shown that this involves complement activation, neutrophil degranulation, and free radical production, together with cytokine release, leading to an increase in capillary permeability, the accumulation of interstitial fluid, and organ dysfunction [3–5]. Proinflammatory cytokines are potent intercellular signaling molecules and, particularly when produced in excess, potential mediators of vascular injury and organ dysfunction [2–5]. Reports on the release of tumor necrosis factor-a (TNF-a) and interleukin-1b
Accepted for publication Aug 20, 1997. Address reprint requests to Dr Ashraf, Cardiothoracic Surgery, Leeds General Infirmary, Great George St, Leeds, UK, LS1 3UQ.
© 1998 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
0003-4975/98/$19.00 PII S0003-4975(97)01349-0
Ann Thorac Surg 1998;65:480 – 4
Material and Methods After ethical committee approval and informed consent were obtained, 41 patients undergoing coronary artery bypass grafting were randomized to either a standard roller pump group (n 5 21) or a nonocclusive centrifugal pump group (n 5 20) for CPB. All patients had effortinduced angina pectoris refractory to maximal antianginal therapy, multivessel coronary artery disease (.70% vessel occlusion), and an ejection fraction of greater than 0.40. Exclusion criteria were unstable angina, myocardial infarction within the previous 3 months, reoperation, diabetes mellitus, liver or kidney failure, chronic obstructive airways disease, and oral anticoagulant or immunosuppressant therapy. The techniques of anesthesia and CPB were standardized. After premedication, anesthesia was induced with fentanyl (30 mg/kg, given intravenously) and muscle relaxation was achieved with pancuronium bromide (0.1 to 0.2 mg/kg, given intravenously). Anesthesia was supported by the inhalation of 1% isoflurane. The extracorporeal circuit consisted of either a Stockert roller pump (Stockert Instrumente, Munich, Germany) or a centrifugal vortex pump (Medtronic Biomedicus Inc, Eden Prairie, MN), a hollow-fiber membrane oxygenator (D703A; Dideco), polyvinylchloride tubing, a two-stage venous cannula, and a venous reservoir. The only difference in the entire perfusion circuit was the arterial pump. In both groups, the cardiotomy reservoir was integral to the venous reservoir and the cardiotomy suction was provided by three roller pumps. Patients were heparinized just before the institution of CPB with 300 IU/kg, with additional doses given as necessary to maintain the activated clotting time at greater than 480 seconds. Nonpulsatile extracorporeal circulation was initiated at flows of 2.4 to 2.6 L z m22 z min21. Moderate systemic hypothermia 28° to 30°C by nasopharyngeal probe) was used uniformly. After aortic cross-clamping, cardiac arrest was achieved by the antegrade infusion of 1 L of cold-blood cardioplegic solution supplemented by topical saline slush. All distal anastomoses were performed, the cross-clamp was removed, and the proximal anastomoses to the aorta were completed during the rewarming period. Extracorporeal circulation was terminated at a nasopharyngeal temperature of 37°C. Heparin was neutralized after CPB with protamine sulfate (1 mg/100 IU of heparin). The operations were performed by two surgeons using a similar operative technique and the same perfusion protocol.
Biochemical Measurements Venous blood samples (20 mL) were drawn and taken into sodium citrate (0.32% wt/vol, 1:9 parts blood) immediately after the induction of anesthesia, 5 minutes after the onset of CPB, at the end of CPB, and 30 minutes, 2 hours, and 24 hours after the administration of protamine sulfate. Specimens were centrifuged (10 minutes, 3,000 rpm, 4°C) immediately to obtain plasma, which was stored at 280°C before batch analysis. Enzyme-linked immunosorbent assay techniques were used to measure
ASHRAF ET AL BYPASS PUMPS AND INFLAMMATORY MEDIATORS
481
Table 1. Clinical Characteristics of the 41 Patients Who Underwent Coronary Artery Bypass Graftinga Study Group Characteristic No. of patients Sex ratio (male:female) Age (y) CPB duration (min) Cross-clamp duration (min) No. of distal grafts
Centrifugal Pump
Roller Pump
20 16:4 61 (37–72) 108 (51–155) 54 (34 – 87) 4 (2– 4)
21 17:4 67 (51–79) 102 (46 –143) 45 (24 – 80) 4 (2–5)
a
Data are presented as medians, with ranges in parentheses, except as noted; p 5 not significant. CPB 5 cardiopulmonary bypass.
each of the cytokines (R & D Systems, Minneapolis, MN) and terminal complement (C5b-9; Quidel, San Diego, CA). Plasma neutrophil elastase concentrations were determined using an autoanalyzer technique (E Merck Diagnostica, Darmstadt, Germany). The limit of sensitivity of each assay undertaken was as follows: TNF-a 5 5 pg/mL, IL-1b and IL-6 5 3 pg/mL, IL-8 5 20 pg/mL, C5b-9 5 16 ng/mL, and elastase 5 20 ng/mL. Neutrophil counts were measured at each sampling point. Results were not adjusted for hemodilution. Blood loss and blood transfusion were recorded until 24 hours after the operation. Packed red blood cells were infused when the hematocrit level was less than 30%. Cardiac output was measured by the thermodilution technique through a Swan-Ganz catheter (Baxter Edwards, Irvine, CA) from the mean of three readings. Cardiac index was calculated and reported in liters per square meter. Systemic and pulmonary vascular resistance were calculated by conventional formulas and were not indexed.
Statistical Analysis Results are expressed as the median, with the range given in parentheses. Area under the curve analysis [15] assessed the overall difference between the groups over the course of the study. Nonparametric tests were used to assess at which time points differences between the groups were most marked. Spearman’s rank correlation coefficient was used to assess association (Statistica software). In all cases, p less than 0.05 was considered statistically significant.
Results The clinical characteristics of the 41 patients are summarized in Table 1. There were no statistically significant demographic differences observed between the centrifugal and roller pump groups in terms of age, CPB and cross-clamp times, or number of grafts. No patient required exploration for postoperative bleeding and there were no operative deaths or adverse complications. All patients were discharged from the intensive care unit on the first postoperative day. The total 24-hour postoperative blood loss into the chest drain was similar in both groups: 727 mL (range,
482
ASHRAF ET AL BYPASS PUMPS AND INFLAMMATORY MEDIATORS
Fig 1. Changes in the circulating median concentration of interleukin-6 (IL-6) with time in patients who underwent cardiopulmonary bypass (CPB) with a roller (squares) or a centrifugal (triangles) pump. Error bars relate to the 25th to 75th interquartile range. (post-p 5 after protamine administration.)
Ann Thorac Surg 1998;65:480 – 4
Fig 2. Changes in the circulating median concentration of interleukin-8 (IL-8) with time in patients who underwent cardiopulmonary bypass (CPB) with a roller (squares) or a centrifugal (triangles) pump. Error bars relate to the 25th to 75th interquartile range. (NS 5 not significant; post-p 5 after protamine administration.)
C5b-9 Complex 135 to 1,900 mL) in the centrifugal group versus 722 mL (range, 345 to 1,140 mL) in the roller group. No statistically significant difference was found between the two groups in terms of transfusion requirements: 380 mL (range, 0 to 820 mL) in the centrifugal group versus 410 mL (range, 0 to 1,120 mL) in the roller group.
Tumor Necrosis Factor-a and Interleukin-1b Plasma levels of TNF-a did not alter significantly with CPB in either group. Low levels of TNF-a were detected in only 3 patients in the centrifugal group and 5 patients in the roller group at variable time points. Interleukin-1b was not detectable at all in either group.
Interleukin-6 The initial baseline level of IL-6 was below the sensitivity threshold of the assay in both groups but rose progressively during CPB (Fig 1). The level peaked 2 hours after protamine administration at 341 pg/mL (range, 109 to 509 pg/mL) in the centrifugal group and 260 pg/mL (range, 103 to 522 pg/mL) in the roller group, then returned toward baseline in most of the patients after 24 hours. Interleukin-6 release was significantly higher in the centrifugal group than in the roller group (p , 0.05). Moreover, at the end of CPB, IL-6 levels were significantly higher (p 5 0.02) in the centrifugal group compared with the roller group.
In both groups, C5b-9 levels increased significantly over baseline during CPB, but they were significantly higher in the centrifugal group than in the roller group during and after CPB (p , 0.05). Moreover, C5b-9 levels in the centrifugal group were significantly higher than those in the roller group both at the end of CPB (p 5 0.03) and 30 minutes after protamine administration (p 5 0.02). The levels peaked at 765 ng/mL (range, 195 to 1,202 ng/mL) in the centrifugal group and 509 ng/mL (range, 92 to 863 ng/mL) in the roller group before returning to preoperative levels after 24 hours (Fig 3).
Neutrophil Counts The initial neutrophil counts did not differ significantly between the two groups. After initial CPB hemodilution, there was a progressive rise in the neutrophil count initiated during rewarming. The neutrophil count reached a significantly higher level in the centrifugal group (8.15 3 109 [range, 4.0 to 14.89 3 109]) than in the roller group (6.2 3 109 [range, 2.4 to 11.6 3 109]) (p 5 0.03). Moreover, the total neutrophil count during and after
Interleukin-8 Changes in plasma concentrations of IL-8 with time are shown in Figure 2. The peak level of IL-8 occurred 2 hours after protamine administration at 143 pg/mL (range, 42 to 538 pg/mL) in the centrifugal group and 150 pg/mL (range, 23 to 987 pg/mL) in the roller group (p 5 not significant), then fell toward the preoperative level in both groups after 24 hours. There was no difference in the release of IL-8 between the two groups (p 5 not significant).
Fig 3. Changes in the circulating median concentration of terminal complement (C5b-9) with time in patients who underwent cardiopulmonary bypass (CPB) with a roller (squares) or a centrifugal (triangles) pump. Error bars relate to the 25th to 75th interquartile range. (post-p 5 after protamine administration.)
Ann Thorac Surg 1998;65:480 – 4
ASHRAF ET AL BYPASS PUMPS AND INFLAMMATORY MEDIATORS
483
Comment
Fig 4. Changes in the circulating concentration of elastase with time in patients who underwent cardiopulmonary bypass (CPB) with a roller (squares) or a centrifugal (triangles) pump. Error bars relate to the 25th to 75th interquartile range. (post-p 5 after protamine administration.)
CPB was significantly higher in the centrifugal group than in the roller group (p , 0.05)
Plasma Leukocyte Elastase In both groups, elastase levels increased during and after CPB, peaking 2 hours after protamine administration at 202 ng/mL (range, 96 to 560 ng/mL) in the centrifugal group and 183 ng/mL (range, 67 to 532 ng/mL) in the roller group (Fig 4). Plasma leukocyte elastase release during and after CPB was significantly higher in the centrifugal group than in the roller group (p , 0.05). Moreover, in the centrifugal group, the elastase level was significantly higher than in the roller group both at the end of CPB (p 5 0.03) and 30 minutes after protamine administration (p 5 0.02). In the centrifugal group, peak levels of elastase and C5b-9 were correlated significantly with each other (r 5 0.64, p , 0.01) (Fig 5). There was no correlation between elastase and IL-8 levels in either group, and there was no correlation between IL-6 levels and hemodynamic parameters such as systemic vascular resistance and cardiac index.
Fig 5. Correlation of circulating peak levels of terminal complement (C5b-9) and elastase in the centrifugal group.
This study confirms the proinflammatory response to CPB in adults. The plasma markers IL-6, C5b-9, and leukocyte elastase are more pronounced immediately after CPB with the use of centrifugal compared with roller pumps. Plasma TNF-a and IL-1b have been detected in some studies [6 – 8], but not in others [4, 9, 10]. Our study did not detect circulating IL-1b and did not find significant levels of TNF-a in either group. The short half-life of TNF-a and IL-1b, the presence of soluble receptors, and differences in the methods of cytokine measurement used may explain the discrepancies between different studies. Further, failure to detect plasma levels of these cytokines does not exclude a local production and paracrine function, although it makes it less likely that they play an important role in systemic inflammatory responses to CPB. Interleukin-6 is a key mediator in the acute-phase response to injury or infection, inducing the synthesis of hepatic acute-phase proteins [2, 16, 17]. It is thought to be a useful marker of the degree of tissue injury. Elevation of plasma IL-6 levels occurs both after cardiac operations with CPB [16] and after major noncardiac operations [17–19]. In the present study, IL-6 release was significantly higher in the centrifugal group compared with the roller group. This implies the induction of a greater acute-phase response and suggests a greater degree of tissue trauma with the centrifugal pump. The role of neutrophil activation in lung and myocardial injury after CPB has been well documented. This led us to measure IL-8, a potent neutrophil chemotactic and activating factor [3, 5, 20]. Unlike IL-6, there was no significant difference in IL-8 release between the centrifugal and roller groups. A possible explanation for this disparity in the release of these two cytokines is that IL-8 release occurs mainly in situations of ischemiareperfusion injury [4, 20, 21], whereas IL-6 release reflects the body’s response to any kind of acute insult. Given comparable CPB durations, the production of IL-8 after reperfusion would be expected to be similar in the two groups. Activation of the complement cascade, predominantly through the alternative pathway, occurs during CPB [1, 2] and results in the formation of the terminal complement complex C5b-9. Additional activation of the complement cascade by heparin-protamine complexes is reflected by further increments in C5b-9 levels after protamine infusion. In the present study, concentrations of C5b-9 were higher at the end of CPB and 30 minutes after protamine administration in patients perfused with centrifugal rather than standard roller pumps. Complement fractions activate neutrophils with release of their proteolytic enzymes, which can contribute to the adverse effects of extracorporeal circulation. To monitor neutrophil activation, leukocyte elastase was measured in plasma, where it is present in an inactive form complexed with a1-antiprotease [2, 4, 22]. The release of elastase was significantly higher in the centrifugal group compared with the roller group, and elastase
484
ASHRAF ET AL BYPASS PUMPS AND INFLAMMATORY MEDIATORS
levels were more pronounced immediately after CPB. Further, peak levels of elastase correlated with peak levels of C5b-9 in the centrifugal group. This is consistent with complement stimulation of neutrophils leading to elastase release. If this occurs at the endothelium (ie, that of the reperfused lung) as a result of IL-8 chemotaxis, it could contribute to endothelial injury and capillary leak after CPB [3–5]. Although we observed no statistically significant correlation between elastase levels and IL-8 levels, an association has been found by others [3, 4]. Clearly, the complex interactions of these molecules and cells at the endothelial level cannot be reflected accurately in statistical correlations of plasma levels. In contrast to our results, another clinical study [12] with a similar duration of CPB demonstrated less complement activation with the centrifugal pump, although in a much smaller number of patients. An in vitro study [13] showed less C5b-9 release with the centrifugal pump during extracorporeal circulation that exceeded 8 hours. It is possible that a roller pump induces less of an inflammatory response over a short duration of extracorporeal circulation, whereas a centrifugal pump is less traumatic over a longer duration. There were no differences in clinical outcomes between the two study groups. This finding does not preclude the potential clinical relevance of our biochemical findings. The magnitude of complement activation and cytokine release during and after CPB has been reported to be correlated significantly with the postoperative clinical outcome [5, 23, 24]. In summary, the greater production of IL-6, C5b-9, and elastase with centrifugal pumps suggests that these pumps are less “biocompatible” for short-term extracorporeal circulation. We thank Mr Martin Ward for excellent technical assistance.
References 1. Kirklin JK, Westaby S, Blackstone EH, Kirklin JW, Chenoweth DE, Pacifico AD. Complement and the damaging effects of cardiopulmonary bypass. J Thorac Cardiovasc Surg 1983;86:845–57. 2. Butler J, Rocker GM, Westaby S. Inflammatory response to cardiopulmonary bypass. Ann Thorac Surg 1993;55:552–9. 3. Finn A, Naik S, Klein N, Levinsky RJ, Strobel S, Elliott M. Interleukin-8 release and neutrophil degranulation after pediatric cardiopulmonary bypass. J Thorac Cardiovasc Surg 1993;105:234– 41. 4. George JF. Cytokines and mechanisms of capillary leakage after cardiopulmonary bypass. J Thorac Cardiovasc Surg 1993;106:566–7. 5. Hennein HA, Ebba H, Rodriguez JL, et al. Relationship of the proinflammatory cytokines to myocardial ischemia and dysfunction after uncomplicated coronary revascularization. J Thorac Cardiovasc Surg 1994;108:626–35. 6. Haeffner-Cavaillon N, Roussellier N, Ponzio O, et al. Induction of interleukin-1 production in patients undergoing cardiopulmonary bypass. J Thorac Cardiovasc Surg 1989;98: 1100– 6.
Ann Thorac Surg 1998;65:480 – 4
7. Jansen NJG, van Oeveren W, Gu YJ, von Vliet MH, Eijsman LL, Wildevuur CRH. Endotoxin release and tumor necrosis factor formation during cardiopulmonary bypass. Ann Thorac Surg 1992;54:744– 8. 8. Tonz M, Mihaljevic T, von Segesser LK, et al. Normothermia versus hypothermia during cardiopulmonary bypass: a randomized, controlled trial. Ann Thorac Surg 1995;59:137– 43. 9. Butler J, Pillai R, Rocker GM, Westaby S, Parker D, Shale DJ. Effect of cardiopulmonary bypass on systemic release of neutrophil elastase and tumor necrosis factor. J Thorac Cardiovasc Surg 1993;105:25–30. 10. Frering B, Philip I, Dehoux M, Rolland C, Langlois JM, Desmonts JM. Circulating cytokines in patients undergoing normothermic cardiopulmonary bypass. J Thorac Cardiovasc Surg 1994;108:636– 41. 11. Hoerr HR, Kraemer MF, Williams JL, et al. In vitro comparison of the blood handling by the constrained vortex and twin roller blood pumps. J Extracorpor Technol 1987;19: 316–20. 12. Weeldon DR, Bethune DW, Gill RD. Vortex pumping for routine cardiac surgery: a comparative study. Perfusion 1990; 5:135– 43. 13. Moen O, Fosse E, Braten J, et al. Roller and centrifugal pumps compared in vitro with regard to haemolysis, granulocyte and complement activation. Perfusion 1994;9:109–17. 14. Nishinaka T, Nishida H, Endo M, Koyanagi H. Less platelet damage in the curved vane centrifugal pump: a comparative study with the roller pump in open heart surgery. Artif Organs 1994;18:687–90. 15. Matthews JNS, Altman DG, Campbell MJ, Royston P. Analysis of serial measurements in medical research. Br Med J 1990;300:230–5. 16. Butler J, Chang GL, Baigrie RJ, Pillai R, Westaby S, Rocker GM. Cytokine responses to cardiopulmonary bypass with membrane and bubble oxygenation. Ann Thorac Surg 1992; 53:833– 8. 17. Nijsten MW, DeGroot ER, Tenduis HJ. Response of serum interleukin-6 in patients undergoing elective surgery of varying severity. Clin Sci (Colch) 1990;79:1611–5. 18. Cruickshank AM, Fraser WD, Burns HJ, Van Damme J, Shenkin A. Response of serum interleukin-6 in patients undergoing elective surgery of varying severity. Clin Sci (Colch) 1990;79:161–5. 19. Ohzato H, Yoshizaki K, Nishimoto N, et al. Interleukin-6 as a new indicator of inflammatory status: detection of serum levels of interleukin-6 and C-reactive protein after surgery. Surgery 1992;111:201–9. 20. Herbert CA, Baker JB. Interleukin-8: a review. Cancer Invest 1993;11:743–50. 21. Kawamura T, Wakusawa R, Okada K, Inada S. Elevation of cytokines during open heart surgery with cardiopulmonary bypass: participation of interleukin-8 and 6 in reperfusion injury. Can J Anaesth 1993;40:1016–21. 22. Faymonville ME, Pincemail J, Duchateau J, et al. Myeloperoxidase and elastase as markers of leukocyte activation during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1991;102:309–17. 23. Seghaye MC, Duchateau J, Grabitz RG, et al. Complement activation in infants and children: relation to postoperative multiple system organ failure. J Thorac Cardiovasc Surg 1993;106:978– 87. 24. Jansen PGM, Velthuis H, Huybregts RAJM, et al. Reduced complement activation and improved postoperative performance after cardiopulmonary bypass with heparin coated circuits. J Thorac Cardiovasc Surg 1995;110:829–34.