| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2004;78:2131-2138
© 2004 The Society of Thoracic Surgeons
a Department of Cardiothoracic Surgery
b Department of Anesthesia and Intensive Care, Sahlgrenska University Hospital, Gothenburg, Sweden
Accepted for publication June 2, 2004.
* Address reprint requests to Dr Jeppsson, Department of Cardiothoracic Surgery, Sahlgrenska University Hospital, SE 413 45 Gothenburg, Sweden (E-mail: anders.jeppsson{at}vgregion.se).
Abstract
BACKGROUND: Cardiopulmonary bypass induces a systemic inflammatory and hemostatic activation, which may contribute to postoperative complications. Our aim was to compare the inflammatory response, coagulation, and fibrinolytic activation between two different perfusion systems: one theoretically more biocompatible with a closed-circuit, complete heparin coating, and a centrifugal pump, and one conventional system with uncoated circuit, roller pump, and a hard-shell venous reservoir.
METHODS: Forty-one elderly patients (mean age, 73 ± 1 years, 66% men) undergoing coronary artery bypass grafting or aortic valve replacement were included in a prospective, randomized study. Plasma concentrations of complement factors (C3a, C4d, Bb, and sC5b-9), proinflammatory cytokines (tumor necrosis factor-
, interleukin-6, and interleukin-8), granulocyte degradation products (polymorphonuclear elastase), and markers of coagulation (thrombin-antithrombin) and fibrinolysis (D-dimer, tissue plasminogen activator antigen and tissue plasminogen activator–plasminogen activator inhibitor-1 complex) were measured preoperatively, at bypass during rewarming (35°C), 60 minutes after bypass, and on day 1 after surgery.
RESULTS: The mean concentrations of C3a (–39%; p = 0.008), Bb (–38%; p < 0.001), sC5b-9 (–70%; p < 0.001), interleukin-8 (–60%; p = 0.009), polymorphonuclear-elastase (–55%; p < 0.003), and tissue plasminogen activator antigen (–51%; p = 0.012) were all significantly lower in the biocompatible group during rewarming. Sixty minutes after bypass, the mean concentrations of sC5b-9 (–39%; p = 0.006) and polymorphonuclear-elastase (–55%; p < 0.001) were lower in the biocompatible group.
CONCLUSIONS: The results suggest that a closed perfusion system with a heparin-coated circuit and a centrifugal pump may improve cardiopulmonary bypass biocompatibility in elderly cardiac surgery patients in comparison with a conventional system.
Cardiac surgery induces a pronounced systemic inflammatory response, which may contribute to postoperative complications [1, 2]. The inflammatory response is complex in its nature, and a number of different factors may contribute. The inflammatory response was initially attributed to the use of cardiopulmonary bypass (CPB), but in recent years a number of other factors, such as the operative trauma per se, regional ischemia-reperfusion injury, endotoxin release, and retransfusion of cardiotomy suction blood and mediastinal shed blood, have been demonstrated to be of importance [1–5].
The inflammatory response is characterized by complement activation followed by activation of different cell populations and cytokine release, which subsequently may lead to tissue injury and organ dysfunction [1, 2]. The magnitude of the inflammatory response, measured as plasma concentrations of complement split products and cytokines, has been related to clinical outcome in both pediatric and adult cardiac surgery [6–8].
There is also a marked activation of the coagulation and fibrinolysis during cardiac surgery. Cardiopulmonary bypass promotes activation of factor XII, kallikrein, and tissue factor and enhances fibrinolytic activation [9]. Cardiopulmonary bypass also induces an increased thrombin-mediated or plasmin-mediated consumption of hemostatic factors and has a direct activating effect on platelets [9–11].
In the present study we evaluated the role of the CPB system for perioperative and postoperative activation of inflammation, coagulation, and fibrinolysis in adult cardiac surgery. Heparin-coated systems, closed circuits, and centrifugal pumps have previously been demonstrated individually to reduce inflammatory [12–15] and hemostatic activation [16–18], although contradictory results exist [19, 20]. The combined use of the three components may potentially have a superior effect compared with the effects of the individual components, and we have previously reported that a similar system improves biocompatibility in pediatric cardiac surgery [21, 22]. In the present study, our aim was to investigate whether a theoretically more biocompatible perfusion system consisting of a closed, completely heparinized CPB circuit and a centrifugal pump reduces systemic inflammatory and hemostatic response in adult patients compared with a conventional CPB system with uncoated surfaces, roller pump, and a hard-shell venous reservoir. For this purpose a prospective, controlled, randomized study was performed in a subgroup of adult patients with an increased risk for postoperative complications (elderly patients with long operation time), and markers of complement activation, cytokine release, granulocyte degradation, coagulation, and fibrinolytic activity were analyzed during and after CPB.
Patients and Methods
Patients
Forty-one elderly patients (mean age, 73 ± 1 years, 66% men) were included in the study. Inclusion criteria were age greater than 60 years, coronary disease with stable angina pectoris or aortic stenosis, left ventricular ejection fraction greater than 0.20, and an expected CPB time of greater than 90 minutes. Exclusion criteria were reoperation and preoperative use of steroids or nonsteroidal antiinflammatory drugs. Aspirin and clopidogrel were discontinued 5 to 7 days before surgery. Patient characteristics are given in Table 1.
|
Cardiopulmonary Bypass Technique
Before cannulation, heparin (Lövens, Ballerup, Denmark) was given and supplemented as required to maintain an activated clotting time of greater than 480 seconds. Hepcon HMS (Medtronic Inc, Minneapolis, MN) automatic dose mode was used for monitoring of anticoagulation and calculation of heparin and protamine doses during the procedure. Full-dose heparin was used in both the biocompatible and conventional groups. Cardiopulmonary bypass was performed with a nonpulsatile flow at a minimum rate of 2.4 L · min–1 · m–2, moderate hypothermia (nasopharyngeal temperature 30°C), and hemodilution (hematocrit, 20% to 30%). The extracorporeal circuit was primed with approximately 1,700 mL of Ringer-acetate (Fresenius-Kabi, Uppsala, Sweden), 200 mL of mannitol (Fresenius-Kabi), 100 mL of Tribonate (Fresenius-Kabi), and 7,500 IU of heparin. Blood gases were continuously monitored in venous and arterial tubing by an inline blood gas monitoring system (CDI400; Cardiovascular Device Instruments, Anaheim, CA). The venous partial pressure of oxygen was kept greater than 4.5 kPa, and the blood gas regimen was performed using -stat principle. No cell-saving devices were used, and all cardiotomy suction blood was retransfused without washing to the perfusate. All patients received tranexamic acid, 2 g before surgery and 2 g after skin closure. Aprotinin was not used. A single period of aortic cross-clamping was used for valve replacement and all coronary anastomoses. Cardioprotection was achieved with cold-blood cardioplegia. Rewarming was performed with a maximum temperature gradient between heat-exchanger water and venous blood of 10°C and with maximum water temperature of 40°C. Weaning from CPB was started after rewarming to a bladder temperature of at least 36°C.
Study Protocol
Forty-one patients were randomly allocated to either of two regimens by a computerized program for randomization with sequential allocation depending on sex, age, weight, and type of operation. The study protocol was approved by the Research Ethics Committee of the Medical Faculty, University of Gothenburg. Informed consent was given by all participating patients.
A closed and totally heparinized extracorporeal system was used in the biocompatible group (n = 21). The circuit consisted of a hollow-fiber membrane oxygenator (Maxima Plus PRF; Medtronic Inc), a Biomedicus centrifugal pump (BP80, Biomedicus; Medtronic Inc), a Biomedicus flowprobe (DP38, Biomedicus; Medtronic Inc), a soft-shell collapsible venous reservoir (CBMVR 1600; Medtronic Inc), a hard-shell cardiotomy reservoir (Intersept; Medtronic Inc), and a blood cardioplegia conducer (Myotherm; Medtronic Inc). All components, including cannulas, tubing connectors, and polyvinyl chloride tubing, were surface-heparinized with Carmeda Bioactive Surface (CBAS; Medtronic Inc).
An open, nonheparinized system was used in the conventional group (n = 20). The circuit consisted of a hollow-fiber membrane oxygenator (Maxima Plus PRF; Medtronic Inc), a hard-shell venous and cardiotomy reservoir (Maxima; Medtronic Inc), and a blood cardioplegia conducer (Myotherm; Medtronic Inc). A roller pump was used as the master pump.
Analyses
Blood samples for measuring the concentrations of C3a, sC5b-9, Bb fragment (Bb), tumor necrosis factor (TNF)-, interleukin 6 (IL-6), interleukin 8 (IL-8), polymorphonuclear (PMN) elastase, thrombin-antithrombin (TAT), D-dimer, tissue plasminogen activator antigen (t-PA-ag), and the complex consisting of tissue plasminogen activator and its inhibitor, tissue plasminogen activator inhibitor-1 (tPA–PAI-1) were drawn from arterial blood on four occasions during the procedure: after induction of anesthesia, after rewarming to 35°C, 1 hour after CPB, and in the morning on the first postoperative day.
Samples for analysis of inflammatory markers were collected into tubes with ethylenediaminetetraacetic acid, and samples for coagulation and fibrinolysis markers were collected into Stabilyte tubes (Biopool AB, Ume
, Sweden) with 0.129 mol/L sodium citrate. All samples were placed immediately on ice and centrifuged. The resultant plasma was stored at –70°C until analysis. All assays were performed in duplicate.
The concentrations of plasma C3a, TNF-, IL-6, and IL-8 were determined by sandwich enzyme-linked immunosorbent assays. The C3a assay is specific for C3adesarg, which has a much longer half-life in plasma than C3a. Modified enzyme-immunoassays were used to quantify sC5b-9, C4d, Bb, and PMN elastase. The assays were performed according to the manufacturer's instructions. The following assays were used: C3a, enzyme-linked immunosorbent assay (Quidel, San Diego, CA); sC5b-9, C4d, Bb, enzyme-immunoassay (Quidel); TNF- IL-6, IL-8, enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN); and PMN elastase enzyme-immunoassay (DPC, Los Angeles, CA).
Plasma concentrations of TAT, D-dimer, t-PA-ag, and tPA–PAI-1 were analyzed with enzyme-linked immunosorbent assays according to the manufacturer's instructions. The following assays were used: TAT (Dade Behring, Marburg, Germany); and D-dimer, t-PA-ag, and tPA–PAI-1 (Biopool AB). Troponin-T and hemoglobin levels were analyzed with clinical standard methods.
Calculations
Hematocrit varied substantially between the measurement points during and after surgery (19% to 51%). Plasma concentrations of the measured inflammatory and hemostatic variables were therefore corrected for hematocrit [4] by relating every measurement to a standard hematocrit value of 40% according to the following formula:
|
| (1) |
Statistics
Results are expressed as the mean ± standard error of the mean. Statistical significance was defined as p less than 0.05. The nonparametric Mann-Whitney U test (continuous variables) or 
2 test (categorical variables) was used to establish whether the randomization process had provided groups that were comparable and to compare clinical variables between the two groups. Two-way analysis of variance with correction for repeated measurements was used to evaluate differences in concentrations of inflammatory markers between the groups, followed by Student's t test if group or interaction between group and time indicated a significant difference. Correlations between inflammatory markers, coagulation, and fibrinolysis markers and clinical variables (CPB time, aortic cross-clamp time, bleeding, time on ventilator, and myocardial injury) were analyzed with Spearman's rank sum test. To avoid random significances because of the large number of analyses, only an r value greater than 0.5 in combination with a p value less than 0.05 was considered significant.
Results
Baseline Data
There were no statistically significant differences between the groups with respect to age, sex, preoperative left ventricular ejection fraction, aortic cross-clamp time, CPB time, or type of operation (Table 1).
Clinical Course
Stroke was diagnosed and verified by computer tomography in 1 patient in the conventional group (Table 2). All other patients recovered without sequelae after surgery and were discharged within 10 days. The number of patients which needed more than 1 day in the intensive care unit was significantly fewer in the biocompatible group (0% versus 20%, p = 0.031), and the number of patients who needed postoperative inotropic support to wean from CPB tended to be fewer in the biocompatible group (29% versus 55%, p = 0.09).
|
COMPLEMENT ACTIVATION
The complement fragments C3a, sC5b-9, and Bb increased during surgery in both groups, whereas C4d remained unchanged (Table 3). The mean concentrations of C3a, sC5b-9, and Bb were significantly lower during rewarming in the patients operated on with the biocompatible system compared with the conventional system (C3a, –39%, p = 0.008; sC5b-9, –70%, p < 0.001; Bb, –38%, p < 0.001). The mean concentration of sC5b-9 was also lower 60 minutes after CPB (–39%, p = 0.006).
|
, IL-6, and IL-8 increased during surgery in both groups (Table 4). The mean concentrations of IL-8 were significantly lower during rewarming in the biocompatible group (–60%, p = 0.009), whereas concentrations of TNF- and IL-6 tended to be lower in the biocompatible group (p = 0.15 and 0.07, respectively).
|
CORRELATION BETWEEN INFLAMMATION AND CLINICAL VARIABLES
There was a statistically significant correlation between CPB time and IL-6 levels during rewarming (r = 0.56, p < 0.001; y = –41 + 0.7x; Fig 1) and between CPB time and Bb fragments 1 hour after CPB (r = 0.51, p < 0.001; y = 0.03 + 0.01x). There was no significant correlation between inflammatory variables and aortic cross-clamp time, bleeding, myocardial injury, or time on respirator.
|
Head 1
Activation of Coagulation and Fibrinolysis
All measured markers (TAT, D-dimer, t-PA-ag, and t-PA–PAI-1) increased during and after surgery (Table 5). The mean concentrations of t-PA-ag was significantly lower in the biocompatible group during rewarming (–51%, p = 0.012) compared with the conventional group. In contrast, there were no intergroup differences in TAT (p = 0.38), D-dimer (p = 0.68), and tPA–PAI-1 (p = 0.21) concentrations.
|
Comment
The results of the study suggest that a closed perfusion system with heparin-coated circuit and a centrifugal pump may improve CPB biocompatibility in adult cardiac surgery.
Inflammatory Response
The inflammatory response after cardiac surgery is characterized by complement activation followed by activation of different cell populations and cytokine release, which subsequently may lead to tissue injury and organ dysfunction [1, 2]. It has been demonstrated that the magnitude of the inflammatory response, measured as the concentrations of complement and cytokines, correlates negatively with postoperative organ function [6–8]. Therefore, it is probably advantageous if these proinflammatory factors could be avoided in the systemic circulation.
The complement system is one of the most important inflammatory pathways. In the present study, complement factors C3a, sC5b-9, and Bb increased during and immediately after surgery in both groups whereas Cd4 levels remained unaltered (Table 3). This confirms previous observations that the complement cascade during CPB is activated mainly through the alternative pathway [21, 23]. The use of a closed heparinized circuit and a centrifugal pump markedly reduced activation of the key complement proteins C3a and sC5b-9 (the latter is also known as the terminal complement complex, TCC, or the complement membrane attack complex). C3a and sC5b-9 have a number of important pathophysiologic effects [24, 25]. C3a is myocardial depressive, contracts smooth muscle, increases vascular permeability, and induces production of cytokines, whereas sC5b-9 is the end product of the complement cascade and may cause irreversible cell membrane damage [24, 25]. Although it is difficult to compare different studies because of variation in patient selection, study design, and equipment, the attenuation of C3a and sC5b-9 concentrations during rewarming (–39% and –70%, respectively) appears to be more pronounced when using our combined approach than what has been noted by single measures [12–15] This suggests that the combined system improves biocompatibility further than the individual methods.
One interesting finding in our study was that patients who subsequently would need inotropic support when weaned from CPB had higher levels of C3a levels during rewarming. This supports previous reports about a negative inotropic effect of C3a [24] and points out the importance of reducing inflammatory activation during CPB.
Circulating concentrations of PMN elastase provides a measure of neutrophil granulation, which may cause tissue damage. In the present study, PMN activation was significantly less pronounced in the combined group, which gives further evidence for an improved biocompatibility. Neutrophil degranulation is triggered by complement C5a [24], and the present investigation supports this association also during CPB, as there was a significant correlation between sC5b-9 and PMN elastase during rewarming (r = 0.68, p < 0.001; y = 132 + 0.24x; Fig 2).
|
, IL-6, and IL-8 are proinflammatory cytokines with important pathophysiologic properties [24]. Tumor necrosis factor- increases stress hormone release and neutrophil adhesion, causes myocardial depression, and stimulates production of other cytokines. Interleukin 6 mediates vasodilation and impairs cardiac function, and IL-8 has been associated with myocardial injury after CPB [1]. Complement proteins can stimulate cytokine production from different cell populations, such as macrophages and endothelial cells [24]. In the present investigation we found that the use of the combined perfusion system reduced IL-8 significantly and tended to reduce both TNF- and IL-6. This was expected because we found a marked reduction in complement production and complement is known to stimulate cytokine production [24]. However, the results are in contrast to our investigation in pediatric cardiac surgery patients [21] in which group differences in cytokine production did not reach statistical significance. One explanation for the divergent results could be the bypass time, which was markedly longer in the adult population (mean, 148 versus 110 minutes). The importance of CPB and bypass time for cytokine production is further emphasized by the significant correlation between CPB time and IL-6 levels during rewarming both in the present study (Fig 1) and in children [21].
Coagulation and Fibrinolysis
Cardiac surgery with cardiopulmonary bypass influences coagulation and fibrinolysis through different mechanisms. Hemodilution, hypothermia, surgical trauma, and the interaction of blood with nonendothelialized surfaces results in impaired hemostasis and may cause excessive postoperative bleeding [9]. Activation of coagulation and fibrinolysis occurs simultaneously during CPB [9]. Physiologic activation of coagulation is mediated almost exclusively through the extrinsic tissue factor pathway, and t-PA is the major physiologic activator of fibrinolysis. However, during CPB, contact activation (intrinsic pathway) may play an important role for coagulation and fibrinolysis, and there are reports indicating reduced activity if a more biocompatible perfusion system is used [16–18]. In the present study, coagulation was evaluated by analyzing TAT. The contact activation and tissue factor pathways converge, resulting in generation of thrombin, which can be quantified by analyzing TAT. We found that TAT increased twofold to threefold during and early after CPB in both groups but without intergroup differences (Table 5). This supports previous comments about CPB as a procoagulant state [26], and the absence of group differences indicates that the contact activation is less important than the tissue factor pathway during cardiac surgery.
Fibrinolysis was assessed in the study by measuring t-PA-ag, t-PA–PAI-1 complex, and D-dimer. The fibrinolysis is activated by kallikrein, thrombin, and fibrin, causing release of t-PA [9]. Tissue plasminogen activator is synthesized from endothelial cells and is inhibited by PAI-1, forming the t-PA–PAI-1 complex. D-dimer is a small, fibrin degradation product that indicates fibrin formation and subsequent lysis. Because all patients received tranexamic acid, t-PA-ag and t-PA–PAI-1 complex were analyzed to detect whether there is a stimulus to fibrinolytic activity upstream of the inhibitor. All fibrinolytic markers increased during and after surgery, which indicates, in accordance with previous studies, an enhanced fibrinolysis during CPB. The use of a combined system reduced the production of t-PA-ag significantly, which again suggests improved biocompatibility. With the same reservations as above about comparing different studies, the fibrinolytic activation appears to be less pronounced with the combined system than with individual alterations of the perfusion system.
Clinical Variables
The study was primarily designed to detect differences in markers of inflammatory and hemostatic activation during and early after surgery and therefore lacks statistical power in clinical variables. However, there are some findings worth mentioning. There was a tendency toward fewer patients who needed inotropic support to wean from CPB in the biocompatible group (29% versus 55%, p = 0.09), and it is notable that none of the patients in the biocompatible group needed more than 1 day in the intensive care unit (p = 0.03 between groups) given that the included patients were high-risk patients (elderly with an expected CPB time > 90 minutes). These results indicate a possible clinical benefit of the more biocompatible perfusion system. This suggestion needs to be validated in a study with clinical end points and sufficient statistical power.
Study Design and Limitations
In most studies about inflammatory activation, only low-risk patients undergoing isolated coronary arterial bypass grafting have been included. In the present study we chose instead to include elderly patients with a long expected CPB time (>90 minutes). We reasoned that this subgroup of patients, with an increased risk for postoperative complications, would have more to gain from a more biocompatibility perfusion system and that subtle differences in inflammatory and hemostatic activation would be easier to detect.
One apparent limitation is that the study design does not allow us to evaluate the role of the individual components of the bypass system. Although desirable, an alternative design would have required significantly larger resources. It was reasoned that if the three-component bypass system proved to be superior to the conventional system, the results could be compared with previous data in which the three components have been evaluated individually. If there appeared to be additional effects of the current system, this would motivate a more elaborate and resource-demanding protocol.
The inflammatory response, coagulation, and fibrinolysis during CPB are extremely complex reactions, and other markers and mediators may respond differently to the perfusion system. It should therefore be emphasized that the conclusions of this study are limited to the investigated population and the analyzed markers only, and different results may be obtained with another study design and other markers. In addition, as mentioned above, the study was designed to detect differences in inflammation, coagulation, and fibrinolysis during and after CPB and therefore lacks statistical power in clinical end points.
Conclusions
The results suggest that the use of a closed, completely heparinized perfusion system and a centrifugal pump may reduce complement activation, neutrophil degradation, cytokine release, and fibrinolytic activation in elderly patients and thus improve biocompatibility in comparison to a conventional system. In addition, we found indications of a positive clinical effect with the more biocompatible system. The importance of these findings needs to be evaluated in larger studies with clinical end points.
Acknowledgments
The study was supported by grants from The Swedish Heart and Lung Foundation and the Gothenburg Medical Association. The authors thank Christina Lövgren, ECCP, Kerstin Björk, ECCP, and Magnus Lundquist, ECCP, for excellent assistance during the operations; Maria Tylman, Christina Linnér, and Elsa Eriksson for laboratory assistance; and Anders Odén for statistical advice.
References
Related Article
This article has been cited by other articles:
|
|
 |
|
  G. S. Murphy, E. A. Hessel II, and R. C. Groom Optimal Perfusion During Cardiopulmonary Bypass: An Evidence-Based Approach Anesth. Analg., May 1, 2009; 108(5): 1394 - 1417. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. L. Ngaage, M. E. Cowen, and A. R. Cale Cardiopulmonary bypass and left ventricular systolic dysfunction impacts operative mortality differently in elderly and young patients Eur. J. Cardiothorac. Surg., February 1, 2009; 35(2): 235 - 240. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Stassano, L. Di Tommaso, M. Monaco, S. Iesu, G. Brando, S. Buonpane, G. Ambrosio, G. Di Benedetto, and P. Pepino Myocardial Revascularization by Left Ventricular Assisted Beating Heart Is Associated With Reduced Systemic Inflammatory Response Ann. Thorac. Surg., January 1, 2009; 87(1): 46 - 52. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Mangoush, S. Purkayastha, S. Haj-Yahia, J. Kinross, M. Hayward, F. Bartolozzi, A. Darzi, and T. Athanasiou Heparin-bonded circuits versus nonheparin-bonded circuits: an evaluation of their effect on clinical outcomes Eur. J. Cardiothorac. Surg., June 1, 2007; 31(6): 1058 - 1069. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Asante-Siaw, J. Tyrrell, A. Hoschtitzky, and J. Dunning Does the use of a centrifugal pump offer any additional benefit for patients having open heart surgery? Interactive CardioVascular and Thoracic Surgery, April 1, 2006; 5(2): 128 - 134. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Jeppsson Reply Ann. Thorac. Surg., February 1, 2006; 81(2): 791 - 791. [Full Text] [PDF]
|
|
|
|
|
|
C. Baufreton More Biocompatibility in Cardiopulmonary Bypass for High-Risk Patients Ann. Thorac. Surg., February 1, 2006; 81(2): 790 - 791. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
