| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2001;71:158-164
© 2001 The Society of Thoracic Surgeons
a Prince of Wales Hospital, Sydney, Australia
Address reprint requests to Dr Dignan, Department of Cardiothoracic Surgery, 2986 The Vanderbilt Clinic, Nashville, TN37232
e-mail: rebecca.dignan{at}surgery.mc.vanderbilt.edu
Presented at the Thirty-sixth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 31–Feb 2, 2000.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsCommentAcknowledgmentsAppendix 1. Transfusion...Appendix 2. Transfusion rate...DiscussionReferences |
|---|
Methods. Patients (n = 202) were randomized to receive either placebo or aprotinin, 0.5 million KIU before incision and 0.5 million KIU during initiation of cardiopulmonary bypass. Differences in quantity of blood transfused were analyzed. Further groups were analyzed to account for the effect of aspirin. Multivariable analysis was performed to determine risk factors for transfusion. Direct costs of blood products and aprotinin were tabulated for each group.
Results. There was an important reduction in the proportion of patients transfused, and number of blood units transfused when aprotinin was given before coronary artery bypass grafting. These differences were even more important in patients on aspirin preoperatively. Independent predictors for increased number of transfusions were aspirin continued before operation, smaller body surface area, and the use of placebo instead of ultra-low dose aprotinin. There was no difference in morbidity between treatment groups. There was a reduction in direct costs associated with the use of aprotinin.
Conclusions. These data support the routine use of aprotinin 1 million KIU in coronary artery bypass grafting with left internal mammary artery to reduce cost and transfusion requirements.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsAppendix 1. Transfusion...Appendix 2. Transfusion rate...DiscussionReferences |
|---|
Aprotinin has been shown to be effective in reducing blood product usage after open-heart operations in a dose-dependent manner [1, 2]. High-dose aprotinin, however, is expensive. In attempting to deliver cost-effective quality surgical care [3], many institutions are unwilling to use high- or even low-dose aprotinin in routine cases. At the study institution, a divided dose of 1 million units of aprotinin has been administered routinely for coronary operations since 1992.
Efficacy of ultra-low dose aprotinin, the minimum dose thought to inhibit plasmin [8], has not been completely defined. The aim of this study was to determine whether 1 million KIU aprotinin administered in a divided dose before two stimuli of the inflammatory response (skin incision and cardiopulmonary bypass [CPB]) would be cost effective in reducing transfusion requirements in routine (elective and urgent) coronary artery bypass grafting.
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsAppendix 1. Transfusion...Appendix 2. Transfusion rate...DiscussionReferences |
|---|
Study drug administration
The study drug was administered by the anesthesiologist as an infusion in a blinded fashion. The first milliliter was infused slowly as a test dose and the patient observed for hypersensitivity or other untoward reaction. The initial dose was completed before skin incision and the second dose during the initiation of CPB. The patients in the aprotinin arm of the study received 0.5 million KIU aprotinin before skin incision and 0.5 million KIU during initiation of CPB. Patients in the placebo arm received the same volume of saline.
Surgical, anticoagulation, and transfusion protocols
The conduct of the operation was performed according to the technique of several anesthetists and surgeons (ie, anesthetic technique, oxygenator type, degree of hypothermia, blood cardioplegia technique, and use of postoperative aspirin). One surgeon used a membrane oxygenator and cardiotomy return suction in every patient (cases distributed equally between the two treatment groups). Two surgeons used bubble oxygenators without cardiotomy return suction. Heparin-bonded tubing was not used. Bovine heparin was administered as a loading dose (300 IU/kg) and 10,000 IU was added to the pump prime. Activated clotting time was maintained at more than 400 seconds as measured by the kaolin-activated system before and after heparinization. Cell-saver suction and postoperative autotransfusion were not used. Homologous red blood cells were transfused during CPB if the patient’s hemoglobin was less than 7.0 g/dL and postoperatively if less than 8.0 g/dL or if the patient’s conditioned warranted (ie, excessive mediastinal bleeding). Platelets and fresh frozen plasma were transfused as judged necessary by the investigators or the participating surgeons (ie, coagulopathy and excessive mediastinal bleeding). Cryoprecipitate was available but never used in these patients. Crossover to aprotinin use for the placebo group and use of additional aprotinin postoperatively in the aprotinin group was allowed by the participating surgeon or investigator if mediastinal hemorrhage occurred, identifying the treatment group to which the patient belonged in only 1 patient.
Data collection and criteria for evaluation of efficacy and morbidity
Demographic data, body surface area, heparin doses, activated clotting times, untoward reactions, blood products transfused intraoperatively and postoperatively, and mediastinal tube drainage at 4 and 12 hours were recorded on a study protocol data sheet for each patient. Changes between preoperative and postoperative electrocardiograms (performed on day 1, and as indicated thereafter by clinical changes), hemodynamic instability, neurologic events, and any other complications were noted on the data sheet. Pre- and postoperative hemoglobin, creatine kinase and myocardial isoenzymes (measured twice postoperatively), troponin (which became available at the study institution in June 1998), pre- and postoperative creatinine, and liver function tests were also recorded.
Evaluation of efficacy included the quantity of mediastinal tube drainage at 4 and 12 hours, incidence of reexploration for diffuse nonsurgical bleeding, proportion of patients requiring transfusion, and comparison of the total number of the blood units and of each type of blood product (packed red blood cells, fresh frozen plasma, and platelets) transfused.
Analysis of morbidity was based on surgeons’ and investigators’ clinical diagnosis of myocardial infarction, neurologic evaluation, and increase in creatinine level measured preoperatively and postoperatively (highest of day 1, 2, and 4 or 5) of more than 0.05 mg/dL. Diagnosis of clinically significant myocardial infarction was made before unblinding and based on the surgeons’ and investigators’ evaluation of electrocardiographic changes, hemodynamic instability, and increase in cardiac isoenzymes or tropronin.
Direct costs of aprotinin (US $74 per 1 million units), cross match (US $66), and blood products (US $66 per unit) were calculated and compared between placebo and aprotinin-treated groups. These were calculated using study institution-specific costs converted from Australian to US dollars. Indirect costs were not available at the study institution because costing is not itemized (ie, cost of blood product administration, operating room costs, intensive care bed costs, and hospital bed costs).
To translate cost efficacy to an institution in the US, cost and charges per patient were calculated for the corresponding author’s institution based on the number of units transfused in the two study groups. Costs were calculated using aprotinin cost ($146 per 1 million units), blood processing fees contracted (per unit) with the American Red Cross (red cells, $119.50; fresh frozen plasma, $46; platelets, $200), platelet pooling fee ($18 per unit), blood administration cost ($4.36 per unit), and cross match ($21.33 per unit) at the corresponding author’s institution. Charges were calculated using aprotinin cost plus pharmacy administrative fee ($411.80 per 1 million units), the blood product processing fee contracted (per unit) with the American Red Cross plus blood bank administrative fee (red cells, $216; fresh frozen plasma, $150; platelets per 4 U pack, $344) and platelet pooling fee ($18), blood administration cost ($4.36 per unit), and charge for cross match ($124) at the corresponding author’s institution. Aprotinin was considered cost effective if the cost per patient was less than or equal to the cost per patient in the placebo group.
Data collection was performed in a blinded fashion; the investigators were unblinded as to treatment after the patient’s discharge so that the medical record could be marked appropriately.
Data analysis
Data were tabulated using the database Summit (Minnetonka, MN). Patients requiring mediastinal reexploration were excluded only if a documented site amenable to surgical repair was located. Comparisons of demographics and mediastinal drainage were made using an independent t test or Wilcoxon test of rank sums. Analysis of the proportion of patients transfused in each group and comparison of postoperative complication rates (neurologic events, renal insufficiency, and myocardial infarction) were made using 
2 analysis. Comparisons of median units of blood products transfused for all patients and also for only those who were transfused were made using analysis of rank sums (Wilcoxon test for nonparametric data). The pre- and postoperative cardiac enzymes for all patients, patients with a clinical diagnosis of myocardial infarction, and patients without myocardial infarction were tested for differences in medians between treatment groups using Wilcoxon test for nonparametric data.
Although not stratified from the outset, groups were analyzed to determine the effect of aspirin on the proportion of patients transfused and units of blood products transfused in each treatment group. The "aspirin" group was defined as those patients who received aspirin on days 1 through 6 before operation. The "no aspirin" group included those patients who did not receive aspirin or received aspirin on day 7 or more before operation.
Multiariable linear regression analyses were performed to determine independent predictors of greater number of transfusions and logistic regression to determine risk factors for transfusion (yes/no transfusion). A two-stage logistic regression analysis was performed, with the covariates entered on the first step and the treatment (placebo versus aprotinin) offered in a stepwise method on the second step, adjusting for the covariates to predict transfusion risk. The first step of the multiple linear regression analysis was performed by forcing variables into the model in a step-wise fashion. In the second step of the analysis, the treatment group was added to the model after adjusting for the other variables.
Statistical tests were performed using the software package JMP (Altura Software, SAS Institute, Inc, Cary, NC) and SPSS v9.0 (SPSS, Inc, Chicago, IL) and were considered significant if p value was 0.05 or less.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsAppendix 1. Transfusion...Appendix 2. Transfusion rate...DiscussionReferences |
|---|
Table 1 shows the results of the demographic analysis. Treatment groups were similar for age, body surface area, gender distribution, and rate of urgent cases per group. There was no difference between groups in number of units of blood given on CPB, operative activated clotting time, time on CPB, or preoperative and discharge hemoglobin levels.
|
|
|
Independent predictors for both risk of transfusion (Table 4) and greater number of transfusions (Table 5) were smaller body surface area and aspirin administered preoperatively within 6 days of surgery. Preoperative hemoglobin below 14.3 g/dL and age more than 65 years predicted transfusion but not a greater number of transfusions. Use of cardiotomy return suction and membrane oxygenator was not a risk factor for either. Treatment without aprotinin was found to be a statistically significant predictor of transfusion and greater number of transfusions independent of the covariates. Patients treated without aprotinin had an increase in the odds of transfusion by a factor of 4.25. Treatment with aprotinin had a significant, independent effect of reducing transfusion requirements and decreasing risk of transfusion (yes/no).
|
|
As mentioned, accounting was not performed by individual departments in the study institution as to actual cost of operating room time, length of hospital stay, blood products or administration of transfusions by nursing staff. The Sydney Red Cross Blood Service estimated the cost to acquire and process each unit of blood product to be A$ 100 (equal to approximately US $66). The difference in cost at the study institution (Table 6) between the two groups represents a 24% reduction in cost for the aprotinin group. The difference in cost at the corresponding author’s institution represents a 26% reduction per patient in the aprotinin group. The difference in charges at the corresponding author’s institution represents a 12% reduction per patient in the aptinin group.
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsAppendix 1. Transfusion...Appendix 2. Transfusion rate...DiscussionReferences |
|---|
The effectiveness of this dose has a physiologic basis. The antifibrinolytic effect of aprotinin is thought to result from the direct inhibition of plasmin, whereas inhibition of kallikrein is involved to a lesser extent [9, 10], or possibly absent at this dose. Although aprotinin plasma concentrations of 200 KIU/mL or greater are needed to inhibit kallikrein, plasma concentrations of 50 KIU/mL are required to inhibit plasmin. This level can be achieved with a loading dose of 1,000,000 KIU [11], the dose used in this study. Hayashida and colleagues [12] showed that when minimal dose aprotinin (1 million units in the pump prime) was used, increased levels of 
2-plasmin inhibitor, plasminogen activator-1, and decreased levels of D-dimer were measured after CPB as compared to controls, thus supporting an antifibrinolytic effect [12].
This reduction in transfusion requirements was true whether or not the patient was on aspirin. The benefit of ultra-low dose aprotinin, however, was more marked in patients who had received recent aspirin. This has been confirmed in another study but at a higher dose of aprotinin [1]. The number of transfusions was also reduced in those patients who received aprotinin and were at reduced risk for bleeding (ie, those patients off aspirin for 7 or more days preoperatively) compared to those who received placebo. In addition, no patient treated with aprotinin and not on aspirin received a platelet transfusion.
The similarity in rate of transient ischemic attack, stroke, myocardial infarction, increase in creatinine level, and acute renal failure between groups, although not proved, support the safety of this aprotinin dose. Other studies have shown high-dose but not low-dose aprotinin to reduce the risk of stroke [13]. The risk of myocardial infarction and graft thrombosis is still controversial. Recent studies, after the importance of adequate heparinization has been acknowledged, have shown no difference in graft patency [4, 14] between patients treated with aprotinin or placebo.
Some investigators have suggested that aprotinin is safe and effective only at the full-dose regimen [15]. The large number of study patients required to show the small difference in morbidity and mortality between a group treated with and without aprotinin at half or lower doses of aprotinin is prohibitive. The benefit of aprotinin-induced reduction in transfusion requirements, and transfusion-associated morbidity and mortality calculated over time, far outweigh the extremely low risk of mortality or morbidity associated with aprotinin administration.
Although satisfactory determination of detailed and actual hospital costs was not available in this public hospital, this study showed a reduction in direct cost when ultra-low dose aprotinin was used. In a retrospective study, low-dose aprotinin was shown to be effective in redo coronary operations in reducing cost [16]. In another prospective, nonblinded study comparing high-dose, low-dose, and no aprotinin used in patients undergoing open heart operation, costs were significantly reduced when low-dose but not high-dose aprotinin was used [3].
Many variables can be considered when cost analysis is undertaken and may vary between countries and institutions. The costs of blood products as well as aprotinin may vary widely between hospitals. In one study, the cost of a pack of platelets was listed as US $545 [17]. At the study institution, the cost of a pack of platelets was estimated to be the same as any other unit of blood product despite the scarcity of platelets. The effect of aprotinin in reducing cost in this study would only increase if the higher estimated cost of platelets were used in the analysis. Depending on the cost of each unit of blood product, especially platelets, ultra-low dose aprotinin (approximately US $80 to US $180) [17] would be effective in reducing cost in most institutions if even one unit less was required in aprotinin-treated patients compared to nontreated patients. Comparison of charges between institutions and pharmacies is unreliable because frequently pharmacy costs are added to the total charges for a particular drug.
In another trial, which included patients taking aspirin within 48 hours of primary coronary operation, another low-dose aprotinin regimen (28 mg or 200,000 KIU then 28 mg/h) was compared to tranexamic acid (10 mg/kg then 1 mg · kg-1 · h-1), and aminocaproic acid (5 g then 1 g/h). All significantly reduced both blood loss and the use of blood products compared with no treatment [18]. Studies have compared high-dose aprotinin to aminocaproic acid and have shown aprotinin to be superior in reducing transfusion requirements [17, 19]. Comparison of the effect of ultra-low dose aprotinin versus aminocaproic acid treatment in reducing transfusion requirements and cost should now be undertaken in a double-blinded, randomized fashion.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsAppendix 1. Transfusion...Appendix 2. Transfusion rate...DiscussionReferences |
|---|
|
Appendix 1. Transfusion requirements after CABG with LIMA (medians) |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsAppendix 1. Transfusion...Appendix 2. Transfusion rate...DiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||
a All patients included in analysis.
b Patients who received aspirin within 6 days before operation.
c Patients who did not receive aspirin within 6 days before operation.
d Analysis using 2.
e Medians for fresh frozen plasma and platelet transfusion not included because frequency of transfusion was low.
CABG = coronary artery bypass grafting; LIMA = left internal mammary artery; PRBCs = packed red blood cells.
|
Appendix 2. Transfusion rate after CABG with LIMA |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsAppendix 1. Transfusion...Appendix 2. Transfusion rate...DiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
a All patients included in analysis.
b Patients who received aspirin within 6 days before operation.
c Patients who did not receive aspirin within 6 days before operation.
d Analysis using 2.
CABG = coronary artery bypass grafting; FFP = fresh frozen plasma;LIMA = left internal mammary artery; PRBCs = packed red blood cells.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsAppendix 1. Transfusion...Appendix 2. Transfusion rate...DiscussionReferences |
|---|
When we consider the original Hammersmith, or high-dose regime in aprotinin, it should not be forgotten that this dose regime (a test dose, a loading dose of 2 million KIU, a pump prime dose of 2 million KIU, and a constant infusion of 0.5 million KIU during the period of the operation) was not calculated on the hypothesis that it would be hemostatic, rather that it would be kallikrein inhibiting throughout the period of the operation. The hypothesis being, at that time, that kallikrein inhibition might reduce the inflammatory response in the lungs.
The issue then of the appropriate dose for hemostasis with aprotinin has continued to the present day. I agree with Dr Dignan and her colleagues that lower doses are effective in patients at lower risk of excessive bleeding.
Before this meeting, I pulled that data from around 500 of my past patients. I use aprotinin in approximately 40% of my surgical practice. I continue to use full Hammersmith dose for all redos and septic patients. This accounts for around 20%. In a further 20%, I use a 2 million KIU regime, for the most part, in patients who have unstable angina who have been maintained on aspirin therapy right until their operation.
However, I believe that the hemostatic response to aprotinin is in fact dose dependent, as proposed by David Royston and colleagues. And that although lower doses may be appropriate and adequate in a low-risk population, increased dose gives increased efficacy in those patients at a much higher risk of excessive bleeding.
I have three questions and one final comment. Did the 2% reexploration rate include or exclude the two identified patients in the aprotinin group? And was the reentry rate actually 4% for aprotinin patients? Second, what do you believe is the appropriate dosage for reoperations and other high-risk bleeding procedures? And third, do you agree that as far as we know at present, and this may change with increased understanding, the antiinflammatory effects of aprotinin are only seen at higher dosage levels? My final comment relates to the cost issue. I checked with our pharmacist at the Hammersmith in London to obtain accurate costs to British hospitals for the full Hammersmith, 2 million, and the proposed 1 million KIU dosage regimes. They are roughly one-third of the costs incurred here in the United States. One might be tempted to say that there may have been an economic downside to the Declaration of Independence after all.
I thank the Society for the privilege of discussing this very interesting paper.
DR SALIM RATNANI (Charleston, WV): I want to ask two questions.
First question is how many of your patients were septuagenarian and octogenarian diabetic patients where the arteries are usually small and diffusely diseased and, as a result, even with the low dose or ultra low dose of aprotinin in these patients, we are seeing some ST changes and ischemic changes on the electrocardiogram? Because those are the arteries where the flows are low and the arteries are small, those are the patients that are more prone for myocardial infarction with even this ultra dose of aprotinin.
And the second question is how many of your patients have received even ultra dose, low continuous infusion postoperatively?
DR JEFFREY E. GEORGE (Huntington, WV): My question has two parts. Could you comment on your having an extremely high incidence of transfusion, up to 70% to 80% for your patients, and could you discuss your protocol for transfusion.
DR DIGNAN: Thank you, Professor Taylor, for those interesting comments and questions, and for the other questions.
As to the reexploration rate, the 2% reexploration rate did not include those two patients that were excluded. The two that were excluded, however, did have surgical bleeding. One was a branch of the saphenous vein and one was a branch on the pericardium.
The dose for reoperation I think could possibly be lower than the full Hammersmith dose; even a half dose would seem to work as well as the full dose in my brief experience. I believe that the antiinflammatory effects are more pronounced with the high dose as compared to half dose, and possibly this should be used with endocarditis patients. I agree that the antiinflammatory effects are probably not seen when ultra low dose aprotinin is used.
As to the question of how many of these patients were octogenarians and diabetics, the mean age of the group was similar, about 66. We did not have very many octogenarians in the group. The incidence of diabetes was not analyzed. However, we did analyze cardiac enzymes and the rate of myocardial infarction and they were not different between groups. In fact, the aprotinin group had only a 3% as opposed to 5% rate of myocardial infarction in the placebo group.
We did not run aprotinin infusions postoperatively. However, 3 patients in the placebo group had postoperative aprotinin and that was only a one time dose of 500,000 U. Only 1 patient in the aprotinin group received postoperative aprotinin.
The transfusion rate is fairly significant in the placebo group but significantly lower in the aprotinin group. This is fairly consistent, however, with transfusion rates in the literature.
Protocol for transfusion: if the hemoglobin was less than 8, blood could be given. It was at the discretion of the investigator or the surgeon to administer blood if the patient had mediastinal hemorrhage.
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsAppendix 1. Transfusion...Appendix 2. Transfusion rate...DiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. R. McIlroy, P. S. Myles, L. E. Phillips, and J. A. Smith Antifibrinolytics in cardiac surgical patients receiving aspirin: a systematic review and meta-analysis Br. J. Anaesth., February 1, 2009; 102(2): 168 - 178. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 The Society of Thoracic Surgeons Blood Conservatio, V. A. Ferraris, S. P. Ferraris, S. P. Saha, E. A. Hessel II, C. K. Haan, B. D. Royston, C. R. Bridges, R. S.D. Higgins, G. Despotis, et al. Perioperative Blood Transfusion and Blood Conservation in Cardiac Surgery: The Society of Thoracic Surgeons and The Society of Cardiovascular Anesthesiologists Clinical Practice Guideline Ann. Thorac. Surg., May 1, 2007; 83(5_Supplement): S27 - S86. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Ronald and J. Dunning Does use of aprotinin decrease the incidence of stroke and neurological complications in adult patients undergoing cardiac surgery? Interactive CardioVascular and Thoracic Surgery, December 1, 2006; 5(6): 767 - 773. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Sedrakyan, A. Wu, G. Sedrakyan, M. Diener-West, M. Tranquilli, and J. Elefteriades Aprotinin use in thoracic aortic surgery: Safety and outcomes J. Thorac. Cardiovasc. Surg., October 1, 2006; 132(4): 909 - 917. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Fergusson, K. C. Glass, B. Hutton, and S. Shapiro Randomized controlled trials of aprotinin in cardiac surgery: could clinical equipoise have stopped the bleeding? Clinical Trials, June 1, 2005; 2(3): 218 - 232. [Abstract] [PDF]
|
|
|
|
|
|
Y. Kaminishi, Y. Hiramatsu, Y. Watanabe, Y. Yoshimura, and Y. Sakakibara Effects of nafamostat mesilate and minimal-dose aprotinin on blood-foreign surface interactions in cardiopulmonary bypass Ann. Thorac. Surg., February 1, 2004; 77(2): 644 - 650. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Zabeeda, B. Medalion, M. Sverdlov, S. Ezra, A. Schachner, T. Ezri, and A. J. Cohen Tranexamic acid reduces bleeding and the need for blood transfusion in primary myocardial revascularization Ann. Thorac. Surg., September 1, 2002; 74(3): 733 - 738. [Abstract] [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 |
