| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2003;75:S721-S728
© 2003 The Society of Thoracic Surgeons
a Department of Cardiovascular Surgery, Albert Ludwigs University of Freiburg, Freiburg, Germany
* Address reprint requests to Dr Doenst, Department of Cardiovascular Surgery, University of Freiburg, Hugstetter Str. 55, 79106 Freiburg i. Br., Germany
e-mail: doenst{at}ch11.ukl.uni-freiburg.de
Presented at the 3rd International Symposium on Myocardial Protection From Surgical Ischemic-Reperfusion Injury, Asheville, NC, June 2–6, 2002.
Abstract
Insulin has been used in the treatment of patients undergoing cardiac surgery or suffering from acute myocardial infarction. Most of these investigations have demonstrated that the metabolic cocktail consisting of glucose-insulin-potassium (GIK) improves recovery of function and outcome after cardiac surgery and substantially reduces mortality of patients with acute myocardial infarction. There is also evidence suggesting that insulin is not effective under these conditions, as demonstrated in a recent large randomized trial in cardiac surgery. It is therefore not surprising that insulin or GIK is not used routinely in clinical practice. Many hypotheses have been advanced to explain the effects of insulin and GIK but none of them has enjoyed convincing support. In cardiac surgery the many different application protocols described make it difficult to compare the results. The application of GIK after cardiac surgery may be complicated by severe disturbances in glucose or potassium homeostasis. In this article we review the literature in this field, addressing the areas of controversy. We discuss the different mechanisms suggested and we propose potential solutions. We conclude that a multifactorial mechanism is likely to explain the effects of insulin or GIK after ischemia and we propose that in a practical sense the application of high-dose insulin during reperfusion, utilizing a newly described, direct nonmetabolic effect, is a convincing concept. We will further demonstrate our clinical experience in establishing a protocol for putting this concept into clinical practice.
Metabolic approaches, specifically glucose-insulin-potassium (GIK) infusion, have been used to improve efficiency of energy production and recovery of contractile function [1]. Glucose-insulin-potassium infusion was first used by Sodi-Pallares and associates [2] for the treatment of acute myocardial infarction. Braimbridge and associates [3] were the first to apply the solution in cardiac surgery. In 1969 they reported that "GIK was successfully used to treat patients with low cardiac output after mitral valve replacement not responding to isoprinaline, digoxin, or pacemaking." The potential enthusiasm that may have been sparked by this study was dampened by a clinical trial of the British Medical Research Council [4] that failed to demonstrate any benefit of GIK in the treatment of acute myocardial infarction. However, concerns were raised based on the dosage of GIK and the inconsistencies of the protocols used in that clinical trial [5]. Nonetheless this setback and the advent of thrombolysis therapy in the treatment of acute myocardial infarction have prevented glucose insulin potassium from becoming standard treatment for patients with acute myocardial infarction.
Although neither of these limitations applies to the use of GIK in cardiac surgery, a breakthrough in this discipline was not noted. The reasons for this observation are not clear especially since most studies, clinical or experimental, demonstrate a benefit of GIK. Furthermore the belief that the benefits of GIK for the treatment of acute myocardial infarction would be overcome by thrombolysis therapy did not materialize [6, 7]. A review of the main studies using GIK for conditions of acute ischemia, clinical or experimental (not all referenced in this article) illustrates the above statement. Of 91 studies reviewed, 74 report a benefit of insulin or GIK. It is important to emphasize that none of the studies failing to demonstrate a benefit reported the occurrence of life-threatening complications.
The concept of GIK has attracted renewed attention lately for two reasons. First, recent studies demonstrated a reduction in 1-year mortality for patients with acute myocardial infarction that was comparable in its magnitude to the initial effects of thrombolytic therapy [6, 8]. Second, we demonstrated a direct positive inotropic effect of insulin on the postischemic rat heart that was independent of insulin’s metabolic effects [9]. These findings were confirmed by others [10, 11] in different experimental models. Thus the results suggest that there may be an additional, new mechanism for insulin after ischemia that may improve the heart’s contractile function.
GIK is not used routinely in clinical practice however. Three main reasons may explain this situation. First, there are no large, randomized multicenter trials investigating the effects of this metabolic therapy. Second, the mechanisms by which glucose and insulin mediate their effects remain controversial. And third, there is no standardized protocol for the application of GIK.
The last reason is especially so for randomized trials in cardiac surgery. Most studies are small, with patient numbers of 10 to 50 per group [12, 13]. Approximately 30 trials have been identified that may or may not have been randomized, have used different endpoints in assessing clinical benefit, and differed significantly in the protocols applied. GIK was administered preoperatively [14], perioperatively [13, 15–18], postoperatively [12, 19] or perioperatively as well as postoperatively [20–22]. The duration of GIK treatment ranged from a few seconds (given as a bolus in the cardiopulmonary bypass circuit [19]) to approximately 18 hours (applied perioperative as well as postoperatively [20]). The total amounts of glucose, insulin, and potassium given differed by a factor greater than 100 among studies. It is impressively difficult to find two or three studies using identical application protocols. This heterogeneity invites a plethora of open questions. Table 1 lists those most pertinent. The ability to rationally utilize insulin for therapeutic gain is intimately linked to understanding the mechanisms of insulin and GIK in the setting of acute ischemia.
|
After Sodi-Pallares and associates [2] used GIK as as "polarizing solution," the focus has quickly shifted toward several metabolic effects of glucose or insulin, or both, and nonmetabolic effects have been recently described. The most important of the proposed mechanisms are reviewed below.
Increasing glycogen content
In the late 1970s and early 1980s many studies attributed the effects of GIK to its potential to increase preischemic glycogen content. As glycogen serves as fuel for anaerobic adenoside triphosphate (ATP) production, the preoperative application of GIK was supposed to increase recovery of contractile function by increasing preischemic glycogen content [1, 23–25]. There has been a long-standing debate on the role of preischemic glycogen content in improving ischemia tolerance [26–28]. Only recently has experimental work on the isolated working rat heart demonstrated that glycogen turnover rather than glycogen content may be associated with ischemia tolerance [29]. From a clinical standpoint this hypothesis would not apply to those studies where GIK was administered during or even after ischemia.
Decreasing free fatty acid levels
High concentrations of free fatty acids in the serum may have negative effects on postischemic contractile function for three main reasons [30, 31]. First, high concentrations of free fatty acids in the cytosol may result in the accumulation of acyl-carnitine, which in turn may cause membrane damage. Second, incomplete oxidation of free fatty acids during periods of oxygen deprivation may cause peroxide formation, which may also impair cellular function. Finally, the production of ATP from fatty acids is associated with relatively high oxygen consumption compared with ATP production from glucose oxidation [32, 33]. Since ATP generation in heart muscle is governed by substrate competition, reducing free fatty acids in the serum may shift ATP production from fatty acid to glucose oxidation and therefore render the heart more oxygen efficient. This mechanism may apply particularly to the setting of cardiac surgery in which free fatty acids are liberated by the administration of heparin [34]. Insulin has the potential to decrease serum free fatty acid levels by inhibiting hormone-sensitive lipase in adipose tissue and by directly inhibiting free fatty acid oxidation in the mitochondrion through the activation of acetyl-CoA-carboxylase [35].
Increasing glycolytic ATP production
A third metabolic mechanism discussed is the stimulation of glycolysis by insulin and glucose. This mechanism may apply during low-flow ischemia in particular. It has been suggested that increasing glycolysis results in cytosolic ATP production [36], which may be used in the maintenance of basic cell function such as ion pumps on the plasma membrane [37]. While this mechanism has been considered a key principle in GIK action, it has also been the center of a debate on using GIK. Increasing glycolysis through GIK would result in greater ATP production. The increase in anaerobic glycolysis also increases lactate production and results in acidosis [38]. Neely and Grotyohann [26] suggested that reduction in anaerobic glycolytic activity would reduce ischemic injury and result in improved postischemic function, concluding that reducing intracellular acidosis may be beneficial. Many potential users of GIK have shied away from using this solution owing to this controversy. Others have tested Neely’s hypothesis under similar or different conditions and were unable to repeat the results [39, 40]. Uncertainty still remains concerning this issue. However, because GIK as an exogenously applied solution must be delivered to the site of action through the blood stream, low-flow ischemia or reperfusion must be present rather than total ischemia. It is likely that if protons and lactate accumulate, this low flow will be able to discard the waste products.
Hyperglycemia
Increasing glucose serum levels has been suggested to be beneficial for the ischemic and reperfused myocardium in two ways. First, hyperglycemia increases the osmolarity of the extracellular space. This increase in osmotic pressure may prevent or reduce the formation of significant cellular edema [41, 42]. Second, increasing substrate availability may increase glucose uptake [43]. The result would be the shift already described in substrate utilization towards glucose. Lazar and associates [21] demonstrated that patients benefited from a GIK protocol that raised glucose levels up to 270 mg/dL. However, that study and others were not able to establish a direct causal relationship between those two observations. Interestingly, another study by Lazar and associates [22] in diabetic patients provides arguments to the contrary. In this study the authors demonstrate that a group of diabetic patients receiving GIK showed no sternal wound infections whereas two infections occurred in the diabetic control group, which had the higher postoperative glucose levels. The role of diabetes in this particular metabolic therapy is still an enigma (see below). Irrespective of the presence or absence of diabetes mellitus, there is convincing evidence that postoperative hyperglycemia is linked to an increased rate of wound infections [44, 45] and that maintaining euglycemia during the postoperative recovery period may reduce mortality and morbidity [46].
Anaplerosis
The Taegtmeyer group suggested an anaplerotic mechanism to explain the effects of GIK on recovery of function [47, 48]. The term anaplerosis was coined by Kornberg [49] in 1966 and refers to the replenishment of ischemically depleted substrates of the glyoxylate cycle in bacteria. A similar depletion of substrate occurs in the citric acid cycle of the mammalian heart during ischemia. This substrate depletion can impair the Krebs cycle’s normal turnover rate and thus significantly interfere with energy production in the heart muscle during the recovery phase after ischemia [50]. The same group demonstrated that anaplerotic pathways are surprisingly active in the myocardium [51].
Other mechanisms
Other less supported mechanisms include the possibility that insulin increases the activation of sympathetic nerves by hyperinsulinemia during GIK, which may then result in increased contractile function [52] Insulin is also a known vasodilator and thereby results in a decrease in peripheral vascular resistance [53, 54]. This decrease in peripheral vascular resistance would then improve cardiac output. The association of a decrease in peripheral vascular resistance by insulin and an increase in cardiac output has been demonstrated in several GIK studies [12, 55]. Finally insulin has been demonstrated to activate plasminogen activator inhibitor 1 [56], the significance of which in the setting of acute application in cardiac surgery of for acute myocardial infarction is still unclear.
Influence of diabetes mellitus
As stated above the role of diabetes mellitus in the metabolic treatment of heart disease and specifically GIK or insulin in cardiac surgery is not clear. In patients with diabetes mellitus type II insulin sensitivity is impaired to varying degreees [57]. Therefore infusion of GIK or insulin may require adjustments in the protocol or may not even be effective at all. There is ample evidence that GIK is a formidable treatment for diabetic patients with acute myocardial infarction, however, [6] as well as for those patients undergoing cardiac surgery [22]. It is conceivable that the changes in insulin sensitivity observed in the entire organism are also present in the heart. The heart has been described as a specific target of diabetes mellitus and mechanisms have been set forth to explain a "diabetic cardiomyopathy" [58]. Interestingly, recent evidence suggests that insulin’s metabolic effects on the heart may be normal. Jagasia and associates [59] demonstrated that patients with diabetes showed lower glucose extraction from the coronary arteries but that the total amount of glucose uptake was the same as in healthy persons. Assuming that GIK utilizes one or more of the metabolic mechanisms, the observation by Jagasia and associates [59] could explain why diabetic patients respond well to GIK treatment. Taking this new evidence and the results of clinical trials on diabetic patients into account it appears reasonable to apply GIK or insulin in diabetic patients. From a mechanistic perspective, the question whether the insulin dosage requires adjustment must remain open at this time. Our data presented below include the application of high dosages of insulin to diabetic patients without showing major differences from nondiabetic patients.
The aftermath
While the mechanisms stated above may be more or less convincing, they all encounter serious limitations by conflicting experimental or clinical evidence. The major concerns involving all the metabolic mechanisms proposed is the presence of insulin resistance after ischemia and especially after cardiac surgery [60]. It is no secret that large amounts of insulin can be administered after cardiac surgery without significantly affecting serum glucose [19]. Such insulin resistance may explain why protocols supplying low doses of GIK have failed to demonstrate significant effects [61]. As in the diabetic heart, it is not yet clear whether such whole-body insulin resistance translates to the heart.
Thus the question as to how insulin or GIK deliver their actions remains unanswered. The heterogeneity of the clinical application protocols used in the context of cardiac surgery suggest a multifactorial mechanism. The most comprehensive explanation at this time may be the recently described, direct, nonmetabolic positive inotropic effect of insulin on the postischemic heart [9] because it appears not to be affected by diabetic or postischemic insulin resistance.
Finding the right solution
In an effort to utilize the beneficial effects of GIK or insulin after cardiac surgery in our patients, we faced questions 3 to 6 posed in Table 1. Reviewing the literature we discovered trends revealing four aspects important in a practical sense. First, the beneficial effects of insulin or GIK were most prevalent when glucose and insulin had been given in high dosages. This statement is based on observations where studies applying low dose GIK after myocardial infarction did not report any beneficial effects [61] whereas studies using high doses demonstrated impressive reductions in mortality [7, 55, 62].
Second, patients with preoperatively or postoperatively impaired contractile function benefit most. This reasoning is supported by clinical studies of GIK in cardiac surgery where no benefit of GIK on recovery of function was observed in patients with normal contractile function [63] but significant effects were observed in patients with impaired preoperative contractile function [14] or those suffering from low output syndrome after surgery [15, 55]. This analysis falls short of addressing other potential benefits of GIK or insulin for patients with preoperatively normal contractile function (eg, the reduction of atrial arrhythmias [22]) and may therefore underestimate the clinical potential.
Third, adminstration of insulin during reperfusion (ie, postoperatively) appears critical. Several investigators have addressed the issue of timing in animal models, again with conflicting evidence [40, 64]. Lazar and associates [64] demonstrated that a perioperative protocol was superior compared with a postoperative one in pigs. While the subsequent clinical studies demonstrated benefits of two perioperative protocols (one for diabetic patients and one for nondiabetic patients), a strictly postoperative protocol has not been compared. The only large randomized clinical trial of insulin application in cardiac surgery is a recent study by Rao and associates [18], where insulin was given in low concentrations into the cardioplegic solution. The authors were unable to demonstrate a reduction in postoperative low output syndrome or mortality. However, in a smaller study the same investigators demonstrated improved recovery of function by the same treatment in one group of that investigation [65]. In addition, other smaller studies demonstrated significant effects when insulin was given postoperatively [12, 55].
Fourth, the use of glucose-insulin-potassium may cause severe disturbances in glucose homeostasis. We therefore reasoned that a high-dose insulin postoperative protocol avoiding hyperglycemia is a convincing concept. In attempting to choose a protocol to put our concept into practice, we noted the scarcity of detailed documentation of postoperative serum glucose levels in studies using high dosages of insulin and glucose. We therefore carried out a pilot study with the goal of developing a safe application protocol for metabolic therapy with GIK in cardiac surgery.
Development of high-dose insulin therapy
We developed a safe protocol for the application of high doses of insulin in three phases. We included 27 patients (aged 48 to 83 years, 8 with diabetes mellitus type II) with an ejection fraction of 30% ± 5%, a creatinine level below 2.0 mg/dL, and similar perioperative characteristics. All patients were undergoing coronary artery bypass grafting or valve replacement or both with cardioplegic arrest. Written informed consent was obtained from each patient. The pilot study was approved by the Ethics Committee of the University of Freiburg. We did not randomize patients as the goal of this study was to evaluate the effects of this metabolic therapy on glucose and potassium homeostasis rather than to assess the effects on recovery of contractile function. We measured free fatty acid levels in the serum to assess the timing of insulin’s suppressive effect. In phases 2 and 3 we also determined insulin and c-peptide in the serum because in phase 1 we noted a long-lasting requirement for glucose infusion after GIK had been discontinued and we wanted to assess whether this requirement was due to an extension of the normally short elimination half-life of insulin or to a reactive endogenous hyperinsulinemia.
Phase 1
We first used a high-dose GIK protocol described in the literature [55]. We aimed to infuse 0.5 g · kg-1 · h-1 of glucose and to maintain blood glucose levels below 350 mg/dL. We started by infusing 20% of the desired glucose dose, 0.15IU · kg-1 · h-1 of insulin and 0.15 mval · kg-1 · h-1 of potassium. Maximal insulin infusion was set at 5 IU · kg-1 · h-1 and infusions were adjusted hourly. After 12 hours, all infusions were to be discontinued slowly. Glucose infusion was not further increased or decreased once the maximal insulin infusion rate and the upper limit of the serum glucose range had been reached. Figure 1A
shows postoperative serum glucose levels as well as glucose and insulin infusion rates of the patients treated with GIK according to protocol 1. At the start of GIK infusion, glucose levels rose to a maximal mean value of 350 mg/dL and returned to near normal values within 6 to 12 hours. There was tremendous scatter and some patients experienced hyperglycemia as high as 600 mg/dL during GIK infusion and hypoglycemia as low as 28 mg/dL after GIK infusion. After discontinuation there was an approximately threefold increase in basal glucose infusion rate compared with patients not treated with GIK. None of the patients experienced any complications during or after GIK treatment. The same was true for all other patients in this study. Maximal insulin infusion rate was 1 IU · kg-1 · h-1, which was only 20% of the upper limit given for insulin.
|
|
Phase 3
We therefore developed a new protocol whereby an insulin infusion rate similar to the maximum in protocol 2 was chosen as the fixed rate (2.5 IU · kg-1 · h-1). Owing to the prolonged elevation of serum insulin after GIK, we shortened the application time to 8 hours. Glucose infusion was started at 0.25 g · kg-1 · h-1. Whenever serum glucose at the beginning of insulin therapy was above 200 mg/dL, there was no initial glucose infusion. Thereafter, both glucose and potassium infusions were adjusted as needed. After 8 hours insulin infusion was ceased and glucose infusion was maintained at 100% of the last rate for 1 hour before tapering off. Before we applied this protocol to patients it was important to assess whether a dose-response effect of insulin on potassium elimination existed. This was important to avoid potentially dangerous hypokalemias at the commencement of insulin infusion. Figure 3
shows the elimination of potassium from the plasma as a function of the initial insulin infusion rates of all patients treated with glucose insulin and potassium. There was no dose-response relationship and maximal insulin-induced potassium elimination was 20 mval/h. Figure 4
shows glucose serum levels and glucose and insulin infusion rates for (A) those patients treated with insulin therapy (protocol 3) and (B) those who received standard postoperative care. Insulin therapy did not cause any further increase in serum glucose and was not associated with hypoglycemia after insulin administration. Postoperative hyperglycemia was briefer than that in the group of patients not treated with high dose-insulin or GIK.
|
|
|
Insulin is a valuable tool for the treatment of patients undergoing cardiac surgery with cardioplegic arrest. High dosages should be applied early postoperatively. High doses of insulin can be safely administered to patients (diabetic or not) after cardiac surgery without adversely affecting glucose or potassium homeostasis. High-dose insulin therapy could be a valuable adjunct in the postoperative care of patients undergoing cardiac surgery. A large randomized study is needed to assess this hypothesis.
Acknowledgments
Doctor Doenst was supported by the Emmy Noether-Program of the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG, Do602/2). We wish to thank Novo Nordisk (Germany) for providing the insulin and funds to cover the study’s liability insurance charges.
References
This article has been cited by other articles:
|
|
 |
|
  T. Albacker, G. Carvalho, T. Schricker, and K. Lachapelle High-Dose Insulin Therapy Attenuates Systemic Inflammatory Response in Coronary Artery Bypass Grafting Patients Ann. Thorac. Surg., July 1, 2008; 86(1): 20 - 27. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. A. Kloner and R. W. Nesto Glucose-Insulin-Potassium for Acute Myocardial Infarction: Continuing Controversy Over Cardioprotection Circulation, May 13, 2008; 117(19): 2523 - 2533. [Full Text] [PDF]
|
|
|
|
 |
|
 T. Doenst, H. Bugger, M. Schwarzer, G. Faerber, M. A. Borger, and F. W. Mohr Three good reasons for heart surgeons to understand cardiac metabolism Eur. J. Cardiothorac. Surg., May 1, 2008; 33(5): 862 - 871. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. D. Evans and Y. Niu Hypolipidaemic effects of high-dose insulin therapy Br. J. Anaesth., April 1, 2008; 100(4): 429 - 433. [Full Text] [PDF]
|
|
|
|
 |
|
 M. T. Dirksen, G. J. Laarman, M. L. Simoons, and D. J.G.M. Duncker Reperfusion injury in humans: A review of clinical trials on reperfusion injury inhibitory strategies Cardiovasc Res, June 1, 2007; 74(3): 343 - 355. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. Ranasinghe, C. J. McCabe, D. W. Quinn, S. R. James, D. Pagano, J. A. Franklyn, and R. S. Bonser How Does Glucose Insulin Potassium Improve Hemodynamic Performance?: Evidence for Altered Expression of Beta-Adrenoreceptor and Calcium Handling Genes Circulation, July 4, 2006; 114(1_suppl): I-239 - I-244. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. Ranasinghe, D. W. Quinn, D. Pagano, N. Edwards, M. Faroqui, T. R. Graham, B. E. Keogh, J. Mascaro, D. W. Riddington, S. J. Rooney, et al. Glucose-Insulin-Potassium and Tri-Iodothyronine Individually Improve Hemodynamic Performance and Are Associated With Reduced Troponin I Release After On-Pump Coronary Artery Bypass Grafting Circulation, July 4, 2006; 114(1_suppl): I-245 - I-250. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. D. Schipke, R. Friebe, and E. Gams Forty years of glucose-insulin-potassium (GIK) in cardiac surgery: a review of randomized, controlled trials. Eur. J. Cardiothorac. Surg., April 1, 2006; 29(4): 479 - 485. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Doenst, D. Wijeysundera, K. Karkouti, C. Zechner, M. Maganti, V. Rao, and M. A. Borger Hyperglycemia during cardiopulmonary bypass is an independent risk factor for mortality in patients undergoing cardiac surgery J. Thorac. Cardiovasc. Surg., October 1, 2005; 130(4): 1144 - 1144. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. W. Quinn, D. Pagano, and R. S. Bonser Glucose and Insulin Influences on Heart and Brain in Cardiac Surgery Seminars in Cardiothoracic and Vascular Anesthesia, June 1, 2005; 9(2): 173 - 178. [Abstract] [PDF]
|
|
|
|
|
|
W. Bothe, M. Olschewski, F. Beyersdorf, and T. Doenst Glucose-Insulin-Potassium in Cardiac Surgery: A Meta-Analysis Ann. Thorac. Surg., November 1, 2004; 78(5): 1650 - 1657. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. G. Pittas, R. D. Siegel, and J. Lau Insulin Therapy for Critically Ill Hospitalized Patients: A Meta-analysis of Randomized Controlled Trials Arch Intern Med, October 11, 2004; 164(18): 2005 - 2011. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. S Lewis, S. L Kane-Gill, M. B. Bobek, and J. F Dasta Intensive Insulin Therapy for Critically Ill Patients Ann. Pharmacother., July 1, 2004; 38(7): 1243 - 1251. [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 |
