| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 1995;60:1203-1209
© 1995 The Society of Thoracic Surgeons
Institute for Biodiagnostics, National Research Council, Winnipeg, Manitoba, Canada, and Department of Surgery, State University of New York at Buffalo, Buffalo, New York
Accepted for publication May 26, 1995.
 |
Abstract |
|---|
 TopFootnotes Abstract IntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Methods. Isolated, beating porcine hearts were perfused antegradely (aortic root pressure = 75 to 95 mm Hg) for 30 minutes. Arrest was induced and maintained for 60 minutes with high K+ blood cardioplegia delivered either antegradely (n = 8) or retrogradely (n = 8) (coronary sinus pressure = 35 to 55 mm Hg). Perfusate was switched to normokalemic blood for recovery of sinus rhythm (30 minutes). Intracellular pH, creatine phosphate, inorganic phosphate, and adenosine triphosphate were monitored continuously and noninvasively with phosphorus 31 magnetic resonance spectroscopy throughout the experiment, and functional variables (rate–pressure product and the positive and negative first derivatives of left ventricular pressure) were assessed concurrently.
Results. Antegrade cardioplegia maintained high-energy metabolites, intracellular pH, and myocardial function. Retrograde normothermic blood cardioplegia resulted in an increase in inorganic phosphate (197% ± 15% of control) and a decrease in creatine phosphate (51% ± 6% of control). There was no significant difference in myocardial function between the two groups (p > 0.05). The magnetic resonance spectroscopy data indicate ischemia occurred within 2 minutes of the initiation of retrograde perfusion.
Conclusions. This study suggests that retrograde normothermic blood cardioplegia causes a transition of the myocardium to ischemic metabolism in the normal porcine heart.
|
Introduction |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
Material and Methods |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
After a midline sternotomy, the brachiocephalic artery was isolated. Intravenous heparin sodium was administered (660 U/kg), and the aorta was cannulated with an RMI 9F, triple-lumen antegrade cardioplegic cannula (Canadian Cardiovascular, Mississauga, ON, Canada). Three lumens simultaneously provided a pressure line, perfusion line, and vent or venous sampling during retrograde perfusion. Umbilical tape was placed around the brachiocephalic artery and around the aortic arch between the braciochepalic and left subclavian arteries. Perfusion down the aortic root was initiated with the perfusion pressure in the aortic cannula kept between 75 and 95 mm Hg, which required a flow rate of 250 to 550 mL/min.
The heart and lungs were isolated en bloc with care taken not to cut the esophagus. The weight of the heart was distributed equally through suspension from the aorta. This was done to ensure that aortic distortion would not cause aortic valve incompetence. The hemiazygos vein was closed with a pursestring suture, and the coronary sinus was cannulated with a 13F dual-lumen retrograde cardioplegia cannula (Sarns, 3M, London, ON, Canada), thus providing concurrent perfusion and pressure monitoring.
A stabilization period began with antegrade flow, and blood content was adjusted whenever necessary to ensure potassium and calcium concentrations of approximately 4 mmol/L and 1 mmol/L, respectively. Hematocrit was maintained at 15% to 20%. Carbon dioxide tension and oxygen tension were maintained at 35 to 40 mm Hg and higher than 200 mm Hg, respectively. The latter was held higher than 200 mm Hg to ensure sufficiently high oxygen saturation of the blood (>99%).
All animals received humane care in compliance with the ``Guide to the Care and Use of Experimental Animals'' published by the Canadian Council on Animal Care [2].
Preparation for Physiologic Monitoring
A phenylphosphonic acid–charged compliant balloon (unstressed volume > 50 mL) was inserted in the left ventricle by way of the left atrium for assessment of diastolic and systolic function before and after the arrest period. The balloon was held in place with a pursestring suture enclosing the mitral leaflets. Phenylphosphonic acid was used as a chemical shift and peak area standard for the 31P MRS. The aortic root, coronary sinus, and left ventricular pressure lines were connected to independent pressure transducers (Cobe Canada Ltd, Scarborough, ON, Canada) and cleared of all in-line air. The transducers were calibrated and zeroed, and the entire preparation was placed in the nuclear magnetic resonance probe. The coronary blood flow was 1 to 1.5 mL • g-1 • min-1 to maintain aortic root pressure of 75 to 95 mm Hg. The whole heart weight was estimated from the previously measured body weight.
The perfusion system was designed to allow easy switching from normokalemic (4 mmol/L) to hyperkalemic (18 mmol/L) blood to arrest the heart and back to normokalemic blood to restore sinus rhythm. The cardioplegia infusion lines were designed for remote switching to either retrograde or antegrade from outside the Faraday cage of the nuclear magnetic resonance instrument. In addition, this design allowed sampling of venous effluent during either retrograde or antegrade cardioplegia infusion. Placement of the retrograde cardioplegia cannula was assured by monitoring the coronary sinus pressure, which was maintained at 35 to 55 mm Hg, requiring a flow rate of 100 to 150 mL/min.
Functional Measurements
Pressure transducers were connected to a multichannel Gould TA 5000 polygraph recorder with digital output (Gould, Valley View, OH). End-diastolic pressure was set at 5 to 10 mm Hg by filling the ventricular balloon with a known quantity of phenylphosphonic acid. Cardiac function was assessed using the first derivative of left ventricular pressure, giving rates of myocardial contraction (+dP/dt) and relaxation (-dP/dt). Heart rate and developed pressure were incorporated into the functional variable rate–pressure product.
Protocol for Nuclear Magnetic Resonance Studies
After the nuclear magnetic resonance probe was tuned and the magnetic field of the heart optimized, a 30-minute series of 2-minute 31P spectra were acquired on the beating heart perfused with normokalemic blood at 37°C (30 minutes). The normothermic blood cardioplegia (blood mixed with lactated Ringer's for a hematocrit of 15% to 25% and adjusted to obtain a final KCl concentration of 18 mmol/L) was initiated antegradely and maintained (antegrade normothermic blood cardioplegia [ANBC] group) or switched to retrograde delivery (RNBC group) for 1 hour during which a series of 31P spectra were acquired. After the arrest period, recovery of normal sinus rhythm was achieved by antegrade reperfusion of normokalemic blood, and a final series of nuclear magnetic resonance spectra were acquired (30 minutes), for a total acquisition time of 2 hours.
Data Analysis
The heights of MR spectral peaks were used to determine the relative changes in inorganic phosphate (Pi), creatine phosphate (CP), and adenosine triphosphate (ATP) (the resonance of the ß-peak of ATP). The values were obtained through the use of in-house analysis software (Allfit). The values were further processed using Microsoft Excel, version 4.0 (Microsoft Corp), appended to functional data, and exported to the software package Statistica 4.5 (StatSoft, Tulsa, OK) for statistical analysis. Data are presented as the mean ± the standard error of the mean. The data were tested for significant differences using both Student's unpaired t test for independent samples and the nonparametric Mann-Whitney U test. Data were considered significant at p values of less than 0.05.
|
Results |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
. Antegrade cardioplegia caused no significant change in the high-energy metabolites Pi, CP, and ATP. Retrograde cardioplegia caused immediate changes (occurring within 2 minutes) in Pi and CP and resulted in a gradual drop in ATP (over a period of 10 to 15 minutes). When compared with changes during ANBC, all changes in high-energy metabolites during RNBC were significant (p < 0.05). The changes in high-energy metabolites during recovery are shown in Figure 2. The alterations in Pi and CP that occurred during RNBC returned to baseline immediately (within 2 minutes) after restoration of function with no significant differences compared with the ANBC group (Figs 3, 4). The levels of ATP did not completely recover in the RNBC group and remained at 79% ± 2.4% of the initial control value. The difference in recovery of ATP between the two groups (ANBC and RNBC) was significant (p < 0.05).
|
|
|
|
. An example of the changes in pHi over time for both groups is shown in Figures 6 and 7. Antegrade cardioplegia maintained normal pHi values, whereas retrograde cardioplegia caused a gradual decline in pHi of almost 0.4 unit to 6.8 ± 0.2 after 60 minutes of perfusion. When normal sinus rhythm was restored after 1 hour of RNBC, pHi returned to normal and was not significantly different from the value in the ANBC group.
|
|
|
shows a comparison of the recovery in +dP/dt, -dP/dt, and rate–pressure product between the ANBC and RNBC groups. All values are expressed as a percentage of the control, recorded at the end of 30 minutes of function prior to initiation of cardioplegia. The mean recovery of +dP/dt and -dP/dt for the antegrade group was 68% and 65%, respectively. In the retrograde group, the recovery of +dP/dt and -dP/dt was 61% and 57%, respectively. The rate–pressure product was 52% in the RNBC group and 64% in the ANBC group. Although retrograde cardioplegia caused a consistently lower recovery of myocardial function, the changes between the two groups were not significant.
|
|
Comment |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Phosphorus 31 MRS can provide quantitive measuring of Pi, CP, ATP, and pHi. Both Pi and CP have distinct resonance frequencies producing sharp peaks in a spectrum that can be quantified. There are three ATP peaks, representing the three phosphate groups. However, the ß-ATP peak was used to follow changes in ATP, as it is least influenced by other merging resonances such as adenosine diphosphate and nicotinamide adenine dinucleotide phosphate. It has been well established that pHi can be determined from the chemical shift of Pi [12]. The protocol used for acquiring data allowed for a time resolution of 2 minutes throughout the experiment. This time resolution was chosen as optimal with a good signal-to-noise ratio and an adequate time frame for observing changes in metabolism.
The model for this study was developed to allow isolation of a perfused, beating heart without any incidence of ischemia, followed by its placement into the bore of a 7.0 T horizontal magnet for observation of changes in metabolites under varying physiologic conditions. This 31P MRS model allowed us to impose RNBC on the heart while continuously monitoring myocardial metabolic status. The porcine heart model was chosen over other models because it is more similar to that of the human heart in terms of biochemistry, size, and susceptibility to fibrillation than that of the more commonly used rat or dog. The isolated heart was chosen to remove unwanted secondary neural and humoral influences on myocardial metabolism and function. In this way, observed changes could be attributed to the effects of a particular method of cardioplegia on myocardial metabolism and function instead of secondary corporeal effects.
Interpretation of Results
High-energy metabolites such as Pi, CP, and ATP were preserved in the myocardium during antegrade cardioplegia. However, homeostasis failed to be preserved when retrograde cardioplegia was used, and within minutes, there was nearly a twofold increase in Pi, and CP decreased by half. This is a strong indication of transition to ischemic metabolism [13]. Under anaerobic conditions, ATP is primarily generated with the breakdown of CP through the creatine kinase–catalyzed reaction:
 |
 |
Accordingly, the decline of mitochondrial ATP production means that Pi will accumulate and CP, decrease. This makes the CP to Pi ratio a very sensitive indicator of ischemic metabolism. The buffering effect of ATP production with the breakdown of CP explains why the changes seen in ATP in response to the onset of ischemia are gradual compared with those of Pi and CP, which alter dramatically within minutes.
The observed rapid transition from aerobic to anaerobic metabolism after initiation of retrograde cardioplegia is indicative of inadequate supply relative to demand typical of hypoperfusion ischemia. The oxygen demand of a beating, unstressed heart is normally approximately 10 mL • 100 g-1 • min-1. Normothermic cardioplegia reduces the demand to approximately 1 mL • 100 g-1 • min-1 [14]. In our study, the typical flow for RNBC was between 80 and 150 mL/min with an arterial oxygen tension of approximately 500 mmol/L. The hematocrit was maintained between 15% and 20% with a hemoglobin concentration of approximately 5 to 7 g/dL.
Total arterial oxygen delivery capacity is equivalent to the total blood oxygen content multiplied by the flow rate. Arterial content can be calculated using the following formula: Hb (g/dL) x SaO2/100 x 1.39 mL/g + 2.35/760 mm Hg x PO2, where 1.39 mL/g is the binding coefficient for milliliters of oxygen to grams of hemoglobin (Hb), 2.35 is the carrying coefficient for milliliters of oxygen per 100 mL of blood per atmosphere of pressure, and SaO2 is arterial oxygen saturation. Therefore, assuming an arterial blood flow rate of 150 mL/min, a hematocrit of 20%, a hemoglobin concentration of 7 g/dL, an arterial oxygen tension of 500 mmol/L, and an arterial oxygen saturation of 100%, then the arterial oxygen content is as follows: (7 g/dL x 1 x 1.39 mL/g + 2.35/760 mm Hg x 500 mmol/L) = 11.3 mL O2/dL. The oxygen delivery capacity is the arterial oxygen content multiplied by the flow rate: 0.113 mL O2/mL x 150 mL/min = 16.95 mL/min. For a heart weighing 300 g, the oxygen delivery would be 5.6 mL • 100 g-1• min-1. However, if 65% to 75% of that flow is nonnutritive, ie, it is shunted through the thebesian vessels, then only approximately 1.4 mL • 100 g-1 • min-1 of oxygen is delivered.
This calculation suggests a sufficient nutrient flow to meet the metabolic demands of the arrested myocardium. However, this calculation makes the basic assumption that there is homogeneous flow distribution, which has been shown not to be the case during retrograde flow [9, 15]. This suggests that there are inadequately perfused areas of the myocardium despite the total oxygen delivery of 1.4 mL • 100 g-1 • min-1.
Antegrade cardioplegia was able to maintain aerobic metabolism, as indicated by the lack of significant change in pHi. Retrograde cardioplegia, however, caused a significant decline in pHi (p < 0.05) to a value of 6.84 ± 0.17. The pHi recovered on antegrade reperfusion with ultimately no significant differences between the two groups. The decrease in pHi during retrograde cardioplegia indicates a moderate intracellular accumulation of protons and is further evidence that RNBC does not maintain aerobic metabolism. There seems to be sufficient coronary flow during RNBC to wash out metabolic wastes, thereby limiting the accumulation of protons and preventing the onset of severe ischemia. These observations are consistent with work by Stahl and associates [8], who found that initiation of RNBC caused a decrease in pHi (<6.8 after 2 hours), a concurrent rise in myocardial lactate concentration, and a significant decrease in recovery of left ventricular function.
The rates of isovolumic contraction (+dP/dt) and relaxation (-dP/dt) are given by the first derivative of the left ventricular pressure curve (dP/dt). In this model the preload is constant; therefore, +dP/dt and -dP/dt are dependent on myocardial contractility and reflect cardiac function [16]. The model involves the insertion of a compliant balloon containing a set volume in the left ventricle; therefore all contraction and relaxation is isovolumic. The rate–pressure product is heart rate multiplied by developed pressure and is an established indicator of cardiac function [17]. Although the recovery of all functional variables in the retrograde group was consistently lower than that in the antegrade group, there was no significant difference between the groups. This may explain why some clinical studies have not seen any significant difference in function between the two cardioplegia methods [4].
Retrograde flow requires lower flow rates (0.3 to 0.5 mL • min-1 • g-1) to avoid damaging the coronary sinus. In addition, a large percentage of the perfusate flowing through the coronary sinus is diverted through collateral veins and is therefore nonnutritive. This study has shown that in the normal isolated porcine heart, RNBC causes a transition to ischemic metabolism within minutes of initiation, a transition indicated by a rise in Pi, a decrease in CP and pHi, and a gradual decline in ATP. In this model, RNBC did not provide sufficient nutritive flow to maintain aerobic metabolism. The RNBC reduced pHi to 6.8 after 1 hour; therefore, ischemia was only mild, and with the exception of ATP, there was no significant difference in recovery between the two groups.
Limitations of the Model
Inorganic phosphate exists chiefly as either HPO42- or H2PO4-. In solution, these two forms give rise to one resonance with a frequency dependent on the pHi [12]. It is possible to form a standard titration curve with which the chemical shift can be compared and pHi calculated. Therefore using 31P MRS, one can monitor pHi continuously and noninvasively, but there is a limit to the accuracy for pHi when the Pi peak is minimal, ie, during normal metabolism. This inaccuracy has to do with the similar resonance of the second phosphate of 2,3 diacylphosphoglycerate in blood, which overlaps the Pi peak [18]. During ischemia when the Pi peak is augmented and 2,3 diacylphosphoglycerate has a minimal contribution to the overall peak, the pHi may be calculated more accurately.
The coil used for the acquisition of MRS was large and encompassed the majority of both the right and left ventricles. Therefore, information about the metabolic status of the myocardium from this study was limited to an average of both ventricles. As a result, little information is available concerning the flow distribution during either cardioplegia technique. Although this model is limited because it does not give localized information, the results obtained from this study provide a useful indication of the overall metabolic status of the myocardium during either retrograde or antegrade perfusion and correlate those effects to function. The metabolic and functional data given here, combined with previous flow distribution studies, allow us to draw a number of conclusions about the efficacy of RNBC in normal porcine hearts. Cardioplegia delivery could not be confirmed other than by monitoring the coronary sinus pressure and observing the venous return from the aortic root during RNBC. This may be improved in the future by the combination of this model with the use of radiolabeled microspheres, which would confirm the delivery of cardioplegia to the end-organ target.
The limitation in using a porcine heart involves species specificity. Other than the primate heart, the porcine heart is thought to be the closest model to the human heart. However, there are limitations in extrapolating data obtained with this model for application to humans. The isolated heart model has the advantage of removing external influences such as neural and hormonal responses, but the isolation is relatively invasive and maintenance on bypass without any hormonal protection can result in both cell damage and, in combination with the lower hematocrit used for blood cardioplegia, extracellular edema. In addition, perfusion for 1 hour with hyperkalemic cardioplegia may cause reversible disturbances in cardiomyocyte handling of calcium and irreversible alterations in myofibril function causing the relatively low functional recovery observed in both groups [19]. Use of a normal, healthy heart provides a less clinically relevant model than one that has been ischemically injured or one on which acute coronary occlusion has been imposed.
Although the results of this study have shown that RNBC causes a transition to ischemic metabolism, it is possible that in the more clinically relevant, coronary-obstructed model, RNBC could provide superior protection to the myocardial tissue supplied by the occluded artery than ANBC. This advantage of RNBC in the occluded left anterior descending coronary artery model has been demonstrated by a number of studies [20, 21]. One hypothesis for this observed advantage of RNBC in superiorly protecting myocardial tissue distal to an occluded left anterior descending coronary artery is the observed preferential perfusion of the more susceptible subendocardial tissue distal to an occluded coronary artery [21]. This issue is currently being addressed with the development of a model in which the left anterior descending branch of the coronary artery is occluded, and localized 31P MRS is used to evaluate and compare the metabolic state of the myocardium distal to the occlusion. In addition, localized spectroscopy can be used transmurally with a surface gradient coil to observe distinct layers of the ventricular wall [22]. This approach should resolve whether RNBC is advantageous in protecting the susceptible regions of both ventricles in cases involving coronary occlusion. This study suggests that further research into RNBC as a means of cardioplegia delivery should be considered.
The possible failings of RNBC to adequately perfuse certain areas of the myocardium have been addressed by Ihnken and co-workers [23], who showed that RNBC and ANBC could be safely combined to have simultaneous perfusion of the coronary sinus and the proximal aorta. This was a comprehensive study including both an experimental animal model and a clinical trial. Porcine hearts had 1 hour of aortic clamping, during which simultaneous retrograde and antegrade cardioplegia was delivered at 200 mL/min. Coronary sinus pressure was maintained at less than 30 mm Hg. There was no right or left ventricular edema, lactate production, or lipid peroxidation, and there was good recovery of left ventricular end-systolic elastance and preload-recruitable stroke work index (101% ± 3% and 109% ± 90%, respectively) after bypass [23]. The clinical segment of the study included 155 high-risk patients with an average clamp time of 94 minutes. Eighteen patients required postoperative circulatory assistance, 16 of whom were in cardiogenic shock preoperatively. The incidence of postoperative myocardial infarction was 2%, and the mortality rate was 4% [23]. This study demonstrated that simultaneous aortic root and coronary sinus perfusion is a viable cardioplegic strategy that takes advantage of the benefits of both antegrade and retrograde delivery.
|
Acknowledgments |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Technical assistance was provided by Lori Gregorash and Rachelle Perchaluk. Statistical analysis was performed in consultation with Randy Summers.
|
Footnotes |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  O. P. Elvenes, C. Korvald, R. Myklebust, and D. Sorlie Warm retrograde blood cardioplegia saves more ischemic myocardium but may cause a functional impairment compared to cold crystalloid Eur. J. Cardiothorac. Surg., September 1, 2002; 22(3): 402 - 409. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Ye, J. Sun, E. F. Hoffenberg, J. Shen, L. Yang, R. Summers, T. A. Salerno, and R. Deslauriers DOES RETROGRADE WARM BLOOD CARDIOPLEGIC PERFUSION PROVIDE BETTER PROTECTION OF ISCHEMIC AREAS THAN ANTEGRADE WARM BLOOD CARDIOPLEGIC PERFUSION? A MAGNETIC RESONANCE STUDY IN PIG HEARTS J. Thorac. Cardiovasc. Surg., May 1, 1999; 117(5): 994 - 1003. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. S. Aldea Complementary use of antegrade and retrograde cardioplegia Ann. Thorac. Surg., September 1, 1998; 66(3): 697 - 698. [Full Text] [PDF]
|
|
|
|
|
|
P. K. Kaukoranta, M. V.K. Lepojarvi, K. T. Kiviluoma, K. V. Ylitalo, and K. J. Peuhkurinen Myocardial protection during antegrade versus retrograde cardioplegia Ann. Thorac. Surg., September 1, 1998; 66(3): 755 - 761. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. C. de Oliveira, T. J. Boeve, D. F. Torchiana, H. L. Kantor, J. S. Titus, C. J. Schmidt, C.-z. Lu, J. Kim, W. M. Daggett, and G. A. Geffin ISCHEMIC INTERVALS DURING WARM BLOOD CARDIOPLEGIA IN THE CANINE HEART EVALUATED BY PHOSPHORUS 31-MAGNETIC RESONANCE SPECTROSCOPY J. Thorac. Cardiovasc. Surg., December 1, 1997; 114(6): 1070 - 1080. [Abstract] [Full Text]
|
|
|
|
|
|
M. A Borger, G. Cohn, R. D Weisel, S. E Fremes, T. M Yau, and R. Deslauriers COMMENTARY ON ISCHEMIC INTERVALS DURING WARM BLOOD CARDIOPLEGIA IN THE CANINE HEART EVALUATED BY PHOSPHORUS 31-MAGNETIC RESONANCE SPECTROSCOPY J. Thorac. Cardiovasc. Surg., December 1, 1997; 114(6): 1079 - 1080. [Full Text]
|
|
|
|
|
|
M. Carrier, L. C. Pelletier, and N. R. Searle Does Retrograde Administration of Blood Cardioplegia Improve Myocardial Protection During First Operation for Coronary Artery Bypass Grafting? Ann. Thorac. Surg., November 1, 1997; 64(5): 1256 - 1262. [Abstract] [Full Text]
|
|
|
|
|
|
M. Tonz, O. N. Krogmann, O. M. Hess, B. Leskosek, T. Mihaljevic, L. K. von Segesser, and M. I. Turina Effect of Intermittent Warm Blood Cardioplegia on Functional Recovery After Prolonged Cardiac Arrest Ann. Thorac. Surg., October 1, 1996; 62(4): 1146 - 1151. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
