| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2005;130:1236
© 2005 The American Association for Thoracic Surgery
Editorial |
Department of Cardiothoracic Surgery, Stanford University, Stanford, Calif
Received for publication June 14, 2005; accepted for publication July 29, 2005. * Address for reprints: Frank L. Hanley, MD, Department of Cardiothoracic Surgery, Stanford University, 300 Pasteur Dr, CRVB 2nd Floor North, Stanford, CA 94305-5407 (Email: mcharles{at}lpch.org).
Opinions among pediatric heart surgeons vary regarding the routine use of deep hypothermic circulatory arrest (DHCA) during the surgical management of neonates and infants with complex congenital heart disease. Currently, some surgeons use DHCA routinely, others use it selectively, and still others have essentially abandoned the technique, opting for a perfusion strategy that uses continuous cardiopulmonary bypass (CPB). While acknowledging that both DHCA and CPB are abnormal physiologic states associated with risk, surgeons in each camp believe that their particular management scheme provides the best opportunity for morbidity-free outcome. This subject has become controversial in recent years, sometimes it seems to the degree that it is poised to join the ranks of other well-known contentious topics, like religion and politics, which have become increasingly unsuitable for polite and civilized discussion. There are three basic reasons for this development. First, both the quality of cardiac repair and neurodevelopmental status are at issue. High stakes promote controversy. Second, whereas in the past DHCA was a necessity, advances in surgical technique and basic technology have now relegated DHCA to one alternative among several. Having options leads to controversy. Finally, although DHCA generates strong and differing opinions among surgeons, there is not nearly enough factual information to definitively support any one view. Uncertainty invites controversy.
What are the hard data that might assist us in determining the relative efficacy of performing neonatal heart surgery with circulatory support strategies that either incorporate DHCA or use continuous CPB alone? Because the rationale for choosing one support technique over the other is to achieve the best technical repair with the least morbidity, it follows that the pertinent data will relate to two main areas: the quality of the cardiac surgical reconstruction achievable and the total body morbidity that is incurred with each strategy.
The Quality of Cardiac Reconstruction With DHCA and Continuous CPB
The point has been made by some advocates that DHCA promotes superior technical outcomes. The main reason given is that the surgical field is unencumbered by blood and bypass paraphernalia, allowing better visualization of critical structures. It is commonplace to see statements in publications that advocate the use of DHCA, referring to "the extraordinary means necessary to maintain continuous CPB" as a reason to use DHCA. On the other hand, those who prefer continuous CPB argue that neonatal operations can be performed routinely without compromising visualization, and when the occasional particularly challenging technical problem or complication arises, the absence of the time limitation that is associated with DHCA actually might allow a superior technical result. These differing views might reflect personal styles and preferences, comfort levels with different surgical skill sets, prior training, or simply attempts to justify the use of one particular support strategy. In the final analysis, all of the views on both sides of the issue are purely subjective.
Anecdotal experience, although not necessarily objective, does carry more weight than subjective opinion alone and can be compelling when it reaches critical mass. Anecdotal evidence is mounting from many groups that effective repairs can be achieved with continuous CPB. A personal observation on the basis of extensive experience with DHCA before 1995 and then exclusive use of continuous CPB since 1995 is that technical outcomes and mortality rates are no different between the two support strategies, regardless of patient age and size or whether aortic arch reconstruction was involved. This experience involves 8 surgeons and includes the use of continuous CPB for all neonatal cases, including more than 300 involving aortic arch repair.
What about objective data? Are there any that support one perfusion strategy over the other? Available publications (both prospective and retrospective studies) in which a range of operations were performed with both support strategies fail to identify any instance in which technically unsatisfactory reconstructions or higher mortality rates were observed more frequently in one circulatory support group or the other. Many of these publications document outcomes after neonatal repair with continuous CPB for a number of lesions, ranging from hypoplastic left heart syndrome to transposition of the great arteries, that are equivalent to those achieved with DHCA. Thus there appear to be no objective data to support the position that DHCA is justified to achieve optimal technical reconstructions, and in fact, there are a good deal of indirect data and experience that run counter to this position.
Morbidity Related to DHCA and Continuous CPB
Both DHCA and continuous CPB represent abnormal physiologic states, and thus both have the potential for causing morbidity, particularly central nervous system (CNS) morbidity. Although it is easy to demonstrate that the basic physiologic perturbations related to these 2 support techniques are substantially different, it is much more difficult to demonstrate differences in CNS clinical outcomes with each strategy.
Pathophysiology
DHCA causes an immediate cellular energy supply-demand imbalance, which results in easily identifiable and progressive intracellular metabolic derangements. As stated succinctly by Ungerleider and Gaynor
1
in their editorial in the May 2004 issue of this Journal, all evidence points to a linear dose response between the duration of DHCA and neurophysiologic derangement. After some duration, these derangements will cause cell death. Before that duration is reached, if energy supply and demand are brought back into balance, the metabolic derangements will reverse, and the cell will recover to survive. The relationship between cell survival after a sublethal insult and return to cell normality is, however, complex and not well understood. The issue is complicated further by many independent influences, both intrinsic and extrinsic to the individual patient, that will influence exactly when permanent damage occurs. The most important of these might well be biologic variability regarding vulnerability to DHCA; however, others include the effects of deep hypothermia itself, the propensity for neurocellular excitotoxicity, altered cerebrovascular autoregulation, critical closing pressure, and the no-reflow phenomenon after DHCA. Thus the likelihood of permanent CNS injury increases with the duration of DHCA, and at some point, which varies among individuals, that likelihood becomes inevitable.
The physiologic perturbations associated with continuous CPB are in many ways more complex than those related to DHCA. Because the circulation is maintained, the potential for damage is not related to an obligatory energy supply-demand imbalance but rather to variable and uncertain events associated with extracorporeal circulation. Three hours of CPB, or even 3 days for that matter, might or might not be associated with irreversible cell damage. The fact that it is not unusual for a patient to undergo extracorporeal membrane oxygenation (ECMO) for 3 days or longer and not have CNS morbidity is a very important observation. It implies that the physiologic perturbations that lead to morbidity are not inevitable and that duration of CPB, although a factor, does not play a mandatory role. In other words, there is no linear dose-response curve. Rather, CPB exposes the patient to the possibility of events that create morbidity. It is both logical and quite reasonable to assume that these morbid events tend to accumulate with increasing duration of CPB; however, factors other than duration appear to play equally important, if not more important, roles. Data from near-infrared spectroscopy cerebral monitoring and transcranial cerebrovascular Doppler monitoring indicate that microemboli and cerebral malperfusion occur commonly during many CPB maneuvers that are unrelated to duration of CPB, such as aortic clamping, starting and stopping CPB, venting and deairing maneuvers, matching of perfusion flow rate to body temperature, and arterial and venous cannula insertion, positioning, and removal.
Despite these effects, there is no denying the ongoing constant but low instantaneous risk that accumulates with the duration of CPB. It has been suggested that the causative link in this association is that duration allows accumulation of noxious stimuli, such as the microembolic load and the degree of inflammatory response. Although it seems somewhat intuitive that these factors are causative, on closer evaluation, this conclusion can be challenged. The duration of CPB is intimately linked to a number of other variables that might be difficult to separate out, even with the most careful multivariate modeling. These variables include the complexity of the operative procedure itself, technical problems during CPB that can be difficult to define and document, surgical bleeding and magnitude of transfusion, postpump hemodynamic instability, and perioperative cardiac reserve. All of these variables and others might influence neurologic outcome, whereas duration of CPB itself might only be a marker.
One can ask whether there are situations in which the duration of CPB is not linked to these other potentially obfuscating cardiac and technical variables. If we look just outside our own field to the classic neonatal ECMO experience for lung disease, we might find some insights. ECMO is normothermic CPB; however, it is performed in the absence of cardiac disease, and the duration of ECMO is unrelated to the cardiac surgical variables mentioned above. The duration of ECMO for neonatal lung disease typically lasts 100 to 150 hours. A number of carefully performed late follow-up studies in this patient population focusing on neurodevelopmental outcome indicate that the majority of patients are normal at follow-up. Admittedly, deriving inferences from these data and applying them to cardiac surgery–related CPB must be done with circumspection; however, the simple observation that many patients who are subjected to 150 hours of extracorporeal circulation experience detectable neurodevelopmental consequences is very important. It implies that in the setting of cardiac surgery, where variations in CPB duration are measured in minutes from case to case rather than in hours or days, duration itself might not be a very important factor in determining neurologic injury.
Does DHCA Replace the Risks of CPB or Add to Them?
DHCA and continuous CPB are not used simultaneously; rather, a period of DHCA replaces an equal period of continuous CPB. Implicit in many discussions is the assumption that when DHCA is used, all of the risks associated with continuous CPB are eliminated and replaced by those of DHCA. This is not the case because CPB is always necessary, even when DHCA is used as an adjunct to replace part of the CPB run. When DHCA is used, the only component of continuous CPB that is eliminated from the overall perfusion management is precisely the duration of CPB equal to the duration of DHCA. This duration will, by necessity, be short, measured in minutes. All of the other continuous CPB maneuvers (and associated risks) remain and are added to the intrinsic risk related to DHCA. It can even be argued that some of the continuous CPB risks, for example those related to the additional cannula manipulation and additional stopping and restarting the pump associated with DHCA, actually increase when DHCA is used. Thus when DHCA is used, it appears that it is most accurate to view the risk of DHCA as additive to that of CPB, with very little to be gained by the minor reduction in CPB duration.
Is It Possible to Define a Clinically Useful Safe Period of DHCA?
This requires having the ability to prospectively determine, in an individual patient, the duration of DHCA that results in morbidity equal to or less than that of the alternative perfusion strategy with continuous CPB. This is a daunting task, and it is easy to argue that we are not even close. Several factors contribute to this: lack of data, difficulty with outcome measures, and biologic variability.
Despite some commendable efforts, the data gap is wide. Two clinical studies from outstanding groups have attempted to characterize the relationship between the duration of DHCA and CNS morbidity. Both studies are convincing in demonstrating the general relationship between CNS damage and increasing duration of DHCA. However, these studies, the best we have in our literature, simply underscore the limitations that exist in trying to go beyond this general relationship. In the study by Gaynor and colleagues
2
in this issue of the Journal, continuous electroencephalographic monitoring is used to identify seizure activity in neonates undergoing cardiac surgery with either DHCA or continuous CPB. The authors argue convincingly that the likelihood of seizure activity increases with the duration of DHCA. The authors go on to compare the seizure activity rate in the patients undergoing DHCA with that in the patients undergoing continuous CPB in an attempt to shed some light on the following important question: Are there durations of DHCA for which the CNS risk is no greater than that for continuous CPB? Unfortunately, a combination of limited raw data and a poor choice of statistical modeling invite the casual reader to draw very misleading conclusions. The raw data consist of 18 outcome events (presence of electroencephalographic seizure activity) distributed among 117 patients undergoing DHCA and 2 events among 61 patients undergoing continuous CPB. In recognition of the limited raw data, the authors have chosen to separate the patients undergoing DHCA into only two groups, one in which DHCA duration was less than 40 minutes and the second in which DHCA duration was greater than 40 minutes. The rates of events are then compared among the continuous CPB group and each of the two DHCA groups. The analysis shows no difference in the rate of events between the continuous CPB group and the group undergoing less than 40 minutes of DHCA, whereas the group undergoing more than 40 minutes of DHCA showed a higher rate of events. Strictly speaking, the statistical analysis is performed correctly; however, the modeling of the data into the groups undergoing DHCA of greater than 40 minutes and less than 40 minutes is unfortunate, in that it carries the potential to be unintentionally misleading. In reality, the lack of difference between the group undergoing continuous CPB and the group undergoing less than 40 minutes of DHCA does not identify 40 minutes and shorter as a safe duration of DHCA for which the risk of CNS injury is indistinguishable from that with continuous CPB.
To emphasize the arbitrary nature of dividing the patients undergoing DHCA into two groups, with the split occurring at 40 minutes' duration, one can alternatively break the patients undergoing DHCA in the study by Gaynor and colleagues
2
into three separate groups rather than two: 1 to 25 minutes of DHCA, 26 to 50 minutes of DHCA, and more than 50 minutes of DHCA. Analysis with this model and insertion of the actual raw data from the article suggests that DHCA of 1 to 25 minutes' duration is indistinguishable from continuous CPB, but DHCA of 26 to 50 minutes' duration and more than 51 minutes' duration are both of higher risk than continuous CPB. One is tempted (incorrectly) to conclude that DHCA of up to 25 minutes' duration is just as safe as continuous CPB. The purpose of this exercise is not to claim a superior modeling of the data, because the three groups of DHCA treatment have all the same inadequacies as the two groups of DHCA treatment chosen by the authors. Neither analysis provides information that can be used to define a safe period of DHCA, and both invite the casual reader to draw incorrect inferences.
The other study that attempts to define a safe period of DHCA is by Wypij and colleagues,
3
published in the November 2003 issue of this Journal. This analysis, along with its companion analysis by Bellinger and associates
4
in the same issue, has already been the focus of one mostly favorable editorial, the one by Ungerleider and Gaynor
1
previously cited here. The acknowledgements are well deserved because these analyses represent the latest contributions from the highly regarded ongoing neurodevelopmental study of 171 neonatal patients undergoing arterial switch originally reported by Newburger and coworkers
5
in 1993. The study by Wypij and colleagues
3
attempts to define the relationship between duration of DHCA and neurodevelopmental morbidity. The basic message of the analysis is that the duration of DHCA correlates with increasing neurodevelopmental deficit in a nonlinear fashion. From 0 to 40 minutes of DHCA, there is no correlation, and at greater than 40 minutes, there is a linear direct correlation. The 40-minute point itself represents the cutoff point. Durations of DHCA of less than this have no effect on neurodevelopmental outcome and thus are stated to be safe. The statistically derived 40-minute point has confidence limits, and the 95% lower confidence limit is 32 minutes. The authors argue that the conservative estimate for the safe period of DHCA would be less than this 32-minute duration. In many regards, the analysis represents a landmark article in attempting to define this relationship, and it does provide important insights. Nonetheless, it has a number of limitations that combine to make it unconvincing in its attempt to define a safe period of DHCA.
First, as already noted in the Ungerleider and Gaynor
1
editorial, it is difficult to accept the validity of the piecewise linear regression modeling and its attendant cutoff points when all available physiologic data suggest a monotonic linear dose response between the duration of DHCA and neurometabolic derangement. One is left to wonder whether the cutoff points represent the threshold of our ability to test for neurodevelopmental deficit rather than the threshold for the deficits themselves.
Second, quantitative assessment of neurodevelopment is not only multidimensional, but it is also extraordinarily complex within each dimension. As a result, it is orders of magnitude more difficult than measuring, for example, renal filtration. Wypij and colleagues
3
make a laudable attempt to assess an array of neurodevelopmental parameters, but even within this limited and selective array, the data are confounding if judged on the basis of the criterion that they define a safe period of DHCA. Specifically, the 32-minute value referred to above is based on an analysis that combines the results from 6 different neurodevelopmental tests. However, when one examines the lower confidence limit for the cutoff point for each of the 6 neurodevelopmental tests individually, as shown in Table 2 in the article, the values ranged much more widely. For example, the value for the test of fine motor skill was 13 minutes, which is quite different from 32 minutes. How does one use these data in the real-life clinical setting? If 29 minutes is safe for speech development, and 13 minutes is safe for fine motor skill, but 32 minutes is safe if we combine the information from 6 disparate neurodevelopmental tests, what does the cardiac surgeon tell the parents of a prospective patient? Even more fundamentally, is it valid to combine the information from 6 disparate neurodevelopmental tests to achieve more data points so that the 95% confidence limit is narrowed, when performance on these disparate tests might depend on very specialized function in different brain areas that might well have different vulnerabilities to DHCA? Is this combining of data akin to adding apples and oranges?
Finally, it is also instructive to revisit the raw data in the analysis of Wypij and colleagues.
3
Although patients are characterized as undergoing DHCA or undergoing low-flow bypass, all patients in both groups underwent a period of DHCA. Because of the study design, the large majority of patients were clustered in two separate ranges of DHCA duration, roughly 0 to 20 minutes (the low-flow group) and 40 to 70 minutes (the DHCA group). By using estimates derived from the panels in Figure 1 in this article, it can be determined that only approximately 16% (25/155) of patients underwent DHCA for durations of between 21 and 40 minutes. It is not particularly reassuring that the duration of DHCA that this analysis recommends as the safe period, 32 minutes, falls precisely in the middle of this no man's land of raw data.
In both the studies by Gaynor and associates
2
and Wypij and colleagues,
3
both the raw data and the modeling are too limited to define a duration of DHCA that can be considered safe. These studies do support, however, the general conclusion that the risk of CNS morbidity increases with increasing duration of DHCA.
Obtaining adequate amounts of raw data and modeling it appropriately to answer the important questions might ultimately be achievable. A more vexing problem is defining outcome measures of CNS injury. This issue has already been touched on above. Quantitative measurement of CNS deficit in all its nuances is difficult under any conditions but particularly in the very young. Thus CNS deficit or damage is often assessed indirectly by using a number of methods. These include functional assessment by physical examination and interactive testing, electroencephalography, imaging studies to assess structural changes, and biochemical markers. No single test or method is definitive or quantitative with respect to assessing CNS deficit.
This difficulty in assessing outcome is only made more problematic when attempts are made to use these methods to project future CNS deficits. It is known that there are potentially serious long-term effects that result from transient CNS perturbations detected at the time of extracorporeal perfusion. These associations are, however, not completely understood. As one example, there is accumulating evidence that transient CNS electrical instability (seizures) in neonates after DHCA or CPB predicts CNS deficit that is identifiable only later in life. What we do not yet know is whether these affected individuals will show even more profound limitations in CNS reserve later in life, say as elderly individuals?
Considering what is currently being learned about the longer-term implications of transient neonatal CNS electrical perturbations in association with DHCA and CPB, should we be concerned about other transient CNS perturbations that occur during extracorporeal circulation? For example, the cellular energy supply-demand imbalance associated with DHCA results in a profound decrease in high-energy phosphates within minutes of the initiation of DHCA. These energy stores recover if DHCA is not prolonged. Will this transient metabolic CNS perturbation be shown to be a marker of longer-term CNS deficit in coming years, just as transient electrical instability has?
The hallmark of this discussion regarding markers of CNS deficit is uncertainty. Given the uncertainties as described, careful consideration must be given to avoiding techniques that cause measurable CNS perturbations, however transient, especially when other techniques exist that do not cause these same perturbations.
Finally, there is the matter of biologic variability. Current medical knowledge has difficulty predicting prospectively why an aspirin is tolerated well by one person and causes gastrointestinal bleeding in another. Considering the complexity of the CNS, it is almost a certainty that tolerance to the energy supply-demand imbalance created during DHCA will show substantial individual variability. Even with sufficient raw data, ideal outcome measures, and appropriate modeling, a statistically derived safe duration of DHCA would have, ironically, limited value in the very situation that is the catalyst for its derivation, that is, as a prospective predictor in an individual patient. Use of a duration of DHCA that is derived statistically and is considered to be, on average, safe might well, in the sensitive individual, have negative consequences. Prospective utility of such a value requires the development and mastery of individual physiologic CNS profiling, something that is not currently available. Again, the key word is uncertainty. Uncertainty is acceptable in association with obligatory techniques but not when alternatives exist.
Summary
When consideration is given to the basic pathophysiology of DHCA, the many uncertainties inherent in defining a safe period of DHCA, and the observation that DHCA largely adds to, rather than replaces, CPB-related risk, it would seem that there should be some relatively compelling reason to justify the routine use of DHCA. Yet compelling justification does not seem to exist at present. There are alternative perfusion strategies that are available for essentially every neonatal operation that the pediatric heart surgeon performs. These alternative strategies can be used routinely and reliably without concerns about compromising the technical surgical repair. Furthermore, the emerging ability to monitor brain perfusion in real time during continuous CPB by using various technologies provides the opportunity to react to unwanted physiologic states, essentially providing ongoing quality control with immediate feedback.
Having taken this position, I would emphasize that currently there is no definitive evidence that points to superior CNS outcomes with either continuous CPB alone or CPB with short periods of adjunctive DHCA. For the foreseeable future, we will have to deal with one of those real-life situations for which complete information is not available to guide our decision. Thus any decision regarding this choice will involve weighing some mix of limited directly applicable evidence, peripheral evidence, logic, and experience. A lack of definitive evidence favoring one strategy or the other, however, is not proof that the two strategies are equal. A lack of such evidence can have many causes, only one being that in reality the two perfusion strategies are equally safe. Other plausible causes have been discussed in this editorial.
Author's Note
The arguments made in this editorial are based on a review of more than 80 pertinent publications. A more extensive bibliography is available in the online version of this editorial.
References
Related Article
This article has been cited by other articles:
|
|
 |
|
  C. A. Joynt, C. M.T. Robertson, P.-Y. Cheung, A. Nettel-Aguirre, A. R. Joffe, R. S. Sauve, W. S.G. Biggs, N. J. Leonard, D. B. Ross, I. M. Rebeyka, et al. Two-year neurodevelopmental outcomes of infants undergoing neonatal cardiac surgery for interrupted aortic arch: A descriptive analysis J. Thorac. Cardiovasc. Surg., October 1, 2009; 138(4): 924 - 932. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. D. Durandy, M. Younes, and B. Mahut Pediatric Warm Open Heart Surgery and Prolonged Cross-Clamp Time Ann. Thorac. Surg., December 1, 2008; 86(6): 1941 - 1947. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Lacour-Gayet Invited Commentary Ann. Thorac. Surg., December 1, 2008; 86(6): 1947 - 1947. [Full Text] [PDF]
|
|
|
|
 |
|
 A. Straub, D. Schiebold, H. P. Wendel, C. Hamilton, T. Wagner, E. Schmid, K. Dietz, and G. Ziemer Using reagent-supported thromboelastometry (ROTEM(R)) to monitor haemostatic changes in congenital heart surgery employing deep hypothermic circulatory arrest Eur. J. Cardiothorac. Surg., September 1, 2008; 34(3): 641 - 647. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J. Rastan, T. Walther, N. A. Alam, I. Daehnert, M. A. Borger, F. W. Mohr, J. Janousek, and M. Kostelka Moderate versus deep hypothermia for the arterial switch operation -- experience with 100 consecutive patients Eur. J. Cardiothorac. Surg., April 1, 2008; 33(4): 619 - 625. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Lacour-Gayet and S. Goldberg Surgical repair of truncus arteriosus associated with interrupted aortic arch MMCTS, March 28, 2008; 2008(0328): 2451. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Pretre and M. I. Turina Deep Hypothermic Circulatory Arrest Card. Surg. Adult, January 1, 2008; 3(2008): 431 - 442. [Full Text]
|
|
|
|
|
|
J. G.T. Augoustides The conduct of experimental circulatory arrest: the search for clinical relevance. J. Thorac. Cardiovasc. Surg., November 1, 2007; 134(5): 1381 - 1382. [Full Text] [PDF]
|
|
|
|
|
|
A. F. Corno Systemic venous drainage: can we help Newton? Eur. J. Cardiothorac. Surg., June 1, 2007; 31(6): 1044 - 1051. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. D. Williams and C. Ramamoorthy Brain monitoring and protection during pediatric cardiac surgery. Seminars in Cardiothoracic and Vascular Anesthesia, March 1, 2007; 11(1): 23 - 33. [Abstract] [PDF]
|
|
|
|
|
|
M. R d. Leval 'Because we can, should we ...?' Eur. J. Cardiothorac. Surg., November 1, 2006; 30(5): 693 - 694. [Full Text] [PDF]
|
|
|
|
|
|
R. L. Hannan, M. A. Ybarra, J. W. Ojito, F. A. Alonso, A. F. Rossi, and R. P. Burke Complex Neonatal Single Ventricle Palliation Using Antegrade Cerebral Perfusion. Ann. Thorac. Surg., October 1, 2006; 82(4): 1278 - 1285. [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 |
