Ann Thorac Surg 1998;65:1003-1008
© 1998 The Society of Thoracic Surgeons
Arrest Duration Influences Postcardioplegia Electrophysiologic Recovery and Reperfusion Arrhythmias
C. Patrick Murrah, MDa,
Edward R. Ferguson, MDa,
Russell D. Spruell, BSEEa,
William L. Holman, MDa
a Division of Cardiothoracic Surgery, University of Alabama at Birmingham, Birmingham, Alabama, USA
Accepted for publication October 23, 1997.
Address reprint requests to Dr Holman, Department of Surgery, University of Alabama at Birmingham, University Station, Birmingham, AL 35294-0007
e-mail: (wholman{at}holman.cvsr.uab.edu)
 |
Abstract
|
|---|
Background. This study tests the hypothesis that postcardioplegia electrophysiologic recovery is influenced by the duration of cardioplegic arrest.
Methods. Pigs were randomized to various durations of cardioplegic arrest (group I, 15 minutes; group II, 60 minutes; group III, 120 minutes). Electrophysiologic data included limb lead, atrial and ventricular epicardial, and ventricular endocardial electrocardiograms. Variables included times for earliest electrical activity and sinus rhythm; number of defibrillations; mechanism for reperfusion ventricular fibrillation; and time until last ventricular fibrillation.
Results. Time to last ventricular fibrillation was 73 ± 8, 134 ± 23, and 238 ± 23 seconds for groups I, II, and III (mean ± standard error of the mean; p < 0.05 between group III versus groups I and II). The number of defibrillations was 1.0 ± 0.3, 5.8 ± 1.2, and 10.5 ± 1.1 for groups I, II, and III (p < 0.05 between groups). The time to sinus rhythm was 66 ± 8, 192 ± 27, and 249 ± 23 seconds for groups I, II, and III (p < 0.05 group I versus groups II and III). The most common mechanism for reperfusion arrhythmias was an accelerating ventricular tachycardia that initiated fibrillation (79 of 167 episodes). However, in many instances postdefibrillation amplifier saturation masked the initiation of reperfusion arrhythmias.
Conclusions. Electrophysiologic recovery after cardioplegic arrest is influenced by the duration of cardioplegic arrest.
|
Introduction
|
|---|
Arrhythmias commonly occur when blood flow returns to the heart after a period of cardioplegic arrest. These arrhythmias are defined as postcardioplegia reperfusion arrhythmias. Postcardioplegia reperfusion arrhythmias occur spontaneously and include ventricular tachycardia and ventricular fibrillation. Defibrillation or cardioversion of these arrhythmias is often unsuccessful during the initial period of postcardioplegia reperfusion. However, as reperfusion progresses and the heart recovers from cardioplegic arrest, defibrillation or cardioversion is usually successful.
A better understanding of postcardioplegia reperfusion arrhythmias will increase our knowledge about ischemia–reperfusion injury that occurs during cardiac operations. This information and further related studies may also suggest new ways to optimize postcardioplegia electrophysiologic recovery and treat the occasional patient who develops persistent ventricular arrhythmias during reperfusion.
The present study correlates electrophysiologic variables measured during postcardioplegia reperfusion with the duration of cardioplegic arrest. The hypothesis is that postcardioplegia reperfusion arrhythmias and electrophysiologic recovery are influenced by the duration of cardioplegic arrest.
|
Material and methods
|
|---|
The interaction of postcardioplegia reperfusion electrophysiology and antecedent duration of cardioplegic arrest was studied in an intact porcine model of cardiopulmonary bypass. All animals in this study received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (NIH publication 85-23, revised 1985).
Thirty pigs of both sexes, weighing 25 to 30 kg, were anesthetized, intubated, and mechanically ventilated with nitrous oxide and oxygen at a ratio of 2:1. A constant infusion of pancuronium and sodium pentobarbital was used to maintain anesthesia.
Catheters were placed in the right external jugular vein and left common carotid artery. Systemic blood pressure (phasic and mean) was measured continuously through the common carotid artery catheter. The left hemiazygos vein was ligated, and a catheter for coronary venous blood sampling was placed with the tip 2 cm past the ostium of the coronary sinus. Right ventricular, left ventricular, and right atrial bipolar epicardial electrodes were placed. Right and left ventricular bipolar endocardial electrodes were also placed. The bipolar endocardial electrodes were placed using 28-gauge wires that were insulated except at the tip, where they were bent into a small barb. The wires were passed across the ventricular wall through a large-bore needle. The needle was then withdrawn and the wires pulled up to engage endocardial trabeculations.
Heparin (3 mg/kg) was given, then cardiopulmonary bypass was established at a flow of 2.2 L · min-1 · m-2 through an ascending aortic cannula and a right atrial cannula. The cardiopulmonary bypass circuit consisted of a bubble oxygenator (model Bentley-5 oxygenator; American Bentley Corp, Irvine, CA), crystalloid priming solution, and a calibrated roller pump.
The left ventricle was vented through the apex. The systemic perfusate temperature was decreased to 30°C, and systemic flow was decreased to 1.6 L · min-1 · m-2. A thermistor was placed into the interventricular septum. An infusion cannula with a pressure monitoring port (cannula model 23009; DLP, Inc, Grand Rapids, MI) was placed in the aortic root. A 4°C hyperkalemic, hypocalcemic, blood cardioplegia solution (Appendix 1) was infused through the aortic root cannula at a mean pressure of 70 mm Hg after the aorta was cross-clamped, and topical saline slush was placed around the heart. The initial cardioplegia infusion (solution II, see Appendix 1) had a potassium concentration of 22 to 25 mmol/L and lasted 3 minutes. The cardioplegia protocol produced prompt electromechanical quiescence that persisted throughout the ischemic interval in all animals. At 15-minute intervals during cardioplegic arrest, a less hyperkalemic blood cardioplegia solution (potassium, 8 to 10 mmol/L) (solution I, see Appendix 1) was infused for 1 minute at a mean aortic root pressure of 70 mm Hg. The line carrying cardioplegia solution or unmodified blood from the pump to the aortic root was cleared immediately before each infusion. This was accomplished by pumping cardioplegia solution or unmodified blood into an empty syringe attached to a stopcock on the aortic root cannula. This maneuver ensured that each aortic root infusion began at the appropriate temperature and with the correct electrolyte composition.
The pigs were randomized to various durations of hypothermic cardioplegic arrest: 15 minutes (group I), 60 minutes (group II), or 120 minutes (group III) (n = 10 pigs per group). All hearts were reperfused for 20 minutes with unmodified 38°C blood (temperature measured at the heat exchanger outflow) at a mean pressure of 70 mm Hg.
The carotid artery catheter and, later in the study, the aortic root infusion line were used for arterial blood sampling. The coronary sinus catheter was used for cardiac venous sampling. Arterial blood gases and electrolytes (sodium, potassium, and ionized calcium) were measured at the following times: before cardiopulmonary bypass; on cardiopulmonary bypass before initiating cardioplegia; 1 minute after initiating cardioplegia in all groups, 45 minutes after cardioplegia in groups II and III, and 1 hour 45 minutes after cardioplegia in group III; 30 seconds after reperfusion in all groups; and 1, 2, 3, 4, 5, 6, 8, 10, 15, and 20 minutes after reperfusion in all groups.
Electrocardiograms were acquired from surface limb leads; right ventricular, left ventricular, and right atrial bipolar epicardial electrodes; and right and left ventricular bipolar endocardial electrodes. The limb lead electrocardiograms were sent through a band pass filter with lower and upper cutoff frequencies of 0.5 Hz and 30 Hz, respectively. The epicardial and endocardial electrograms were sent through a band pass filter with lower and upper cutoff frequencies of 10 Hz and 1 kHz, respectively. The signals were amplified (amplifier models 13-G4615-64A and 20-4615-58; Gould Electronics, Cleveland, OH) and periodically sampled (Strip Chart Recorder model 1400; Gould Electronics). Data were digitized at a rate of 100 Hz per channel and stored on a hard disk or optical disk for subsequent analysis (CODAS Software; DATAQ, Inc, Akron, OH).
Postcardioplegia reperfusion ventricular fibrillation was defined as chaotic ventricular electrical activity without discernible isoelectric segments that occurred after the initiation of reperfusion. Internal defibrillations (7 J delivered energy) were given as often as every 20 seconds for each episode of reperfusion ventricular fibrillation until a nonfibrillating rhythm was established.
Electrophysiologic variables analyzed included mechanism for the initiation of each episode of postcardioplegia reperfusion ventricular fibrillation; site of earliest fibrillation for each episode of ventricular fibrillation during reperfusion; number of ventricular fibrillation episodes during reperfusion; time to return of electrical activity; time until final episode of reperfusion ventricular fibrillation; time until return of sinus rhythm; QRS duration during reperfusion; and first derivative of electrocardiogram voltage with respect to time (dV/dt) during reperfusion.
Blood gases and electrolytes (sodium, potassium, and ionized calcium) were measured with a model BGE-1400 blood gas analyzer (Instrumentation Laboratories, Lexington, MA). Hemoglobin concentration was measured with an IL model 482 co-oximeter (Instrumentation Laboratories). Other measured variables included (1) myocardial septal temperature (model NTM-100 Digital Thermometer, Webster Laboratories, Inc, Baldwin Park, CA); (2) bypass flow and aortic root flow (calibrated roller pumps); and (3) myocardial tissue weight (obtained postmortem).
The data were analyzed using SAS-PC software (SAS Institute, Inc, Cary, NC), then displayed using Sigma Plot software (Jandel Scientific, Corte Madera, CA). Statistical comparisons were made with analysis designs contained in the General Linear Models and Means procedures of SAS-PC. Duncan’s multiple range test was used to compare the variables for significant differences between groups at any given sample time. A least square means test corrected for multiple comparisons by the method of Bonferoni was used to evaluate specific preplanned comparisons between groups I, II, and III. Within-group comparisons with control values were performed with a paired Student’s t test. The level of significance chosen for this study was a p value of less than 0.05.
|
Results
|
|---|
A total of 167 episodes of reperfusion ventricular fibrillation occurred in the three study groups. The electrophysiologic mechanism was evaluated for each of these episodes by examining the electrophysiologic data before each episode of fibrillation. The most common mechanism identified was an accelerating ventricular tachycardia that led to continuous ventricular electrical activity (ventricular fibrillation) (Table 1; Fig 1). This occurred in 79 of 167 cases. In 7 cases, reperfusion ventricular fibrillation occurred spontaneously (Fig 2). The mechanism could not be determined in 81 instances, most often because of saturation of the amplifiers after a defibrillation obscured the electrocardiograms during the initial moments of recurrence of ventricular fibrillation.
View larger version (40K):
[in this window]
[in a new window]
|
Fig 1. The most common mechanism for postcardioplegia reperfusion ventricular fibrillation is an accelerating ventricular tachycardia that leads to continuous ventricular electrical activity. (II = limb lead II; aVL = electrocardiographic lead aVL; RAE = right atrial electrogram; RV epi, LV epi = right and left ventricular epicardial electrograms; RV endo, LV endo = right and left ventricular endocardial electrograms.)
|
|
View larger version (43K):
[in this window]
[in a new window]
|
Fig 2. In 7 of 167 episodes, postcardioplegia reperfusion ventricular fibrillation occurred without antecedent ventricular tachycardia. (See Fig 1 for abbreviations.)
|
|
The earliest region of reperfusion ventricular fibrillation was determined for each episode of reperfusion ventricular fibrillation by inspecting the right and left ventricular epicardial and endocardial electrocardiograms recorded immediately before each episode. Reperfusion ventricular fibrillation was seen earliest in the left ventricular electrocardiogram in 61 of 167 instances. Reperfusion ventricular fibrillation occurred simultaneously in left and right ventricular electrocardiograms in 41 instances. Presumably, this represents initial fibrillation in or near the ventricular septum. Reperfusion ventricular fibrillation was seen earliest in the right ventricular electrocardiograms in 24 instances. The site of origin of reperfusion ventricular fibrillation could not be determined for the remaining 41 episodes.
The time from the initiation of postcardioplegia reperfusion until the last episode of reperfusion ventricular fibrillation was compared between groups. These results, which are summarized in Table 2, show that the time from the initiation of postcardioplegia reperfusion until the last episode of reperfusion ventricular fibrillation increased as the duration of cardioplegic arrest increased. The number of reperfusion ventricular fibrillation episodes was compared between groups. There were 1.0 ± 0.3 episodes for group I, 5.8 ± 1.2 episodes for group II, and 10.5 ± 1.1 episodes for group III. The p values for all between-group comparisons of this variable were less than 0.05 (see tables for specific p values).
Other variables analyzed included the time to return of electrical activity, the total duration of ventricular fibrillation, and the time to return of normal sinus rhythm. These results are summarized in Table 3.
View this table:
[in this window]
[in a new window]
|
Table 3. Time to Initial Electrical Activity, Duration of Ventricular Fibrillation, and Time Until Sinus Rhythm by Group.
|
|
The time to return of electrical activity was defined as the time between the start of reperfusion and the onset of electrical activity in the ventricular electrocardiograms. The relatively long period of asystole during initial reperfusion in group I is probably attributable to the relatively high concentration of potassium in the initial infusion of cardioplegia solution (solution II, see Appendix 1), and the consequently high potassium concentration present in the heart during the initial seconds of reperfusion (Fig 3). Specifically, the potassium concentration in the coronary sinus effluent in group I is higher after the initial 30 seconds of reperfusion (sample 6) than in groups 2 and 3 (p = 0.046). This reflects the relatively high potassium concentration in the initial infusion of cardioplegia (sample 3), which is higher than the potassium concentration in subsequent cardioplegia infusions (samples 4 and 5). Group I animals only received the relatively high potassium cardioplegia solution (solution II) during their arrest interval of 15 minutes.
View larger version (15K):
[in this window]
[in a new window]
|
Fig 3. The concentration of potassium in coronary sinus blood was measured before, during, and after cardioplegic arrest. (K+ = potassium concentration of coronary sinus blood; sample number: 1 = before bypass; 2 = before cardioplegia infusion on bypass; 3 = 1 minute after initial cardioplegia infusion [all groups]; 4 = 1 minute after cardioplegia at 45 minutes of arrest [groups II and III only]; 5 = 1 minute after cardioplegia at 105 minutes of arrest [group III only]; 6 through 16 = 0.5, 1, 2, 3, 4, 5, 6, 8, 10, 15, and 20 minutes after start of reperfusion.)
|
|
Myocardial (septal) temperature was measured and is shown in Figure 4. The blood temperature during reperfusion was 38°C as measured at the outlet port of the oxygenator’s heat exchanger, but the septal temperature remained below 38°C in all groups throughout the 6 minutes of reperfusion. The temperature increased toward an asymptote, and was above 32°C in all groups within 120 seconds of initiating reperfusion.
View larger version (23K):
[in this window]
[in a new window]
|
Fig 4. The septal myocardial temperature was measured. During reperfusion the myocardial temperature increased rapidly toward normothermia (37°C). There were no statistically significant differences in temperature between groups, although the return of electrophysiologic function differed between groups according to the duration of cardioplegic arrest. (Time = seconds after initiation of postcardioplegia reperfusion.)
|
|
The total duration of ventricular fibrillation was defined as the time from the onset of the first episode of ventricular fibrillation to the time of the last defibrillation. The duration of ventricular fibrillation increased as the duration of cardioplegic arrest increased. The time to return of sinus rhythm was 66.2 ± 8.2 seconds for group I. This was significantly less time than for groups II and III (191.9 ± 26.8 and 249.4 ± 23.3 seconds).
QRS durations immediately before cardioplegic arrest, 1 minute after the start of reperfusion, and 20 minutes after the start of reperfusion were measured. There was a rapid recovery of QRS duration in the limb lead and left ventricular epicardial electrograms after reperfusion(Tables 4 and 5). The dV/dt was measured in the left ventricular epicardial electrocardiogram immediately before cardioplegic arrest, 1 minute after the start of reperfusion, and 20 minutes after the start of reperfusion. There was a similar rapid recovery to precardioplegia values.
|
Comment
|
|---|
This study varied the duration of cardioplegic arrest while reperfusion conditions remained constant. Thus, the experimental design separated the effect of arrest duration from the effect of reperfusion conditions on postcardioplegia arrhythmias and electrophysiologic recovery. The findings confirmed the hypothesis of the study by showing that the time required for postcardioplegia electrophysiologic recovery and the number of postcardioplegia reperfusion arrhythmias increase as the duration of cardioplegic arrest increases from 15 minutes to 2 hours.
Reperfusion arrhythmias are one manifestation of ischemia–reperfusion injury [1–3]. Therefore, it is plausible that knowledge of postcardioplegia reperfusion arrhythmias and electrophysiologic recovery after cardioplegic arrest will prove useful for understanding ischemia–reperfusion injury that occurs during cardiac operations.
To date, only a few studies of postcardioplegia reperfusion arrhythmias have been published. Clinical studies have treated postcardioplegia reperfusion ventricular fibrillation as a dichotomous variable (ie, present or not present) [4–7]. No attempts were made in these studies to measure the duration of ventricular fibrillation, or quantify the number of defibrillations required to establish a nonfibrillating rhythm. Furthermore, none of these studies accounted for patient or procedure-related risk factors that may have had an important effect on the prevalence of postcardioplegia reperfusion arrhythmias (eg, presence of incomplete revascularization or ventricular hypertrophy; differences in duration of cardioplegic arrest or in the conditions of reperfusion). At present, the central question of what postcardioplegia reperfusion arrhythmias represent remains unanswered [8].
The prompt return of sinus rhythm without episodes of ventricular arrhythmias after cardioplegic arrest is generally considered to be an indicator of optimal myocardial protection. In laboratory and clinical studies, the prevalence of reperfusion ventricular fibrillation has been used as a variable for comparing different methods of cardioplegic arrest, although the meaning of any differences in the prevalence of reperfusion ventricular fibrillation remains intuitive. It is known that continuous normothermic retrograde cardioplegia and controlled reperfusion techniques substantially decrease the prevalence of postcardioplegia reperfusion arrhythmias; however, the reason for this is unknown.
In contrast to postcardioplegia reperfusion arrhythmias, reperfusion arrhythmias that occur after a period of ischemia without cardioplegia have received considerable attention from investigators [3, 9, 10]. Manning and Hurst [9] have shown that there is a relationship between the intensity of regional ischemia–reperfusion injury (intensity as defined by the duration of ischemia, mass of ischemic tissue, and presence of collateral circulation) and the duration and prevalence of reperfusion arrhythmias. Specifically, the duration and prevalence of reperfusion arrhythmias increase as the intensity of ischemia increases. In the present study, the prevalence and duration of postcardioplegia reperfusion arrhythmias increased as the duration of arrest increased.
Pogwizd and Corr [11] used a three-dimensional 232-channel digital mapping system to define the electrophysiologic mechanisms underlying regional ischemia–reperfusion arrhythmias in feline hearts. They demonstrated that repetitive and accelerating ventricular depolarizations typically initiated ventricular fibrillation. The present study identified a similar initiating mechanism for 47% of postcardioplegia reperfusion arrhythmias. Unfortunately, there was a large number of episodes in which the initiation of ventricular fibrillation was masked by postdefibrillation amplifier saturation. The electrophysiologic mechanisms for these episodes remain unknown, thus further comparisons with the mechanisms for regional ischemia–reperfusion arrhythmias is not possible. Use of an electrophysiologic data acquisition system with rapid gain switching capability will improve on this problem and allow more thorough analysis of the mechanisms for failure of defibrillation early during postcardioplegia reperfusion.
The conditions after hypothermic cardioplegic arrest are quite different from the conditions that exist after global or regional ischemia at normothermia without cardioplegia. The time necessary to normalize temperature and electrolyte concentrations in the myocardium after hypothermic cardioplegic arrest probably play a role in determining the time required for postcardioplegia electrophysiologic recovery. However, the findings of the present study suggest that other factors, such as the restoration of normal metabolic and membrane function in myocytes, are more important than temperature and electrolyte concentrations in determining postcardioplegia electrophysiologic recovery. Information from a previous study is consistent with this notion [12].
The finding of the present experiment—the duration of cardioplegic arrest effects reperfusion electrophysiology—corroborates the inference from our previous work [13] that postcardioplegia reperfusion arrhythmias are similar to other ischemia–reperfusion arrhythmias. This is an important finding as ischemia–reperfusion arrhythmias in normothermic hearts have been intensively examined, and it may be possible to apply some of this information to the understanding and further investigation of postcardioplegia reperfusion arrhythmias and ischemia–reperfusion injury.
|
Acknowledgments
|
|---|
This work was performed during the tenure of Dr Holman as an Established Investigator of the American Heart Association, and with support from PHS grants HL43213 (William L. Holman) and HL09192 (C. Patrick Murrah).
|
Appendix 1. Composition of blood cardioplegia solutions
|
|---|
All blood cardioplegia solutions were made by mixing oxygenator blood with a premixed crystalloid solution in a 4:1 ratio. The composition of the premixed crystalloid solution was as follows: KCl (2 mmol/L/mL), 5 mL (solution I) or 20 mL (solution II); Tham (0.3 mol/L), 100 mL; CPD, 25 mL; D51/4 normal saline, 275 mL. Solution I: potassium (mmol/L), 8 to 10; pH, 7.7 to 7.8; calcium (mmol/L), 0.5 to 0.7; osmolarity (mOsm/kg), 340 to 360. Solution II: potassium (mmol/L), 22 to 25; pH, 7.7 to 7.8; calcium (mmol/L), 0.5 to 0.7; osmolarity (mOsm/kg), 340 to 360.
Blood cardioplegia solution II was used for the initial infusion of cardioplegia, whereas solution I was used for all subsequent infusions. Reperfusion was carried out with blood withdrawn directly from the oxygenator without additives.
|
References
|
|---|
- Hearse D.J. Myocardial injury during ischemia and reperfusion. In: Yellon D.M., Jennings R.B., eds. Myocardial protection, 1 ed. New York: Raven, 1992:13-33.
- Kloner R.A. Does reperfusion injury exist in humans?. J Am Coll Cardiol 1993;21:537-545.[Abstract]
- Bolli R., Patel B. Factors that determine the occurrence of reperfusion arrhythmias. Am Heart J 1988;115:20-29.[Medline]
- Gorlach G., Podzuweit T., Borsutzky B., Lohmann E., Dapper F. Factors determining ventricular fibrillation after induced cardiac arrest. Thorac Cardiovasc Surg 1991;39:140-142.[Medline]
- Spanos P.K., Brown A.L., McGoon D.C. The significance of intraoperative ventricular fibrillation during aortic valve replacement. J Thorac Cardiovasc Surg 1977;73:605-610.[Abstract]
- Kinoshita K., Mitani A., Tsuruhara Y., Kanegae Y., Tokunaga K. Analysis of determinants of ventricular fibrillation induced by reperfusion: dissociation between electrical instability and myocardial damage. Ann Thorac Surg 1992;53:999-1005.[Abstract]
- Lake C.L., Sellers T.D., Nolan S.P., Crosby I.K., Wellons H.A., Crampton R.S. Determinants of reperfusion cardiac electrical activity after cold cardioplegic arrest during coronary bypass surgery. Am J Cardiol 1984;54:519-525.[Medline]
- Holman W.L. Electrical instability during reperfusion [Letter]. Ann Thorac Surg 1992;54:1245-1246.[Medline]
- Manning A.S., Hearse D.J. Reperfusion-induced arrhythmias: mechanisms and prevention. J Mol Cell Cardiol 1984;16:497-518.[Medline]
- Curtis M.J., Pugsley M.K., Walker M.J.A. Endogenous chemical mediators of ventricular arrhythmias in ischaemic heart disease. Cardiovasc Res 1993;27:703-719.[Free Full Text]
- Pogwizd S.M., Corr P.B. Electrophysiologic mechanisms underlying arrhythmias due to reperfusion of ischemic myocardium. Circulation 1987;76:404-426.[Abstract/Free Full Text]
- Holman W.L., Spruell R.D., Pacifico A.D. Duration of asystolic reperfusion and reperfusate electrolyte composition influence postcardioplegia ventricular fibrillation. J Thorac Cardiovasc Surg 1993;106:511-519.[Abstract]
- Holman W.L., Spruell R.D., Vicente W.A., Pacifico A.D. Electrophysiologic mechanisms for postcardioplegia reperfusion ventricular fibrillation. Circulation 1994;90(Suppl 2):293-298.
This article has been cited by other articles:
|
|
|
|
 
D. Pereda, M. Castella, J.-L. Pomar, R. Cartana, M. Josa, C. Barriuso, J. Roman, and J. Mulet
Elective cardiac surgery using Celsior or St. Thomas No. 2 solution: a prospective, single-center, randomized pilot study
Eur. J. Cardiothorac. Surg.,
September 1, 2007;
32(3):
501 - 506.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
