Cold crystalloid versus warm blood cardioplegia in patients undergoing aortic valve replacement

Introduction

The use of cardioplegia solution is aimed to avoid myocardial muscle damage leading poor contraction and abnormal increased release of cardiac biomarkers enzymes during cardiac arrest (1).

Cold crystalloid cardioplegia (CCC) associated with mild-to-moderate hypothermia has the advantage to decrease the oxygen consumption, offers some degree of myocardial protection during period of low flow or low perfusion pressure, and it is inexpensive and simple to use.

Alternatively, warm blood cardioplegia (WBC) is been proposed as a safe technique for myocardial protection based on the rationale that blood, as opposed to crystalloid solution, could potentially improve postoperative (PO) cardiac outcomes, because it more closely approximates the normal physiology, i.e., carrying oxygen to the myocardium or ensuring a less hemodilution.

In order to protect the hypertrophic myocardium, i.e., during cardiac surgery for the replacement of a diseased aortic valve, several studies published in 80s and 90s have not showed substantial differences between type (crystalloid versus blood), temperature of solution (cold, tepid, or warm), or via administration (antegrade versus retrograde) of the cardioplegia (2-6).

The aim of this study was to evaluate retrospectively whether in patients undergoing aortic valve replacement, isolated or in association with coronary artery bypass grafting, myocardial protection and in-hospital outcome were differently affected by two types of cardioplegic solution currently used in cardiac surgery: cold crystalloid St. Thomas cardioplegia and WBC.

Methods

From January 2015 to October 2016 at the Cardiac Surgery Unit of the Tor Vergata University Hospital of Rome 191 consecutive patients (mean age of 72±9 years) underwent aortic valve replacement, isolated or in association with coronary artery bypass grafting. Cardiac arrest during normothermic cardiopulmonary bypass was obtained using CCC group (n=32) or WBC group (n=159) according to the choice of the surgeon, both administered intermittently and via antegrade. The two groups of patients represented the object of the present study.

All patients performed preoperatively trans-thoracic echocardiography and cardiac catheterization with selective coronary angiography, and postoperatively a trans-thoracic echocardiography on the 3rd to 4th PO day.

Patients affected by other significant concomitant heart valve diseases, operated in emergency, or needing a redo operation were excluded from the study. The study was approved by the Institutional Review Board of the Tor Vergata University Hospital, which waived the need for patient consent. The study was designed to be as retrospective one. All patients gave informed surgical consent.

Data collection

In all patients were evaluated serum levels myocardial enzymes, i.e., total creatine-kinase (CK), creatine-kinase-MB (CK-MB), aspartate aminotransferase (AST), cardiac troponin I (cTnI), and their peak levels, at the end of operation [time 0, i.e., at the admission in intensive care unit (ICU)], 24 and 48 PO hours after surgery. The value of CK-MB/CK ratio greater than (>) 10% was measured at the same times. Perioperative myocardial infarction was defined as an increase of cTnI above 10 ng/mL in association with an increase of CK-MB more than 10% of the total CK and the onset of electrocardiogram (ECG) anomalies.

Complete revascularization was defined when each of three major vascular territories subtended by a significant coronary artery stenosis was grafted. PO low output cardiac syndrome was defined by a cardiac index value less than 2.0 L/min/m², requiring the inotropic support (i.e., epinephrine, enoximone or levosimendan with or without norepinephrine intravenous infusion), for a period greater than 24 hours, and or the use of intra-aortic balloon pump. Preoperative and at weaning form cardiopulmonary bypass hematocrit level and haemoglobin value were evaluated.

Major non-cardiac complications were also analysed. Pulmonary complication was defined as an episode of primary respiratory failure requiring mechanical ventilation for more than 48 hours, re-intubation, or intermittent application of non-invasive positive-pressure ventilation. Permanent neurological complication due to focal or general cerebral lesion was defined as a stroke; transient ischemic attack was defined when neurological symptoms lasted less than 24 hours before disappearing. Acute kidney injury was defined as a two-fold increase of preoperative creatinine serum level or oliguria requiring need of continuous veno-venous hemofiltration.

Operative mortality included death for any causes in-hospital after operation, at anytime or within 30 days after discharge.

Surgical procedure

Surgery was carried out in all patients by means of a complete longitudinal sternotomy, cardiopulmonary bypass with right atrial or bi-caval cannulation and arterial cannulation in the ascending aorta, and aortic cross-clamping. The composition of the crystalloid cardioplegia was referring to the St. Thomas Hospital solution II, with the adding of HCO₃ 8.4 mmol.

Type of cardioplegia was given in accordance with the surgeons’ choice: one surgeon (GR) used in 45 patients WBC from January 2015 to April 2016, and subsequently, in the other 32 patients cold crystalloid St. Thomas cardioplegia. All the other surgeons have adopted WBC throughout the entire period of the study.

In the CCC group intermittent antegrade cold (4 °C) crystalloid cardioplegia (10 mL/Kg the first dose, followed by doses of 5 mL/Kg) was administered every 25–30 minutes. In the WBC group intermittent antegrade warm (35–36 °C) blood cardioplegia (600 mL the first dose, followed by doses of 400 mL, each in 2 minutes) was administered every 20–25 minutes.

Cardiopulmonary bypass temperature in all patients was maintained at 35–36 °C. In presence of isolated aortic valve stenosis cardioplegia was administered into the aortic root, in presence of mixed pathology or pure insufficiency of the aortic valve it was done through the selective infusion into the coronary ostia.

When required, coronary artery bypass was performed using in most cases the left internal mammary artery graft to the left anterior descending artery, in association with saphenous vein grafts, single or in Y-graft composition, to the right coronary artery and/or to the left circumflex artery branches. The aortic valve was replaced according with the standard surgical technique, with the implant of a biological pericardial or a mechanical aortic valve prosthesis, based on patient’s age and choice.

Statistical analysis

Statistical Analysis was performed with Stat View 4.5 (SAS Institute Inc., Abacus Concepts, Berkeley, CA, USA). The continuous variables were calculated as mean values plus minus standard deviation, the categorical variables as number and percentage. The ANOVA test was used to calculate the repeated measures of myocardial enzymes levels at the time 0, and at 24 and 48 hours after surgery. The differences between the two groups of patients were detected by means of the Student’s t-test for continuous data, and the χ² or Fisher’s exact test for categorical data. All P values less than 0.05 were considered statistically significant.

Results

Preoperative characteristics of the two groups of patients are reported in Table 1. Both groups were similar for preoperative characteristics, except for a greater incidence in CCC group of hypertension (P=0.034) and concomitant aortic valve insufficiency (P=0.001). Intraoperative data are summarized in Table 2. In comparison with WBC group, cardiopulmonary bypass time was longer in WBC group (P=0.017). As expected, the mean number of cardioplegia’s doses per patient was higher in WBC group (2.5±1.0 versus 1.8±0.7; P<0.0001), despite the mean number of distal coronary artery anastomoses per patient was similar in the two groups (0.5±0.8 versus 0.4±0.7). Complete revascularization was achieved in all patients of both groups.

Table 1 Preoperative characteristics
Full table

Table 2 Intraoperative variables
Full table

Before cardiopulmonary bypass, mean hematocrit level and hemoglobin value were not statistically different in both groups, whereas after the weaning from cardiopulmonary bypass their values remained favorably higher in WBC group, with a slight levels of statistical significance (P=0.082, and P=0.047, respectively) (Table 2).

Serum levels of total CK, CK-MB, AST and cTnI were lower in CCC group in each time of evaluation. In particular, the greater statistically significant differences were found at time 0, either for the serum level of the enzymes (mean values and their peaks) (P<0.05, for all comparisons) than for the value of the CK-MB/CK ratio >10% (5.9% versus 7.8% of the patients; P<0.0001) (Figures 1-5) (Table 3). In the subgroup of patients undergoing isolated AVR, the CK-MB/CK ratio >10% at time 0 was 6.0% in CCC group versus 8.0% in WBC group (P<0.01). Similar statistically significant differences of the values of CK-MB/CK ratio >10% at time 0 was observed for aortic cross-clamp times greater than 60 minutes (5.0% versus 7.0%, P<0.01) or less (6.0% versus 8.0%, P=0.001).

Figure 1 Postoperatively (at time 0, 24, 48 hours) release of creatine-phospho-kinase (CPK) expressed as U/L, and peak value. WBC, warm blood cardioplegia; CCC, cold crystalloid cardioplegia.

Figure 2 Postoperatively (at time 0, 24, 48 hours) release of creatine-kinase-MB (CK-MB) expressed as ng/mL, and peak value. WBC, warm blood cardioplegia; CCC, cold crystalloid cardioplegia; CPK, creatine-phospho-kinase.

Figure 3 Postoperatively (at time 0, 24, 48 hours) release of cardiac troponin I (cTnI) expressed as ng/mL, and peak value. WBC, warm blood cardioplegia; CCC, cold crystalloid cardioplegia.

Figure 4 Postoperatively (at time 0, 24, 48 hours) release of aspartate aminotransferase (AST) expressed as ng/mL, and peak value. WBC, warm blood cardioplegia; CCC, cold crystalloid cardioplegia.

Figure 5 Mean value (%) of the ratio greater than 10% between creatine kinase MB (CK-MB) and total creatine-phospho-kinase (CPK) postoperatively measured (at time 0, 24 and 48 hours). WBC, warm blood cardioplegia; CCC, cold crystalloid cardioplegia.

Table 3 Serum levels of cardiac enzymes
Full table

In the CCC group was found a significantly higher incidence of the use of norepinephrine intravenous infusion for a time greater than 24 hours after operation (9.4% versus 0.6%; P=0.002) (Table 4).

Table 4 Postoperative results
Full table

The incidence of PO low output cardiac syndrome, perioperative myocardial infarction, overall PO incidence of major complications (i.e., neurological damage, acute kidney injury, pulmonary complications), re-exploration for bleeding, atrio-ventricular blocks requiring need for pacemaker implantation, mean number of units of blood transfused per patient, in-ICU and PO length of stay were similar in both groups (Table 4).

In the CCC group was observed a major incidence of primary respiratory pulmonary failure (9.4% versus 1.3%) with a slight statistical relevance (P<0.05).

Operative mortality was 1.3% (n=2) in WBC group, 3.1% in CCC group (P=0.438). Two deaths occurred for cardiac causes, one was due to neurological damage, respectively.

As compared with CCC group, the incidence of PO paroxysmal atrial fibrillation was higher in WBC group (53.5% versus 34.4%), although this difference did reach a slightly level of statistical significance (P=0.053) (Table 4).

Discussion

Over the past decades, the question of which solution, temperature or mode of administration of different types of cardioplegia provides a better myocardial protection during cardiac surgery has been widely discussed. Experimental studies have suggested a more favorable outcome with the use of blood cardioplegia in comparison with CCC. Several clinical studies with or without randomization have been performed to assess which cardioplegic solution guarantees a better myocardial protection, but some studies have reported a favorable outcome of the blood (cold or warm) cardioplegia (2-5,7,8), others have not been able to demonstrate any difference (9-11). In another study a worse outcome for patients receiving cold blood cardioplegia has been reported (12).

Also with respect to the surgical treatment of congenital heart diseases, no clear demonstrably clinical advantage has been shown with the use of antegrade cold blood cardioplegia over CCC (13). However, data from a more recently published clinical meta-analyses show a more favorable outcome with the use of cold blood cardioplegia in comparison with cold crystalloid solutions, although also these studies have not been conclusive regarding which type of cardioplegia provides the best protection method. Several possible reasons for these findings can be depend on the heterogeneous groups of patients included in these studies, i.e., affected by coronary artery disease, different valve pathologies, other diseases, to the different surgical procedures, and to the duration of the aortic cross clamp time. Guru, Jacob and their coworkers (14,15) in two meta-analyses showed that cold blood cardioplegia provides a favorable outcome in terms of lower incidence of low cardiac output syndrome, CK-MB release in comparison with CCC, but the incidence of myocardial infarction and death were similar. In another two meta-analyses Fan, Abah, and their colleagues (16,17) showed no difference between warm and cold cardioplegia on the short-term mortality. Jacquet and Coworkers (18), on 200 patients undergoing coronary artery bypass grafting showed as a better method of myocardial protection the antegrade warm cardioplegia in comparison with combined antegrade and retrograde CCC on the PO cardiac enzymes release, although ischemic arrest duration was significantly shorter in the warm group. In a large prospective randomized study Ovrum and coworkers (19) on 1,440 patients undergoing coronary artery bypass did not detect any difference in clinical outcome. However, the mean cross clamp time was only 34 minutes, and this could be too short to show a potential difference. For these reasons, we select a relatively homogeneous group of patients, less frequently investigated in the literature, who underwent cardiac surgery with a relatively long cross clamp time and affected by myocardial hypertrophy, that are two relevant variables to test the efficacy of myocardial protection achieved by two types of cardioplegia routinely used. The degree of the myocardium hypertrophy measured on the left ventricular septum and the posterior wall, the mean duration of the aortic cross-clamping time, and the incidence of associated coronary artery bypass grafting, i.e., about one third of patients in each group, were similar in both groups (Table 2). We observed that the mean duration of the aortic cross-clamping time was slightly longer in the WBC group, but, as expected, it was likely due to the greater required mean number of cardioplegia’s doses administered. A similar difference was recorded for the duration of the aortic cross-clamping time in the subgroups of patients undergoing isolated aortic valve replacement only (Table 2).

The main aspect highlighted in our study is that a better myocardial protection in the presence of myocardial hypertrophy can be achieved by using a type of cold rather than hot cardioplegia, either in all patients of the two groups of patients, than in patients undergoing isolated aortic valve replacement. Improved protection of the hypertrophic myocardium by cold rather than warm cardioplegia has also been demonstrated in the prospective randomized study of Boening et al. (5) in terms of oxygen consumption during aortic valve replacement surgery. Except that, unlike our study, the comparison was performed using cold and WBC.

During the first 24 PO hours after operation we observed that CK-MB, cTnI release, their serum peak levels, and the ratio of the CK-MB were lower in the CCC group, whereas the release of the AST remained lower also up to 48 PO hours (Figures 1-4) (Table 3). Also the difference in the CK-MB/CK ratio >10% between the two groups was about 2% at the time 0 of investigation, in favor of CCC group (Figure 5). PO left ventricular ejection fraction was also better in CCC group in comparison with WBC group, but this difference was likely due to its preoperative mean value, that was higher in CCC group. In literature there are little data comparing warm blood and CCC in the aortic valve replacement surgery. Braathen and Tønnessen in 80 patients undergoing isolated aortic valve replacement showed a better myocardial protection with the use of cold blood cardioplegia in comparison with CCC (20). In contrast to the observed value of CK-MB in our study that was less than 20 ng/mL at 24 PO hours in the CCC group, they found a CK-MB value greater than 45 micrograms (mcg)/L at 20 PO hours with the use of the CCC. Whereas, CK-MB value in the cold blood group was similar to that observed in our study in the cold crystalloid group. On the contrary, Bouchart and colleagues on 60 patients undergone isolated aortic valve replacement reported a statistically significant high level of CK-MB at 6 hours (70 mcg/L) with the use of intermittent cold blood cardioplegia in comparison with continuous WBC (33 mcg/L) and CCC (45 mcg/L) (12). Kaul et al. (21) on 123 patients undergoing combined valve (aortic or mitral) and coronary artery bypass surgery found a significant reduced release in AST enzyme in favor of cold blood cardioplegia in comparison with cold crystalloid or with the use of ischemic cardiac arrest. Ascione and coworkers (22) found a significant reduced release of cTnI at 1, 24, 48 hours postoperatively in favor of cold blood cardioplegia in comparison with WBC. Interestingly, the mean value of cTnI at 24 and 48 hours reported using cold and WBC was similar to that observed in our study in the CCC group and in the WBC group, respectively. However, although the reported study and other studies have underlined some different degree of myocardial protection in favor of one or of the another type of cardioplegia, no differences have been observed in terms of perioperative myocardial infarction, low output cardiac syndrome, PO non-cardiac major complications, and death (12,20-23). These findings are in accordance with those observed in our study: in fact, although the release of cardiac enzymes was in favor of CCC, the perioperative outcome was similar in both groups of patients. Ovrum and Colleagues (24), in a prospective randomized study on 345 aortic valve patients, found no difference between retrograde cold blood cardioplegia and retrograde CCC in terms of bleeding, need for blood transfusions, perioperative myocardial infarction, stroke, renal function, infections or mortality. On the contrary, Calafiore and colleagues on 271 patients undergoing aortic valve replacement showed a lower cardiac-related mortality in favor of WBC in comparison with cold blood cardioplegia (25).

As expected, and also as reported in previous studies (2,26), in the CCC group was observed lower hematocrit level and hemoglobin value at the end of cardiopulmonary bypass (Table 2). The mean number of units of blood transfused per patient was similar (Table 4). Consequently, the lower level of hematocrit could justify the greater use of norepinephrine infusion in the PO period, as also reported in other studies (8,14,16,17), although this need did not cause any difference in the in-ICU and in-hospital stay in comparison with the WBC group of patients.

Finally, in a propensity-matched analysis of more than 7,000 patented patients undergoing aortic valve replacement recently published by Hoyer and colleagues, no significant differences were demonstrated for operative mortality, PO complications and long-term survival with the use of cold blood cardioplegia and CCC Custodiol (27).

The greater incidence of primary respiratory pulmonary complications observed in the CCC group appears more difficult to be explain in our experience, nor even in literature we did found consistent data that demonstrate a correlation with the use of one type of cardioplegia compared to another.

The study has several limitations. First, it is a retrospective and observational analysis. Second, the analyzed sample that constituted the CCC group was quite small. However, this study took into account a short surgical period in order to avoid bias. Moreover, the population in question regarded a quite homogeneous group of patients undergoing aortic valve replacement in association or not with coronary artery surgery. Third, we have compared only two types of cardioplegia, intermittently and antegrade administered.

Conclusions

From the data obtained in this study, we cannot try any definitive conclusion about what is the best method of myocardial protection during aortic valve surgery. The clinical bottom line of our findings is that in patients undergoing aortic valve replacement CCC or WBC during normothermic cardiopulmonary bypass have resulted in similar rate of perioperative myocardial infarction, PO complications, and death. Therefore, the statistically significant difference observed in the myocardial enzymes release in favor of crystalloid cardioplegia did not translate into different clinical outcome.

However, this study suggests that in presence of cardiac surgical conditions associated with a significant myocardium hypertrophy, i.e., in presence of aortic valve disease, a better protection can be achieved with the use of a cold rather than a warm cardioplegia. Therefore, CCC can be still safely used.

Acknowledgements

None.

Footnote

Conflicts of Interest: The authors have no conflicts of interest to declare.

Ethical Statement: The study was approved by the Institutional Review Board of the Tor Vergata University Hospital, which waived the need for patient consent. The study was designed to be as retrospective one. All patients gave informed surgical consent.

References

1. Buckberg GD, Athanasuleas CL. Cardioplegia: solution or strategies? Eur J Cardiothorac Surg 2016;50:787-91. [Crossref] [PubMed]
2. Barner HB. Blood cardioplegia: a review and comparison with crystalloid cardioplegia. Ann Thorac Surg 1991;52:1354-67. [Crossref] [PubMed]
3. Catinella FP, Cunningham JN Jr, Spencer FC. Myocardial protection during prolonged aortic cross-clamping. Comparison of blood and crystalloid cardioplegia. J Thorac Cardiovasc Surg 1984;88:411-23. [PubMed]
4. Feindel CM, Tait GA, Wilson GJ, et al. Multidose blood versus crystalloid cardioplegia. Comparison by quantitative assessment of irreversible myocardial injury. J Thorac Cardiovasc Surg 1984;87:585-95. [PubMed]
5. Boening A, Sanuri M, Buchwald D, et al. Aortic valve replacement: better myocardial protection by cold or warm retrograde cardioplegia? J Heart Valve Dis 1996;5:273-80. [PubMed]
6. Hendry PJ, Masters RG, Haspect A. Is there a place for cold crystalloid cardioplegia in the 1990s? Ann Thorac Surg 1994;58:1690-4. [Crossref] [PubMed]
7. Fremes SE, Christakis GT, Weisel RD, et al. A clinical trial of blood and crystalloid cardioplegia. J Thorac Cardiovasc Surg 1984;88:726-41. [PubMed]
8. Ibrahim MF, Venn GE, Young CP, et al. A clinical comparative study between crystalloid and blood-based St Thomas’ hospital cardioplegic solution. Eur J Cardiothorac Surg 1999;15:75-83. [Crossref] [PubMed]
9. Caputo M, Dihmis W, Birdi I, et al. Cardiac troponin T and troponin I release during coronary artery surgery using cold crystalloid and cold blood cardioplegia. Eur J Cardiothorac Surg 1997;12:254-60. [Crossref] [PubMed]
10. Hendrikx M, Jiang H, Gutermann H, et al. Release of cardiac troponin I in antegrade crystalloid versus cold blood cardioplegia. J Thorac Cardiovasc Surg 1999;118:452-9. [Crossref] [PubMed]
11. Martin TD, Craver JM, Gott JP, et al. Prospective, randomized trial of retrograde warm blood cardioplegia: myocardial benefit and neurological threat. Ann Thorac Surg 1994;57:298-302; discussion 302-4. [Crossref] [PubMed]
12. Bouchart F, Bessou JP, Tabley A, et al. How to protect hypertrophied myocardium? A prospective clinical trial of three preservation techniques. Int J Artif Organs 1997;20:440-6. [Crossref] [PubMed]
13. Young JN, Choy IO, Silva NK, et al. Antegrade cold blood cardioplegia is not demonstrably advantageous over cold crystalloid cardioplegia in surgery for congenital heart disease. J Thorac Cardiovasc Surg 1997;114:1002-8; discussion 1008-9. [Crossref] [PubMed]
14. Guru V, Omura J, Alghamdi AA, et al. Is blood superior to crystalloid cardioplegia? A meta-analysis of randomized clinical trials. Circulation 2006;114:I331-8. [Crossref] [PubMed]
15. Jacob S, Kallikourdis A, Sellke F, et al. Is blood cardioplegia superior to crystalloid cardioplegia? Interact Cardiovasc Thorac Surg 2008;7:491-8. [Crossref] [PubMed]
16. Fan Y, Zhang AM, Xiao YB, et al. Warm versus cold cardioplegia for heart surgery: a meta-analysis. Eur J Cardiothorac Surg 2010;37:912-9. [Crossref] [PubMed]
17. Abah U, Garfjeld Roberts P, Ishaq M, et al. Is cold or warm cardioplegia superior for myocardial protection? Interact Cardiovasc Thorac Surg 2012;14:848-55. [Crossref] [PubMed]
18. Jacquet LM, Noirhomme PH, Van Dyck MJ, et al. Randomized trial of intermittent antegrade warm blood versus cold crystalloid cardioplegia. Ann Thorac Surg 1999;67:471-7. [Crossref] [PubMed]
19. Øvrum E, Tangen G, Tølløfsrud S, et al. Cold blood cardioplegia versus crystalloid cardioplegia: a prospective randomized study of 1440 patients undergoing coronary artery bypass grafting. J Thorac Cardiovasc Surg 2004;128:860-5. [Crossref] [PubMed]
20. Braathen B, Tønnessen T. Cold blood cardioplegia reduces the increase in cardiac enzyme levels compared with cold crystalloid cardioplegia in patients undergoing aortic valve replacement for isolated aortic stenosis. J Thorac Cardiovasc Surg 2010;139:874-80. [Crossref] [PubMed]
21. Kaul TK, al Khadimi R, Sharif H, et al. Results of combined valve replacement and myocardial revascularization. Relation to method of myocardial protection. J Cardiovasc Surg (Torino) 1989;30:322-7. [PubMed]
22. Ascione R, Caputo M, Gomes WJ, et al. Myocardial injury in hypertrophic hearts of patients undergoing aortic valve surgery using cold or warm cardioplegia. Eur J Cardiothorac Surg 2002;21:440-6. [Crossref] [PubMed]
23. Lee JH, Jeong DS, Sung K, et al. Clinical results of different myocardial protection techniques in aortic stenosis. Korean J Thorac Cardiovasc Surg 2015;48:164-73. [Crossref] [PubMed]
24. Øvrum E, Tangen G, Tølløfsrud S, et al. Cold blood versus cold crystalloid cardioplegia: a prospective randomized study of 345 aortic valve patients. Eur J Cardiothorac Surg 2010;38:745-9. [Crossref] [PubMed]
25. Calafiore AM, Teodori G, Bosco G, et al. Intermittent antegrade warm blood cardioplegia in aortic valve replacement. J Card Surg 1996;11:348-54. [Crossref] [PubMed]
26. Günday M, Bingöl H. Is crystalloid a strong predictor of intra-operative hemodilution? J Cardiothorac Surg 2014;9:23. [Crossref] [PubMed]
27. Hoyer A, Lehmann S, Mende M, et al. Custodiol versus cold Calafiore for elective cardiac arrest in isolated aortic valve replacement: a propensity-matched analysis of 7263 patients. Eur J Cardiothorac Surg 2017;52:303-9. [Crossref] [PubMed]