Transesophageal Echocardiography Guidance of Antegrade Cardioplegia Delivery for Cardiac Surgery David J. Canty, PhD,*†§ Prashant Joshi, MD,‡ Colin F. Royse, MD,*§ James McMillan, CCP,‖ Sara Tayeh, CCP,‖ and Julian A. Smith, MS*‡ Objectives: The initial volume of antegrade cardioplegia used to induce asystole during aortic cross-clamp still is based on empiric methods and may be excessive, potentially leading to hyperkalemia, myocardial edema, and acute left ventricular distention from aortic regurgitation. The objectives were to determine whether the volume of cardioplegia required to induce asystole is proportional to left ventricular mass, and whether the degree of left ventricular distention is proportional to the severity of aortic regurgitation. Design: Prospective observational study. Setting: Two tertiary university hospitals. Interventions: Transesophageal echocardiography was used to estimate left ventricular mass (prolate ellipse revolution formula), quantify aortic regurgitation, and monitor for distention during initial antegrade cardioplegia delivery. The volume of cardioplegia required for asystole was recorded. Participants: Fifty-eight patients aged over 18 years scheduled for cardiac surgery requiring aortic cross-clamping.

Measurements and Main Results: There was a weak correlation of left ventricular mass and antegrade cardioplegia volume required for asystole (r ¼ 0.35, p ¼ 0.047). The degree of left ventricular distention correlated moderately with the severity of aortic regurgitation (r ¼ 0.55, p ¼ 0.007) and was excessive and stopped early (aborted) in 24% of all patients, including 18% of 39 patients without aortic regurgitation. An aortic regurgitation vena contracta of 0.3 cm predicted aborted cardioplegia with modest accuracy (AUC 0.81, 0.66-0.99, p ¼ 0.02, sensitivity 71%, specifity 81%). Conclusions: Estimated left ventricular mass is not a useful predictor of the initial volume of antegrade cardioplegia required to induce asystole. However transesophageal echocardiography can predict and monitor for left ventricular distention, which is common. & 2015 Elsevier Inc. All rights reserved.

M

guide the volume of cardioplegia, potentially avoiding morbidity from overdosage. The aims of the study were to determine whether the volume of initial antegrade cardioplegia required to induce asystole correlates with LV mass estimated with TEE and to determine whether or not severity of AR correlates with the degree of LV distention during antegrade cardioplegia delivery.

YOCARDIAL DELIVERY of potassium-containing cardioplegia solution causes asystole, enabling surgery on a motionless heart and protection from ischemia during interruption of coronary arterial perfusion. While other techniques of myocardial management have been described,1 the use of antegrade cardioplegia is widely used, effective, and performed by infusion through an ascending aortic cannula placed proximal to the aortic cross-clamping, thus allowing cardioplegia delivery into the coronary circulation via the coronary ostia. Asystole usually occurs after delivery of less than 300 to 400 mL. However, it is common practice to deliver up to 3 times this volume to ensure delivery to all regions of the myocardium, especially in patients with coronary stenosis or occlusion. Clinically important consequences of excessive cardioplegia delivery include arrhythmia, hyperkalemia, and hypermagnesemia with severe hypotension and metabolic acidosis, hemodilution, and myocardial edema.1 In the presence of aortic regurgitation (AR), acute left ventricular (LV) distention may occur, rendering cardioplegia delivery less effective or ineffective, and risking mechanical and ischemic injury to the LV.2,3 There are sparse published data to guide the initial dosage, which tends to be empiric (not based on patient characteristics), and potentially excessive, ranging from 750 to 1,000 mL. Matsuda et al4 demonstrated an inverse relationship between postoperative troponin leak and cardioplegia dose and LV mass after aortic valve replacement in patients with chronic AR and LV hypertrophy. This suggests that the amount of cardioplegia required for myocardial protection may correlate with muscle mass. Left ventricular muscle mass can be estimated with reasonable accuracy with transesophageal echocardiography (TEE), which is used as routine intraoperative monitoring in many centers. If LV muscle mass correlates with the cardioplegia dose required for myocardial protection, then TEE may be used to more accurately

KEY WORDS: cardiac surgery, cardiac arrest, transesophageal echocardiography, aortic valve insufficiency, cardioplegia

METHODS

This prospective cohort study received ethics approval from the Monash Health (12399Q) and Melbourne Health (QA2013047) Human Research Ethics Committees as a quality assurance project, and patient consent was not required. Patients were screened when the primary investigator was available (convenience sampling) from consecutive patients older than 18 years presenting for cardiac surgery requiring cardiopulmonary bypass (CPB) and myocardial protection during aortic cross-clamping with antegrade cardioplegia at Monash Medical Centre and The Royal Melbourne Hospital.

From the *Department of Surgery, Level 6 Center for Medical Research, University of Melbourne, Royal Parade, Parkville, Australia; †Department of Anaesthesia and Perioperative Medicine; ‡Department of Surgery and Department of Cardiothoracic Surgery, Monash Medical Center, Clayton, Australia; §Department of Anaesthesia and Pain Management, Royal Melbourne Hospital, Parkville, Australia; and ‖Perfusion Services, Unit 2, Tulip Street, Cheltenham, Australia. Address reprint requests to David J. Canty, PhD, Ultrasound Education Group, Department of Surgery, Level 6 Centre for Medical Research, University of Melbourne, Royal Parade, Parkville, Victoria, Australia 3050. E-mail: [email protected] © 2015 Elsevier Inc. All rights reserved. 1053-0770/2601-0001$36.00/0 http://dx.doi.org/10.1053/j.jvca.2015.03.009

Journal of Cardiothoracic and Vascular Anesthesia, Vol ], No ] (Month), 2015: pp ]]]–]]]

1

2

Patients were excluded if cardioplegia was not delivered via an ascending aortic cannula or if TEE was contraindicated or not possible. Intraoperative TEE was used to estimate LV mass and severity of AR and to monitor for LV distention during delivery of the initial antegrade cardioplegia. The volume of cardioplegia required to cause asystole was recorded and explored for correlation with estimated LV mass. The degree of LV distention, including failure of antegrade cardioplegia to induce asystole, was correlated with severity of AR. Anesthesia was conducted according to anesthesiologist preference, which included either volatile inhalation anesthesia (sevoflurane or desflurane) or total intravenous anesthesia (propofol or fentanyl and midazolam) or both. Initial analgesia included intravenous fentanyl (5-10 μ/kg). The neuromuscular blockade was maintained with either intravenous pancuronium or rocuronium or both. Routine intraoperative monitoring included arterial, central venous and pulmonary artery pressures, intermittent cardiac output by pulmonary artery thermodilution, 5-electrode electrocardiogram (ECG) (II and V), pulse oximetry, capnography, and TEE. TEE data were recorded prospectively on a standardized report form by a single researcher (D.C.) board certified in perioperative TEE using either an iE33, HD-11 or HD-15 (Phillips Medical Systems, Andover, MA) echocardiography machine with a 7.2 MHz transesophageal probe. Aortic regurgitation was graded using the vena contracta width taken from the midesophageal long-axis view as recommended by the American Society of Echocardiography.5 Vena contracta was chosen over jet-to-left ventricular outflow tract ratio due to its simplicity and feasibility and that it appears to be more robust.5 The left ventricular mass was estimated using the method recommended by The American Society of Echocardiography,6 which is based on modelling the LV as a prolate ellipse of revolution, which has been validated with necropsy (r ¼ 0.90, p o 0.001).7  LV mass ¼0:8  1:04 ðLVIDdþ PWTdþ SWTdÞ3  ðLVIDdÞ3 Þþ 0:6 g where LVIDd is the maximal internal LV diameter at enddiastole, PWTd and SWTd are the thicknesses of the posterior LV wall and interventricular septum at end-diastole, respectively. As per recommendations,6 these measurements were recorded using M-mode placed perpendicularly to the LV walls at the chordae level intersecting the interventricular septum below the left ventricular outflow tract rather than at the mitral leaflet tips, as most patients had significant coronary artery disease. As the transgastric 2-chamber view frequently is easier to obtain than the long-axis view, the anterior and inferior wall thicknesses and LVIDd were recorded for comparison to the long-axis view. To reduce measurement bias, the calculation of LV mass was performed off-line and, hence, the researchers and treating staff were blinded to the LV mass estimation. The transgastric mid-view of the LV was used to assess for LV distention during cardioplegia delivery and was displayed (in readiness) prior to application of the aortic cross-clamp. The LV endocardial area (excluding the papillary muscles) was measured after clamping of the aorta, both before (baseline) and at the end (the maximum LV cavity size) of cardioplegia

CANTY ET AL

delivery. The decision of when to cease antegrade cardioplegia due to excessive LV distention and failure to attain asystole was left to the discretion of the surgeon. Digital echocardiographic measurements were measured off-line by 2 independent echocardiographers (using Synapse Cardiovascular software [Fujifilm, Akasaka, Minato, Tokyo, Japan]) who were blinded to the cardioplegia data. Reported values were the average of 3 consecutive beats per observer. Cardiopulmonary bypass was achieved with aortocaval cannulation following median sternotomy at a flow rate of 2.3 to 2.5 L/min/m2 and mean arterial pressure maintained between 50 mmHg to 80 mmHg. After aortic cross-clamping, cardioplegia was delivered via roller pump (Sorin, Milan, Italy) via a 9-French cannula (DLP, Medtronic, Minneapolis, MN) sited into the ascending aorta proximal to the aortic cross-clamp perfusing the myocardium via the coronary arterial circulation (antegrade) at a pressure of less than 200 mmHg (measured distal to the heat exchanger). Perfusion data were recorded prospectively by an assistant clinical perfusionist on a standardized report form. The cardioplegia flow rate, aortic root (driving) pressure, and volume of cardioplegia infused were recorded every 5 seconds until asystole was confirmed. It was requested that the surgeons not touch the heart during delivery of cardioplegia to avoid unwanted ECG artefact, which could interfere with monitoring for asystole. Asystole was defined as no detectable cardiac electric activity from 2-electrode ECG (II and V) for 1 full monitor screen cycle (5 seconds), which was confirmed by an independent observer and by the absence of LV movement with TEE. The antegrade cardioplegia used consisted of cardioplegia electrolyte solution (Cardioplegia A solution, Baxter, Toongabbie, NSW, Australia), buffered with sodium bicarbonate to a pH 7.4-7.8 and mixed with blood (blood:cardioplegia 4:1, Kþ 20 mmol/L) at room temperature. The primary endpoint was the volume of cardioplegia delivered to obtain asystole. Secondary endpoints included the amount of LV distention during delivery of antegrade cardioplegia as estimated by TEE and whether or not the surgeon deemed this as unsafe and recommended stopping further antegrade delivery (aborted delivery). STATISTICAL ANALYSIS

The estimated sample size of 60 patients was based on unpublished data from a pilot study. In the pilot study, a moderate correlation (r ¼ 0.67) between estimated LV mass and volume of cardioplegia required for asystole was found by a single observer in 14 consecutive patients undergoing CABG surgery without detectable AR. Using an estimated incidence of at least trivial AR in cardiac surgical patients of 33%, this would require 21 patients to enroll 14 patients without AR. However, in the pilot cohort, there were only 3 patients out of 14 with criteria for LV hypertrophy (4131 g/m2).8 To recruit 10 patients with LVH, a sample size of 47 was calculated, which was rounded up to 60 to account for exclusions from inadequate echocardiography imaging and problems with cardioplegia delivery. The data from the pilot study patients were not used in this study.

3

TEE GUIDANCE OF CARDIOPLEGIA

All statistical analysis was performed in SPSS version 21 (IBM, Greenwood, SC). All graphical displays were produced in GraphPad Prism (GraphPad Software, La Jolla, CA). Statistical significance was defined as a p value of less than 0.05. Pearson’s correlations were used to explore the relationship among LV mass, AR vena contracta, and cardioplegia volume. Group comparisons of the degree of left ventricular distension and vena contracta were performed using independent samples t-tests. To assess the role of TEE outcome (severity of AR) as a diagnostic test (unacceptable LV distention), contingency tables were created at various AR vena contracta cut-off values between 0.025 and 0.635 cm. The sensitivity, specificity, and area under the receiver-operator characteristic curve were calculated. Interobserver and intraobserver variability were assessed for the key echocardiography measurements by measuring the mean difference and limits of agreement (⫾2 SDs of the difference). The authors considered the agreement between observers to be acceptable if the limits of agreement were less than 30% of the mean value of the variable being measured. RESULTS

Out of 60 patients recruited between December 2012 and February 2014, 2 patients were excluded, leaving 58 patients with data for analysis. One patient was excluded due to inadequate echocardiography imaging (no transgastric views), and another patient was excluded as the surgery was performed without cardioplegia. Patient characteristics and surgical data are shown in Table 1. Aortic regurgitation (of any severity) was detected in 19 patients who were excluded from the correlation analysis of cardioplegia volume with LV mass (Table 2). In the remaining Table 1. Patient Characteristics and Surgical Data Patient Characteristics Number of patients Age, mean (SD) years Female (%) Urgent surgery (%) Left ventricular function, median score (IQR) Previous cardiac surgery (%) Height, mean (SD) cm Weight, mean (SD) kg Body surface area, (SD) m.kg-2 Surgical Data Institution Monash Medical Centre (%) Royal Melbourne Hospital (%) Type of surgery Isolated CABG (%) Aortic valve replacement (%) Mitral valve surgery (%) Tricuspid valve surgery (%) Thoracic aorta (%) Combined valve þ CABG (%) Aortic cross-clamp time, mean (SD) min Total bypass time, mean (SD) min

58 68 (11) 20.7 36.2 1 (1-2) 1.7 168 (9) 80 (17) 1.92 (0.22)

71 29 69 5.2 1.7 1.7 3.4 19 91 (39) 119 (43)

Abbreviations: CABG, coronary artery bypass graft; SD, standard deviation; IQR, interquartile range.

39 patients with no detectable aortic regurgitation prior to aortic cross-clamping, there were 7 patients who developed sufficient LV distention during cardioplegia delivery (despite no aortic regurgitation prior to cross-clamp) for whom retrograde delivery via a coronary sinus catheter was required to achieve asystole. After exclusion of these 7 patients, correlation of LV mass (calculated with the transgastric long-axis view) had only a weak (r ¼ 0.35, p ¼ 0.047) correlation with the volume of initial antegrade cardioplegia required to cause asystole (Table 3) (Fig 1). There was also a weak correlation using the transgastric 2-chamber view (r ¼ 0.33). However, this did not reach statistical significance (p ¼ 0.066). There was no significant correlation of cardioplegia volume required for cardiac arrest with patient height, weight, or body surface area. The degree of LV distention (change in LV cavity area) during initial antegrade cardioplegia delivery correlated moderately with the vena contracta of AR (r ¼ 0.55, 95% CI 0.17 to 0.78, p ¼ 0.007) as shown in Figure 2. The amount of cardioplegia volume required to cause cardiac arrest was increased in patients with AR compared to those without (Table 2). Unacceptable LV distention resulting in abortion of antegrade delivery increased from 31% of patients with mild AR to 50% of patients with moderate or severe AR (Table 2). Even in patients without detectable AR, unacceptable LV distention occurred in 18% of patients. The degree of LV distention required to abort antegrade delivery was not defined but was significantly greater than in patients in whom antegrade delivery was completed (mean change in LV area 11.0 ⫾ 1.5 cm2 and 3.5 ⫾ 0.4 cm2, respectively, mean difference in LV area 7.5 cm2, 95% CI 5.2 to 9.8, p o 0.0001), as shown in Figure 3A. The vena contracta was significantly larger in patients with unacceptable LV distention compared to those with acceptable LV distention and successful antegrade cardioplegia (Fig 3B) (unpaired 2-tailed Student t test p ¼ 0.046). A cut-off value for vena contracta of 0.3 cm predicted failure of antegrade cardioplegia delivery with reasonable accuracy (Fig 4) with an area under the curve of 0.81 (95% CI 0.66 to 0.99, p ¼ 0.02, sensitivity 71%, specifity 81%, likelihood ratio 3.8).

Table 2. Incidence of Prebypass Aortic Regurgitation and Unacceptable Left Ventricular Distention During Initial Antegrade Cardioplegia Delivery in 58 Patients Undergoing Cardiac Surgery Volume of antegrade

Enlargement of end-

Aborted ACP due

Completed ACP as LV

cardioplegia to

diastolic area

to LV

distention

Aortic

achieve arrest

Mean (SD)

distention

acceptable

regurgitation

Mean (SD) mL

cm2

n (%)

(n)

(n)

None Mild AR, VC o 0.3 cm More than mild AR, VC 4 0.3 cm Total

300 (108) 381 (139)

3.7 (3.6) 6.8 (6.6)

7 (18%) 4 (31%)

32 9

39 13

608 (247)

11.5 (4.3)

3(50%)

3

6

350 (162)

5.2 (5.1)

14 (24%)

44

58

Total

Abbreviations: ACP, initial antegrade cardioplegia; AR, aortic regurgitation; LV, left ventricular.

4

CANTY ET AL

Table 3. Correlation of Volume of Antegrade Cardioplegia Required to Induce Asystole in 32 Patients Without Aortic Regurgitation (Excluding 19 Patients With Aortic Regurgitation and 7 Patients With Aborted Delivery)

Variable

Pearson correlation (95% confidence interval)

p value

0.354 0.329 -0.04 0.095 0.096 0.141

0.047 0.066 0.829 0.606 0.601 0.440

LV mass (long-axis view) LV mass (2-chamber view) Height Weight Body surface area Coronary vascular resistance

The intraobserver variation for both observers and the interobserver variability showed acceptable agreement for the AR vena contracta and internal LV chamber dimensions but not the LV wall thicknesses. The limits of agreement were less than 30% of the mean value for all comparisons other than the wall thicknesses where the agreements were 32%, 34%, 43%, and 44% for the inferolateral, inferior, anteroseptal and anterior walls, respectively. DISCUSSION

The authors’ study showed that the estimation of LV mass by intraoperative TEE was not clinically useful for estimating the amount of initial volume of antegrade cardioplegia required to achieve asystole. However, it was shown that TEE was useful for routine monitoring for left ventricular distention during antegrade cardioplegia delivery as excessive LV distention was common, requiring conversion to retrograde delivery to achieve asystole in a quarter of patients and occurs in nearly 20% of patients without detectable AR prior to crossclamp application. An AR vena contracta of 0.3 cm measured during pre-bypass TEE is a reasonable predictor of LV distention occurring, but given the unpredictable occurrence of LV distention without AR, the authors recommend routine monitoring for LV distention with TEE.

Fig 1. Scatterplot with line of best fit demonstrating Correlation of initial antegrade cardioplegia volume required for asystole with left ventricular mass estimated from transesophageal echocardiography using the transgastric long-axis view, excluding patients with aortic regurgitation or in whom cardioplegia delivery was aborted due to excessive left ventricular distention.

Fig 2. Correlation of severity of aortic regurgitation and degree of left ventricular distention during antegrade cardioplegia delivery in 19 patients with aortic regurgitation. Patients in whom cardioplegia was aborted and sustained were analyzed separately.

Excessive delivery of cardioplegia may have adverse side effects, and it is desirable to minimize the dose required to achieve asystole and protect the myocardium. Although it seems logical that the dose of cardioplegia required to cause asystole should be proportional to the mass of myocardium, in this study, there was only a poor correlation with LV mass estimated with TEE in patients without AR. Intraoperative estimation of LV mass with TEE is, therefore, not a clinically useful measurement to calculate the amount of cardioplegia required to arrest the heart. Even in the absence of detectable AR by TEE, 7 out of 39 patients (18%) suffered significant LV distention during antegrade cardioplegia delivery, requiring retrograde delivery to induce asystole. This implies newonset AR after commencing CPB, aortic cross-clamping, and the commencement of cardioplegia delivery. The mechanism for new AR is not known but is presumably from reduced aortic cusp coaptation due to changes in aortic root geometry resulting from the cross-clamp or hemodynamic effects of the jet of cardioplegia onto the aortic leaflets. It was likely that there were also a significant proportion of patients who also developed new-onset AR but in whom antegrade delivery of cardioplegia was able to induce asystole as shown by the number of patients with LV distention (mean 3.7 ⫾ 3.6 cm2, points above the mean bar in Figure 3A) measured with TEE. This would represent a likely reason why a lack of correlation demonstrated between LV mass and volume of cardioplegia required to cause asystole, as a significant proportion of cardioplegia volume may not have been delivered to the coronary circulation. More importantly, if the LV is not monitored for distention, it may go unnoticed, potentially causing mechanical and ischemic damage to the LV, compounded by the effect of inadequate delivery and myocardial protection from cardioplegia. Resulting ventricular fibrillation could worsen LV ischemia from a resulting rise in oxygen consumption. Downing et al demonstrated impaired contractility in sheep after passive distention of the LV protected with cardioplegia3 compared with controls. Intraoperative TEE provides a convenient method for monitoring for LV distention in contrast to visual examination of the heart, which predominantly monitors the right ventricle.

TEE GUIDANCE OF CARDIOPLEGIA

5

Fig 4. Receiver operating curve of using aortic regurgitation vena contracta to predict unacceptable left ventricular distention during antegrade cardioplegia delivery (AUC 0.81, 95% CI 0.66 to 0.99, p ¼ 0.02, cut-off VC 0.29: sens 71% spec. 81% likelihood ratio 3.8).

Fig 3. (A) Degree of left ventricular distention compared between patients in whom delivery of antegrade cardioplegia was completed (non-aborted group) and terminated early due to excessive distention (aborted group). (B) Severity of aortic regurgitation was significantly different between patients with acceptable and unacceptable left ventricular distention during antegrade cardioplegia delivery (VC 0.3 unpaired 2-tailed t-test, p ¼ 0.046). Aborted; initial antegrade cardioplegia stopped due to unacceptable left ventricular distention.

Unlike palpation of the LV, TEE is able to directly measure and track the LV cavity size, potentially alerting the surgeon to significant AR before a rise in LV pressure occurs. It is possible that a correlation was not detected due to measurement error in estimation of LV mass by TEE. Left ventricular mass is calculated from the maximum LV internal diastolic diameter and the thicknesses of the 2 opposing walls. The interobserver variability of the internal diameter was acceptable, but the variability of interobserver agreement of the LV wall thickness was higher. The reasons for variability in LV wall measurement were due to difficulty in discriminating the endocardial border due to frequent interference with papillary muscles and the external border due to interference

with the pericardium and the right ventricle. A limitation of the prolate ellipse method of LV mass estimation is that it presumes normal geometry, such as in patients with hypertension, which might not be the case for cardiac surgical patients.6 Additionally, since the formula requires cubing primary measurements, even small errors in measurements may result in significant error. An alternative method for estimating LV mass using tracings of the endocardial and epicardial border may avoid these issues but was not found to be feasible due to inadequate imaging of the lateral borders of the LV with TEE.6 Coronary artery stenoses and occlusion have been demonstrated to impair regional perfusion of antegrade cardioplegia9 and may have impacted on the correlation of cardioplegia volume and LV mass as 69% of patients in this study underwent CABG. However, there was no correlation between the cardioplegia volume required and the relative coronary vascular resistance calculated from the aortic root pressure and cardioplegia flow rate (r ¼ 0.119). This demonstrated that a coronary stenosis affects the time required for cardioplegia to induce asystole but not the volume required. In addition to the issues already raised, there were several important limitations to this study. Although the LV cavity dimensions were measured during antegrade delivery, the degree of LV distention for the decision to abort antegrade delivery was not defined and was at the discretion of the surgeon. Therefore, conclusions from this must be made with caution. However it was shown that the increase in LV cavity area was significantly greater in the patients in whom antegrade cardioplegia was aborted compared to those in whom antegrade cardioplegia was sustained. What has been demonstrated is the high frequency of significant LV distention during cardioplegia delivery, even in the absence of detectable aortic regurgitation, which could go unnoticed if not monitored. TEE was used to estimate the LV mass and did not account for the mass of the right ventricle, which, although it is usually much smaller than the LV, may have varied considerably and represents a potential confounder to the correlation of the LV mass and volume of cardioplegia required for asystole. There were a number of technical problems with measuring the cardioplegia volume required for asystole.

6

CANTY ET AL

As the cardioplegia volume is displayed in real time, it requires careful scrutiny from a dedicated observer who relies on verbal prompting as to when to take the measurement from another observer watching the ECG for asystole. An error in reading the volume of 2 seconds would result in an error of 6 mL at a flow rate of 200 mL/minute. In conclusion, intraoperative TEE estimation of LV mass using the prolate ellipse formula does not correlate with the amount of antegrade cardioplegia required to cause initial cardiac asystole. Routine intraoperative monitoring for LV distention with TEE during antegrade cardioplegia delivery

is recommended due to frequent and unpredictable LV distention. ACKNOWLEDGMENTS

The authors are grateful for the assistance from the cardiac surgeons, anesthetists, and clinical perfusionists at Monash Medical Centre and The Royal Melbourne Hospital who assisted with data collection. They also thank Sandy Clarke and Darsim Haji for statistical advice. There was no funding provided or competing interests declared.

REFERENCES 1. Maruyama Y, Chambers DJ, Ochi M: Future perspective of cardioplegic protection in cardiac surgery. J Nippon Med Sch 80: 328-341, 2013 2. Robicsek F: Administration of hypothermic cardioplegia in the presence of aortic regurgitation. Ann Thorac Surg 39:192-193, 1985 3. Downing SW, Savage EB, Streicher JS, et al: The stretched ventricle. Myocardial creep and contractile dysfunction after acute nonischemic ventricular distention. J Thorac Cardiovasc Surg 104: 996-1005, 1992 4. Matsuda H, Maeda S, Hirose H, et al: Optimum dose of cold potassium cardioplegia for patients with chronic aortic valve disease: Determination by left ventricular mass. Ann Thorac Surg 41: 22-26, 1986

5. Zoghbi WA, Enriquez-Sarano M, Foster E, et al: Recommendations for evaluation of the severity of native valvular regurgitation with two-dimensional and Doppler echocardiography. J Am Soc Echocardiogr 16:777-802, 2003 6. Lang RM, Bierig M, Devereux RB, et al: Recommendations for chamber quantification. Eur J Echocardiogr 7:79-108, 2006 7. Devereux RB, Alonso DR, Lutas EM, et al: Echocardiographic assessment of left ventricular hypertrophy: Comparison to necropsy findings. Am J Cardiol 57:450-458, 1986 8. Levy D, Savage DD, Garrison RJ, et al: Echocardiographic criteria for left ventricular hypertrophy: The Framingham Heart Study. Am J Cardiol 59:956-960, 1987 9. Aronson S, Lee Bk, Liddicoat JR, et al: Assessment of retrograde cardioplegia distribution using contrast. Ann Thorac Surg 52:810-814, 1991