Does gender bias affect outcomes in mitral valve surgery for degenerative mitral regurgitation?

Abstract

 

OBJECTIVES

This study was conducted to determine if gender bias explains the worse outcomes in women than in men who undergo mitral valve surgery for degenerative mitral regurgitation.

METHODS

Patients who underwent mitral valve surgery for degenerative mitral regurgitation with or without concomitant ablation surgery for atrial fibrillation were identified from the Cardiovascular Research Database of the Clinical Trial Unit of the Bluhm Cardiovascular Institute at Northwestern Memorial Hospital and were defined according to the Society of Thoracic Surgery National Adult Cardiac Surgery Database. Of the 1004 patients (33% female, mean age 62.1 ± 12.4 years; 67% male, mean age 60.1 ± 12.4 years) who met this criteria, propensity score matching was utilized to compare sex-related differences.

RESULTS

Propensity score matching of 540 patients (270 females, mean age 61.0 ± 12.2; 270 males, mean age 60.9 ± 12.3) demonstrated that 98% of mitral valve surgery performed in both groups was mitral valve repair and 2% was mitral valve replacement. Preoperative CHA₂DS₂-VASc scores were higher in women and fewer women were discharged directly to their homes. Before surgery, women had smaller left heart chambers, lower cardiac outputs, higher diastolic filling pressures and higher volume responsiveness than men. However, preoperative left ventricular and right ventricular strain values, which are normally higher in women, were similar in the 2 groups, indicating worse global strain in women prior to surgery.

CONCLUSIONS

The worse outcomes reported in women compared to men undergoing surgery for degenerative mitral regurgitation are misleading and not based on gender bias except in terms of referral patterns. Men and women who present with the same type and degree of mitral valve disease and similar comorbidities receive the same types of surgical procedures and experience similar postoperative outcomes. Speckle-tracking echocardiography to assess global longitudinal strain of the left and right ventricles should be utilized to monitor for myocardial dysfunction related to chronic mitral regurgitation.

INTRODUCTION

The overall mortality rates associated with cardiac surgery have dropped steadily over the past 5 decades, but several recent studies have suggested that the improved surgical outcomes are not equally shared among men and women. A recent study demonstrated that female sex was associated with increased operative mortality and a 15% greater likelihood of dying within 1 year of surgery [1]. Other studies have suggested that women often present for degenerative mitral regurgitation (DMR) surgery at an older age and with more advanced disease [2–4]. A previous study from our group found that women are referred for mitral valve (MV) surgery later in the course of their disease [5], and registry data suggest that valvular heart disease in women may be underdiagnosed [6]. Women are also more likely to undergo MV replacement rather than MV repair, even though repair has better long-term outcomes [7, 8]. In addition, MV repair restores normal life expectancy in men but not in women [8].

At the time of surgery, women have smaller left atria, smaller left ventriculars (LVs) and greater MV regurgitant volumes than men [9], but current recommendations for the timing of MV surgery are based primarily on male populations [10]. Consequently, it can be hypothesized that women may meet the criteria for surgical intervention later in the course of their disease than men. Despite these observed differences, no studies have adequately assessed preoperative myocardial function and corresponding postoperative outcomes in DMR patients using conventional echocardiography and speckle-tracking analysis of preoperative atrial and ventricular strain to distinguish between the impact of potential sex bias, preoperative disease status, and intraoperative surgical practice on outcomes in women and men.

LV global longitudinal strain (LV-GLS) is an accurate measure of LV myocardial function [5], and sex has a significant impact on both LV and right ventricular (RV) strains. In addition, LV-GLS, RV free wall strain (RV-FWS) and RV global longitudinal strain (RV-GLS) are normally higher in women than in men, although sex-specific reference values for LV-GLS and RV-GLS in degenerative mitral valve disease are unknown. In this study, we utilized conventional echocardiography and speckle-tracking analysis to compare preoperative myocardial function in women and men to identify possible causes for the apparent gender differences in postoperative outcomes for patients undergoing MV surgery for DMR.

PATIENTS AND METHODS

Study population

Data were obtained from the Cardiovascular Research Database of the Bluhm Cardiovascular Institute at Northwestern University and retrospective chart review (Institutional Review Board project STU00012288). Between April 2004 and June 2017, 1004 patients underwent MV surgery for type II DMR. Thirty-three percentage were women (mean age 62.1 ± 12.4 years) and 67% were men (mean age 60.1 ± 12.4 years). Patients who had previously undergone surgical procedures for transcatheter aortic valve (AV) replacement, ventricular assist device, cardiac transplant, cardiac trauma, previous MV intervention, emergency intervention or an undocumented atrial fibrillation (AF) type were excluded, as were patients who refused to provide consent for enrolment into the Cardiovascular Research Database. Propensity score (PS) matching was performed in 270 females (mean age 61.0 ± 12.2 years) and 270 males (mean age 60.9 ± 12.3 years).

Data were defined per the Society of Thoracic Surgeons National Adult Cardiac Surgery Database (www.sts.org). Multiple preoperative clinical variables were included (Supplementary Material, Table S1) and the grade of MV insufficiency and underlying pathophysiology were classified according to Carpentier. Operative variables included MV and AV measurements, perfusion time, cross-clamp time, concomitant surgical procedures and operative urgency (Supplementary Material, Table S2). All of the surgical procedures were performed by the senior author (Patrick M. McCarthy). Postoperative variables included total ICU time, length of hospital stay, Society of Thoracic Surgeons (STS) risk score, readmission rate and predischarge complications.

Surgical technique

All procedures were performed by a single surgeon and the primary technique of repair was based on resection with leaflet reconstruction to return the normal 2:1 ratio of anterior to posterior leaflet length using direct measurements and has been previously described in detail [11].

Standard echocardiography and tissue Doppler

All patients had either a transthoracic echocardiogram or a transoesophageal echocardiogram prior to surgical intervention. LV end-diastolic volume, LV end-systolic volume and ejection fraction were derived from apical two- and four-chamber views using Simpson’s biplane method. RV function was assessed using tricuspid annular plane systolic excursion (Table 1).

Table 1:

Preoperative echocardiography and tissue Doppler characteristics of propensity score-matched cohort

Variable	Male (n = 270)		Female (n = 270)		P-value
Left atrium, LV, aortic valve, mitral valve assessment
Heart rate (bpm), mean ± SD	68.9	± 14.2	72.5	± 20.1	0.10
Left atrial size (cm), median (Q1, Q3)	4.5	(4.0, 4.9)	4.2	(3.7, 4.7)	<0.001
Left atrial volume index (ml/m2), mean ± SD	51.2	± 22.0	51.3	± 18.6	1.00
LV end-systolic dimension, median (Q1, Q3)	33.0	(30.0, 38.0)	32.0	(28.0, 36.0)	0.002
LV end-diastolic dimension, median (Q1, Q3)	54.0	(49.0, 58.0)	51.0	(46.0, 56.0)	<0.001
LV end-diastolic volume (ml), mean ± SD	138.8	± 49.8	103.9	± 32.7	<0.001
LV end-systolic volume (ml), mean ± SD	52.5	± 24.7	39.5	± 14.9	<0.001
LV stroke volume (ml), mean ± SD	87.0	± 31.0	63.8	± 22.1	<0.001
LV cardiac output (l/min), mean ± SD	5.9	± 2.4	4.5	± 2.1	<0.001
Ejection fraction, median (Q1, Q3)	62.0	(60.0, 65.0)	62.0	(58.0, 65.0)	0.98
LVOT stroke volume Vmax, mean ± SD	1.0	± 0.2	1.0	± 0.2	0.47
LVOT stroke volume Vmean, mean ± SD	1.5	± 6.9	1.2	± 6.3	0.72
LVOT velocity time integral, mean ± SD	17.0	± 5.4	18.3	± 5.1	0.08
MV area, median (Q1, Q3)	3.3	(2.9, 4.3)	3.7	(3.2, 4.4)	0.06
MV E wave velocity, mean ± SD	1.2	± 0.3	1.1	± 0.3	0.56
MV A wave velocity, mean ± SD	1.1	± 5.0	1.0	± 3.3	0.87
MV E/A, mean ± SD	7.6	± 23.2	5.5	± 17.3	0.43
MV mean velocity, mean ± SD	0.9	± 0.7	0.9	± 0.7	0.88
MV peak velocity, mean ± SD	1.7	± 1.1	1.6	± 0.8	0.60
MV mean gradient (mmHg), median (Q1, Q3)	2.7	(2.0, 4.0)	3.0	(2.0, 4.1)	0.037
MV peak gradient (mmHg), median (Q1, Q3)	8.3	(5.7, 11.0)	8.5	(6.0, 14.0)	0.20
MV velocity time integral (cm), mean ± SD	40.6	± 28.9	39.4	± 21.9	0.83
E′ septal, mean ± SD	0.2	± 1.0	0.2	± 0.8	0.78
E/E′ septal ratio, mean ± SD	14.7	± 5.7	17.3	± 9.9	0.12
E′ lateral, mean ± SD	0.2	± 1.0	0.3	± 1.3	0.56
E/E′ lateral ratio, mean ± SD	11.4	± 4.3	13.9	± 7.3	0.053
E/E′ mean, mean ± SD	13.0	± 4.6	15.6	± 8.0	0.06
Doppler-derived mitral deceleration time (ms), mean ± SD	206.2	± 55.3	203.1	± 67.5	0.71
MR volume (ml), median (Q1, Q3)	66.0	(58.0, 88.0)	66.0	(45.0, 88.0)	0.65
MR radius (cm), mean ± SD	1.0	± 0.4	0.9	± 0.3	0.16
MR mean velocity, mean ± SD	3.9	± 0.6	15.8	± 65.2	0.13
MR mean gradient (mmHg), mean ± SD	70.0	± 21.3	75.3	± 20.9	0.16
MR peak gradient (mmHg), mean ± SD	110.4	± 29.5	119.5	± 30.4	0.083
MR velocity time integral (cm), mean ± SD	142.6	± 53.6	161.6	± 44.0	0.018
ERO, median (Q1, Q3)	0.6	(0.4, 0.9)	0.5	(0.4, 0.7)	0.075
AV area (V, D), median (Q1, Q3)	2.6	(2.2, 3.3)	2.3	(1.9, 2.9)	0.012
AV mean gradient, median (Q1, Q3)	3.9	(3.0, 5.0)	3.7	(3.0, 4.8)	0.63
AV peak gradient, median (Q1, Q3)	7.0	(5.6, 10.0)	7.0	(5.7, 9.4)	0.61
AV velocity time integral (cm), mean ± SD	25.5	± 12.8	25.8	± 9.1	0.83
Right ventricle, tricuspid valve assessment
PA mean, median (Q1, Q3)	21.0	(18.0, 25.0)	23.0	(18.0, 30.0)	0.12
PASP, median (Q1, Q3)	31.0	(26.0, 42.0)	33.0	(26.0, 40.0)	0.90
Tricuspid diameter, median (Q1, Q3)	4.2	(4.1, 4.5)	3.8	(3.2, 4.5)	0.082
TV mean gradient, median (Q1, Q3)	1.0	(1.0, 1.0)	1.0	(0.7, 1.0)	0.33
TV peak gradient, median (Q1, Q3)	23.5	(2.0, 27.0)	23.0	(3.0, 25.0)	0.78
TAPSE (cm), mean ± SD	2.4	± 0.5	2.3	± 0.5	0.025

AV: aortic valve; ERO: effective regurgitant orifice; LV: left ventricular; LVOT: left ventricular outflow tract; MV: mitral valve; MR: mitral regurgitation; PA: pulmonary artery; PASP: pulmonary artery systolic pressure; SD: standard deviation; TAPSE: tricuspid annular plane systolic excursion; TV: tricuspid valve. The bold text in table indicates statistical significance.

Speckle-tracking echocardiography

Myocardial function was further investigated using speckle-tracking analysis for strain calculations in all PS-matched patients. Speckle-tracking analysis takes advantage of the fact that ultrasound-myocardial tissue interactions cause visible white ‘speckles’ within the myocardial images that can be seen on routine two-dimensional digital echocardiography. Because these echocardiographic ‘speckles’ within the myocardium change position during each cardiac cycle, the regional function of individual segments of the myocardium can be determined by recording the degree and direction of their movement using appropriate image-processing algorithms. Since ventricular strain is the % decrease in the length of myocardial fibres between diastole and systole, longitudinal strain is expressed as a negative percentage. Note that positive strain means elongation, whereas negative strain denotes shortening.

For LV-GLS analysis, two-dimensional grey-scale images were acquired from the apical four-chamber, apical three-chamber and apical two-chamber views (Fig. 1). Specifically, peak longitudinal strain based on two-dimensional analysis was used to compute LV-GLS (Fig. 1). Archived digitally acquired images were uploaded to a TomTec system (Image Arena version 4.6, TomTec, Munich, Germany) for strain calculations. All echocardiographic images had frame rates no less than 40 fps. The region of interest for GLS determination included the mid-wall and was recorded at end-diastole, i.e. the frame before the mitral valve completely closed, using the QRS complex. Systolic length was recorded during the first frame after complete AV closure in the apical three-chamber view. The software system then automatically constructed a strain curve from the standard 2D echo grey-scale images. All patients were in sinus rhythm at the time of strain analysis.

Figure 1:

Methodology of speckle-tracking echocardiography to assess left ventricular global longitudinal strain. Global longitudinal strain of the left ventricular is calculated using a combination of the two-chamber, three-chamber and four-chamber echo views of the heart. A Bulls-Eye map is automatically constructed by the software using multiple views of the left ventricular. The Bulls-Eye map simulates looking straight up the left ventricular ‘cone’ from the apex.

Similar to the LV, the region of interest of the RV is defined as its mid-wall border. For RV strain determinations, we used the RV-focused apical four-chamber view, which was optimized for orientation, depth and gain to maximize the RV size and to visualize the RV apex throughout the cardiac cycle. RV-GLS was recorded as the average of strain in the ventricular septum and strain in the RV free wall. RV-FWS measurements were performed according to the Guidelines of the American Society of Echocardiography [5]. RV global longitudinal strain (RV-GLS) was derived from an automatically traced border in the modified apical four-chamber view. After adjusting tracking points and myocardial penetration manually, 2D RV-GLS was obtained for each RV myocardial segment. The RV-FWS was automatically subdivided by the software system into apical, mid and basal segments and was automatically recorded for each segment of the RV free wall. The software system also provided the mean of 3 segments of RV-FWS (Table 2).

Table 2:

Preoperative speckle-tracking strain analysis of propensity score-matched cohort

Variable	Male (n = 270)	Female (n = 270)	P-value
LV global longitudinal strain	−20.7 ± 4.4	−20.0 ± 4.1	0.13
RV free wall strain	−20.2 ± 6.5	−20.3 ± 6.7	0.87

LV: left ventricular; RV: right ventricular.

Statistical analysis

Continuously distributed data are summarized and presented using mean ± standard deviation or median (first quartile, third quartile) values. Discrete data are presented based on counts and percentages. Continuously distributed data gender comparisons used one-way analysis of variance or the Mann–Whitney test. Gender comparisons of discrete data were based on the chi-square or Fisher’s exact test.

To estimate the PS (probability to be a female), we constructed a multivariable logistic regression model with the following explanatory variables: age, body mass index (kg/m²), LV ejection fraction, MV implant size, New York Heart Association functional class, diabetes, hypertension, hypercholesterolaemia, COPD, repeat sternotomy, AF history, concomitant AF ablation surgery, coronary artery bypass grafting, AV surgery, tricuspid valve surgery, Alfieri commissuroplasty and chordal transfer. PS matching (1-to-1) was used to alleviate baseline covariate imbalances between the 2 groups (male and female). PS matching based on a greedy algorithm with a calliper of size 0.1 logit PS standard deviation units (Fig. 2) was used. To assess covariate balance after PS matching, we used standardized mean differences, with values <0.2 in absolute value being considered indicative of adequate balance [12]. Overall survival following surgery was summarized using Kaplan–Meier curves, with gender comparisons based on log-rank test. Freedom from MV reoperation was estimated based on cumulative incidence functions for competing risks models, while gender comparisons involved Gray’s test. Analyses of the entire MR patient trajectory were used to estimate freedom from late moderate-severe or greater MR (3–4+) (as yes/no outcomes) via generalized linear modelling. Estimation involved generalized estimating equations, assuming an independent within-patient correlation structure. Statistical significance was established at a two-sided 0.05 level. Statistical analyses were performed in SAS software (v 9.4, SAS Institute, Cary, NC).

Figure 2:

Dotplot of standardized mean differences between genders in original and propensity score-matched groups (orange dots = original groups, violet dots = propensity score-matched groups). AF: atrial fibrillation; CABG: coronary artery bypass grafting; NYHA: New York Heart Association.

RESULTS

Figure 3 shows no statistically significant overall survival differences by gender in the original (panel A1), as well as the PS-matched groups (panel A2). Freedom from MV reoperation was high at 10 years post-discharge (panel B1: 99% men, 100% women, in original groups; panel B2: 98% men, 100% women, in the PS-matched groups) and not statistically significantly different by gender. Freedom from MR3+ or higher was superior among women (99.7% at 10 years post-discharge versus 98.9% for men, panel C1), and this overall difference was marginally statistically significant (P = 0.0498). After PS matching, freedom from MR3+ or higher did not differ by gender (P = 0.119, panel C2), while 10-year rates remained excellent (99.6% women, 98.6% men).

Figure 3:

Overall survival (first row), freedom from mitral valve reoperation (second row) and freedom from MR3+ or higher mitral regurgitation (third row) in the original groups (first column) and propensity score-matched groups (second column) for men compared to women who underwent surgery for degenerative mitral regurgitation. Pointwise 95% confidence intervals are indicated by the shaded areas.

A total of 540 patients (270 women, 270 men) were PS matched, resulting in well-balanced groups, as assessed by standardized mean differences being <0.2 after PS matching (violet dots) (Fig. 2). Preoperative CHA₂DS₂-VASc scores were higher for women than for men (2.2 ± 1.3 vs.1.3 ± 1.3, P < 0.001) and creatinine levels were lower in women than in men (0.8 vs 1.1, P < 0.001). All other baseline parameters were similar.

Our PS-matched analysis compared men and women presenting with a similar form of DMR and elucidated sex-based differences in those referred for MV surgery. Women presented with smaller (in median value) left atria (4.2 vs 4.5 cm)—the Left Atrial Volume Index (LAVI) was the same (51.3 vs 51.2 ml/m²)—lower LV end-diastolic (51.0 vs 54.0 mm) and end-systolic (32 vs 33 mm) dimensions, lower end-diastolic volumes (103.9 vs 138.8 ml), end-systolic volumes (39.5 vs 52.5 ml), stroke volumes (63.8 vs 87.0 ml) and lower cardiac outputs (4.5 vs 5.9 l/min) than men, respectively. Notably, preoperative LV-GLS and RV-FWS were the same in both groups (Table 2).

The operative characteristics were similar in women and men in the PS-matched groups. Women and men received the same type of mitral valve surgery, with mitral valve repair being performed in 98% of both groups. Comparison of the length of hospital stay, postoperative mortality and readmission rates also revealed no differences. Fewer women were discharged to their homes (236 vs 253, respectively) and their median length of ICU stay was longer for women than for men (29.1 vs 26.8 days, respectively). In addition, the proportion of men readmitted to the ICU was higher than that of women (10 vs 3, respectively), even though postoperative complications were similar in men and women. Overall survival was similar in women and men at 1, 5 and 10 years (Fig. 3).

DISCUSSION

Previous studies have suggested that women have worse outcomes than men following MV surgery for DMR [13]. This is the first study to evaluate potential sex bias in patients undergoing MV surgery for DMR in which preoperative conventional echocardiography with tissue Doppler and speckle-tracking analysis was used to determine preoperative atrial and ventricular strains. The results suggest that neither the surgical treatment of DMR nor its outcomes are based on gender differences but, rather, are dictated by the severity of mitral disease at the time of surgery.

At the time of presentation for MV surgery in our study, women had a higher preoperative CHA₂DS₂-VASc score, which would normally suggest that women were more susceptible to thromboembolic events and mortality [14]. However, since this score is primarily used as a risk indicator for patients with non-rheumatic AF, it is difficult to draw any definitive conclusions from the differences in the CHA₂DS₂-VASc scores between women and men in this study, especially since gender is a part of the CHA₂DS₂-VASc score. Creatinine levels and body surface area were additional preoperative values found to be slightly different, but these variations were likely due to the intrinsic differences in physiology and size between males and females.

Several studies that have considered differences in strain associated with sex and age have demonstrated that, among healthy individuals, women have higher absolute values of LV and RV strains compared to males [15, 16]. Interestingly, we found that LV and RV strains were similar in our PS-matched cohorts. This suggests a greater preoperative decrease in LV and RV strains in women than in men that was undetectable without analysing preoperative strain. Therefore, even though statistically ‘insignificant’, these findings suggest that women present for MV surgery later in the disease course compared to men, an observation that has been noted in previous studies. For example, a study that investigated MV surgery in Medicare patients found that the significant disparity in surgical mortality between women and men was associated with the fact that women presented for mitral valve operations later in the disease process [8]. Seeburger et al. [17] observed a similar relationship in patients who presented for minimally invasive MV surgery. In addition, tricuspid annular plane systolic excursion was found to be significantly lower in females, highlighting the fact that global RV function was worse in females than in males at the time of surgery.

Although conventional echocardiography with tissue Doppler is the most common diagnostic method for assessing myocardial function, it is limited by angle-dependency, low signal-to-noise and poor spatial resolution [18, 19]. For this reason, we sought to utilize speckle-tracking techniques to understand the function of the myocardium more precisely since this technique provides highly accurate quantification of myocardial fibre deformation and torsion [20].

Since the subtle and progressive changes in ventricular function secondary to DMR cannot be detected reliably by standard echocardiography, the ability of strain measurements to document these differences in men and women earlier should result in more optimal timing of surgical intervention in women and promises to overcome the current problem of women being referred for surgery significantly later in the course of their disease progression than men. This study documents that if women undergo surgery at a similar stage of their disease progression as men, the differences in their surgical outcomes can be expected to disappear.

Limitations

There are several limitations of this study, the most obvious being that it is a single-institution, retrospective, observational study. Even though PS matching was used to alleviate baseline covariate imbalances, there is still a possibility of residual covariate imbalance. This study did not detect sex-based bias in surgical referrals for the treatment of DMR at our institution. Three-dimensional (3D) echocardiography was not used in our study, so 3D speckle-tracking was not available and 3D parametric computations could not be used as controls.

CONCLUSION

This study confirms several previous reports showing that preoperative conventional echocardiographic parameters predict worse outcomes in women than in men following MV surgery for DMR. However, while this is clearly due to women being referred later in the course of their disease than men, this later referral of women is not necessarily due to a sex bias on the part of the referring physicians and/or operating surgeons. Our study documents that the surgical outcomes for DMR are dependent on the extent of MV disease and preoperative myocardial dysfunction at the time of surgery. Importantly, speckle-tracking echo analysis reveals that more subclinical preoperative ventricular dysfunction exists in women with DMR than in men at what appears by conventional echocardiography to be similar stages of the disease. Thus, speckle-tracking echocardiography, in addition to conventional echocardiography, should be utilized to monitor cardiac function in women to enable the earlier detection of their mitral disease progression and should result in the earlier referral of women for surgery than in the current practice. Since the outcomes of MV surgery are similar for males and females who present with the same type and degree of DMR, the additional preoperative ventricular strain data gathered from speckle-tracking should result in similar operative outcomes in women and men following mitral valve surgery for DMR in the future.

SUPPLEMENTARY MATERIAL

Supplementary material is available at ICVTS online.

Funding

No funding was received for this study.

Conflict of interest: The author(s) have the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Dr Patrick M. McCarthy is a consultant for Edwards Lifesciences and Abbott Vascular and receives royalties for intellectual properties with Edwards Lifesciences. Dr S. Chris Malaisrie is a consultant for Edwards Lifesciences and Medtronic. Dr James L. Cox is a consultant for Atricure and is a member of the Board of Directors of Adagio Medical, PAVmed and Lucid Diagnostics.

ABBREVIATIONS

AF

Atrial fibrillation

AV

Aortic valve

CHA2DS2-VASc

Congestive heart failure, hypertension, age ≥75 years, diabetes, stroke, vascular disease, age 65–74 years, sex category

3D

Three-dimensional

DMR

Degenerative mitral regurgitation

LV

Left ventricular

LV-GLS

Left ventricular global longitudinal strain

MV

Mitral valve

PS

Propensity score

RV

Right ventricular

RV-4CLS

Right ventricular four-chamber longitudinal strain

RV-FWS

Right ventricular free wall strain

RV-GLS

Right ventricular global longitudinal strain

 

Presented at the 34th Annual Meeting of the European Association for Cardio-Thoracic Surgery, Barcelona, Spain, 8–10 October 2020.

Author contributions

Viswajit Kandula: Data curation; Investigation; Writing—original draft; Writing—review & editing. Olga N. Kislitsina: Conceptualization; Investigation; Writing—review & editing. Vera H. Rigolin: Supervision; Writing—review & editing. James D. Thomas: Supervision; Writing—review & editing. S. Chris Malaisrie: Supervision; Writing—review & editing. Adin-Cristian Andrei: Data curation; Formal analysis; Visualization. Ashvita Ramesh: Data curation; Investigation; Writing—review & editing. Jane Kruse: Project administration; Supervision; Writing—review & editing. James L. Cox: Conceptualization; Methodology; Supervision; Writing—review & editing. Patrick M. McCarthy: Conceptualization; Methodology; Resources; Supervision; Writing—review & editing.

Reviewer information

Interactive CardioVascular and Thoracic Surgery thanks Panagiotis Artemiou, Eduardo Bernabeu, Carmelo Mignosa and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.