The study was approved by the Institutional Review Board of the Pusan National University Yangsan Hospital (05-2021-209). The requirement for informed consent was waived owing to the retrospective nature of the study.

This retrospective observational study included patients who underwent HTx at a single center between June 2014 and March 2021. All data were collected from an institutional prospective surgical database and electronic charts. During the study period, 89 patients underwent HTx at a single institution. Of these, 77 (86.5%) cases of HTx were conducted in adult recipients, and 73 patients were included in present study after excluding four multiorgan transplantation cases. MD was defined as based on the following criteria: donor age >55 years, left ventricular ejection fraction (LVEF) <50% on preoperative echocardiography, cold ischemic time >240 minutes, or significant cardiac structural problems on any imaging study. Thirty-two (43.8%) recipients were classified into the MD group, and the remaining 41 (56.2%) were included in the non-marginal donor (NMD) group (Figure 1).

The primary and secondary endpoints were mortality and primary graft dysfunction (PGD), respectively. The definition of PGD was any graft dysfunction that occurs within 24 hours of HTx, and complies with the International Society for Heart and Lung Transplantation (ISHLT) meeting in 2014 (Supplementary Table 1).7) Graft function after HTx was evaluated based on vasoactive inotropic score (VIS). VIS was derived from summing of inotropic dose rates: Dopamine Dose (μg/kg/min) + Dobutamine Dose (μg/kg/min) + 100 × Epinephrine Dose (μg/kg/min) + 10 × Milrinone Dose (μg/kg/min) + 10,000 × Vasopressin Dose (unit/kg/min) + 100 × Norepinephrine Dose (μg/kg/min).

Additionally, to quantify preoperative recipient risk, we employed the Index for Mortality Prediction after Cardiac Transplantation (IMPACT) score.8)9) The details of IMPACT score ere described in Supplementary Table 2.

We performed a conventional median sternotomy with selective vena caval cannulation for HTx. Inferior vena cava was drained cannulation via the femoral vein access. A recipient who underwent preoperative ECMO was converted to cardiopulmonary bypass (CPB) using existing cannulas. In such cases, cannulation of the superior vena cava was performed before aortic cross clamping (ACC). For patients who did not receive ECMO, arterial cannulation was routinely performed in the distal aorta. To minimize the CPB time, the heart of the recipient was explanted just before the donor heart arrived. Donor hearts were protected with infusion of a 4°C HTK solution (Custodiol^®; Köhler Chemie GmbH, Bensheim, Germany) for 8 minutes before harvesting. The harvested heart was washed with the HTK solution and transported using an ice package. The standard bicaval anastomosis technique was routinely performed.

A Swan-Ganz catheter was inserted in the operating room. All patients were transferred to the cardiac intensive care unit (ICU) after surgery. The postoperative target ranges of the monitoring indicators were as follows: cardiac output, 4–7 L/min; target heart rate, 95–110 beats/min; pulmonary capillary wedge pressure (PCWP), 15–20 mmHg; central venous pressure, <15 mmHg; <1/3 of pulmonary arterial pressure over systemic arterial pressure; lactate level, <2.0 mmol/L; mixed venous oxygen, >70 mmHg; pulmonary partial pressure of oxygen, > 100 mmHg; urine output >100 mL/hr; and potassium level >4.0 mmEq. Low dose dobutamine and norepinephrine were the most frequently used agents to maintain vital signs. If the patient’s demand for inotropes increased and isolated pulmonary arterial hypertension was detected, milrinone was given, and if decreased systemic vascular resistance was detected with norepinephrine demand >0.2 mcg/kg/min, vasopressin was administered. Isoproterenol was used to achieve the target heart rate in cases of poor response to chronotropic drugs or complete atrioventricular block, and temporary or epicardial pacing was applied. After confirming consciousness and controlled bleeding, the patients were extubated. They were sent to the general ward from 3 days after the operation according to their medical conditions. Cardiac rehabilitation was started after chest tube removal and continued until discharge.

An anti-interleukin-2 receptor monoclonal antibody (basiliximab) was used for induction from June 2014 to July 2018. Antithymoglobulin (rabbit) was replaced with basiliximab from August 2018. However, in cases of recipients older than 65 years of age or those with a high risk of infection, induction therapy was omitted. A bolus of 500 mg of methylprednisolone was administered when allograft left atrial anastomosis was initiated. The initial postoperative dose of prednisolone for maintenance therapy was 1 mg/kg/day, which was tapered to 0.25 mg/kg/day at 1 month and then to 2.5–5 mg/day at 1 year. Prednisolone was discontinued at approximately 1 year after HTx, except in the patients who experienced any type of rejection. Maintenance immunosuppression consisted of tacrolimus, mycophenolate mofetil (500–1,000 mg), and corticosteroids. Twelve months postoperatively, tacrolimus and everolimus, with or without prednisolone, were administered. In cases of cross-matching positive transplantation, pre-/post-desensitization with plasma pheresis and intravenous immunoglobulin (total, 1 mg/kg for 3 days) were administered. Rituximab, with or without bortezomib, was used to prevent antibody-mediated rejection.

Variables were summarized by frequency and percentage for categorical data and mean ± standard deviation and median (range) for numeric data. Group differences were tested using the χ² test or Fisher’s exact for categorical data and independent t-test or Mann-Whitney U test for numeric data as appropriate. To check if its distribution is normal, we used Shapiro-Wilk’s test.

Univariate and multivariate logistic regression analysis were performed to identify prognostic factors which are independently related to event. Due to the small number of deaths, risk factor analysis for the primary endpoint has been judged to be inappropriate and was not described. Instead, risk factor analysis was performed on the PGD, a secondary endpoint. Variables for analysis were selected as factors considered clinically meaningful, including MD. And variables known as risk factors in previous reports were also included.7)10) Variables that were too small to be suitable for analysis were excluded (such as race, weight mismatch ≥30%, etc.) To check the multicollinearity problem, the variance inflation factor (VIF) was also estimated. VIF quantifies the severity of multicollinearity in regression analysis and VIF less than 10 indicated that there was no problematic multicollinearity among the independent variables (VIF <10).

Overall survival and PGD-free survival were estimated using Kaplan-Meier curve. Survival curves were compared between groups using the log-rank test. Univariate and multivariate analyses, using Cox proportional hazard regression analysis, were performed to identify prognostic factors which are independently related to time to event (death or PGD).

Considering the nature of the repeated measured data of echocardiography (LVEF), a generalized linear mixed model (GLMM) with random intercepts was used to fit a model. The GLMM model included repeated measures of numeric variables as dependent variables; group, time, and group × time interaction as fixed effects; and subject as a random effect. To avoid making any assumptions about the covariance structure, we used an unstructured covariance matrix that was allowed to differ across groups for the GLMM analysis.

All statistical analyses were carried out using SPSS 26.0 statistical software (IBM Corp. Released 2019, IBM SPSS Statistics for Windows, version 26.0; IBM Corp., Armonk, NY, USA) and R statistical software (version 3.4.0; R Foundation, Vienna, Austria, http://www.r-project.org/). The p values less than 0.05 was considered statistically significant.

Donors in the MD group were older than those in the NMD group. LVEF was significantly lower in the MD group (Table 1). Twenty-five percent of patients in the MD group were rejected more than 5 times by other hospitals with a higher HTx allocation priority. Sex and medical history were not significantly different between the two groups. However, the rates of preoperative ECMO and ICU admission, the degree of emergency were higher in the NMD group. Although the preoperative creatinine level was higher in the NMD group than in the MD group, other laboratory results were comparable between the groups. Donor and recipient characteristics are described in detail in Table 1.

Cold ischemic and total ischemic times were numerically longer in the MD group. The ACC and CPB times in the recipients were also similar. There was one case of early mortality in each group. The incidence of other adverse events during the perioperative period between the groups. The surgical and perioperative details are presented in Table 2.

Immediate postoperative creatine kinase-MB isoenzyme levels were similar between the two groups, and vasoactive inotropic scores were not different within 24 hours (Figure 2A). The cardiac index showed a slightly lower trend in the MD group compared to the NMD group without a statistical difference immediately postoperatively; however, it became similar between groups after 24 hours (Figure 2B). Moderate or higher grade PGD developed in 13 (17.8%) patients during the perioperative period, and there was no difference in the incidence of PGD between the groups. On the first echocardiography performed 1 week postoperatively, there was no difference in the LVEF between groups (Table 2). The 30-day and 1-year survival rates were 97.6±2.4% and 90.2±4.6% in the NMD group and 96.9±3.0% and 87.5±5.9% in the MD group (p=0.87, 0.73 respectively). There was no difference in the long-term mortality rate between groups: 12.5% vs. 9.8% in MD vs. NMD during follow up (mean 26.3±24.6 vs. 29.1±17.6 months, p=0.75; Figure 3A). The LVEF on serial echocardiography showed no difference between 2 groups during long-term follow-up (Figure 3B).

In the multivariate logistic regression analysis for PGD, there were no meaningful predictors. MD hearts or IMPACT scores were not significant predictors of PGD. Logistic regression analysis failed to converge since there was no case of PGD in female to male transplantation. The detailed values of the univariate analysis are shown in Table 3.

Of the MD group, 19 recipients (59.3%) were classified as low-risk groups (IMPACT Risk Score <10) and 13 as high-risk groups (IMPACT Risk Score ≥10). There was no difference in donor characteristics or surgical profiles between the low- and high-risk groups (Table 4). The high-risk group showed higher incidence of prolonged ICU stay (Table 4). Other clinical outcomes were not different between two groups. However, the low-risk group showed better long-term survival than high-risk group (Figure 4). Particularly, there was no mortality in low-risk group during follow-up regardless of donor’s marginality (Figure 4A). On the other hand, high-risk recipients transplanted with marginal donor showed lower survival during follow up period (Figure 4B). In the NMD group, recipient risk status did not affect the survival (low-risk vs. high-risk recipient group, 100% vs. 86.2%, p=0.19; Figure 4C). However, in the MD group, the survival was significantly different according to the recipient risk status (low-risk group vs. high-risk group, 100% vs. 69.2%, p=0.01; Figure 4D).

Like other donor allocation programs, Korean Network for Organ Sharing (KONOS) re-allocates declined donor heart to the patient with highest priority among the remaining candidates. We collected the specific reasons for turning down donor hearts among candidates with higher priority than our patients. We considered that if a donor heart was declined more than 5 times by other candidates due to the ‘donor factor,’ it was discerned as a marginal donor by other institutions meaningfully. Supplementary Table 3 summarizes the characteristics and results of seven transplant cases utilizing donor hearts that were declined more than 5 times by other listed candidates. The specific reasons for ‘medical unsuitability’ of donor hearts were various, including old age (61 years), bradycardia, eccentric left ventricular (LV) hypertrophy, LVEF of 47% with regional wall motion abnormalities, right ventricular dysfunction (suspicious intracardiac right to left shunt), LVEF of 35% with global hypokinesia, and one case with an unidentifiable reason. In two donors whose donations were delayed for remnant electroencephalogram signals, the LVEF recovered to the normal range after transplantation. Donor #4, whose diastolic LV septal and posterior wall thicknesses exceeded 13 mm, required permanent pacemaker implantation for atrioventricular dyssynchrony despite 1 month of waiting for recovery after transplantation. Sinus rhythm was completely restored 2 weeks after permanent pacemaker insertion. All recipients had excellent 1-year survival rates.