All patients with Fontan failure who underwent cardiac transplantation at Children’s Wisconsin between 2003 and 2021 were included in this study. Patients were divided into two groups, those who underwent a prior Norwood (FFN, n = 22) and those who did not undergo a prior Norwood (FF, n = 11). The Children’s Wisconsin Institutional Review Board determined this study to meet exempt criteria (ID: 1883417-1) and waiver of the Health Insurance Portability and Accountability Act was granted on May 13, 2022.

Patient information was obtained from the general medical and cardiology records, as well as our transplantation records and data submitted to the United Network for Organ Sharing (UNOS). De-identified patient information was collected utilizing a password-protected web-based server with Excel capability.

Patient charts were systematically reviewed for predetermined variables and outcomes. Donor demographics and the ejection fraction of donor hearts were taken from UNOS. Prior sternotomies were considered any entry into a closed sternum regardless of the time between procedures. Ventricular dysfunction was determined by the most immediate echocardiogram preceding cardiac transplantation, with mildly diminished or worse considered dysfunctional. Atrioventricular valve insufficiency was met if graded moderate or worse on echocardiogram. Pre-transplant circulatory support included patients who were taken to the operating room on ventricular assist devices. We intended to gather all panel reactive antibody (PRA) results immediately pre-transplant, however, for several patients, this was not possible and we only had access to PRAs that were taken months in advance. Moreover, we could not obtain PRA data for some patients in the earlier period of this cohort.

Intraoperative variables and outcomes were taken from the operative report, cardiopulmonary perfusion record, and anesthesia record. Any significant injury or event that occurred during re-entry into the sternum was considered a sternal re-entry event.

Post-operative variables and outcomes were attained from progress reports, post-operative procedure notes, and biopsy results. For this study, acute rejection was considered any grade of rejection within 1 year of transplant. Primary graft non-function (PGNF) was defined as the use of extra-corporeal membrane oxygenation (ECMO) within 48 h of transplant. Reoperations were defined as any operation that occurred post-transplant during transplant admission. Appropriate management (mediastinal exploration for simple removal of packing or vents), vacuum dressing changes, routine catheterizations, and sternal closure were not counted as a reoperation. Catheterizations were only counted if they included ballooning, stenting, or coiling of collaterals. Similarly, other procedures that did not involve the mediastinum were not counted with the exception of diaphragm plication and vocal fold injections. All patients who met the 1-year survival outcome had a last known alive date beyond 1-year post-transplant.

Medians and ranges were used to summarize continuous variables while counts and percentages were used for categorical variables. Wilcoxon rank-sum test and Fisher’s exact test were used to compare continuous variables and categorical variables, respectively, between different groups of patients. Analyses were performed using SAS 9.4 (SAS Institute, Cary, NC, USA). A threshold p-value of 0.05 was used to determine statistical significance.

Table 1 outlines preoperative characteristics and demographics between the two groups. The only significant differences between the groups were their pre-operative cardiac diagnoses (p < 0.0001), as well as median age and height at transplant. Notably, the vast majority of the FFN cohort was diagnosed with hypoplastic left heart syndrome (95.5%), while the majority of FF patients were diagnosed with double inlet left ventricle (Fig. 1). Furthermore, 6 FFN patients (27.3%) had a PA stent pre-transplant, while only 1 FF patient (9.1%) had a PA stent pre-transplant (p = 0.378). The groups were similar with respect to sensitization, ventricular function, creatinine, and bilirubin.

Table 2 demonstrates intraoperative data between the two groups. Compared to the FF group, the FFN group had a longer duration of surgery (median of 682 min vs. 575.5 min, p = 0.092) and required longer circulatory arrest times (median of 25.5 min vs. 12.5 min, p = 0.123). Moreover, the FFN cohort exclusively experienced sternal re-entry events [3 patients (13.6%) vs. 0 patients, p = 0.534] and concomitant neo-aortic reconstruction [6 patients (27.3%) vs. 0 patients, p = 0.077]. The sternal re-entry events consisted of a problem with the pacemaker that caused a drop in blood pressure, a drop in blood pressure that required cardiac massage, and direct aortic injury. Fig. 2 demonstrates the cannulation strategy for the two groups. The majority of patients were managed with central canulation (90.9% of FF patients and 81.8% of FFN patients), while a larger portion of FFN patients required initial peripheral cannulation or a combination of peripheral and central cannulation (18.1% vs. 9%, overall p = 0.583). It should be noted the above-mentioned results did not reach our threshold for statistical significance.

Table 3 demonstrates similar discharge survival between groups. The causes of death for the three FFN patients were multi-organ failure and global hypoxic ischemic injury two weeks post-transplant, multi-organ failure and diffuse watershed injury 3 months post-transplant, and fungemia progressing to cardiogenic and septic shock 3.5 months post-transplant. In the FF group, one patient died due to severe end-organ dysfunction despite adequate flows on ECMO 2 weeks post-transplant and the other patient died of multi-organ failure in the context of septic shock 3.5 months following transplant.

Figs. 3 and 4 show a graphical representation of post-operative management and morbidity trends. The FFN cohort experienced a longer duration of mechanical ventilation post-operatively (median of 5 days vs. 2 days, p = 0.033). While not statistically significant the FFN cohort demonstrated more use of open sternal management [11 patients (50%) vs. 3 patients (27.3%), p = 0.278] and longer length of stay (median of 30 days vs. 19 days, p = 0.363). Additionally, more patients in the FFN cohort required at least 1 reoperation [12 patients (54.5%) vs. 4 patients (36.4%), p = 0.465]—a larger number of FFN patients specifically required mediastinal exploration [9 patients (40.9%) vs. 2 patients (18.2%), p = 0.258] and non-routine catheterization [5 patients (22.7%) vs. 1 patient (9.1%), p = 0.637].

This study was underpowered, hence descriptive univariable statistics only, and failed to meet statistical significance for many outcomes. It must be acknowledged that due to the lack of statistical significance, our findings may be related to chance. However, we still regard many of the results as clinically relevant and believe further investigation is warranted. Surgeon bias in specific case strategy likely existed in some form and the potential impact of patient selection bias in cardiac transplantation remains unclear. It is also worth mentioning the lack of data regarding blood product administration and collateral flow. The cohort spans a long period of time during which clinical practice likely changed. A low number of patients precluded a rigorous analysis of this issue. We will attempt to overcome some of these limitations in the future by collaborating with other institutions to expand the cohort size.