Insights into Bridging with ECMO to Heart Transplantation Based on Heart Failure Etiology

M. Shah¹, B. L. Shou², A. Kilic^3
1Johns Hopkins University School of Medicine, Baltimore, Maryland ²Johns Hopkins School of Medicine, Baltimore, Maryland ³The Johns Hopkins Hospital, Baltimore, Maryland

Purpose: Extracorporeal membrane oxygenation (ECMO) as a bridge to transplantation has dramatically increased, following the changes in the heart transplant allocation policy. We aimed to investigate waitlist outcomes in patients bridged with ECMO as it relates to their underlying heart failure (HF) etiology.

Methods: The United Network of Organ Sharing (UNOS) database was queried for all adults listed for heart transplantation after the allocation change. Candidates were categorized into the ECMO group if they were on ECMO at listing and/or at transplantation. Baseline characteristics and demographics were compared between the two cohorts using appropriate statistical testing. The study population was then further stratified according to HF etiology into six subgroups: ischemic cardiomyopathy (CMP), dilated nonischemic CMP, restrictive CMP, retransplantation, congenital heart disease (CHD), and other. The Fine-Gray competing risks model was implemented with the primary outcome as death and competing risks as transplantation or clinical recovery. Fine-Gray model results were computed and presented as subhazard distribution ratios (SHR). Cox proportional hazards model was utilized to evaluate the association of ECMO with mortality upon adjustment for age, BMI, history of diabetes, and tobacco use.

Results: The study population included 18,431 patients of whom 1,069 (5.8%) were supported by ECMO at listing and/or at transplantation. Patients on ECMO were statistically different than patients not on ECMO in regard to age (43.9 vs. 50.7, p< 0.001), BMI (26.76 vs. 27.50, p< 0.001), diabetes (22.4% vs. 29.2%, p< 0.001), tobacco use (28.0% vs 39.3%, p< 0.001), previous malignancy (6.5% vs. 8.7%, p=0.010), prior cardiac surgery (27.7% vs. 35.1%, p< 0.001), preoperative dialysis (18.1% vs. 5.0%, p< 0.001), and serum creatinine (1.33 vs. 1.42, p=0.020) (Table).

Using Fine-Gray modeling, mortality trends were analyzed for 1 year after initial listing. In our full study cohort, ECMO was associated with an increased risk of waitlist mortality (SHR = 2.184, p< 0.001). After stratifying candidates into diagnosis subgroups, this trend persisted in every subgroup with varying degrees of increased risk: ischemic (SHR=1.668, p=0.006), dilated nonischemic (SHR=1.464, p=0.007), restrictive (SHR=3.026, p=0.005), retransplant (SHR=2.062, p=0.005), CHD (SHR=3.323, p< 0.001), and other (4.530, p< 0.001) (Figure). Cox modeling demonstrated that ECMO was associated with an increased hazard of death (HR=5.578, p< 0.001) after multivariate adjustment. Again, this association persisted in all subgroups with varying degrees: ischemic (HR=7.236, p< 0.001), dilated nonischemic (HR=4.510, p< 0.001), restrictive (HR=9.227, p< 0.001), retransplant (HR=6.473, p< 0.001), CHD (HR=4.022, p< 0.001), and other (HR=10.760, p< 0.001).

Conclusion: In evaluating HF etiology and preoperative ECMO, we found that ECMO was consistently associated with increased mortality risk. The degree of impact appears to vary by diagnosis, trending towards greater risk for non-dilated CMP candidates. This warrants future research into ECMO's implications which can help inform decision-making regarding cannulation.

Identify the source of the funding for this research project: There was no funding obtained for this research project.