The Impact of the 2018 Allocation System Change on Patients Bridged With Durable Left Ventricular Assist Device: An Updated UNOS Registry Analysis

Heart transplantation (HT) remains the gold-standard therapy for end-stage heart failure (HF). Conversely, left ventricular assist device (LVAD) technology made significant strides, reaching 2 year overall survival comparable to HT.1 In the old HT allocation system, ~50% of HT occurred in LVAD patients with lower waitlist mortality, and their 5 year post-HT outcomes were equivalent to those without LVAD.2 On October 18, 2018, the heart transplant allocation system changed (ASC) into a six tiered classification to prioritize the most unstable patients. Uncomplicated LVAD patients were placed in lower priority status, rapidly changing clinical practice in the United States. More and more patients who would be earlier implanted with LVAD were now bridged with temporary mechanical circulatory devices.3,4 However, preliminary studies of the ASC in LVAD patients showed reduced wait time and similar waitlist outcomes but worse post-HT survival.3 In contrast, subsequent analyses did not reveal inferior post-HT outcomes under the new policy.1,5 Most recently, a united network for organ sharing (UNOS) registry analysis showed analogous 2 year post-HT survival between policy eras.6 Despite numerous efforts to understand the impact of the ASC, overall findings are still conflicting. We leveraged the UNOS registry to address these inconsistencies in reported outcomes using a different statistical approach to minimize risk of bias from temporal outcome changes. Two cohorts of patients (≥18 years) from the UNOS database were identified based on listing period (pre-ASC: October 18, 2014, to October 18, 2017, and post-ASC: October 18, 2018, to October 18, 2021). We included patients with HeartMate II, HeartWare, or HeartMate 3 (HM3) and excluded patients i) listed before and transplanted after ASC, ii) listed for multiorgan or retransplant, iii) implanted with LVAD after listing given uncertainty of LVAD implant timing, and iv) lost to post-HT follow-up. The primary outcome was 1 year post-HT mortality. Secondary outcomes included waitlist mortality and HT rates. To account for possible temporal changes in the HT rates, comparisons were performed within each era. Cox regression models were used to evaluate the association of clinically relevant variables (including blood type and listing status at HT) with study outcomes. Of 17,689 listed patients, we identified 3,656 LVAD patients (n = 1,963 post-ASC), of which 2,853 were ultimately transplanted (n = 1,519 post-ASC). Baseline characteristics are shown in Table S1, Supplemental Digital Content, https://links.lww.com/ASAIO/B176. After the ASC, LVAD patients were more frequently implanted with HM3 and experienced both an average decrease in wait time of 35 days and an average increase in ischemic time of 52 minutes (p < 0.001, for all). Despite listing post-ASC being associated with higher HT rates, the likelihood of HT was lower among patients with versus without LVAD in the post-ASC cohort (pre-ASC: hazard ratio [HR], 0.97 vs. post-ASC: HR, 0.75; p interaction <0.001; Table 1 and Figure 1). Left ventricular assist device support was associated with a lower hazard of waitlist death at 1 year follow-up in both eras; however, the survival benefit from LVAD support was significantly lower in the post-ASC period (pre-ASC: HR, 0.28 vs. post-ASC: HR, 0.55; p interaction = 0.006, Table 1). Table 1. - Hazard Ratios of 1 y Likelihood of Heart Transplant, Waitlist Mortality, and Post-Transplant Mortality Between LVAD vs. Non-LVAD Patients in the Old and New Heart Transplant Allocation Policy Eras Outcome Pre-ASCHR (95% CI) Post-ASCHR (95% CI) p interaction Heart transplantation (model 1) 0.97 (0.91–1.04) 0.75 (0.71–0.80) <0.001 Waitlist mortality (model 2) 0.28 (0.19–0.38) 0.55 (0.38–0.79) 0.006 Post-transplant mortality (model 3) 1.18 (0.95–1.45) 1.58 (1.31–1.91) 0.039 P<0.05 are in bold.HR reported for LVAD patients compared with non-LVAD patients (reference group) within each allocation system era. pinteraction between LVAD support and ASC. Adjusted for the following covariates: Model 1: recipient age, gender, race, ECMO at listing or transplant, blood type O along with body mass index >30, transplant region (North East includes regions 1, 2, and 9; South includes region 3, 4, and 11; Mid-West includes regions 7, 8, and 10; and West includes regions 5 and 6). Model 2: covariates used for model 1 in addition to diabetes mellitus, and tobacco use. Model 3: covariates used for models 1 and 2, in addition to estimated glomerular filtration rate, ventilator use at time of transplant, ischemic cardiomyopathy, ischemic time, predicted heart mass ratio, donor age, high listing status at transplant (1A in pre-ASC or ≤3 in post-ASC), and female donor to male recipient mismatch. Individual covariates estimates are listed in Table S2, Supplemental Digital Content, https://links.lww.com/ASAIO/B176.ASC, allocation system change; CI, confidence interval; HR, hazard ratio; LVAD, Pre-ASC, before October 18, 2018; Post-ASC, after October 18, 2018. Figure 1.: Cumulative incidence of transplantation in waitlisted patients with vs. without LVAD in the old and new Heart Transplant Allocation Policy eras. ASC, allocation system change; LVAD, left ventricular assist device; Post-ASC, after October 18, 2018; Pre-ASC, before October 18, 2018; w/, with; w/o, without.One year post-HT mortality was not different between patients with versus without LVAD in the pre-ASC era (HR, 1.19; p = 0.11, Table 1), whereas it was 58% higher among patients with versus without LVAD after the ASC (HR, 1.58; p < 0.001; p interaction = 0.043, Table 1 and Figure 2). Risk estimates of individual covariates included in the multivariable models are shown in Table S2, Supplemental Digital Content, https://links.lww.com/ASAIO/B176.Figure 2.: Kaplan-Meier analysis of 1 y survival after heart transplantation in patients with vs. without LVAD support in the old and new Heart Transplant Allocation Policy eras. ASC, allocation system change; LVAD, left ventricular assist device; Post-ASC, after October 18, 2018; Pre-ASC, before October 18, 2018; w/, with; w/o, without.This is the most contemporary UNOS registry analysis addressing the impact of ASC in LVAD patients. We found that i) post-HT mortality increased substantially after the ASC in LVAD recipients compared with non-LVAD patients and ii) LVAD is associated with lower HT probability and risk of waitlist mortality across both eras, but to a different extent after the ASC. The prior allocation system contributed to a significant allocation imbalance by incentivizing HT among LVAD patients.3 An ASC was proposed, downgrading LVAD patients without an expected decrease in waitlist or post-HT survival given improved LVAD technology.7 Despite preliminary findings of worse post-HT survival after the ASC in LVAD patients, possibly explained by prolonged ischemic times,3 subsequent studies with longer follow-up noted no difference in post-HT survival perhaps due to larger geographic donor pool and shorter wait time counterbalancing the consistently elevated ischemic times.5,6 However, our study, which comprised a larger number of patients with more updated follow-up than others,5,6 aligns with some of the preliminary studies.3 Our analyses were also controlled for several potential confounders and temporal outcome changes were accounted for by using non-LVAD patients within the same era as control group. Hess et al.6 demonstrated similar survival from initial waitlisting between eras, but both waitlist and post-HT mortality were grouped in their waitlist analysis. In our study, while patients on LVAD are still at a significant survival advantage, the protective effect is significantly diminished in the new era. Furthermore, HT rates were significantly lower among patients supported with LVAD compared with those without LVAD in the new era. Our study shows that the ASC led to a shorter HT wait time for LVAD patients, offset by higher post-HT mortality. Like previous studies, our findings highlight the importance of continuously monitoring the impact of policy changes, especially in subgroups that could be disadvantaged. Mounting cumulative evidence on worse post-HT outcomes in LVAD patients urges for critical reassessment of current listing criteria.