Transplantation for alcohol-associated liver disease in the modern era: Benefits and issues unresolved

Liver transplantation (LT) for alcohol-associated liver disease (ALD) has experienced substantial changes in practice during the past decade.^[1,2] Most transplant centers no longer have strict cutoffs for the duration of abstinence prior to LT for ALD.^[3] In combination with LT policy changes at transplant centers across the United States, the growing burden of AUD and ALD has contributed to a rise in transplantation for ALD, such that it is now the leading indication for LT in the United States.^[4,5]

In 2020, a new liver allocation policy based on acuity circles (AC) was implemented with the goal of reducing the geographic variation in the median Model for End-Stage Liver Disease (MELD) score at transplant and increasing access for patients with the highest MELD scores. AC allocates liver allografts using concentric circles with radiuses of 150, 250, and 500 nautical miles around donor hospitals based on MELD. In 2022, Wey et al^[6] found that candidates in higher MELD categories without MELD exceptions had better access to transplants after AC implementation. Though the authors did not stratify results by disease categories, it is possible that this policy change may favor individuals with ALD, who have been found to have a higher median MELD at the time of listing.^[7]

Using nationally available data from the Organ Procurement and Transplantation Network Scientific Registry for Transplant Recipients, Vock et al^[8] evaluated the impact of AC implementation on LT access and outcomes for patients with ALD compared with other liver disease etiologies. The authors demonstrated that after AC, adjusted transplant rates increased more in patients with ALD compared with patients without ALD, which exacerbated an already present disparity. When they were stratified by MELD score, these trends persisted. The study also revealed that post-AC implementation, death on the transplant waitlist (WL) decreased in patients with ALD but remained stable in patients without ALD. Post-LT survival was higher in patients with ALD before AC, and this disparity increased after AC. The authors acknowledged these widening disparities as a possible marker of discrepant advantages of the AC system on patients with ALD compared with others. Thus, the authors proposed the possibility of an ALD diagnosis–based coefficient to provide similar LT opportunities to patients with and without ALD.

While we acknowledge the disparities in LT access and outcomes when comparing patients with ALD to those without ALD, we would like to address several important factors that need additional discussion. First, the relationship between AC implementation, transplant rates, WL outcomes, and post-transplant survival may be confounded by changing trends in transplant center policies for ALD. Around the time of AC implementation, we experienced a dramatic shift in how we evaluate and list patients with ALD, such that more patients with severe alcohol-associated hepatitis were being considered for LT. Furthermore, the study period overlapped with the onset of the COVID pandemic, which was accompanied by rising alcohol use and ALD disease burden. These temporal changes may contribute to the trends observed in this paper.

A second point to consider is how unique disease characteristics impact waitlist and transplant dynamics. At very high MELD scores (> 30), mortality without transplant is high, and many of these patients may have acute-on-chronic liver failure. The authors acknowledge that patients with ALD may be more stable and able to undergo LT at higher MELD scores due to a lower prevalence of other medical comorbidities compared with other disease etiologies. For instance, patients with long-standing metabolic dysfunction–associated liver disease (MASLD) and progressive frailty may be less likely to undergo LT because they have less physiologic reserve and become too unstable for surgery over time, particularly at higher MELD scores or in the setting of acute-on-chronic liver failure. A diagnosis-based coefficient may penalize patients with ALD because they do not have long-standing metabolic syndrome, cardiac disease, or other medical comorbidities, thereby compromising post-transplant outcomes for transplant programs. Third, we should consider how recent changes in transplant practice and policy might affect observed trends after MELD 3.0 adoption and increasing use of deceased after cardiac death organs with machine perfusion.

We also wonder about how individual patient contribution to waitlist days factored into the analysis, whether the non-ALD group may be skewed by a smaller number of patients who contribute more waitlist days (ie, patients with MASLD), and how MELD trajectory for these patients is incorporated into these analyses. Patients with ALD may spend less time on the WL than other disease etiologies due to unique delisting situations (relapse, recompensation, or rapid deterioration in the very acute population with AH) and often more acute presentations, and as such, contribute fewer total waitlist days to the denominator. Although the authors included supplemental figures demonstrating WL removal/death, it is not clear how many waitlist days each individual patient contributed to the denominator; this should be further explored. Prior research on the implementation of AC demonstrated that candidates with MELD exception points had lower organ offers and transplant rates after AC implementation.^[6] This may impact the results of this manuscript because patients with MELD exceptions may contribute more waitlist days, particularly for patients with HCC who have long waitlist times.

United Network for Organ Sharing (UNOS) does not have pre-WL data on referral or evaluation and does not account for total disease burden or referral bias. Patients with ALD experience stigma and reduced referrals for LT evaluation. Even once they are referred, it is possible that the stigma associated with alcohol use disorder may impact listing. We know that the ALD waitlist is not reflective of the national ALD disease burden. Kaplan et al^[9] found that across all demographics, patients with ALD actually had reduced access to LT listing and overall liver transplant compared with other chronic liver disease etiologies. A negative coefficient might further exacerbate disparities in LT access for ALD. As the national experience builds, it will be important to study if there is differential policy impact in geographic areas with high ALD mortality burden and disproportionately low ALD transplant rates.

Lastly, we need to consider the downstream impact of an added coefficient on the MELD score. First, an ALD-based coefficient may be oversimplified. Many patients classified with ALD have underlying MASLD or could even be classified as metabolic dysfunction alcohol-associated liver disease. These patients may be disproportionately harmed by a coefficient, particularly if their disease progression and prognosis align more with the population without ALD described in this study. A negative coefficient could incentivize centers to classify patients as anything other than ALD when able (ie, if metabolic syndrome is present, list it as MASLD), thus compromising the integrity of transplant epidemiology. This has been seen in the past for subjective variables, as the decisions in metabolic dysfunction alcohol-associated liver disease could be. In summary, although an ALD diagnosis–based coefficient may appear to reduce the disparities observed in this study, we believe further exploration is needed prior to considering any policy changes.

Abbreviations:

AC

acuity circles

AH

alcohol-associated hepatitis

ALD

alcohol-associated liver disease

LT

liver transplantation

MASLD

metabolic dysfunction–associated liver disease

MELD

Model for End-Stage Liver Disease

WL

waitlist