Rising Trend in Waitlisting for Alcoholic Hepatitis With More Favorable Outcomes Than Other High Model for End-stage Liver Disease in the Current Era : Transplantation

Journal Logo

Original Clinical Science—Liver

Rising Trend in Waitlisting for Alcoholic Hepatitis With More Favorable Outcomes Than Other High Model for End-stage Liver Disease in the Current Era

Bittermann, Therese MD, MSCE1,2; Mahmud, Nadim MD1,2; Weinberg, Ethan M. MD1; Reddy, K. Rajender MD1

Author Information
doi: 10.1097/TP.0000000000004049
  • Free
  • SDC
  • Infographic



In the last decade, 2 important developments have markedly altered the landscape of liver transplantation (LT) in the United States. First, the widespread approval of highly effective direct-acting antiviral (DAA) therapies has decreased new LT waitlistings for decompensated cirrhosis because of hepatitis C virus (HCV).1 Concurrently, the number of LT waitlist entrants with alcohol-related liver disease (ALD) has markedly increased, with ALD now being the leading indication for LT in the United States.2,3 The increased recognition of early LT as a viable treatment option for severe alcoholic hepatitis (AH) may have yielded a broader acceptance of LT in ALD candidates overall.4,5

Despite a clearer understanding of which AH patients are most appropriate for LT listing, they frequently remain controversial.4 This partly stems from the need to rely predominantly on the psychosocial assessment with limited opportunity for longitudinal follow-up pre-LT, and ongoing challenges in predicting which candidates will ultimately improve with medical management alone.6 Severe AH is also associated with significant liver dysfunction and renal failure, leading to a high Model for End-stage Liver Disease (MELD) score and prompt access to LT once listed. Nevertheless, early LT is believed to offer a survival benefit over delaying LT to achieve a period of sobriety, particularly among those with high MELD score.7

Patients with AH on the LT waitlist compete for organ access against other candidates with marked severity of illness because of decompensated liver disease from other causes. In the pre-DAA era, these high MELD candidates primarily had cirrhosis because of HCV.8 Although AH candidates have reduced waitlist mortality compared with other liver diseases, this is in the context of broad comparator groups that span a range of MELD scores.9 In this study, we explore the potential effects of the recent expansion of early LT for AH and specifically compare the waitlist outcomes of AH candidates with those of other high MELD candidates in the United States.


Study Population, Exposures, And Outcomes

This was a retrospective cohort study using the United Network for Organ Sharing (UNOS) database. Adults waitlisted for first LT alone or simultaneous liver-kidney (SLK) transplantation between January 1, 2008, and June 12, 2020, were included. Multiorgan and prior transplant recipients and emergent LT (eg, acute liver failure) listings were excluded.

Patients listed for AH were identified using the UNOS diagnosis code 4217. The primary comparator population included all other patients with listing native MELD score ≥30, given comparable risk of waitlist mortality. Although we did not specifically exclude patients with exception points, the study cohort had very few (0.4%). Secondary comparators included: non-AH patients listed with other ranges of native MELD (≥25, 20–29, 20–25), patients with HCV and listing MELD score ≥30 given the recent impact of HCV therapies, and patients without AH waitlisted with acute-on-chronic liver failure grade 3 (ACLF-3) given comparable medical complexity and resource utilization.10

ACLF-3 was defined in accordance with the European Association for the Study of the Liver criteria11 and involved an adapted classification of 6 potential organ failures: renal (creatinine ≥2 mg/dL or need for renal replacement therapy), brain (grade 3–and 4 hepatic encephalopathy), liver (bilirubin ≥12 mg/dL), coagulation (international normalized ratio ≥2.5), circulation (need for vasopressor therapy), and respiratory (need for mechanical ventilation). Patients with ≥3 organ failures were categorized as ACLF-3, consistent with prior studies using national registry data.12,13

As previously demonstrated, the sensitivity of the UNOS diagnosis code for AH is only 35%, leading to gross underestimation of the potential waitlisted population with AH in the United States.14 Therefore, for the purpose of a sensitivity analysis of our main findings (see Statistical Analyses), we developed an alternate algorithm to identify “definite or probable AH.” This included patients with definite AH (through UNOS coding), and any candidate ≤50 y old with ALD and a native listing MELD score ≥30. These criteria were developed from the reported characteristics of patients missed by the UNOS code in the American Consortium of Early Liver Transplantation for AH multicenter cohort (median age of 43 y [interquartile range, IQR, 37–50] and median MELD at LT of 39 [IQR, 35–40]).14 Additionally, several seminal clinical trials in AH have indicated that the typical age of AH patients is between 40 and 50 y.4,15,16 To assess the validity of algorithm, we performed a chart review of our center’s candidates during the study period with these criteria but lacking the UNOS AH code (N = 22). The “definite or probable AH” criteria by the American Association for the Study of Liver Diseases were used to adjudicate AH cases.17 The positive predictive value (PPV) of this expanded AH algorithm at our center was 86.4%. To date, there remains no consistent definition of AH in clinical trials, nor is there a UNOS definition of AH or any other validated algorithm to identify AH patients in administrative databases.4,18

Exposures included age, sex, race/ethnicity, liver disease cause, native MELD score and its components, ascites severity, encephalopathy grade, need for dialysis support, need for mechanical ventilation, and Karnofsky Performance Status. Socioeconomic variables included highest education level and insurance type. Primary waitlist outcomes included the waitlist endpoints of LT and removal because of death or being too sick. Secondary outcomes included delisting because of clinical improvement, post-LT mortality, and allograft failure.

Statistical Analyses

Characteristics of AH patients were compared with non-AH candidates with MELD ≥30 at listing. The other MELD ≥30 group was further subdivided into patients with non-AH ALD and non-ALD. Descriptive comparisons used chi-square tests for categorical variables and Kruskal-Wallis tests for continuous variables. Differences between patients with definite AH by UNOS coding and those with probable AH using our algorithm (ie, ALD with listing native MELD ≥30 and age ≤50 y in the absence of a UNOS AH diagnosis code) were also assessed.

Competing risks (Fine and Gray) models evaluated AH as an independent predictor waitlist mortality and LT. When ascertaining waitlist mortality, LT was defined as a competing risk, whereas removals because of clinical improvement or other reasons were defined as censoring events.19 In the model evaluating LT, waitlist mortality was set as a competing risk, whereas other waitlist removal reasons were censoring events. The primary analysis compared the waitlist outcomes of patients with AH to those listed for other indications with native MELD ≥30. The waitlist outcomes of AH were then evaluated relative to the secondary comparators listed in Study Population, Exposures, and Outcomes.

We performed several sensitivity analyses to ascertain the robustness of these findings. For the primary analysis (ie, AH versus other MELD ≥30), additional multivariable models for each outcome were developed that separately: (1) excluded patients with SLK waitlisting, and (2) used our expanded “definite or probable AH” algorithm to identify AH patients. Additionally, given the potential overlap in listing MELD between the AH group and the secondary comparator groups with listing native MELD 20–29 and 20–25, we repeated waitlist mortality analyses, restricting the AH group to those with listing native MELD ≥30.

Differences in waitlist removal because of clinical improvement were described between AH and high MELD non-AH groups. In a subsequent analytic phase, the post-LT patient and graft survival of patients with definite AH were compared with those with listing native MELD ≥30. Adjusted Cox proportional hazards models were developed and reported with patient and graft survival curves.

All multivariable models evaluating waitlist outcomes were adjusted for age, sex, race/ethnicity, native MELD score at listing, ascites severity at listing, encephalopathy grade at listing, dialysis at listing, mechanical ventilation at listing, Karnofsky Performance Status at listing, education level, insurance type at listing, and UNOS region. Multivariable models evaluating posttransplant outcomes were adjusted for these covariates, but captured at waitlist removal, as well as patient location at LT (unavailable at listing), donor age, and receipt of a donation after circulatory death allograft. To focus the analyses on the most current era and acknowledge the timing of the rapid expansion of LT listings for AH, all multivariable models were restricted to patients listed since 2014.

This study was approved by the Institutional Review Board of the University of Pennsylvania. All analyses were performed using STATA version 16 (College Station, TX).


Study Population and Baseline Characteristics

Between January 1, 2008, and June 12, 2020, 719 (0.6%) patients with definite AH were identified among 120 708 adults waitlisted for first LT alone or SLK in the United States. Concurrently, 15 472 (12.8%) were listed with native MELD ≥30, of whom 38.4% had ALD without reported AH, 20.5% hepatitis C, 20.0% nonalcoholic steatohepatitis, 9.4% autoimmune liver disease, 3.3% hepatitis B virus, and 8.3% other.

Table 1 compares the characteristics of AH patients with high MELD non-AH patients with and without ALD. Of the 3 groups, AH patients were the youngest (median age 43 versus 51 y for ALD MELD ≥30 and 56 y for non-ALD MELD ≥30; P < 0.001), the least racially and ethnically diverse (79.6% White versus 72.6% and 61.3%, respectively; P < 0.001), and the most educated (35.9% college or more versus 25.0% and 19.9%, respectively; P < 0.001). Although AH patients had high rates of private insurance coverage (65% versus 55.9% for ALD MELD ≥30 and 51.2% for non-ALD MELD ≥30), they also were frequently Medicaid-insured (24.2% versus 25.7% and 20.6%, respectively; P < 0.001 in overall pairwise comparisons). With regards to liver disease severity, MELD score was highest among AH candidates (P < 0.001), a difference that was entirely driven by their higher serum bilirubin (median 23.2 versus 15.1 mg/dL for ALD MELD ≥30 and 17.5 mg/dL for non-ALD MELD ≥30; P < 0.001) given other MELD components were more favorable. AH candidates also had lower rates of moderate-severe ascites (P < 0.001), grade 3–4 encephalopathy (P < 0.001), dialysis at listing (P < 0.001), and mechanical ventilation (P = 0.032). Rates of ACLF-3 (vs all other ACLF grades) were not different among the 3 groups (P = 0.264).

TABLE 1. - Characteristics at waitlisting of patients listed with AH, non-AH ALD with native MELD 30+, and non-ALD MELD 30+ between 2008 and 2020 (N = 16 191)
AH (n = 719) Non-AH ALD MELD 30+ (n = 5941) Non-ALD MELD 30+ (n = 9531) P
Male, N (%) 466 (64.8) 4150 (69.9) 5361 (56.3) <0.001
Age, median (IQR), y 43 (35–52) 51 (44–58) 56 (49–62) <0.001
Race/ethnicity, N (%) <0.001
 White 571 (79.6) 4311 (72.6) 5843 (61.3)
 Black 29 (4.0) 334 (5.6) 1269 (13.3)
 Hispanic 77 (10.7) 1069 (18.0) 1738 (18.2)
 Asian 22 (3.1) 108 (1.8) 513 (5.4)
 Other 19 (2.6) 119 (2.0) 168 (1.8)
Ascites, N (%) <0.001
 None 101 (14.1) 329 (5.5) 919 (9.7)
 Mild 316 (43.9) 2190 (36.9) 3708 (38.9)
 Moderate-severe 302 (42.0) 3420 (57.6) 4900 (51.4)
Encephalopathy, N (%) 190 (26.4) 965 (16.3) 2061 (21.6) <0.001
 None 394 (54.8) 3607 (60.7) 5510 (57.8)
 Grade 1–2 135 (18.8) 1367 (23.0) 1956 (20.5)
 Grade 3–4
Creatinine, median (IQR), mg/dL 2.0 (1.0–3.6) 2.1 (1.2–3.4) 2.2 (1.4–3.4) <0.001
Bilirubin, median (IQR), mg/dL 23.2 (10.6–33.3) 15.1 (8–26.1) 17.5 (8–29.5) <0.001
INR, median (IQR) 2.0 (1.7–2.5) 2.5 (2.0–3.1) 2.5 (2.0–3.1) <0.001
Sodium, median (IQR), mEq/L 136 (132–139) 134 (130–138) 135 (131–139) <0.001
Laboratory MELD, median (IQR) 36 (29–40) 35 (32–39) 35 (32–40) 0.003
Dialysis, N (%) 210 (29.2) 2032 (34.2) 2999 (31.5) <0.001
Mechanical ventilation, N (%) 55 (7.7) 526 (8.9) 938 (9.8) 0.032
ACLF grade, N (%) <0.001
 0 149 (20.7) 222 (3.7) 349 (3.7)
 1 95 (13.2) 1516 (25.5) 2388 (25.1)
 2 273 (38.0) 2457 (41.4) 3904 (41.0)
 3 202 (28.1) 1746 (29.4) 2890 (30.3)
KPS, N (%) <0.001
 80%–100% 60 (8.4) 290 (4.9) 525 (5.5)
 50%–70% 163 (22.7) 1321 (22.2) 2016 (21.2)
 10%–40% 490 (68.2) 4213 (70.9) 6666 (69.9)
 Missing 6 (0.8) 117 (2.0) 324 (3.4)
Education level, N (%) <0.001
 HS or less 229 (31.9) 2484 (41.8) 4597 (48.2)
 Less than college 191 (26.6) 1482 (25.0) 2101 (22.0)
 College or more 258 (35.9) 1485 (25.0) 1893 (19.9)
 Missing 41 (5.7) 490 (8.3) 940 (9.9)
Insurance, N (%) <0.001
 Private 467 (65.0) 3319 (55.9) 4833 (51.2)
 Medicaid 174 (24.2) 1525 (25.7) 1942 (20.6)
 Medicare 51 (7.1) 835 (14.1) 2140 (22.7)
 Other 27 (3.8) 261 (4.4) 526 (5.6)
Missingness other than those reported as individual strata was <1%.
ACLF, acute-on-chronic liver failure; AH, alcoholic hepatitis; ALD, alcohol-related liver disease; HS, high school; INR, international normalized ratio; IQR, interquartile range; KPS, Karnofsky Performance Status; MELD, Model for End-stage Liver Disease.

Table S1 (SDC, https://links.lww.com/TP/C351) highlights the clinical characteristics of the AH group identified by UNOS coding (ie, “definite AH”) versus those captured through the “probable AH” algorithm. Notable differences among those meeting the probable AH criteria included: greater racial/ethnic diversity (P < 0.001), higher rates of high school or less education (P < 0.001), and higher rates of Medicaid-insured (P < 0.001). Severity of illness was modestly greater in the probable than the definite AH group (Table S1, SDC, https://links.lww.com/TP/C351).

Temporal and Geographic Trends in Waitlist Additions

Between 2008 and 2019, the proportion of new waitlist additions with AH increased 6.5-fold, with a more rapid increase since 2014 (Figure 1A). In 2019, AH accounted for 7.2% of all waitlist additions with listing MELD ≥30, whereas non-AH ALD accounted for 50.0% (Figure 1B). The prevalence of AH and non-AH ALD candidates among listings with ALCF-3 in 2019 was 8.6% and 58.8%, respectively (Figure 1C). The increased listing rate for AH has coincided with the decline of HCV MELD ≥30 in the setting of DAAs, with AH surpassing high MELD HCV as a listing indication in 2019 (Figure 1D).

A, Rate of waitlistings for AH, non-AH ALD with MELD ≥30, and non-ALD MELD ≥30 between 2008 and 2019. B, Percent of listings with native MELD ≥30 accounted for by AH and other ALD candidates. C, Trends in liver disease cause among patients with ACLF-3 between 2008 and 2019. D, Rate of waitlistings for AH as compared to HCV with MELD ≥30 and other ACLF-3 between 2008 and 2019. ACLF-3, acute-on-chronic liver failure grade 3; AH, alcoholic hepatitis; AI, autoimmune; ALD, alcohol-related liver disease; HBV, hepatitis B virus; HCV, hepatitis C virus; MELD, Model for End-stage Liver Disease; NASH, nonalcoholic steatohepatitis.

Of the 139 transplant centers included, 40 (28.8%) had no listings AH during the study period. Median overall listing volume during the study period was lower at these 40 centers (66 versus 966 at the remaining 99 centers; P < 0.001), as was their proportion of waitlist entrants with ALD (22.0% versus 24.3%; P < 0.001). Among the 99 centers with ≥1 AH listing, the prevalence of AH ranged from 0.2% to 18.2% among ALD candidates and 0.6% to 25.0% among those with native listing MELD ≥30.

Waitlist Outcomes

Waiting time was short for AH and non-AH patients with MELD ≥30: median 10 (IQR, 4–36) and 8 (IQR, 4–19) d, respectively (P < 0.001). Waiting time was significantly correlated with waitlist outcome (P < 0.001 for both groups). For example, median waiting time for AH candidates transplanted was 7 (IQR, 3–16) d, whereas this was 17 (IQR, 7–49) d for those who died or were too sick and 289 (IQR, 142–591) d for those delisted because of clinical improvement. This trend was also noted among non-AH MELD ≥30 candidates, though waiting times before removal because of death/deterioration or clinical improvement were shorter (median 12 and 141 d, respectively). MELD at waitlist removal was comparable between AH and non-AH MELD ≥30 candidates (median 37 versus 37, respectively; P < 0.001). When stratified by waitlist outcome, no difference in ΔMELD from listing to waitlist removal was observed (P > 0.05 for each outcome), suggesting no difference in patients’ evolution of disease while listed.

Figure 2 highlights waitlist outcome rates by high MELD group and era. In 2018–2020, waitlist removals for LT increased by approximately 10% among AH candidates compared with the prior decade. LT has also become more frequent for all high MELD ALD candidates compared with non-ALD candidates. Waitlist mortality was consistently lower in the AH group. A pronounced decline in waitlist mortality recently occurred for AH patients (16.3% in 2014–2017 to 6.3% in 2018–2020), whereas more subtle decreases occurred in other ALD MELD ≥30 patients (17.9%–13.7%, respectively) and non-ALD high MELD candidates (22.4%–20.8%, respectively). Overall rates of delisting because of clinical improvement were low, though highest in the AH group: 3.3% versus 0.5% for other ALD MELD ≥30 and 1.0% for non-ALD MELD ≥30 (P < 0.001).

Temporal trends in transplantation and waitlist mortality rates by group and waitlist era. P < 0.001 for all within diagnosis temporal trends. AH, alcoholic hepatitis; ALD, alcohol-related liver disease; MELD, Model for End-stage Liver Disease.

In the current era, AH is associated with reduced waitlist mortality (adjusted subhazard ratio [SHR], 0.67; P = 0.002) and increased LT rate (adjusted SHR, 1.14; P = 0.006) compared with other high MELD listing indications (Table 2; Figure 3A and B). Excluding patients waitlisted for SLK, the results were unchanged: adjusted SHR 0.66 (95% confidence interval [CI], 0.50-0.87; P = 0.003) for waitlist mortality and adjusted SHR 1.11 (95% CI, 1.01-1.22; P = 0.032) for LT. The expanded AH algorithm also revealed similar findings: waitlist mortality adjusted SHR 0.64 (95% CI, 0.55-0.74; P < 0.001) and LT rate adjusted SHR 1.17 (95% CI, 1.10-1.24; P < 0.001). There was no difference in waitlist outcomes between patients in the definite versus the probable AH groups (P = 0.309 for waitlist mortality and P = 0.265 for LT). AH candidates also experienced lower waitlist mortality and increased access to LT compared with patients with lower MELD scores, HCV candidates with MELD ≥30, and patients listed with ACLF-3 (see Table 2 for all other adjusted HRs). The reduced adjusted waitlist mortality observed for AH compared with patients with MELD 20–29 and 20–25 persisted despite restricting the AH group to those with listing MELD ≥30: adjusted SHRs of 0.48 (95% CI, 0.34-0.68; P < 0.001) and 0.56 (95% CI, 0.23-0.94; P = 0.018), respectively.

TABLE 2. - Results of multivariable models comparing adjusted liver transplantation rate and waitlist mortality between AH and other comparator groups since 2014
Waitlist mortality aSHR, 95% CI P Transplantation aSHR, 95% CI P
Other MELD 30+
(N = 10 344)
0.67 (0.52-0.86)
0.002 Reference
1.14 (1.04-1.25)
Other MELD 25+
(N = 16 297)
0.62 (0.48-0.80)
<0.001 Reference
1.21 (1.09-1.35)
Other MELD 20-29
(N = 16 482)
0.49 (0.36-0.66)
<0.001 Reference
1.39 (1.28-1.63)
Other MELD 20-25
(N = 11 898)
0.53 (0.37-0.77)
0.001 Reference
1.29 (1.07-1.54)
(N = 1999)
0.59 (0.42-0.83)
0.002 Reference
1.26 (1.09-1.45)
Other ACLF-3
(N = 3490)
0.57 (0.43-0.76)
<0.001 Reference
1.18 (1.05-1.32)
ACLF-3, acute-on-chronic liver failure grade 3; AH, alcoholic hepatitis; aSHR, adjusted subhazard ratio; CI, confidence interval; HCV, hepatitis C virus; MELD, Model for End-stage Liver Disease.

A, Adjusted cumulative incidence of WL mortality between 2014 and 2020 for AH vs other MELD ≥30 (N = 10 344). B, Adjusted cumulative incidence of LT between 2014 and 2020 for AH vs other MELD ≥30 (N = 10 344). AH, alcoholic hepatitis; LT, liver transplantation; MELD, Model for End-stage Liver Disease; WL, waitlist.

Post-LT Outcomes

Compared with non-AH patients with MELD ≥30, those transplanted for AH had improved post-LT outcomes after adjusting for differences in severity of illness at LT and donor quality (Figure 4). The adjusted HR for AH for post-LT mortality was 0.68 (0.50-0.91; P = 0.011) and for graft failure was 0.74 (95% CI, 0.57-0.96; P = 0.025). Notably, there was no difference in donor age (median 39 y for AH versus 38 y for MELD ≥30 recipients; P = 0.242) or in the use of donation after circulatory death organs (3.1% versus 2.1%, respectively; P = 0.073) between the 2 groups.

Adjusted patient (A) and graft (B) survival for AH and other MELD ≥30 candidates between 2008 and 2020 (N = 11 376). AH, alcoholic hepatitis; LT, liver transplantation; MELD, Model for End-stage Liver Disease.


The option of LT as an effective treatment modality for severe AH has been a subject of controversy for over 30 y.20,21 A carefully designed prospective study by Mathurin et al4 in 2011 demonstrated good outcomes for patients undergoing LT with severe AH, generating a renewed interest and confidence in the role of LT for severe AH among transplant centers. In this study, we demonstrate that AH is a rapidly growing indication for LT listing in the DAA era, though considerable heterogeneity exists among transplant centers. We further observe that patients with AH experience more favorable waitlist outcomes than other high MELD candidates, and even reduced waitlist mortality compared with candidates at lower MELD scores. This issue of AH candidates having a potential waitlist advantage over other patients at high risk of waitlist mortality warrants further consideration on a policy level in light of the temporal trends presented and also given the new liver allocation policy adopted in 2020.

We surmise that there are several potential explanations underlying the superior waitlist outcomes of AH candidates. From a liver perspective, they are frequently listed with higher MELD scores than others with MELD ≥30. This increased MELD score is entirely driven by changes in bilirubin, and other markers of liver disease severity are less prevalent, such as severe encephalopathy, ascites, or the need for dialysis. AH candidates are also younger and frequently have fewer, if any, non liver comorbidities. Their shorter duration of prelisting disease may have also led to fewer consequences, such as less frailty, sarcopenia, or malnutrition. These factors combined, we suspect that AH patients have increased resilience to stressors and other clinical perturbations while listed. This is evidenced by the fact that AH patients are waitlisted longer than MELD ≥30 patients before being removed because of clinical deterioration. Centers may also be more willing to pursue LT despite acute changes in their clinical status for these same reasons and given their acceptable post-LT outcomes. Although we did not find differences in allograft quality, future studies should evaluate for differences in the frequency and reasons for organ refusals between high MELD groups to understand the impact of center behaviors in facilitating LT for AH.

Several mechanisms could address the potential waitlist advantage of AH candidates. For example, a MELD ceiling or MELD correction factor could be applied, the magnitude of which might differ on the basis of coexistent renal failure.22 Notably, the concept of a MELD correction factor is not new: liver disease cause was factored into the original MELD score equation and several studies have demonstrated differences in MELD performance for different liver diseases and clinical scenarios.23-25 As a related issue, a recent European publication indicated that nearly 9% of waitlisted patients with ALD could be delisted because of clinical improvement.6 An increased rate among women was believed to be explained by the known sex disparities in access to LT, permitting a longer period of observation to allow for recovery. In our study, only 3.3% of AH patients were delisted because of clinical improvement and overall time from listing to delisting was extremely short (median 10 d). Although AH patients are likely afforded a period of observation prelisting, once listed, they may be committed to the pathway of LT and few are allowed the opportunity to demonstrate further recovery. A formal mechanism that takes into consideration longitudinal changes in MELD score and clinical status as justification to maintain AH candidates actively waitlisted could serve as an opportunity to reduce unnecessary transplants in a larger subset.

Although the proportion of AH candidates may seem small overall, this group represented 3.9% of all ALD listings and 5.7% of all listings with MELD ≥30 during the first half of 2020. Given the low sensitivity (35%) of the UNOS diagnosis codes, AH may actually represent >10% and >15% of such candidates, respectively. Moreover, AH is a precipitant of ACLF in patients with ALD and is also one of the more frequent indications for hospitalization pre-LT.26,27 This highlights the likely important role of AH in the recent marked rise in LT waitlisting for ALD, particularly among young and high MELD recipients.3 As part of this research, we developed an algorithm to identify patients with AH who were missed by UNOS coding that demonstrated high PPV (86.4%) at our center. The relationship between AH listing practices and overall LT experience that we observed may be explained by differences in centers’ willingness to consider potentially high-risk candidates, resource availability for pre- and post-LT care (eg, transplant psychiatry), and centers’ familiarity with appropriate UNOS coding, as suggested in the study by Lee et al.14 The trends and challenges that we present here have recently gained interest at the administrative level and plans are recently underway for the Organ Procurement and Transplant Network Liver and Intestinal Committee to improve the characterization of patients with ALD on the LT waitlist and better distinguish those with AH.28 Future studies are also needed to develop validated methods that enhance the identification of AH patients in databases more broadly.

In the last year, consensus recommendations in the United States and internationally have been developed to help ensure consistency in the selection of AH candidates.29-31 Assessing centers’ evolving adoption of this guidance will remain a challenge until an oversight mechanism is developed. As an example, we found that nearly 20% of AH candidates had grade 3–4 encephalopathy at listing—an issue that, if also present during the evaluation process, would prevent an accurate psychosocial assessment and be counter to recommended criteria.29 Given the unique aspects of LT for AH, including their rapid access to transplant once listed, holding centers accountable to minimum performance standards is also warranted. Additionally, our data also show that LT listing for AH has thus far most frequently represented an opportunity for White, well educated, and privately insured individuals. Moving forward, deliberate and systematic efforts will be needed to make certain that access to LT for patients with AH in the United States is fair and equitable, particularly given the rapidly growing interest in this area.

Our study had several limitations. Although we used the UNOS AH diagnosis code to identify patients, there are no predefined UNOS AH criteria. Given the low reported sensitivity of the UNOS diagnosis code,14 a subset are likely included in the MELD ≥30 comparator. Thus, our findings may actually be conservative. Using our expanded AH definition, we did begin to observe this trend. We also suspect demographic, temporal, and center differences exist in the sensitivity and PPV of UNOS coding for AH, though this has yet to be studied. For example, greater racial/ethnic and socioeconomic diversity was noted in probable AH patients with our algorithm. Moreover, there possibly worse performance of the algorithm among women (PPV 72.7% versus 100% in men; P = 0.107). The UNOS database currently does not contain any information related to duration of alcohol abstinence, though this may change.28 AH candidates are highly selected psychosocially and medically with many factors that are not captured in the UNOS database. We attempted to adjust for a diverse set of covariates available to help remedy this. Finally, should centers expand LT to higher risk AH candidates, their waitlist advantage and superior post-LT outcomes may not persist over time.

In conclusion, we demonstrate that AH is a rapidly expanding indication for LT among high MELD candidates, effectively filling the space left by the decrease in transplants for high MELD HCV patients. AH is associated with more favorable waitlist outcomes than other high MELD indications with lower rates of waitlist mortality and higher transplant rates, despite accounting for clinical differences. AH-specific allocation policies and center performance strategies are warranted given the increasing interest in this field to ensure that other high MELD candidates are not disadvantaged. Upcoming efforts to better characterize candidates with AH and others with ALD in UNOS will allow for more effective research in this area on a broad scale. Future studies will also need to evaluate the potential impact of the new liver allocation policy on the differential access to LT among high MELD candidates.


1. Bittermann T, Reddy KR. In the era of direct-acting antivirals, liver transplant delisting due to clinical improvement for hepatitis C remains infrequent. Clin Gastroenterol Hepatol. 2021;19:2389–2397.e2.
2. Cholankeril G, Ahmed A. Alcoholic liver disease replaces hepatitis C virus infection as the leading indication for liver transplantation in the United States. Clin Gastroenterol Hepatol. 2018;16:1356–1358.
3. Cholankeril G, Gadiparthi C, Yoo ER, et al. Temporal trends associated with the rise in alcoholic liver disease-related liver transplantation in the United States. Transplantation. 2019;103:131–139.
4. Mathurin P, Moreno C, Samuel D, et al. Early liver transplantation for severe alcoholic hepatitis. N Engl J Med. 2011;365:1790–1800.
5. Lee BP, Vittinghoff E, Dodge JL, et al. National trends and long-term outcomes of liver transplant for alcohol-associated liver disease in the United States. JAMA Intern Med. 2019;179:340–348.
6. Pose E, Torrents A, Reverter E, et al. A notable proportion of liver transplant candidates with alcohol-related cirrhosis can be delisted because of clinical improvement. J Hepatol. 2021;75:275–283.
7. Lee BP, Samur S, Dalgic OO, et al. Model to calculate harms and benefits of early vs delayed liver transplantation for patients with alcohol-associated hepatitis. Gastroenterology. 2019;157:472–480.e5.
8. Murken DR, Peng AW, Aufhauser DD Jr, et al. Same policy, different impact: center-level effects of share 35 liver allocation. Liver Transpl. 2017;23:741–750.
9. Kitajima T, Moonka D, Yeddula S, et al. Liver transplant waitlist outcomes in alcoholic hepatitis compared with other liver diseases: an analysis of UNOS registry. Clin Transplant. 2020;34:e13837.
10. Singal AK, Arora S, Wong RJ, et al. Increasing burden of acute-on-chronic liver failure among alcohol-associated liver disease in the young population in the United States. Am J Gastroenterol. 2020;115:88–95.
11. Moreau R, Jalan R, Gines P, et al.; CANONIC Study Investigators of the EASL–CLIF Consortium. Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis. Gastroenterology. 2013;144:1426–1437.
12. Sundaram V, Jalan R, Shah P, et al. Acute on chronic liver failure from nonalcoholic fatty liver disease: a growing and aging cohort with rising mortality. Hepatology. 2021;73:1932–1944.
13. Sundaram V, Shah P, Mahmud N, et al. Patients with severe acute-on-chronic liver failure are disadvantaged by model for end-stage liver disease-based organ allocation policy. Aliment Pharmacol Ther. 2020;52:1204–1213.
14. Lee BP, Im GY, Rice JP, et al. Underestimation of liver transplantation for alcoholic hepatitis in the national transplant database. Liver Transpl. 2019;25:706–711.
15. Im GY, Kim-Schluger L, Shenoy A, et al. Early liver transplantation for severe alcoholic hepatitis in the United States—a single-center experience. Am J Transplant. 2016;16:841–849.
16. Thursz MR, Richardson P, Allison M, et al.; STOPAH Trial. Prednisolone or pentoxifylline for alcoholic hepatitis. N Engl J Med. 2015;372:1619–1628.
17. Crabb DW, Im GY, Szabo G, et al. Diagnosis and treatment of alcohol-associated liver diseases: 2019 practice guidance from the American Association for the Study of Liver Diseases. Hepatology. 2020;71:306–333.
18. Louvet A, Labreuche J, Artru F, et al. Combining data from liver disease scoring systems better predicts outcomes of patients with alcoholic hepatitis. Gastroenterology. 2015;149:398–406.e8.
19. Ebel NH, Hsu EK, Berry K, et al. Disparities in waitlist and posttransplantation outcomes in liver transplant registrants and recipients aged 18 to 24 years: analysis of the UNOS database. Transplantation. 2017;101:1616–1627.
20. Starzl TE, Van Thiel D, Tzakis AG, et al. Orthotopic liver transplantation for alcoholic cirrhosis. JAMA. 1988;260:2542–2544.
21. Kumar S, Stauber RE, Gavaler JS, et al. Orthotopic liver transplantation for alcoholic liver disease. Hepatology. 1990;11:159–164.
22. Potts JR, Goubet S, Heneghan MA, et al. Determinants of long-term outcome in severe alcoholic hepatitis. Aliment Pharmacol Ther. 2013;38:584–595.
23. Malinchoc M, Kamath PS, Gordon FD, et al. A model to predict poor survival in patients undergoing transjugular intrahepatic portosystemic shunts. Hepatology. 2000;31:864–871.
24. Godfrey EL, Malik TH, Lai JC, et al. The decreasing predictive power of MELD in an era of changing etiology of liver disease. Am J Transplant. 2019;19:3299–3307.
25. Hernaez R, Liu Y, Kramer JR, et al. Model for end-stage liver disease-sodium underestimates 90-day mortality risk in patients with acute-on-chronic liver failure. J Hepatol. 2020;73:1425–1433.
26. Karvellas CJ, Francoz C, Weiss E. Liver transplantation in acute-on-chronic liver failure. Transplantation. 2021;105:1471–1481.
27. Schaubel DE, Wei G, Dykstra DM, et al. Hospitalization patterns before and after liver transplantation. Transplantation. 2007;84:1590–1594.
28. Organ Procurement and Transplantation Network. OPTN liver and intestinal organ transplantation committee—meeting summary (June 4, 2021). 2021. Available at https://optn.transplant.hrsa.gov/media/4709/20210604_liver_meeting_summary.pdf. Accessed July 23, 2021.
29. Asrani SK, Trotter J, Lake J, et al. Meeting report: the Dallas Consensus Conference on liver transplantation for alcohol associated hepatitis. Liver Transpl. 2020;26:127–140.
30. Greenberg R, Goldberg A, Anthony S, et al. Canadian Society of Transplantation white paper: ethical and legal considerations for alcohol and cannabis use in solid organ listing and allocation. Transplantation. 2021;105:1957–1964.
31. Rodríguez-Perálvarez M, Gómez-Bravo MÁ, Sánchez-Antolín G, et al.; Spanish Society of Liver Transplantation (SETH) Consensus Panel. Expanding indications of liver transplantation in Spain: consensus statement and recommendations by the Spanish Society of Liver Transplantation. Transplantation. 2021;105:602–607.

Supplemental Digital Content

Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.