Approaching Alcohol Use Disorder after Liver Transplantation for Acute Alcoholic Hepatitis

Introduction

Severe alcoholic hepatitis (AH) is an acute manifestation of alcohol-related liver disease (ALD) with distinct histopathologic findings of steatohepatitis, pronounced cholestasis, and sclerosing hyaline necrosis; the latter two features help distinguish severe AH from nonalcoholic steatohepatitis.¹ Severe AH carries a grim prognosis even with expert medical management.² Traditional risk factors for severe AH include female sex, younger age, obesity, and presence of other liver disease such as hemochromatosis or hepatitis C.³ On a genetic level, predisposition to severe AH involves a complex interplay between genes regulating inflammatory mediators via the commonly cited tumor necrosis factor pathway, genes regulating endotoxin response, genes involved in the balance of oxidative stress, genes involved in hepatic lipid metabolism such as PNPLA3, and epigenetic changes.⁴ Major environmental risk factors include the quantity and pattern of alcohol consumption that may affect the CYP2E1 pathway that metabolizes alcohol to acetaldehyde, gut microbiota alterations leading to gut derived pathogen-associated molecular patterns and subsequent cytokine/chemokine release by Kupffer cells, and poor nutritional status.⁴

Traditionally, severe AH, defined as those with Maddrey discriminant function of ≥32, is associated with a 28-day case-fatality ratio of 50%.^5,6 Despite improvements in supportive care since the turn of the century, at least one-quarter of patients with severe AH still die within 90 days without liver transplantation (LT).⁷ The 28-day case-fatality ratio in corticosteroid non-responders was 47% in a meta-analysis that defined non-responders as patients who had a day 7 Lille score of ≥0.56.⁸

Early LT for AH is defined as LT performed without a 6-month period of pre-LT alcohol abstinence.⁹ The combination of high mortality and lack of durably effective pharmaceuticals led major society guidelines to recommend LT evaluation for patients with severe AH who do not respond to medical therapy.^10,11 Notwithstanding its survival benefit for patients with severe AH, early LT remains controversial among healthcare providers.¹² Meanwhile, alcohol use disorder (AUD) has grown over two decades to a prevalence of 5.8% (14.4 million) in the U.S. adult population in 2018, as reported by the National Survey on Drug Use and Health.¹³ Early market data also showed more alcohol purchases nationwide after the coronavirus pandemic caused a plethora of socioeconomic disruptions.¹⁴ This alcohol epidemic within a pandemic will likely lead to a rise in severe AH.

In the present review, we provide a brief overview of the management of severe AH and outline methods to optimize transplant outcomes in early LT for severe acute AH.

Diagnosis and Pre-transplant Medical Management of Alcoholic Hepatitis

To help standardize definitions in research, the American Association for the Study of Liver Diseases (AASLD) recommends categorizing the clinical syndrome of AH into three groups according to the consensus published in 2016: definite AH (biopsy proven), probable AH, and possible AH (should obtain biopsy).¹⁵ In practice, it is crucial early in the clinical course to distinguish between AH and decompensation of alcohol-related cirrhosis (AC) to guide medical treatment. Prompt administration of corticosteroids and complete abstinence may prevent LT in AH, but has less effect on the disease course of decompensated AC. At this time there are no validated noninvasive tests that can differentiate between AH and decompensated AC. A liver biopsy can help in the event of diagnostic uncertainty.

There are no impressive pharmaceutical agents in the treatment of AH. Nonetheless, key aspects of pre-LT medical management must be emphasized both to limit the number of LT eventually needed for AH – thus preserving limited organ resources – and to optimize the LT candidate for major surgery. First, alcohol abstinence is the only proven treatment that leads to improved long-term outcomes and is also a necessary condition for LT candidacy in AH.¹⁶ Second, corticosteroids are currently the only pharmaceuticals that decrease 28-day mortality, but the mortality benefit dissipates beyond 90 days.⁷ Nonresponse to steroids is a common context within which providers may consider early LT for AH. It is worth noting that steroid therapy has been associated with a higher rate of serious infections, which can complicate LT. Biomarkers such as circulating bacterial DNA to predict infection may be helpful in guiding antibiotic therapy to improve outcomes.¹⁷ Third, a comprehensive screening for infection including blood, urine, and peritoneal cultures at hospital admission is recommended.¹⁸ Finally, a focus on nutrition and ensuring an enteral diet comprising 1–1.5 g protein/kg and 30–40 kcal/kg of body weight per day is a cornerstone of supportive care specifically in patients with encephalopathy.^18,19

Appropriate management of potential LT candidates through the acute phase of severe AH as outlined above will not only optimize patients medically, but it will also maximize the utilization of non-transplant resources before resorting to LT as a life-saving measure.

The Role for Early Liver Transplantation in Alcoholic Hepatitis and Available Clinical Experience

Transplant centers performing early LT in AH report excellent recipient and liver allograft survival, which we summarize in Table 1.^9,20–23 However, many transplant centers in the United States still use the 6-month rule for alcohol abstinence as a prerequisite to LT. The practice is problematic in the context of severe AH, where a significant proportion of patients die within 6 months of diagnosis.²⁴

Table 1:

Cohorts of Early Liver Transplantation for Alcoholic Hepatitis

Author (Year published)	Location (Number of Participating Centers)	Sample size	Proportion of total grafts used for severe AH indication	Survival (months post-LT)	Graft survival (months post-LT)	Alcohol relapse	Definition of alcohol relapse per study	Post-transplant alcohol relapse monitoring
Mathurin et al. (2011)9	France and Belgium (7)	26	2.9%	77% (6), 71% (24)	N/A	12% at 24 months post-LT	Any alcohol use	Informal interviews with the patient (preferably with family present though not required) with a median of 11 visits over 6 months. Any alcohol was considered relapse and the frequency, type and amount of drinking would be recorded.
Im et al. (2016)20 #	United States (1)	9	3%	89% (6)	100% (6)	12.5%	4 or more drinks per day or 1 drink per day in 4 or more consecutive days	Assessment of relapse during outpatient follow-up visits in addition to random alcohol testing.
Weeks et al. (2018)21 #	United States (1)	46	8.8%	98% (6), 97% (12)	95% (6), 93% (12)	28% (any alcohol), 17% (harmful drinking)	Any alcohol use and presence of harmful drinking pattern	Assessed for both any alcohol relapse and presence of harmful drinking, defined as 4 or more days of drinking per week or 5 drinks/day for men or 4 drinks/day for women.
Lee et al. (2018)22 #	United States (12)	147	N/A	94% (12), 84% (36)	N/A	10% at 12 months post-LT, 17% at 36 months post-LT	Sustained alcohol use, defined as a minimum relapse duration of 100 days.	Assessment of alcohol use at every post-LT visit. Seven of 12 centers had routine urinary ethyl glucuronide or blood phosphatidylethanol testing.
Sundaram et al. (2018)23	United States (1)	8	N/A	100% (10)	100% (10)	10% at 10 months post-LT	N/A	Median of 8 office visits at median followup time of 261 days.

AH, alcoholic hepatitis; LT, liver transplantation

All cohorts included only patients with nonresponse to medical management and stipulated pre-LT evaluation by social worker and addiction specialist.

^#Redundancy present in these cohorts in the United States.

In Europe, Mathurin et al. described 26 early LT recipients who had 6-month and 2-year post-LT survivals of 77% and 71%, respectively.⁹ Shortly thereafter, the U.S. began adopting this practice, and Lee et al. described 147 early LT recipients across 12 U.S. transplant centers who had 1-year and 3-year post-LT survivals of 94% and 84%, respectively.²² A mathematical simulation based in part on the aforementioned published U.S. data estimated an average life expectancy of 6.55 years and 1.46 years for patients offered early and delayed LT for severe AH, respectively.²⁵ The increase in life expectancy was most pronounced in recipients with Model for End-Stage Liver Disease (MELD) scores 32 or more and Lille scores of 0.50 to 0.82 prior to LT.

Both the European and U.S. cohorts had strict candidate selection criteria toward early LT, including non-response to medical management, social work evaluation for social support, and specialty assessment for substance use. In the European cohort and most U.S. centers, there was an explicitly stated criterion for a consensus to transplant among all stakeholders on the liver transplant committee. One notable difference between the European and U.S. experiences was the exclusion of patients with severe psychiatric comorbidities by European centers, whereas only two out of 12 U.S. centers explicitly excluded patients with psychiatric comorbidities.²²

Not surprisingly, a measure emphasized on par with mortality in all of the aforementioned studies was post-LT alcohol relapse. The mathematical simulation above calculated a decrease of 7.23 life-years for LT recipients who relapse into sustained alcohol use.²⁵ Mathurin et al. reported three transplant recipients (15%) who returned to any drinking after two years of follow-up, while Lee et al. reported 10% and 17% of sustained alcohol use (defined as a minimum duration of 100 days of alcohol use) at 1-year and 3-years, respectively.^9,22 Despite differences in candidate selection criteria between the studies, the reported incidences of alcohol use were not notably different.

While the past decade of early LT for AH has been encouraging from the patients’ perspective, intense controversy continues to surround the practice. This debate invariably distills down to balancing the prerogative of physicians to save lives (i.e., beneficence) against the other fundamental bioethical principles of justice, utility, and autonomy.^26,27 Ultimately, the patient selection process is paramount in both consolidating trust in the transplant selection process and optimizing post-transplant outcomes to maximize utility of precious graft resources.

Optimizing Transplant Outcomes in Early Liver Transplantation for Severe Alcoholic Hepatitis

Among centers offering early LT, there remain significant regional variations on the practice of early LT.²⁸ A consensus conference convened in Dallas, Texas in April 2019 to establish recommendations for transplant centers.²⁹ The main goals were to 1) avoid LT in patients with AH who may recover with medical management, 2) eliminate non-data-driven barriers to LT (i.e., the 6-month sobriety requirement) in favor of a multidisciplinary evaluation, and 3) avoid creating or worsening health disparities through LT. These recommendations serve as a guide to standardize pre and post-LT practices that ultimately could optimize outcomes for patients with AH.

Pre-Transplant Candidate Selection

An accurate diagnosis and medical management of severe AH, as summarized previously, is the first step in candidate selection. Patients with severe AH and nonresponse to medical management can then be evaluated for early LT. Response to medical management predicts survival and is best assessed by either changes in MELD score alone or the combination of MELD and the Lille Model (a dynamic measure).^29–31

Once the physician identifies a medically appropriate candidate, selection shifts toward minimizing the risk of alcohol relapse post-LT to optimize transplant outcomes. LT recipients who relapse to alcohol use not only develop significant hepatic steatosis and advanced fibrosis on histopathology, but they also have poorer graft and patient survival than those who remain abstinent.^25,32,33

In severe AH, one frequent challenge is conducting an expeditious but accurate risk assessment for post-LT alcohol relapse within a limited window of opportunity for LT. To accomplish the task, the Dallas consensus conference recommended multidisciplinary psychosocial assessments by transplant social workers, addiction specialists, and other mental health professionals as well as a consensus of medical and paramedical staff.²⁹ Existing LT literature supports identifying potential risk factors of adverse post-transplant outcomes. For example, Dew et al. conducted a meta-analysis of 54 studies on solid organ transplantation (including 50 on LT) and found poor social support and family alcohol history associated with post-transplant relapse of alcohol use.³⁴

Numerous studies have proposed prediction models or guides that aim to predict poor psychosocial outcomes after LT, including alcohol relapse.^35,36 We summarize the most pertinent models in Table 2.^21,36–45 Of note, only the sustained alcohol use post-LT (SALT) and Hopkins psychosocial scale (HPSS) were developed specifically in patients who underwent early LT for severe AH; both instruments lack a consistent outcome measure and external validation.^21,37,38 The harmful alcohol use post-LT (HALT) score is a recently developed model that calculates the probability of alcohol relapse instead of providing a risk category.³⁶ The high-risk alcoholism relapse (HRAR) scale, originally developed in a Veterans Affairs (VA) population, was also studied in early LT for severe AH, but it was not predictive of alcohol relapse after LT.^21,38 The alcohol relapse risk assessment (ARRA) was developed in a cohort of 118 LT recipients who also had AUD, but only 38% had ALD as the primary cause of liver disease.⁴¹ The ARRA also has not undergone subsequent validation.

Table 2:

Selected Prediction Models for Alcohol Relapse After Liver Transplantation

Instrument	Scoring	Cohort studied	Performance	Strengths	Limitations/Weaknesses
Sustained alcohol use post-LT (SALT) score37	SALT score <5 predicted abstinence. >10 drinks per day at initial hospitalization (+4 points) multiple prior rehabilitation attempts (+4 points) prior alcohol-related legal issues (+2 points) prior illicit substance abuse (+1 point)	Early liver transplantation for alcoholic hepatitis. Developed in 134 LT recipients.	NPV: 95% PPV: 25% c-statistic of 0.76. c-statistic of 0.73 on internal cross validation.	Cohort is most relevant to early LT for AH Internal validated	Low PPV, which means high SALT score still may need full evaluation. Exclusion based on high SALT score may exclude suitable candidates. Arisen from retrospective data Not externally validated
Hopkins psychosocial scale (HPSS)21,38	HPSS score <0 would predict sustained alcohol relapse.# Calculated according to protective characteristics and at risk characteristics. Protective characteristics: self-admission to hospital, drinks/day pre-abstinence, insight into diagnosis, marital status, abstinence period prior to transplant. At risk characteristics: psychiatric comorbidity, history of other substance abuse, history of failed rehab attempt, family history of alcoholism, employment immediately prior to presentation, legal issues related to alcohol.	Early liver transplantation for alcoholic hepatitis. Developed first in a cohort of 17 LT recipients.38 Re-examined in an expanded cohort of 43 LT recipients including the original 17 patients compared against LT recipients for ALD meeting 6 month sobriety recruited during the same period.21	HPSS <0 had hazard ratio of 3.63 for any alcohol relapse.	Cohort is most relevant to early LT for AH Many variables considered thought to be more easily assessed objectively Combines unique characteristics of the AH population with established risk factors of relapse	Not internally or externally validated Single center Small sample size Complicated calculation
Harmful alcohol use post-LT (HALT) score36	Multivariate regression model that calculates probability of harmful alcohol use after LT using the following variables: age at LT non-alcohol-related criminal history pre-LT abstinence period drinks per day	Developed in 241 LT recipients with ALD, including 59 patients with less than 6 months of sobriety.	Mean c-statistic of 0.74	Cohort includes early LT Internal validated Model output is a probability of relapse	Not externally validated Includes 74 (30.7%) patients with additional cause of liver disease or indication for LT
High-risk alcoholism relapse (HRAR) scale21, 39, 40	HRAR 4 or more predicted readmission within 6 months for alcohol treatment Duration of heavy drinking in years, < 11 years = 0, 11–25 years = 1, > 25 years = 2. Usual daily number of standard drinks, < 9 = 0, 9–17 = 1, > 17 = 2. Number of previous alcoholism inpatient treatments, 0 = 0, 1 = 1, >1 = 2.	Developed in 299 male patients in a VA population admitted to an alcohol treatment unit.39 Re-examined in a prospective cohort of 387 LT recipients for ALD, which found association with harmful drinking.40 Also re-examined in a single center cohort of 46 LT recipients for AH, where HRAR was not predictive.21	HRAR >3: OR = 4.0 (95% CI = 2.4–6.8) for readmission at 6 months. HRAR <4: OR = 2.5 (95% CI = 1.4–4.2). for abstinence at 6 months.	Studied in many different cohorts specifically involving alcohol Easily calculated	Requires 25 years of heavy drinking to classify as highest risk assignment, which would limit utility in AH Has not been shown to be predictive in LT for AH when studied specifically in this population Skewed demographics including 100% male patients, which may not be generalizable
Alcohol relapse risk assessment (ARRA) score41	ARRA score 4 or more predicted any alcohol relapse. Each of these 9 parameters are given 1 point if present, and summated to a total ARRA score: absence of hepatocellular carcinoma (+1), tobacco dependence (+1), continued alcohol use after liver disease diagnosis (+1), low motivation for alcohol treatment (+1), poor stress management skills (+1), no rehabilitation relationship (+1), limited social support (+1), lack of nonmedical behavioral consequences (+1), and continued engagement in social activities with alcohol present (+1)	Developed in 118 LT recipients with history of AUD.	NPV: 92% PPV: 87% c-statistic 0.892 R2=73%	Studied in LT recipients with AUD Classifies patients into 4 tiers of ARRA (I, II, III, and IV) based on score, which could be helpful in predicting intensity in alcohol relapse	Not internally or externally validated Single center Not specifically in the early LT for AH population and only 38% had ALD as primary cause of liver disease leading to LT Skewed demographics including 84% white and 86% male, which may not be generalizable Includes subjective parameters that may be difficult to reproduce/validate
Stanford integrated psychosocial assessment for transplantation (SIPAT)42, 43	SIPAT is used to predict positive or negative post-transplant outcome (rated by social worker and transplant coordinator) taking into account treatment adherence, psychosocial support system, recidivism, development of psychiatric problems, or graft failure. SIPAT score ranges from 0 to 110 and is calculated by assessing four domains: Patient’s readiness level and illness management. Social support system level of readiness. Psychosocial stability and psychopathology. Lifestyle and effect of substance abuse.	Developed in 102 randomly selected cases of liver, heart, or lung transplant recipients.42 Re-examined prospectively in same center that it was developed for 217 solid organ transplant recipients to find higher SIPAT scores predicted higher rates of rejection episodes, hospitalizations, psychiatric decompensation, and support system failure.43	c-statistic 0.70. High SIPAT scores correlated with negative outcomes.	Psychiatry-led development of model High interrater reliability, even between expert and novice raters	Not developed or validated in a pure LT population with alcohol use disorder as the focus Single center Not validated to predict relapse of substance abuse
Michigan alcohol prognostic score (MAPS)44, 45	MAPS ranges from 5 to 20. Scored by summating the following domains: Acceptance of alcoholism: both patient and family (+4), patient only (+3), family only (+2), neither (+1) Prognosis for sobriety: substitute activities (yes: +3, no: +1), behavioral consequences (yes: +3, no: +1), hope/self-esteem (yes: +3, no: +1), social relationship (yes: +3, no: +1) Social stability: steady job (+1), stable residence (+1), does not live alone (+1), stable marriage (+1)	Developed in a cohort of 99 patients with AUD undergoing evaluation for LT, 45 of which was deemed suitable for LT, but a fixed threshold/cutoff was not applied.44 Re-examined in a validation study in a cohort of 50 LT recipients with AUD retrospectively and no differences found in MAPS between patients who relapsed and patients who abstained.45	Higher MAPS theoretically correlated with reduced rate of relapse.	Studied in LT recipients with AUD	Not found to be correlated to alcohol relapse in later validation study No threshold or estimation was published Primarily used as a guide rather than a prediction model

AH, alcholic hepatitis; ALD, alcohol-related liver disease; AUD, alcohol use disorder; LT, liver transplantation; NPV, negative predictive value; PPV, positive predictive value; VA, Veterans Affairs

^#defined as alcohol use for a minimum duration of 100 days post-LT

The Stanford integrated psychosocial assessment for transplantation (SIPAT) was developed in 102 randomly selected solid organ transplant recipients at a single transplant center to predict a range of post-transplant outcomes. In a later validation study, however, the SIPAT was only moderately predictive of general psychosocial outcomes. It did not examine substance use outcomes due to inadequate clinical documentation.^42,43 Finally, the Michigan alcohol prognostic score (MAPS) originated from a cohort of 99 patients with AUD who were also undergoing LT evaluation.⁴⁴ However, a follow-up study by the same authors found no association between MAPS and post-LT alcohol relapse.⁴⁵

Post-Transplant Monitoring and Treatment – A Focus on Alcohol Use Disorder

Most aspects of post-LT medical management for the transplant indication of severe AH and other indications are the same.⁴⁶ However, early LT for severe AH often marks the beginning of the recipient’s relationship with decompensated liver disease, whereas LT recipients for most other indications have struggled with liver disease for many years. The relatively brief disease awareness afforded to patients with severe AH leaves transplant practitioners justifiably concerned about the monitoring and treatment of alcohol relapse post-LT.²⁷ We here review the three pillars of post-LT management specific to severe AH: monitoring for alcohol relapse, pharmacologic therapy, and behavioral treatment.

Monitoring for Alcohol Relapse

The importance of abstinence, or at least harm reduction, cannot be overstated. While every patient with ALD will benefit from abstinence, early LT recipients stand to benefit the most with a new lease on life. Therefore, AUD treatment post-LT is imperative. The optimal care for AUD begins with routine, destigmatized monitoring for relapse to alert clinicians to modify treatment as indicated.

In addition to regular screening for alcohol relapse by interview during scheduled post-LT visits (see Table 1), biomarkers of alcohol use can facilitate the early detection of relapse and initiation of treatment. Direct biomarkers are metabolites of alcohol; in contrast, indirect biomarkers reflect manifestations of the physiologic effects of alcohol.

Most indirect biomarkers of alcohol use such as the aspartate aminotransferase to alanine aminotransferase ratio, and mean corpuscular volume are good screening measures for heavy alcohol use that result in hepatitis or significant hepatic injury, but are insensitive to the occasional slip. Gamma-glutamyl transferase and carbohydrate deficient transferrin (CDT) are indirect biomarkers that can detect alcohol use without hepatitis, but both can be elevated due to other causes and confounders.

Select direct biomarkers such as blood alcohol level or alcohol breath tests are only useful for acute alcohol consumption. Table 3 summarizes commonly used biomarkers in the post-LT setting.^10,47–51 Of note, phosphatidylethanol is a direct biomarker that was evaluated prospectively in 61 LT recipients and had 100% sensitivity for alcohol use in the preceding week.⁴⁸ The same study found ethyl glucuronide in hair to be useful in detecting alcohol use within the past 3 to 6 months.⁴⁸

Table 3:

Biomarkers for Alcohol Relapse in Liver Transplant Recipients

Biomarker	Type of biomarker	Specimens tested	Detection window	Drinking pattern designed to detect	Potential confounders
Ethanol47	Direct	Blood	Several hours, depending on the level of drinking	Any level	- False elevations with blood ketones or non-ethanol alcohols (e.g., isopropyl alcohol or methanol)
Phosphatidylethanol (PEth)47,48	Direct	Blood	2–4 weeks	Moderate drinking (3–4 drinks/day)	- No notable confounders47
Ethyl glucuronide (EtG)47,48	Direct	Hair (must be 3 cm or longer)	Up to 6 months	Moderate drinking (2–3 drinks/day)	- False positives in impaired renal function or EtG-containing hair treatments47 - False negatives in hair treatment involving dye, perm, or bleach47
		Urine	Up to 80 hours	After a sitting of 1–2 drinks	- False positives with non-beverage alcohol use (e.g., mouthwash) or impaired renal function47 - False negatives with diuretic use or urinary tract infection47
Carbohydrate deficient transferrin (CDT)47,49	Indirect	Blood	1–3 weeks	Heavy drinking (3–5 drinks/day)	- False positives with genetic variants, lower body mass index, or end-stage liver disease49 - False negatives with age <30, female sex, or obesity49
Gamma-glutamyl transferase (GGT)50,51	Indirect	Blood	2–4 weeks	Moderate to heavy, chronic	- Elevated in advanced hepatic fibrosis, regardless of cause - False elevations by various medications (e.g., phenytoin, furosemide, heparin)51

Given the differences in the window of detection and the level of drinking each test is designed to detect, a combination of biomarkers to maximize detection of both chronicity and low levels of alcohol consumption is ideal. For example, Staufer et al. utilized the combination of urine ethyl glucuronide and CDT in a cohort of 141 LT recipients to achieve a positive predictive value of 89.3% and negative predictive value of 98.9%.⁵²

There is no consensus on the intensity or duration of monitoring for alcohol relapse post-LT. It may be reasonable to monitor more frequently for the first year after LT, during which 75% of all alcohol relapses occur.²²

Pharmacologic Therapy

Pharmacologic treatments for AUD have seen slow application in patients with concurrent liver disease in part due to a relative scarcity of empirical safety data: studies on AUD pharmacotherapeutics often exclude patients with liver disease.¹¹ Unfortunately, data are even sparser in the post-LT scenario, though adverse effects are unlikely in theory since most transplant recipients have normal or near-normal liver function.

An ideal pharmacologic agent in the post-LT setting would have no hepatotoxicity, outstanding efficacy, and minimal interaction with anti-rejection medications. As abovementioned, none of the presently available AUD pharmacotherapies possess safety or efficacy data in liver transplant recipients. Currently, there are three medications approved by the Food and Drug Administration (FDA) for treating AUD: naltrexone, acamprosate, and disulfiram. Additionally, the American Psychiatric Association (APA) recommends gabapentin and topiramate for off-label use in AUD refractory to naltrexone and acamprosate.⁵³

Naltrexone is a mu-opioid receptor antagonist that blocks euphoric effects of alcohol on the mesolimbic dopaminergic system. Initial naltrexone trials demonstrating reduced rates of relapse and fewer drinking days led to the medication’s approval by the FDA in 1994. Subsequently, a Cochrane review of randomized controlled trials (RCTs) found a 17% reduction in heavy drinking and 4% reduction in drinking days with naltrexone treatment for AUD compared to placebo.⁵⁴ A potential concern in using naltrexone is the FDA boxed warning on naltrexone’s capacity to cause hepatocellular injury “when given in excessive doses.”⁵⁵

Acamprosate is hypothesized to act on inhibitory N-methyl-D-aspartic acid (NMDA) receptors to attenuate pleasurable effects of alcohol and prevent uncomfortable effects of withdrawal and abstinence.⁵⁶ A Cochrane systematic review of 24 RCTs showed acamprosate to significantly increase cumulative abstinence duration compared to placebo.⁵⁷ Side effects of the medication include dose-dependent, typically transient diarrhea and suicidal ideation; however, rates of completed suicide were no higher in patients on acamprosate compared to placebo. Acamprosate is renally eliminated; it is contraindicated in patients with severe renal impairment.

Naltrexone and acamprosate may be good candidates for study in the post-LT setting given their relatively good safety profiles. In a meta-analysis by Jonas et al., the numbers needed to treat (NNT) to prevent return to any drinking were 12 and 20 for acamprosate and oral naltrexone.⁵⁸ The NNT to prevent return to heavy drinking was 12 for oral naltrexone; the use of acamprosate did not result in statistically significant changes. In comparing naltrexone and acamprosate for AUD, another meta-analysis of 64 RCTs suggested naltrexone had a greater effect on the reduction of heavy drinking and acamprosate was more effective in maintaining abstinence.⁵⁹

Disulfiram was the first medication approved by the FDA for treating AUD. Disulfiram inhibits aldehyde dehydrogenase, the enzyme responsible for metabolizing acetaldehyde downstream of the alcohol metabolic pathway. When one consumes alcohol while taking disulfiram, the accumulation of acetaldehyde results in aversive consequences like severe nausea and vomiting. The negative reinforcement theoretically discourages subsequent alcohol use, but it requires patient insight on the risks with alcohol intake while on disulfiram.⁵³ Furthermore, disulfiram can lead to aminotransferase elevations in up to 25% of patients on chronic therapy. The acute liver injury is even occasionally fatal, which limits the use of disulfiram in the post-LT patient.⁶⁰ RCTs on disulfiram have differed greatly by primary outcomes. A meta-analysis on disulfiram efficacy suggested a higher chance of treatment success than different controls (Hedges’ g = 0.58), but the authors detected possible publication bias in the literature.⁶¹

Anticonvulsant medications gabapentin and topiramate are not FDA-approved for treating AUD, but the APA nonetheless recommends them as second line treatments.⁵³ A Cochrane review found gabapentin and topiramate reduced heavy drinking and drinks-per-drinking-day compared to placebo, but there was no improvement on continuous abstinence. The Cochrane authors acknowledged the difficulty in drawing conclusions from their meta-analysis given the heterogeneity of included studies.⁶² A subsequent RCT illustrated a dose-dependent effect of gabapentin up to a daily dose of 1800mg in maintaining abstinence and no heavy drinking.⁶³ One concern with gabapentin is its independent potential for abuse. Care providers should monitor its use by individuals with other substance use disorders in addition to AUD.⁶⁴

Baclofen is the only anti-craving medication that has been studied in advanced liver disease, though not in the post-LT setting. It is neither FDA-approved nor recommended by the APA for treating AUD, though the AASLD and the American College of Gastroenterology have endorsed using the medication in patients with ALD.^2,65 Existing data on baclofen’s efficacy are conflicting; RCTs that examined the medication’s effect on reducing heavy drinking or increasing the duration of abstinence presented divergent conclusions.⁶⁶ The single RCT that enrolled patients with advanced liver disease – excluding those with hepatic encephalopathy or renal failure – suggested that baclofen has a relatively benign safety profile.⁶⁷ However, more recent national administrative data from France demonstrated a dose-dependent, positive association between the use of baclofen for AUD and increased risks of hospitalization and death.⁶⁸ Baclofen also can potentially cause encephalopathy.² Taken together, the risk-benefit ratio of using baclofen to treat AUD remains highly contested.

Table 4 summarizes the currently available AUD pharmacotherapeutics.^{54,55,57–62,66–69} At this time, none can be routinely recommended to LT recipients given the absence of supporting literature. All of the AUD pharmacotherapies need further assessment in the post-LT setting.

Table 4:

Pharmacotherapeutics for Alcohol Use Disorder

Medication	Administration	Dosing	Efficacy compared to placebo	Level of evidence for efficacy	Common side effects	Safety in liver disease and important pharmacokinetics	FDA approval for AUD
Naltrexone	PO, IM depot	PO: 50 mg once a day IM: 380 mg every 4 weeks	Reduces risk of heavy drinking (RR 0.83), drinks/drinking day (decrease by 3.89%), rate of relapse (decrease by 5%), and craving.54,58,59	RCTs and meta-analysis	Nausea, dyspepsia, loss of appetite	Not studied in patients with liver disease. Black box warning for hepatotoxicity.55 Up to 25% on chronic therapy has mild aminotransferase elevation.69	Yes
Acamprosate	PO	666 mg three times a day, may reduce to twice a day for patients <60 kg (Contraindicated in CrCl<30)	Reduces risk of any drinking (RR 0.86), rate of relapse (decrease by 5%), drinking days (decrease by 8.8%).57,58	RCTs and meta-analysis	Diarrhea, suicidality	Not studied in patients with liver disease. Not metabolized by liver. Renally excreted.	Yes
Disulfiram	PO	500 mg once daily for 1–2 weeks, followed by 125 mg to 500 mg daily as maintenance	Maintenance of abstinence, but has mixed results.58,61	RCTs and meta-analysis	Gastrointestinal upset, headache, tremors, metallic taste, skin rash	Not studied in patients with liver disease. Hepatotoxicity. Up to 25% of patients on chronic disulfiram will have mild transaminase elevation.60 Fatality rate is 10% when jaundice present.60	Yes
Gabapentin	PO	Daily dose of 600 mg to 1800 mg (Dosage adjustment required for CrCl<60)	Reduces heavy drinking (decrease by 0.45 SMD) and drinks/drinking day (2.14 less).62	RCTs and meta-analysis	Sedating, euphoria (potential for abuse)	Not studied in patients with liver disease. Rare reports of self-limited cholestatic liver enzyme elevations. Renally excreted.	No
Topiramate	PO	Daily dose of 150mg to 400mg (Half dosing in CrCl<70)	Reduces heavy drinking (decrease by 0.44 SMD) and drinks/drinking day (1.55 less).62	RCTs and meta-analysis	Dizziness, paresthesia, anorexia	Not studied in patients with liver disease. Rare reports of serum transaminase elevations. Approximately 70% renally excreted.	No
Baclofen	PO	Daily dose of 15mg to 60mg	Mixed results. Achieves and maintains abstinence (OR 2.67).66	RCTs and meta-analysis	Sedation, dizziness, confusion	Single RCT in patients with ALD (excluding HE and renal failure) demonstrated safety.67 However, French national claims data showed dose-dependent association with increased hospitalizations and deaths.68 Limited hepatic metabolism. Up to 80% renally excreted.	No

AUD, alcohol use disorder; ALD, alcohol-related liver disease; CrCl, creatinine clearance; FDA, Food and Drug Administration; HE, hepatic encephalopathy; IM, intramuscular; PO, per oral; RCT, randomized controlled trial; SMD, standardized mean differences

In addition to the pharmacologic options discussed earlier, there are also some investigational medications for AUD. Nalmefene shares a similar chemical structure and pharmacologic mechanism with naltrexone and is potentially more efficacious at reducing alcohol consumption. However, its oral formulations used in treating AUD is not available in the United States.^70,71 A recent meta-analysis of placebo-controlled RCTs involving participants with heavy drinking or AUD showed varenicline to attenuate alcohol craving but not reduce heavy drinking days, drinks per drinking day, or days abstinent.⁷² While its effect on AUD appears modest, varenicline may still have a role in patients with co-existing nicotine dependence. Aripiprazole reduced the number of drinks per drinking day but did not improve abstinence in a large multicenter RCT for AUD.⁷³ Similar to varenicline, aripiprazole may still benefit select patients with AUD who have a concurrent psychiatric indication for the prescription. Finally, there are encouraging preliminary data on the use of ondansetron and zonisamide in treating AUD, but both medications require further studies to clarify their roles.^74,75

Behavioral Treatment

While total abstinence post-LT may be the preferred goal, all relapses are not the same. Sustained alcohol use and harmful drinking patterns contribute most to deleterious outcomes.^22,76 Once alcohol relapse occurs, prompt intensive alcohol addiction treatment is paramount to mitigate harmful or sustained drinking and preserve allograft function. Behavioral treatment (BT) is the category of addiction treatment that identifies and modifies maladaptive behaviors. These strategies take root in traditional theories of psychology such as operant conditioning and cognitive theory.

The implementation of BT in the post-LT setting can be conceptualized along three axes: the selected BT strategies, the intensity of treatment, and the integration with the liver transplant clinic.⁷⁷ However, most of the data supporting BT in LT recipients are extrapolations from observational studies of pre-LT patients with ALD.⁷⁸ Figure 1 illustrates our approach to BT.

Figure 1:

Behavioral Treatment For Alcohol Use Disorder

The three axes of overall behavioral treatment strategy is conceptualized in this diagram. The methods under each axis are listed in the order of preference in the posttransplanl setting.

*Trained counselors provide 24-hour care to stabilize imminent danger from addiction.

**An organized outpatient treatment usually during the day. but does not provide 24-hour care.

***Outpatient treatment typically occurring after work or school or on weekends.

#Interviewing style aimed at fostering a constructive partnership between practitioner and patient. Can be part of a formal counseling program, or adapted to the transplant clinic setting.

BT encompasses a spectrum of strategies from brief interventions conducted in the physician’s office to recurring sessions of psychological interventions led by addiction therapists. Two of the best studied psychological interventions for AUD are cognitive behavioral therapy and motivational enhancement therapy. Cognitive behavioral therapy (CBT) is the technique of first identifying a trigger that leads to drinking and then redirecting the response to said trigger towards non-drinking activities. The delivery of CBT often spans 10 or more sessions.⁷⁹ In a meta-analysis of 30 RCTs involving individuals with alcohol and other substance use disorders (but not specifically ALD), CBT was more effective than minimal treatment controls across multiple outcome types and follow-up durations.⁸⁰ Motivational enhancement therapy (MET) is a focused, structured treatment using the principles of motivational interviewing (see below) to evoke a patient’s internal motivation for behavioral change. Instead of specialist-driven, unidirectional instructions, MET applies motivational feedback on a patient’s personal circumstances to encourage concrete abstinence strategies.⁵⁶ In one of the few RCTs evaluating BT in LT candidates, Weinrieb et al. reported similar prevalence of self-reported relapse (~25%) over two years of follow-up between 91 patients who received MET versus usual care (defined as referral to local addiction treatment services). However, among patients who relapsed, the MET group had fewer drinking days and fewer drinks per drinking day.⁸¹

Relative to the more structured MET, motivational interviewing (MI) is a broader counseling philosophy that is nonjudgmental yet directive and focuses on the patient’s internal recognition of the link between alcohol use and its consequences. The complete mastery of MI challenges even addiction specialists, but non-specialists (e.g., transplant clinicians) can nonetheless learn to incorporate the essence of the MI style in routine follow-up visits for patients with AUD.^77,82 As a harm reduction technique, MI helps to establish a democratic, nonpunitive partnership between the clinician and the patient, where the clinician aspires to encourage and support behavioral change. The process facilitates constructive dialogue between the two parties should alcohol cravings arise, which hopefully promotes early recognition and treatment engagement. In a meta-analysis, Vasilaki et al. examined RCTs on MI interventions for excessive drinking and found MI more efficacious than no treatment in reducing alcohol intake over the short term.⁸³

While MI is not a strictly structured BT paradigm, it does take multiple encounters to show effect. The five guiding principles of MI are 1) express empathy to establish common ground with patients, 2) identify discrepancies between patients’ stated goals and their current actions, 3) eschew direct confrontations, 4) rolling with patient resistance, and 5) support self-efficacy.⁸⁴ Open-ended questions and reflective listening initially help the clinician understand the patient’s goals and then recognize discrepancies from actual behavior. Points of contention and resistance may arise as the clinician strives to guide the patient to perceive these discrepancies. It is vital to avoid arguments by shifting focus away from confrontation. Affirming the patient’s inner guide and encouraging self-motivational statements are two strategies to reinforce the patient’s positive self-reflection.

Less structured BT methods such as brief interventions or mutual support have a smaller – but not negligible – role in treating the severe AUD that may lead to severe AH or ALD.⁵⁶ Brief interventions are short sessions usually conducted by a clinician during a routine visit. Jointly, the provider and the patient set an achievable goal either to reduce or stop drinking and work on an action plan to that end. There is typically an agreement to reassess alcohol consumption on subsequent visits.⁸⁵ The United States Preventive Services Task Force (USPSTF) made a grade B recommendation in 2018 for conducting brief behavioral counseling in primary care settings to reduce unhealthy alcohol use, based on a mean reduction of 1.59 standard drinks per week in a meta-analysis of 37 trials involving 15,974 adults.⁸⁶ The summary of evidence notably excluded studies with pharmacologic treatment for severe AUD, which limited its extrapolation to patients with ALD.

Mutual support is peer-led support groups that meet regularly and provide individuals with a beneficial social network and positive experiences in a sober environment.⁸⁷ Mutual support offers advantages like longitudinal therapy and low cost. Alcoholics Anonymous (AA) is perhaps the most visible among mutual support groups, but its spiritual foundations may not suit everyone. SMART (Self-Management And Recovery Training) Recovery is a less spiritual alternative rooted in several of the abovementioned BT theories. The supportive social network adds a safety net not inherently available in CBT or MET, but mutual support alone may be insufficient for patients with severe AUD often seen in LT recipients for severe AH.⁵⁶ One reason may be the frequent lack of standardization between different sites, even among groups supposedly rooted in the same therapeutic tradition. Conversely, in two RCTs (1,936 participants), standardized/manualized mutual support programs compared favorably to formal psychological interventions in achieving and maintaining sustained abstinence for 12 months.⁸⁸

In addition to the BT strategy selected either alone or in combination, treatment intensity should be personalized to each LT recipient. The American Society of Addiction Medicine (ASAM) offers a useful conceptual continuum that encompasses early intervention (ASAM level 0.5), outpatient services (ASAM level 1 to 2.5), and inpatient services (ASAM level 3.1 to 3.5).⁸⁹ Early intervention describes BT before a confirmed AUD – or relapse of AUD for LT recipients – to gather information for assessment and prevent disease progression. Outpatient services comprise different time commitments ranging from less than 9 hours a week (ASAM level 1) to partial hospitalization without 24-hour staff availability (ASAM level 2.5). Inpatient services offer varying levels of structured clinical time, but all patients live on-site with a 24-hour support staff. While higher-intensity arrangements may offer more resources in theory, they are typically costlier and still may be less optimal for a given individual. Factors such as time commitment, distance from home, privacy, and insurance coverage are all important considerations in light of multiple competing priorities in the first year post-LT, including post-surgical complications and financial obligations.^90,91 Thus, the decision of optimal treatment intensity ideally involves a thoughtful discussion between the patient, transplant providers, social workers, and addiction specialists.

Integrated (defined as co-locating in the same physical space) care delivery models that combine BT with the standard transplant practice may be the most effective in treating AUD and reducing post-LT relapse.⁹² The close physical proximity enhances the communication between transplant and addiction care providers. Addolorato et al. compared post-LT alcohol outcomes before versus after the 2002 establishment of an Alcohol Addiction Unit within their Liver Transplant Center in Rome, Italy. The LT recipients who received integrated treatment had significantly lower prevalence of alcohol relapse and death (16.4% and 14.5%) than the recipients who received non-integrated care (35.1% and 37.8%).⁹³ Similarly, the aforementioned RCT by Weinrieb et al. demonstrated the effectiveness of integrated MET on decreasing the quantity and frequency of post-LT drinking relative to a non-integrated approach.⁸¹

In summary, extrapolating from the pre-LT literature, BT interventions likely have the potential to maintain abstinence and mitigate harmful drinking in early LT recipients with AH. Unfortunately, the current literature lacks detailed descriptions of BT strategies in the post-LT setting. A prospective therapeutic trial is under way.^94,95

Conclusion

Liver transplantation is an exercise of public trust bestowed upon the transplant team by society. The present article proposed several strategies to manage AUD after early LT for AH. For the high-stakes scenario of LT where harmful alcohol relapse can lead to devastating post-LT outcomes, frequent follow-ups within an integrated/co-locating clinical practice combining transplant and addiction specialists may prove to be the most effectual approach. Prospective investigations are in progress to elucidate and elevate our care for these vulnerable patients.

Synopsis:

Severe alcoholic hepatitis portends a high risk of mortality without liver transplantation. Transplant outcomes in severe alcoholic hepatitis exhibit a strong inverse association with post-transplant alcohol relapse. The ingredients most central to ameliorating alcohol relapse risk may include destigmatized post-transplant alcohol monitoring, nonpunitive clinician-patient partnership, and multimodal therapies to maintain abstinence and mitigate high-risk drinking. We here review the core principles of post-liver transplant management specific to alcohol use disorder.

Key points:

• Severe alcoholic hepatitis carries a grim prognosis. It is increasingly considered an indication for liver transplantation in select candidates.

• A deliberate pre-transplant assessment is the first step toward mitigating adverse post-transplant outcomes.

• The three pillars of post-liver transplant management specific to alcohol use disorder are destigmatized routine monitoring for alcohol relapse, pharmacologic therapy, and behavioral treatments.

Abbreviations:

AASLD

American Association for the Study of Liver Diseases

AH

alcoholic hepatitis

AC

alcohol-related cirrhosis

ALD

alcohol-related liver disease

APA

American Psychiatric Association

ASAM

American Society of Addiction Medicine

AUD

alcohol use disorder

BT

behavioral treatment

CDT

carbohydrate deficient transferrin

CBT

cognitive behavioral therapy

FDA

Food and Drug Administration

LT

liver transplantation

MELD

Model for End-Stage Liver Disease

MET

motivational enhancement therapy

MI

motivational interviewing

NNT

number needed to treat

RCT

randomized controlled trial

USPSTF

United States Preventive Services Task Force