Recidivism in Liver Transplant Recipients for Alcohol-related Liver Disease

Abstract

Liver transplantation (LT) is the only cure for patients with end-stage liver disease, which offers good long-term survival. The long-term issues after LT affecting survival are cardiovascular disease, chronic kidney disease, de novo malignancies, recurrence of original disease and immunological causes. Alcoholic-related liver disease (ALD) is one of the most common indications for LT worldwide including India. LT for ALD is associated with several unique challenges as compared with other etiologies. Long-term survival after LT in patients with ALD is affected by recidivism. Various studies have shown different predictors of relapse; the main predictors of relapse are pretransplant abstinence, psychiatric comorbidities, and lack of social support. Although several risk scores have been proposed, these scores are not validated. Studies with active involvement of psychiatrist have shown lower relapse rates. The relapse prevention strategy for reducing likelihood and severity of relapse after initial cessation of alcohol uses a combination of pharmacotherapy and cognitive behavioral approach (identifying and addressing high-risk situations for relapse).

Alcoholic-related liver disease (ALD) is a common etiology of end-stage liver disease and contributes to a significant proportion of patients undergoing liver transplantation (LT).^1,2 In a study from North India, ALD was indication of 26% of living donor LTs.¹ LT for ALD presents several challenges in the long term, such as recidivism and a higher risk of de novo malignancies, that may impair long-term survival as compared with other etiologies.^3,4 Several studies have shown that long-term survival is similar to other etiologies, whereas other studies have shown a lower survival in LT recipients with ALD, likely secondary to graft loss (recidivism) or de novo malignancies.^{2,3,5, 6, 7, 8} Pageaux et al analyzed data of 128 LT recipients. A total of 10% were occasional drinkers and 21% were heavy drinkers at a mean follow-up of 53.8 months. Although the actuarial survival rates were not different, 3 of 7 deaths among heavy drinkers were related to alcohol relapse.⁵ The data from European liver transplant registry from 1988 to 2005 showed a 73% 5-year and a 59% 10-year survival rate for ALD in a cohort of 9880 patients; this survival was better as compared to patients with hepatitis C and cryptogenic cirrhosis. The following were causes of death/graft failure that were seen more commonly in LT recipients with ALD: de novo malignancies, cardiovascular, and social causes.³ Grąt et al showed that alcohol relapse increased the risk of death during the first five years. Also, survival of patients with ALD was worse during the 5–10 years period.⁷ In a meta-analysis, patients with recidivism had higher chances of having steatohepatitis, rejection, graft failure, and mortality.⁹ Burra et al showed that fatty change and pericellular fibrosis on liver biopsy were significantly more common in heavy drinkers when compared with occasional and nondrinkers.⁶ A French long-term study, with a median follow-up of 11 years (range 3–18 years) showed a severe alcoholic relapse in 20% (73/369) of LT recipients. A total of 18 patients had recurrent alcoholic cirrhosis at 6 years (range 3–10 years) after LT and 4.5 years (range 2–8 years) after severe alcoholic relapse. The cumulative risk of F4 fibrosis was 54% at 10-years after severe alcoholic relapse.¹⁰

Cuadrado et al showed that the 10-year survival rate for patients with recidivism was 45.1% compared with 85.5% in patients without recidivism.⁸ Egawa et al also found similar results. The 10-year survival rates were 21.9% in relapse group versus 73.8% in patients with no relapse at 18 months after LT.¹¹ Pfitzmann et al showed a significantly worse survival in LT recipients with significant or abusive drinking as compared with recipients with occasional drinking.¹² Satapathy et al found that 1-, 3-, and 5-year patient survival rates, respectively, were as 95%, 87%, and 80% in abstinent group compared with 87%, 49%, and 49% in recidivism group (p = 0.001).¹³ Rice et al showed that continuous heavy drinking was associated with allograft loss [hazard ratio = 2.57) and significant fibrosis.¹⁴ In the study from our center, 9.5% of 408 LT recipients for ALD had alcohol relapse and 23% of these were heavy drinkers. All heavy drinkers presented with features of graft dysfunction.¹

Thus patients with occasional alcohol intake appear to have similar outcomes; however, those with significant alcohol intake are associated with poor survival after LT.

Predictors of relapse

Various predictors of relapse after LT are shown in Table 1 (references ^{7, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31}). Most of these studies are retrospective in nature and have used various definitions of relapse or recidivism, which makes a comparison difficult. Usually relapse is defined as any amount of alcohol intake after LT.^1,32 As discussed earlier, significant drinking (and not occasional alcohol intake) is important for prognosis. There are several definitions of significant drinking used in various studies. Significant drinking is defined as more than 21 units/week (males) or 14 units/week for females,^19,26 drinking associated with medical/social harm, or daily consumption of alcohol,²⁴ >30 g/day for women and >60 g/day for men,¹¹ moderate (1–20 units/week) or severe (>20 units/week),¹ continuous or nearly continuous use without clear intervals of sobriety.¹⁴ Other definitions have also been used.²³

Table 1.

Predictors of Alcohol Relapse After LT in Studies Published After 2000.

Author,ref. year	N	Relapse	Abstinence predicted relapse	Psychiatric factors	Others factors/comments
Bellamy,15 2001	123	13 definite, 3 doubtful relapse	No	No	Daily ethanol consumption
Jauhar,16 2004	111	15%	No	a	Family history of alcoholism
Kelly,17 2004	100	18/90 (20%) harmful	No	Depression	Grams/day alcohol pre-LT, smoking, reliance of family/friends for post LT support, protective factors- stable partner, insight of etiology
Björnsson,18 2005	93	33%, 18% harmful	No	a	Structured management program protective
Perney,19 2005	61	52% relapse, severe in 13%	Yes	History of alcohol withdrawal, alcohol dependence	Younger age, alcohol abuse in first relatives
DiMartini,20 2006	167	42%, 26% binge pattern	Yes	Diagnosis of dependence, depression	Multivariate model for binge drinking: abstinence
De Gottardi,21 2007	387	11.9% harmful relapse	Yes	Psychiatric comorbidities	Age> 50 years, HRAR score higher than 3
Pfitzmann,12 2007	300	19%	Yes	Predicted poor psychosomatic prognosis	Absence of companion in life, presence of young children
Gedaly,22 2008	142	19%	Yes	Participation in rehabilitation	Pretransplant drug abuse predicted drug abuse after transplantation also
Tandon,23 2009	171	24%, 13% problem drinking	Yes	a	a
Karim,24 2010	80	10% harmful	Yes	a	Female sex
Hartl,25 2011	120	16% recidivism	Yes	Nonacceptance of having an alcohol problem before LT	a
Egawa,11 2013	140	22.9%	No	a	History of treatment for psychological diseases other than alcoholism, noncompliance with clinic visits after LT and smoking
Rice,14 2013	300	16%, 33% of these had harmful relapse	Yes	a	Multivariate analysis for relapse not available, young age, biliary complications significantly high in relapse group
Deruytter,26 2013	108	29%, problem drinking in 16%	No	No	Presence of a first degree relative with alcohol abuse
Rodrigue,27 2013	118	33.8%	No	Continued alcohol use after liver disease, low motivation for alcohol treatment, poor stress management skills, no rehabilitation relationship, limited social support, diagnosis	Absence of hepatocellular carcinoma, tobacco dependence, lack of nonmedical behavioral consequences, continued engagement in social activities with alcohol present
Rodrigue,28 2013	118	33.8%	No	Treatment of substance abuse both before and after LT (not before LT alone)	a
Grąt,7 2014	97	33.5%	No	a	Younger age
Satapathy,13 2015	148	10.8%, harmful	No	a	Older age and immediate family support- protective
Saigal,1 2016	408	9.5%, 59% of these moderate to heavy drinkers	Yes	a	Younger age
Attilia,29 2018	69	8.7%	No	History of alcohol withdrawal syndrome	Female gender, shorter time of multidisciplinary support program before LT
Skladany,30 2019	89	26%, 52% of these had harmful drinking	No	a	Smoking, loss of social status, time after LT
Kitajima,31 2019	190	13.7%	Yes (<1.5 years)	Yes	Complications after LT, alcohol relapse before LT

^aNone other/not studied or not important.

Any relapse is shown to be present in 8–52%, and harmful relapse (recidivism) is shown to be present in 10–26% (Table 1). The predictors of relapse can be broadly divided into following categories: pretransplant abstinence, psychiatric comorbidities (depression, substance addiction, prior rehabilitation etc.), lack of social support, and other miscellaneous factors. The most widely studied predictor is pretransplant sobriety, which has not shown to be a significant predictor of post LT relapse in many studies; moreover, various studies have found different cutoffs of pretransplant sobriety, ranging from 3 months²⁵ to 1.5 years.³¹ Owing to organ shortage and concern of sharing limited resources to patients with self-inflicted (alcoholism) disease with risk of recidivism after LT, 6-month abstinence rule was proposed. The 6-month abstinence rule served 2 purposes; it provided some time to a patient to demonstrate sobriety (which is thought to predict sobriety after LT), also a patient might recover on medical management once abstinent from alcohol. Although this 6-month abstinence rule was more of empirical in nature and did not predict post LT relapse with good accuracy.³³ Another limitation of applying 6-month rule is poor prognosis of patients with severe alcoholic hepatitis not responding to medical management.^34,35

Several series have shown good outcomes after LT in patients with severe alcoholic hepatitis, although LT done at short abstinence period. Lee et al showed incidence of sustained alcohol use as 10% at 1 year and 17% at 3 years in a series of 147 patients with alcoholic hepatitis with a median pretransplant abstinence of 55 days.³⁶ We also found a relapse rate of 12.8% in patients who underwent LT for severe alcoholic hepatitis, which is almost similar to overall experience of ALD reported earlier from our center.^1,37 A meta-analysis of 92 studies found alcohol relapse rate as 22% [95% confidence interval (CI): 19–25%] and heavy alcohol relapse rate as 14% (95% CI: 12–16%) after LT during a mean follow-up of 48.4 ± 24.7 months.³ There was significant heterogeneity in results. The presence of psychiatric comorbidities [odds ratio (OR) 3.46, 95%CI: 1.87–6.39), abstinence < 6 months (OR 2.76, 95% CI: 2.10–3.61), unmarried status (OR 1.84, 95% CI: 1.39–2.43) and smoking (OR 1.72, 95% CI: 1.21–2.46) were the factors associated with relapse.³

Several predictive scores have been proposed to predict relapse. Two commonly used are scores are Alcohol Relapse Risk Assessment and High-Risk Alcoholism Relapse (HRAR) as shown in Table 2.^27,38,39 It should be noted that HRAR score was developed for alcohol use disorder and was applied later to LT recipients. Recently, Lee et al developed a score to predict sustained alcohol use (defined as a minimum duration of 100 days) from the ACCELERATE-AH cohort. The Sustained Alcohol Use PostLT (SALT) score (range: 0–11) consisted of 4 variables as shown in Table 2. The C statistic was 0.76. A SALT score ≥5 predicted sustained alcohol use with a sensitivity of 55%, specificity of 84%, positive predictive value of 25%, and negative predictive value of 95%. Thus the score was able to predict low risk of relapse, but was not able to predict high risk of relapse, even a maximum score of 11 had 50% positive predictive value.³⁹

Table 2.

Risk Scores to Predict Relapse of Alcohol After Liver Transplantation.

Risk score, author (year)	Score calculation	Comments
Alcohol Relapse Risk Assessment (ARRA), Rodrigue27 (2013)	One point to each of following 9 parameters; absence of HCC, tobacco dependence, continued alcohol use after liver disease diagnosis, low motivation for alcohol treatment, poor stress management skills, lack of rehabilitation, limited social support, lack of nonmedical behavioral consequences, engagement in social activities with alcohol	Score range 0–9, relapse rates were 0% for the ARRA I (score 0), 8% for the ARRA II (score 1–3), 57% for the ARRA III (score 4–6), and 75% for the ARRA IV (score 7–9), ARRA classification was also associated with intensity of relapse
High-Risk Alcoholism Relapse (HRAR) model, Yates,38 1993	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)	Score 4–6 (HAR group) 61% were re-admitted within 6 months compared to 28% of the low-risk alcoholism relapse (score 1–3)
Sustained Alcohol Use PostLT (SALT) score Lee BP,39 2019	>10 drinks per day at presentation (+4 points), ≥2 rehabilitation attempts (+4 points), any history of alcohol-related legal issues (+2 points), and prior illicit non-tetrahydrocannabinol substance abuse (+1 point)	Score ≥5 had a 25% positive predictive value, and a score of <5 had a 95% negative predictive value for sustained alcohol use post LT

Use of prediction scores in clinical practice

As harmful relapse is not present most LT recipients, so having a good sensitivity and positive predictive value with a single risk score is difficult, as seen with SALT score where sensitivity and positive predictive values are poor. Patients at low risk could be identified better than patients who relapsed to sustained alcohol use.³⁹ It should be noted that only few studies have attempted to validate these scores. The HRAR score predicted relapse in the study by De Gottardi et al,²¹ but failed to predict in studies by Egawa et al and DiMartini et al.^11,40 The HRAR score performed poorly in study be Lee et al also, with c-statistic of 0.56.³⁹ A patient may have higher risk of relapse in presence of multiple risk factors.²¹ De Gottardi et al analyzed 387 patients, 11.9% relapsed to harmful drinking. Multivariate analysis revealed the following as independent factors of relapse: pretransplant abstinence less than 6 months [OR, 3.3]; presence of psychiatric comorbidities (OR, 7.8), and HRAR score higher than 3 (OR, 10.7). Alcohol relapse was only 5% in absence of these factors, relapse increased to 18%, 64%, and 100% in presence of 1, 2, or 3 factors.²¹ Thus, a risk score can help in identifying risk, but other known risk factors (Table 1) should also be considered when deciding a risk of alcohol relapse.

Prevention of relapse

Active involvement with a psychiatry/de-addiction team before onset of relapse is associated with less alcohol relapse. Björnsson et al compared recidivism rates of ALD-LT recipients transplanted between 1988 and 1997, and after 1998 when institution of structured recidivism management was done. A total of 33% ALD recipients had relapse of alcohol intake and 17 patients were 18% were heavy drinkers. The proportion of patients with relapse was 19/40 (48%) before the onset of structured management that decreased to 13/58 (22%) in later period.¹⁸ Addolorato et al also found similar results. A total of 92 LT (1995–2010) recipients were evaluated. These patients had clinical evaluation and management of alcohol use by psychiatrists not affiliated to the liver transplant unit before 2002 (n = 37; group A), or by alcohol addiction unit within the liver transplant center 2002 onward (n = 55; group B). Patients in group B had significantly lower prevalence of alcohol recidivism (16.4 versus 35.1%; p = 0.038) and mortality (14.5 versus 37.8%; p = 0.01).⁴¹

A study by Rodrigue et al found that patients who received substance abuse treatment both before and after LT had lower alcohol relapse (16%) than patients without substance abuse treatment (41%) or with pretransplant treatment only (45%).²⁸ Attilia et al analyzed results of adoption a program of multidisciplinary support to patients undergoing LT for ALD. Sixty-nine patients underwent multidisciplinary support and 8.7% presented alcohol relapse. At multivariate analysis, a shorter time of multidisciplinary support was independent risk factor for relapse along with female gender and alcohol withdrawal syndrome. The rate of relapse was significantly lower than historical group who did not undergo multidisciplinary support (subdistributional hazard, 0.21; p = 0.009).²⁹

Egawa et al showed that noncompliance with clinical visits was associated with relapse;¹¹ thus more active involvement and frequent visits to treating team may help to decrease relapse. DeMartini et al conducted a randomized 8-week pilot study of text message–based alcohol intervention. The participants (n = 15) responded to 81% of messages, reported high rates of satisfaction, found it easy to complete intervention, and also looked forward to receive messages. None of the participants in the text message group had positive urine alcohol tests at 8 weeks, whereas 2 participants in standard care group tested positive. Participants in text message had less stress at 4 and 8 weeks.⁴² These studies suggest that active engagement of patients with a specialist deaddiction team in pretransplant and post-transplant period may decrease alcohol relapse.

Impact of smoking

Smoking affects ALD-LT recipients in several ways. Smoking has been shown be associated with risk of relapse in several studies.^11,17,30 Smoking is common among LT recipients for ALD. Heide et al found that smoking use was more common in patients with ALD, both before and after LT. Forty-four percent of patients with ALD had active smoking after LT.⁴³ DiMartini et al found that ALD recipients resumed smoking early after LT and increased consumption over time.⁴⁴ Smoking is a risk factor for de novo malignancy and cardiovascular disease, which are important causes of morbidity and mortality after LT.^{45, 46, 47, 48, 49, 50}

Relapse prevention in alcohol dependence: what is available for treatment

Relapse prevention (RP) is a strategy for reducing the likelihood and severity of relapse after the cessation of alcohol use. RP is done using a combination of pharmacotherapy and a cognitive behavioral approach with the goal of identifying and addressing high-risk situations for relapse and assisting individuals in maintaining abstinence. RP has two specific aims: preventing an initial lapse (defined as initial drinking after a period of abstinence) and maintaining abstinence and providing lapse management if a lapse occurs such that further relapses can be prevented. An algorithm for RP is shown as Figure 1.

Figure 1.

Management of relapse after liver transplantation.

Pharmacotherapy in relapse prevention and treatment

It should be noted that safety data of pharmacotherapy are not available in patients with advanced liver disease⁵¹ or in post-transplant patients. The initial step in the treatment of alcohol dependence is management of withdrawal, which is usually treated, with the help of benzodiazepines.⁵² Withdrawal management is only the initial step of treatment of alcohol dependence. The main challenge is ensuring abstinence, or preventing a relapse. Acamprosate, disulfiram, baclofen, and naltrexone are the main agents used for RP as shown in Table 3 (references ^{53, 54, 55}).

• A.Acamprosate: FDA approved acamprosate in 2004 for treatment of alcohol use disorder. Its mechanism of action has been ascribed to aspects of glutaminergic and/or GABA-ergic neurotransmission. The results of a large number of RCTs and meta-analysis have shown that acamprosate in conjunction with psychosocial treatment significantly increases the proportion of people who remain abstinent from alcohol at 6 months. It is not metabolized in the liver, does not interact with alcohol, and is particularly useful in patients with impaired hepatic function. The most common side effect is diarrhea. It is contraindicated in patients with severe renal impairment and those with hypersensitivity to the drug.^{53, 54, 55}
• B.Naltrexone: It is used in the management of opioid dependence and alcohol dependence. Naltrexone is believed to reduce the rewarding effects of alcohol by acting on the mesolimbic dopaminergic pathway. Its mechanism of excretion is primarily renal. In combination with psychosocial treatment, it has a modest but significant effect on reducing relapse rates. Use of naltrexone is associated with rise of liver enzymes. The most common adverse events are gastrointestinal. The most important safety consideration is its reaction with opioid drugs. Naltrexone is contraindicated in patients taking or likely to take opioids. It is also contraindicated in patients with acute or chronic hepatic failure.^{53, 54, 55} A meta-analysis has shown a benefit of naltrexone in treatment of alcohol use disorder.⁵⁶
• C.Disulfiram: It has been used in clinical practice for >60 years. Disulfiram is an aldehyde dehydrogenase inhibitor. The high levels of acetaldehyde (primary metabolite of alcohol), which accumulate in people taking disulfiram after alcohol ingestion, leads to a constellation of symptoms (disulfiram reaction). These symptoms appear within minutes after alcohol ingestion and may last for hours. The symptoms act as a deterrent against alcohol consumption, but may prove fatal. There are a number of contraindications to the use of disulfiram including cardiovascular disease, hypertension, suicidal patients, severe personality disorders, and psychosis. Caution should be exercised in patients with liver and renal disease. Disulfiram should not be used in post-transplant patients given adverse events and risk of hepatotoxicity.^{53, 54, 55,57}
• D.Baclofen: Baclofen is a selective GABA-B receptor agonist and is the only pharmacotherapy tried in the setting of cirrhosis. Several placebo control trials^{58, 59, 60, 61} have been undertaken with variable results as discussed in EASL and AASLD guidelines.^62,63 The AASLD guidelines recommend use of acamprosate or baclofen (based on limited data) for the treatment of alcohol use disorder in patients with ALD. However, a recent meta-analysis of 14 double-blind randomized controlled studies (n = 1522 patients) in patients with alcohol use disorder (liver disease not present in all studies) did not show a statistical significant benefit with use of baclofen as compared with placebo.⁶⁴

Table 3.

Pharmacotherapy Options for Alcohol Dependence (References ^{53, 54, 55}).

Agent	Mechanism of action	Adverse events	Hepatotoxicity potential	Comments
Disulfiram	Acetaldehyde dehydrogenase inhibition	Hypotension, flushing, gastrointestinal, psychiatric, neurological	Reports of hepatotoxicity and liver failure	Not preferred, important to explain patients to not take alcohol
Naltrexone	Opioid receptor antagonist	Gastrointestinal, rise of liver enzymes	Hepatotoxicity, in particular in patients with significant liver problems	Should not be used in patients with hepatitis or liver failure, or patients receiving long-term opioid therapy
Acamprosate	Glutamatergic antagonist and GABA agonist	Diarrhea	No liver metabolism, no alcohol pharmacokinetic interactions, can be used in mild to moderate liver disease	No data of use in patients with advanced liver disease, should not be used if hypersensitivity
Baclofen	GABA-B receptor agonist	Sedation, weakness	Not reported	Doubtful efficacy

Given the profile of aforementioned medications, acamprosate appears to be the safest drug to treat patients with post-transplant recidivism.

Psychosocial management

RP is based on the premise that problematic drinking behaviors are likely to re-emerge. It is best to combine psychosocial treatment with pharmacotherapy for best results. The main components of RP include the following:

• •Identifying and coping with high-risk situations. These include trigger for relapse, which may be internal or external;
• •Enhancing self-efficacy;
• •Cognitive restructuring;
• •Balanced lifestyle and positive addictions;
• •Stimulus control techniques;
• •Urge management techniques.

Although LT for ALD is associated with excellent short-term survival, the long-term survival is affected by recidivism. The harmful relapse (recidivism) occurs in 10–26% of patients transplanted for ALD. Multiple risk factors of relapse have been described; pretransplant abstinence, psychiatric comorbidities, and lack of social support are the most important factors that predict relapse of alcohol intake after LT. Although several risk scores have been proposed, these are not validated widely and should be considered along with other known risk factors. There is scarcity of literature in management of post-transplant recidivism; both psychosocial treatment and pharmacotherapy should be used.

CRediT authorship contribution statement

Narendra S. Choudhary: Conceptualization, Writing - original draft. Neeraj Saraf: Writing - review & editing. Saurabh Mehrotra: Conceptualization, Writing - original draft. Sanjiv Saigal: Writing - review & editing. Arvinder S. Soin: Writing - review & editing.

Conflicts of interest

The authors have none to declare.