Alcoholic liver disease (ALD) is the second most common diagnosis for patients undergoing liver transplantation (LT) in the United States and Europe.1, 2 ALD, either alone or in combination with a hepatitis C virus (HCV) infection, accounted for 20% of all primary transplants in the United States between 1988 and 2009 (>19,000 recipients). This is a remarkable number, especially when it is contrasted with the prediction made at the landmark National Institutes of Health consensus conference in 1984 that not many patients with ALD would be selected for LT.3 Moreover, the outcomes for patients who undergo transplantation for ALD are at least as good as those for patients with most other diagnoses and are better than those for patients with HCV.4 However, the apparent success of LT for ALD masks a more complex reality. There are still unresolved controversies about LT for patients with ALD. In this review, I address several of these contentious issues, which include the following: the goals of treatment; the referral, evaluation, and selection of patients for LT; and the impact of the diagnosis of ALD on care after LT.
Although alcoholic liver disease (ALD) is one of the most common indications for liver transplantation (LT), there are still unresolved controversies about the goals of treatment, the referral, evaluation, and selection of patients with ALD for LT, and their care after LT. It is uncertain whether there is a large unmet need for LT among patients with ALD because of the unmeasured effects of recent drinking, relapse, and recovery with abstinence in this population. A careful assessment of the extrahepatic effects of alcohol-related end-organ damage is needed for ALD patients who are referred for an LT evaluation. Although there clearly is a relationship between the length of sobriety and future abstinence, the present methods for predicting future drinking are inexact. The survival of ALD patients after LT is as good as the survival of non-ALD patients, although patients with coincident ALD and hepatitis C virus have higher mortality and morbidity rates. After LT, ALD patients have an increased risk of developing malignancies and cardiovascular disease. These risks appear to be linked to cigarette smoking. Covert drinking occurs both before and after transplantation, and approximately 20% of patients return to harmful drinking after LT. Harmful drinking after LT (instead of slips) causes liver damage and reduces survival. Better therapies for controlling addictions to alcohol and nicotine are needed for ALD patients both before and after LT. Liver Transpl 17:751-759, 2011. © 2011 AASLD.
GOALS OF TREATMENT
The goal of LT is the treatment of life-threatening liver failure or cancer that is intractable to medical management. The medical management of ALD starts with abstinence from alcohol. Patients with alcoholism who remain abstinent can recover from advanced liver failure, and stable liver function can be reestablished with the resolution of portal hypertension.5, 6 Unfortunately, alcoholism is a disease of relapses and remissions, and this pattern persists even after life-threatening episodes such as variceal hemorrhage.7 The frequency of recovery from decompensated liver failure due to ALD is restricted by the frequency of drinking relapses.8 A therapeutic formulation addressing LT for ALD needs to encompass the psychological and somatic health of potential candidates. In other words, LT should be seen as a treatment of end-stage liver failure within a comprehensive care program that addresses the management of addictions to alcohol, cigarettes, and any other addictive drugs.
REFERRAL OF ALD PATIENTS FOR AN LT EVALUATION
The combined prevalence of alcohol abuse and dependence in the United States has been estimated to be 84.5 cases per 1000 persons who are 18 years old or older; this translates into approximately 18 million adults at risk for ALD.9 The estimated age-adjusted death rate related to liver cirrhosis in 2005 was 9.2 deaths per 1000 persons; this translates into approximately 27,000 deaths.9 In 2004, liver disease (not including viral hepatitis) accounted for 2.4 million ambulatory care visits, and it was the third most common digestive diagnosis on hospital discharge records.10 Although these data do not provide a precise estimate of the prevalence of patients with life-threatening ALD who are suitable candidates for LT, previous researchers have asserted that ALD patients are under-referred for LT in the United States.11
On the other hand, data documenting the process of referral and evaluation for patients with problem drinking are inconclusive on this point. Julapalli et al.12 described a cohort of 199 patients with liver disease who received their medical care at a large metropolitan Veteran Affairs medical center, albeit one without an LT program, between October 2001 and September 2003. All members of the cohort met the guidelines for referral for possible LT; nevertheless, despite 300 clinical encounters, only 15 patients were eventually referred for evaluation. Even when those patients with a history of recent alcohol use were removed from consideration, the presence of ALD was a significant negative determinant for the consideration of an LT referral. In contrast, in a retrospective study of patients at a community hospital in South Wales, United Kingdom from 1987 to 1990, although ALD was the most common diagnosis among patients who were not referred to an LT unit, continuing drinking was the usual explanation; the writers considered this to be appropriate.13 Similarly, when Veldt et al.8 undertook a prospective assessment of patients admitted to a French inpatient liver unit on account of alcohol-associated liver failure, the combination of death during the initial hospital stay, recovery with abstinence, and alcoholic relapse during the immediate follow-up meant that very few actual transplant candidates emerged. Thus, whether there is an unmet need for transplantation in patients with liver failure due to ALD remains unanswered.
There may also be a lack of recognition of the contribution of excessive alcohol consumption to liver failure in the general community. In their referral population, Day et al.14 identified patients with liver failure in whom alcohol abuse or addiction was not recognized or acknowledged by the referring physicians. Whether this was a tactical omission on the part of the referring physicians or was due to a genuine lack of recognition is not known, but their data underscore the need for vigilance with respect to the diagnosis of alcoholic disorders in the LT population.
The attitude of physicians to alcohol addiction is likely an influence on the referral of ALD patients for LT. With respect to LT, Neuberger et al.15 showed that at least in the United Kingdom, the general public, general practitioners, and gastroenterologists hold pejorative views of patients who suffer from alcohol abuse and dependence.
It is also possible that many primary care providers and community gastroenterologists are confused about when they should refer their ALD patients for an LT evaluation and about whether a specific interval of abstinence is needed before any referral. The role of an interval of sobriety in the process of selecting patients for LT is considered later in this article. In my experience, LT programs prefer for potential patients to be referred rather than triaged in the community. As a rule of thumb, any ALD patient who is failing to improve after 3 months of abstinence is unlikely to improve with medical management and should be considered for referral to a transplant center.8
EVALUATION FOR LT
A comprehensive evaluation of an ALD patient must assess all tissues at risk from alcohol damage. Cardiac function, kidney function, the central and peripheral nervous system, and the immune system are at risk from chronic alcohol abuse. In addition to the standard assessment of liver function and for hepatocellular carcinoma, each system should be studied carefully. The interpretation of data testing the integrity of extrahepatic organ systems is often complicated by competing explanations of abnormal findings. For example, we have difficulty distinguishing hepatic encephalopathy from Wernicke's encephalopathy and painful peripheral neuropathies due to alcohol from those due to other causes (especially diabetes), and the full effects of the alcohol-associated myopathic heart may be masked by the reduced systemic vascular resistance (afterload) common in patients with advanced liver disease.
ALCOHOL USE BY ALD CANDIDATES FOR LT
The evaluation of the patient with ALD differs from the evaluation of other potential candidates for LT because it must take into account the patient's history of addiction, which may include not only alcohol but also nicotine and other addictive drugs either recently or in the past. One important question is whether the potential ALD candidate is drinking now. By and large, only stable alcoholics who are thought to be abstinent are referred to transplant programs.16 Two recent studies have shown that a not inconsiderable proportion of these supposedly abstinent ALD patients who are undergoing an evaluation for LT or are on the waiting list continue to drink.17, 18 This observation underscores the difficulty in establishing accurate data about alcohol use, especially when it is not in the patient's interest to admit to drinking. Indeed, the inhibition of candor militates against the best interests of the patient because it makes it more difficult for the patient to acknowledge a slip and seek help in reestablishing sobriety.19
PROGNOSIS FOR SOBRIETY, SOBRIETY AS PROGNOSIS
Outside the special circumstances of alcoholic hepatitis, which is discussed later, most ALD patients who are likely to be considered for LT have already established an interval of abstinence. There remains controversy about the minimal acceptable interval for LT. In 1997, a consensus conference of the American Association for the Study of Liver Diseases and the American Society of Transplantation concluded that “there is a strong consensus for requiring that most alcoholic patients should be abstinent from alcohol for at least 6 months before they can be listed for liver transplantation.”20 This is often called the 6-month rule.
In 1997, it was declared that the first purpose of the 6-month interval was to allow recovery with medical management from alcohol-induced liver failure. However, since then, the 6-month rule has mostly been discussed as a prognostic tool for predicting subsequent alcoholic relapse (often called recidivism). In the literature on alcoholism, 6 months appears too short to determine meaningful abstinence, and Vaillant21 suggested that sobriety is robust after 5 years. Data from single-center studies (usually retrospective) have yielded conflicting data about whether 6 months of abstinence is predictive of drinking after transplantation.11, 22–25 These data mostly come from studies in which relapse is defined as any use. The circumstances of abstinence are usually not defined, and abstinence due to hospital admissions is not distinguished from abstinence at home. In recent years, the emphasis has shifted to considerations of the pattern of drinking rather than the frequency of any use.26 Harmful or addictive drinking is defined as more than 5 drinks per day for a man, more than 4 drinks per day for a woman, or more than 4 successive days of drinking for either sex. Although a meta-analysis of studies using the any-use standard found that the 6-month rule was independently predictive (albeit weakly) of future drinking along with poor social support and a family history of alcoholism,27 recent studies are more supportive of the association between the duration of abstinence and subsequent harmful drinking.11, 24, 25 At the same time, the 6-month rule has been challenged as limiting the access of appropriate candidates to LT and admitting inappropriate candidates. Yates et al.28 showed in modeling studies that a requirement for 6 months of abstinence would penalize some patients with short-term sobriety who would not relapse in the future.28
In the 1990s, Beresford16 proposed a broad-based examination of psychological health to assess the risk of relapse after LT. Drawing on studies of alcoholics in the nontransplant setting, he described several positive and negative prognostic factors that could aid in the prediction of abstinence after transplantation. The favorable factors include the acknowledgment by the patient of his or her addiction, the presence of strong social support (eg, a spouse, a job, and a home), and 4 prognostic elements identified by Vaillant21 that are also indicative of social integration: substitute activities, a source of improved self-esteem or hope, a rehabilitation relationship, and a perception by the drinker of the negative consequences of alcoholic relapse. The negative prognostic factors are preexisting psychotic disorders, unstable character disorders, unremitting multidrug abuse, repeated and unsuccessful attempts at rehabilitation, and social isolation. In addition, Beresford believed that patients meeting the criteria for dependency are at higher risk than patients with a diagnosis of alcohol abuse; this observation was later confirmed prospectively by DiMartini et al.29 Beresford advocated a comprehensive psychosocial assessment for 2 reasons. First, nearly 80% of his referred patients had good prognosis scores, and second, “the range among the various factors is too wide to justify using any one as a strict inclusion or exclusion criterion.”16 Instead, upon the completion of the patient's assessment, the addiction specialist should be able to give an estimate of the risk of alcoholic relapse and to recommend treatment when it is appropriate. Then, the transplant team is responsible for integrating this prognostic assessment into the comprehensive clinical review in order to determine whether the patient should be placed on the waiting list. Since then, an evaluation by an addiction specialist has become a common feature of LT programs for ALD patients, and subsequent reviews have supported the importance of social integration as a predictor of posttransplant abstinence.11
More recently, De Gottardi et al. reported a retrospective 5-year follow-up study of 387 ALD patients who underwent LT at 2 European centers.24 All were required to have been abstinent for at least 3 months before transplantation, although most had longer intervals of sobriety; 11.7% reported harmful drinking after LT. Three pretransplant factors were predictive of harmful drinking: less than 6 months of abstinence; a diagnosis of anxiety or depression; and positive findings with the High-Risk Alcoholism Relapse scale, a prognostic scale previously described for US patients in the Veterans Affairs health system. The High-Risk Alcoholism Relapse scale counts the years of drinking, the number of drinks per day, and the number of previous inpatient rehabilitations for alcoholism. The strengths of this study included the large size of its cohort, its length of follow-up, and its concentration on harmful drinking rather than slips. This study was also mindful of the importance of psychiatric comorbidities in the assessment, and it was not as subjective as the Beresford score. On the other hand, De Gottardi et al.'s study was retrospective; it included only patients who were accepted for and proceeded to transplantation, it entailed infrequent measures of the use of alcohol after the first year, and it was largely restricted to good-risk candidates.
Because of the difficulty of obtaining accurate histories of alcohol use from patients who might be disadvantaged by candor, an alternative approach is to use the histological findings of the explanted liver as indicators of recent drinking by LT recipients with ALD. Two studies found that although neither the presence of histological alcoholic hepatitis in the explant nor a history of drinking within 6 months correlated with subsequent relapses, posttransplant relapses were correlated with less than 12 months of abstinence.30, 31
Against the background of these conflicting data, the consensus regarding pretransplant abstinence and the suitability of patients with alcoholic hepatitis for transplantation has varied from country to country. France has moved away from the 6-month rule to a comprehensive psychosocial assessment,32 whereas the United Kingdom precludes transplantation in patients with alcoholic hepatitis,33 although this position has been challenged recently.34 In the United States, the evaluation process usually results in the presentation of a comprehensive clinical and psychosocial assessment to the transplant program's selection committee. When the selection committee decides to recommend transplantation, the approval of the third-party payer is necessary before the patient is placed on the LT waiting list. In the case of patients with ALD, this discussion often turns on the issue of the prognosis for sobriety after LT. The 6-month rule has been widely adopted by the US insurance industry. Although there are no published data on the frequency of coverage denial by a third-party payer when the transplant program wants to proceed or on the frequency with which patients die in the aftermath of that decision, occasional anecdotes have been reported.35, 36 Similarly, there are personal reflections by insurance medical officers on the difficulties that they face in such cases.37
Table 1 outlines some of the lessons that can be drawn about the place of sobriety in gauging prognosis and vice versa. It remains true that the severely ill patient who has been drinking recently but has other favorable prognostic indicators with respect to addiction poses a very difficult question for a transplant program.
|1. No single measure is a reliable prognosticator for future relapses into harmful drinking after transplantation.|
|2. With respect to admitting recent use, patients with alcoholism who are under consideration for LT have a conflict of interest.|
|3. Although the duration of abstinence is associated with subsequent drinking, it is an imprecise prognostic tool.|
|4. Liver biopsy provides an unreliable estimate of recent alcohol use.|
|5. A careful evaluation by a trained addiction specialist with a special interest in transplant medicine is very helpful.|
|6. The psychosocial assessment determines lower and higher risks of relapsing into harmful drinking but does not determine the absolute risk.|
|7. The psychological assessment needs to be incorporated into a more comprehensive consideration of the appropriateness of LT for a particular patient.|
|8. The severely ill patient who has been drinking recently but has other favorable prognostic indicators with respect to addiction poses a very difficult question for a transplant program.|
SELECTION OF ALD PATIENTS FOR LT
The plasticity of ALD, particularly in response to abstinence, makes it difficult to accurately determine the prognosis of alcoholic liver injury independently of LT. This assessment is crucial because all LT programs would prefer to avoid transplantation in those patients whose potential for the recovery of liver function is good with abstinence and medical therapies. Mathematical models and a prospective study from France have suggested that only ALD patients with Child-Pugh class C liver failure derive survival benefits from transplantation.38–40 In contrast, a retrospective analysis of the United Network for Organ Sharing database that estimated survival benefits and encompassed survival and mortality before and after transplantation showed that ALD patients with relatively low Model for End-Stage Liver Disease scores in the range of 9 to 11 derived a survival benefit.4 The present system of organ allocation in the United States ensures that LT is limited to ALD patients with severe liver failure or hepatocellular cancer who are highly likely to die without LT.
LT FOR PATIENTS WITH SEVERE ALCOHOLIC HEPATITIS
Patients with severe alcoholic hepatitis present particular challenges to transplant teams because they have invariably consumed alcohol in the previous month. Patients who fail to respond to corticosteroids have a very high 90-day mortality rate. Previously, as in the 1997 American Society of Transplantation/American Association for the Study of Liver Diseases guidelines, alcoholic hepatitis was an absolute contraindication to placement on the transplant waiting list; this is also the position endorsed by the UK Liver Advisory Group.20, 33 However, data are emerging from a European multicenter study about a carefully selected group of patients suffering from their first episode of severe alcoholic hepatitis. These patients, for whom medical treatment had failed but who had received a favorable psychosocial assessment, had excellent intermediate-term survival and a low frequency of harmful drinking after LT.41 Consequently, transplant groups on both sides of the Atlantic have argued for placement on the LT waiting list for the occasional patients with life-threatening alcoholic hepatitis who meet these stringent criteria.34, 42
MANAGEMENT OF ADDICTIONS (INCLUDING ALCOHOLISM) BEFORE LT
There are few published data about treating addictions during the evaluation or waiting-list period. Some programs have recommended that candidates be required to sign a contract to remain abstinent. Many centers encourage attendance at support groups such as Alcoholics Anonymous, but data on the efficacy of contracts or Alcoholics Anonymous in this particular population are lacking. Georgiou et al.43 reported a pilot study of psychosocial interventions in the form of 3 structured meetings during the preparatory phase for LT. Although they demonstrated that a positive reinforcement approach could be feasible, they did not assess the efficacy of such interventions with respect to subsequent alcohol use. A recently published randomized trial of patients who were being evaluated for LT or were awaiting LT at 2 US centers compared the impact of one positive reinforcement technique, motivational enhancement therapy (MET), against advice to attend Alcoholics Anonymous or local counseling, which was called treatment as usual (TAU).19 The study revealed considerable hidden drinking (including harmful drinking) in ALD patients awaiting LT, whereas the effects of MET were modest at best (see Fig. 1).
Although cigarette smoking is widespread in this population, we lack data on efficacious strategies for combating smoking during the evaluation stages before LT.
MORTALITY AND MORBIDITY AFTER LT IN ALD PATIENTS
ALD patients selected for LT in the United States have pretransplant and posttransplant survival rates similar to those of LT recipients without a diagnosis of ALD.2, 4 An analysis of large multicenter databases from the United States and Europe showed greater mortality in patients with comorbid ALD and HCV,2, 4 although this was not found in a single-center series.44 It is possible that the advent of more efficacious therapies for HCV either before or after transplantation will improve the survival of comorbid patients.
Although ALD patients have survival rates similar to those of LT recipients without ALD, the causes of death after transplantation differ between recipients with ALD and recipients without ALD. A retrospective analysis of the European Liver Transplant Registry by Burra et al.2 showed that cardiovascular causes and de novo malignancies were significantly overrepresented in patients who had undergone transplantation for ALD versus recipients without ALD. Similarly, Watt et al.45 showed that in a prospective cohort of 780 primary graft recipients, ALD was significantly associated with the risk of cardiovascular death 1 year after LT.45 Studies from the European Liver Transplant Registry and several single centers suggest that patients who undergo transplantation for ALD have an increased incidence of de novo cancers after transplantation, and these cancers are associated with worse survival.46–49 These studies do not show an association between new-onset cancers and alcoholic relapse. In some but not all of these studies, new tumors were concentrated in the aerodigestive tract. The stratification of cardiovascular deaths and new-onset cancers of the aerodigestive tract in patients receiving LT for ALD strongly hints at a causal linkage with cigarette smoking. Smoking is prevalent in ALD patients undergoing an evaluation for LT, and DiMartini et al.50 showed that recipients of LT for ALD who were smokers before transplantation quickly reestablish smoking at addictive levels. If the link between cigarette smoking and death from either cancer or cardiovascular disease is correct, it points the way to improving posttransplant health through the promotion of smoking cessation in LT recipients with alcoholism and in all LT recipients.
DRINKING RELAPSE AFTER LT
There are large variations in the reported rates of alcoholic relapse after transplantation (approximately 10%-90%).22 These data mostly come from studies that define a relapse as any use instead of distinguishing between occasional lapses or slips and harmful or addictive drinking.26 An analysis of a prospective, longitudinal cohort of patients who underwent transplantation for ALD by DiMartini et al.51 has yielded 5 patterns of alcohol use after transplantation; these include 3 separate patterns of addictive drinking based on the time to relapse and the subsequent course (see Fig. 2). Approximately 80% of the patients either did not drink or consumed only small amounts occasionally. Conversely, among the remaining patients (approximately 20%), there were 3 patterns of harmful drinking. The patterns varied according to the time to relapse and the consumption of alcohol (sustained, heavy use or subsequently modified drinking). These data are similar to retrospective data from Tang et al.,52 who found harmful drinking in 16% of the members of a smaller cohort.52
Data on the consequences of alcoholic relapse have tended to be anecdotal and are often based on retrospective accounts from single centers. These anecdotal reports suggest that patients who relapse into harmful drinking are at risk of alcoholic liver injury, which includes alcoholic hepatitis, delirium tremens, alcoholic pancreatitis, and pneumonia.53–55 Two single-center studies have suggested that relapsing into harmful drinking is associated with reduced survival.25, 56 Furthermore, the causes of death in patients who return to heavy alcohol consumption tend to be liver-related, whereas abstinent patients with ALD die from cardiovascular disease and malignant tumors. These data strongly suggest that controlling heavy drinking is an appropriate goal after LT in ALD patients.
Evidence of alcoholic relapse in conjunction with a failure to take immunosuppression is patchy. Some accounts have suggested an association between alcoholic relapse and so-called noncompliance, whereas others have contended that noncompliance is no more common in these patients than in patients without ALD.53–55, 57 In addition, it is likely that the negative social and familial consequences of abusive drinking found in alcohol-dependent persons will be seen when alcoholic relapse occurs after transplantation, although good data on this point are not readily available. There is a need for a more systematic assessment of the consequences of relapsing into harmful drinking in LT patients.
QUALITY OF LIFE (QOL) AFTER LT
QOL assessments after LT have shown that although quality measures improve in LT patients in most domains in comparison with their status before transplantation, LT recipients continue to have many deficits in comparison with age-matched control populations.58, 59 These observations are drawn from studies with many limitations, such as cross-sectional study designs, inconsistent, nonspecific, or unvalidated study instruments, and short follow-up intervals. Because of the long-term survival of many LT patients, the question remains whether these short-term increments in QOL are sustained or even improved over the longer term. Unfortunately, the reverse may be true. In a 12-year longitudinal follow-up study of 381 LT recipients in 3 US centers who survived for at least 4 years after LT, Ruppert et al.60 found that the early QOL gains gradually diminished over time in all diagnostic groups, including patients with ALD; this was manifested by worsening physical symptoms, fatigue, and a greater sense of being unwell. Neither the presence of the pre-LT diagnosis of ALD nor the use of alcohol or cigarettes after LT was associated with worse long-term QOL. These data on the lack of impact of the diagnosis of ALD on post-LT QOL are in agreement with most but not all previous studies.59 In contrast, Ruppert et al.'s study indicates that the patients with the combined diagnosis of ALD and HCV had worse long-term QOL. This observation of the dual-diagnosis group complements the worse survival outcomes reported for this group.2, 4
TREATMENT OF ADDICTION AFTER LT
Just as there are few studies of therapy for alcoholism before LT, there is a paucity of investigations of the treatment of alcoholism after LT. There are several difficulties to overcome when structured treatment studies in this population are being contemplated. First, patients with alcoholism who undergo LT are probably not homogeneous with respect to the risk of addiction relapse. Many have a strong sense of having recovered from alcoholism, deny cravings, and consequently express little motivation for treatment.61 This is in contrast to patients in alcoholism treatment units, in whom endorsing the treatment model is considered a favorable prognostic indicator. This resistance to treatment may reflect the fear that a declaration of a desire for alcohol will be interpreted by the transplant team as a sign of poor candidacy or a lack of commitment to sobriety. It is likely that some of these responses are genuine and indicate an absence of internal prompts to consume alcohol. A lack of interest in receiving treatment for alcoholism was a reason for the failure to recruit participants for a trial of naltrexone for LT recipients with alcoholism.62 An additional impediment was that LT recipients were unwilling to take a potentially hepatotoxic medication such as naltrexone. In contrast, Björnsson et al.57 introduced into their transplant program a plan for the structured management of alcoholism, which comprised an assessment by a psychiatrist skilled in the care of alcoholics, the initiation of treatment in patients who had not been treated in the past, encouragement to participate in motivational enhancement, and the use of an abstinence contract. The protocol was started before transplantation and continued after transplantation with interviews at 3 months and at 1, 3, and 5 years. In consecutive patients, they observed a reduction in the prevalence of alcohol use in comparison with a matched historical control group (48% versus 22%), although they did not report their data in terms of harmful drinking. Future treatment initiatives should be targeted to the subcohort of ALD patients with persistent cravings, and the goal should be the prevention of harmful drinking (eg, the acceleration in drinking found in some of DiMartini et al.'s cohort51).
Similarly, we need studies designed to enable LT recipients to stop smoking cigarettes. Whether the consumption of alcohol or smoking by the ALD LT recipient is being considered, the treatment under study should not be hepatotoxic, and the endpoint of the study should be control of the addictive behavior rather than patient or graft survival.