The National Institutes of Health consensus development conference on liver transplantation, conducted in Bethesda, MD, in 1983, was an important milestone in the path to recognition of liver transplantation as an appropriate rather than experimental therapy.1 It was also the starting point for a continuing debate on the suitability for, and consequences of liver transplantation for patients with alcoholic liver disease (ALD). This dialogue has evolved over the ensuing 23 years. The recent articles in Liver Transplantation by Pfitzmann et al.2 and DiMartini et al.3 constitute important contributions to this debate; I think they will help further refine our understanding of the central questions about the process of selection for and long-term outcome after transplantation in the ALD population, and ultimately allow a better understanding of where to place liver transplantation in array of treatments available for patients with disorders related to alcohol abuse and addiction.
The consensus view in 1983 was that “only a small proportion of alcoholic patients with liver disease would be expected to meet the rigorous (selection) criteria.”1 This naive view was challenged by Starzl et al.,4 who reported successful transplantation of patients with ALD. In addition, these workers suggested that few ALD patients returned to alcohol use after liver transplantation. Since then, many groups have reported successful transplantation in alcoholics,5 and ALD is one of the most common indications for transplantation in adults. Most studies up to now have shown that patient and allograft outcome are similar in alcoholic and nonalcoholic transplant recipients alike. This orthodoxy is now challenged by Pfitzmann et al.2
Alcoholic relapse, often termed “recidivism,” and assessing the risk of relapse have been concerns from the start. Thomas Beresford,6 an addiction psychiatrist working with the transplant group at the University of Michigan, introduced the concept of psychosocial assessment of ALD transplant candidates. Beresford6 stressed the importance of assessment by an expert in addiction medicine to make an accurate diagnosis of alcohol dependence or abuse and of psychiatric comorbid conditions. Beresford6 alerted the transplant community to the clinical insights into the factors associated with maintaining sobriety reported in the addiction medicine literature. Thus, based on the studies of George Vaillant7 and of Strauss and Bacon,8 he constructed a panel of negative prognostic factors that he used to assess prospective ALD candidates for liver transplantation. The negative prognostic factors included psychiatric comorbid conditions, such as uncontrolled polysubstance abuse or unstable character disorder, a history of many failed rehabilitation attempts, social isolation as shown by lack of employment, absence of a fixed abode, and living alone without a spouse or companion. As a result, psychosocial assessment of candidates for liver transplantation with ALD has become the norm in transplant programs.
At the same time, it has become commonplace, often at the behest of third party payers, to require a fixed interval of abstinence, usually 6 months, as a prerequisite to transplantation. The clinical utility of the “6-month rule” as a instrument to predict future alcoholic relapse remains controversial. Both the Pfitzmann et al.2 and DiMartini et al.3 articles provide useful insights into pretransplant assessment. Pfitzmann et al.'s2 data might be interpreted as supportive of the 6-month rule. In their study, half of the recipients who were abstinent for less than 6 months before transplantation resumed drinking after transplantation, compared with relapse in less than 16% of recipients who were abstinent more than 12 months prior to transplantation. However, the actuality is less clear. In Pfitzmann et al.'s2 study, all recipients had been selected according to criteria that included a requirement for a minimum of 6 months pretransplantation sobriety. Consequently, the 18 of 300 recipients described as having less than 6 months sobriety were presumably thought to have meet the rule at time of transplant. How were the discrepant data on duration of sobriety discovered? This retrospective attribution of duration of sobriety is fraught with the risk of observational bias, in which the patients who relapse receive additional scrutiny about their pretransplantation drinking behavior, whereas those subjects who do not relapse are not questioned as to whether their pretransplantation history was accurate. For this reason, prospective data are more likely to be informative.
DiMartini et al.'s3 recent prospective study provides new information on drinking by alcoholics after transplantation, and on predictive instruments such as pretransplantation sobriety. A total of 167 of a possible 178 consecutive ALD liver transplant recipients at the University of Pittsburgh were recruited into a prospective observational study. Using structured interviews such as the alcohol time-line follow-back questionnaire, the National Institute on Alcohol Abuse and Alcoholism quantity-frequency questionnaire, and an unstructured interview with Dr. DiMartini, a nuanced account of drinking behavior after transplantation was developed, and the 5-year data are now reported. After 5 years had elapsed after transplantation, 42% of alcoholic recipients had used alcohol at least once, and the first drink occurred in the first year in 22%. Furthermore, by 5 years, 26% had drunk in binges (6 drinks a day for men, 4 drinks a day for women), and 22% had drunk on 4 consecutive days.
Similar to the Pfitzmann et al.2 study, all subjects admitted to liver transplantation in Dr. DiMartini's study3 had undergone an extensive pretransplantation psychosocial evaluation. Only 4 recipients had a recorded sobriety of less than 6 months, rendering this cohort unsuitable to assess the 6-month rule. However, the authors assessed whether there was a specific sobriety threshold greater than 6 months that would predict relapse. They calculated receiver operating curves comparing months sober prior to transplantation to relapse, defined as any use, binge, or frequent drinking. The analysis demonstrated that while the sensitivity of prediction increased with longer pretransplantation sobriety, the false-positive rate (1 − specificity) also increased. For example, 36 months of recorded sobriety was 80% sensitive as a predictor of future abstinence, but only 40% specific. In other words, while longer sobriety was associated with less likelihood of drinking, even 3-years abstinence was a poor predictor. These data are supported by Vaillant's9 longitudinal observations in alcoholic men, wherein abstinence was secure only after 5 years.
The most telling observation derived from Pfitzmann et al.'s2 study is that mortality after transplantation is greatly influenced by drinking patterns after transplantation. This important advance, which confirms anecdotal reports or small single-center series that documented alcoholic injury in ALD liver transplant recipients,10–12 expands our understanding because of the cohort size and the clarity of the result. Taken together, these studies challenge the view that relapse to abusive drinking after transplantation is free of harm to the ALD recipient.13
Why has it taken until now to confirm that recurrent drinking by ALD patients after transplantation is harmful? First, it has been necessary to have sufficient duration of follow-up in a large cohort of ALD patients. The median follow-up in Pfitzmann et al.'s2 study was 7.4 yr with a range of 2 to 15 years. Previous studies have tended to have insufficient numbers of patients followed closely for sufficient time.
Perhaps even more important is distinguishing patterns of drinking into “slips” and “harmful drinking.” Pfitzmann et al.2 acknowledge the workshop on “liver transplantation in alcoholic liver disease,” sponsored by the National Institutes of Health in 1997, in which R.K. Fuller14 made a plea to incorporate this nomenclature into liver transplantation assessment. It is only after the stratification of severity of drinking behavior that Pfitzmann et al.2 were able to determine that harmful drinking, as opposed to any alcohol use, by ALD liver transplant recipients, is associated with significant long-term mortality, and that the increased mortality is due to alcoholic damage to the allograft. In the light of Pfitzmann et al.'s2 study, we now have data to help us educate ALD recipients that continued harmful drinking will damage their allograft and lead to death.
A second conclusion, which follows from the adoption of the nomenclature distinguishing slips from harmful drinking, is that slips do not affect mortality, at least in the time frame covered by this study. This is reassuring. A slip can be seen as a lapse in maintaining sobriety. When it is followed by insight and a desire to restore sobriety, it can be a healthy element in regaining sobriety. Indeed Weinrieb et al.15 have argued that transplant teams, in addition to advocating sobriety, should also encourage ALD transplant recipients to come forward to the transplant team after a slip. Weinrieb et al.15 argued that without the reassurance that such an admission will not generate negative consequences (e.g., disapproval by the transplant team, dismissal from the transplant program) ALD patients in transplant programs may fail to get the assistance needed to convert a slip into a valuable experience that initiates sustained abstinence rather than let it deteriorate into harmful drinking.
How do these 2 studies point to the future? First, DiMartini et al.'s3 study reaffirms Beresford's6 advocacy of full psychosocial assessment as part of the pretransplantation evaluation of the ALD candidate, rather than blind application of the 6-month rule. Second, both studies indicate the importance of distinguishing harmful drinking from slips. Furthermore, DiMartini et al.'s3 data suggest that the frequency of both minor and harmful drinking is considerable in the first 5 yr after transplantation even after careful pretransplant assessment. These data mirror the data from a prospective study of ALD patients awaiting transplantation that found any alcohol use by 25%, and harmful drinking by 10% (Weinrieb RM, unpublished observations).
Given that alcohol use by ALD liver transplant patients is more common than previously recognized, that transplant patients have a vested interest in maintaining secrecy about drinking, and that abusive drinking is associated with graft loss and death, it is clear that we in the transplant community need to promote and maintain sobriety by ALD transplant patients, and find ways to restore sobriety in those patients who have relapsed. Consequently, both the studies of Pfitzmann et al.2 and DiMartinin et al.3 are a helpful foundation to treatment of alcoholism in ALD patients before or after transplantation. Unfortunately, this remains a clinical area in which the questions outweigh the answers. Weinrieb et al.16, 17 found that many ALD transplant candidates or recipients are resistant to entering alcoholism treatment programs, often because they perceive a lack of need. Curiously, these patients often claimed to no longer have any craving for alcohol. Therefore, we may need to be more sophisticated in our understanding of alcoholism, and of the mechanisms of treatment in order to affect drinking behavior in the minority of ALD liver transplant recipients who relapse into harmful drinking. Among our goals, I would suggest that we should identify those subjects at greater risk of relapse, devise programs for early recognition and therapeutic intervention in patients who “slip,” and institute alcoholism treatment in patients who resume harmful drinking.