The Alcohol Relapse Risk Assessment:

a scoring system to predict the risk of relapse

to any alcohol use after liver transplant
Context—Alcohol relapse after liver transplant heightens concern about recurrent James R. Rodrigue, PhD,
disease, nonadherence to the immunosuppression regimen, and death. Douglas W. Hanto, MD, PhD,
Objectives—To develop a scoring system to stratify risk of alcohol relapse after
liver transplant.
Michael P. Curry, MD
Beth Israel Deaconess Medical Center
Design—Retrospective medical record review. and the Harvard Medical School,
Setting and Participants—All adult liver transplants performed from May 2002 Boston, Massachusetts (JRR, MPC),
to February 2011 at a single center in the United States. Washington University School of
Main Outcome Measure—The incidence of return to any alcohol consumption Medicine, St Louis, Missouri (MPC)
after liver transplant.
Results—Thirty-four percent (40/118) of patients with a history of alcohol abuse/ Corresponding author: James R.
dependency relapsed to use of any alcohol after liver transplant. Nine of 25 hypothe- Rodrigue, PhD, The Transplant Institute,
sized risk factors were predictive of alcohol relapse after liver transplant: absence Beth Israel Deaconess Medical Center,
110 Francis Street, 7th Floor,
of hepatocellular carcinoma, tobacco dependence, continued alcohol use after liver
Boston, MA 02215
disease diagnosis, low motivation for alcohol treatment, poor stress management
(e-mail: jrrodrig@bidmc.harvard.edu)
skills, no rehabilitation relationship, limited social support, lack of nonmedical behav-
ioral consequences, and continued engagement in social activities with alcohol pres- To purchase electronic or print reprints,
ent. Each independent predictor was assigned an Alcohol Relapse Risk Assessment contact:
(ARRA) risk value of 1 point, and patients were classified into 1 of 4 groups by American Association of Critical-Care Nurses
ARRA score: ARRA I = 0, ARRA II = 1 to 3, ARRA III = 4 to 6, and ARRA IV 101 Columbia, Aliso Viejo, CA 92656
= 7 to 9. Patients in the 2 higher ARRA classifications had significantly higher rates Phone (800) 899-1712 (ext 532) or
of alcohol relapse and were more likely to return to pretransplant levels of drinking. (949) 448-7370 (ext 532)
Conclusion—Alcohol relapse rates are moderately high after liver transplant. The Fax (949) 362-2049
E-mail reprints@aacn.org
ARRA is a valid and practical tool for identifying pretransplant patients with alco-
hol abuse or dependency at elevated risk of any alcohol use after liver transplant.
(Progress in Transplantation. 2013;23:310-318)

©2013 NATCO, The Organization for Transplant Professionals

doi: http://dx.doi.org/10.7182/pit2013604

A pproximately half of all patients presenting for

liver transplant are estimated to have a history of
alcohol abuse or dependency, whether or not alcohol is
Since the early years of liver transplant, the iden-
tification of patients at high risk of relapse to alcohol
use after liver transplant has been of clinical impor-
the primary cause of their liver disease.1,2 Favorable liver tance.5,13,15 Numerous sociodemographic, psychiatric,
transplant outcomes have been reported for patients and social variables have been examined, but few reli-
with alcohol-related liver disease who maintain alco- able predictors have been identified across studies.4,16,17
hol abstinence after transplant.3-8 These outcomes may This lack of reliable predictors makes it difficult to
be attributable, in part, to the careful selection of patients identify high-risk candidates for whom preventive and
thought to be at low risk of alcohol relapse, minimum therapeutic interventions may be useful. Duration of
abstinence duration requirements, provision of sub- abstinence before transplant was thought to be an impor-
stance abuse treatment, and/or regular monitoring via tant predictor as well, which led to the widespread
toxicology testing. Despite such policies and practices, adoption of the so-called “6-month rule.”18 However,
however, relapse to alcohol use after liver transplant short-term abstinence duration has not proven to be a
occurs in one-third to one-half of recipients and height- reliable predictor of posttransplant relapse,8,19-23 and the
ens concern about recurrent disease, nonadherence to 6-month rule has come under considerable scrutiny
the immunosuppression regimen, and death.8-14 and criticism.24-26

310 Progress in Transplantation, Vol 23, No. 4, December 2013

 The Alcohol Relapse Risk Assessment

A clinical tool to guide the alcohol relapse risk assessment questionnaire that is completed annually
assessment during the pretransplant period would help by liver transplant recipients.
to identify patients at heightened risk of posttransplant
relapse and guide the development of program policy Predictor Variables
and clinical interventions. Several researchers have Based on a review of the liver transplant and gen-
advanced the concept of a scoring algorithm for pre- eral addictions literatures, consultation with substance
dicting alcohol relapse in the context of liver trans- abuse researchers and treatment specialists, and our own
plant,5,8,27-29 but these methods have not been widely clinical experiences, we developed a list of 25 dichoto-
evaluated or clinically adopted by most transplant pro- mized variables hypothesized to be associated with
grams. Therefore, the primary objective of this study risk of alcohol relapse after liver transplant (Table 1).
was to conduct a more comprehensive evaluation of We chose to assess 3 different durations of alcohol
pretransplant predictors and to develop a scoring sys- abstinence, based on our program policy (3 months),
tem to stratify relapse risk to any alcohol use after the commonly cited 6-month rule, and the longer
liver transplant. abstinence period (12 months) required by some pro-
grams. Alcohol abuse/dependence, polysubstance abuse/
Methods dependence, and psychiatric comorbid conditions were
Data Collection determined on the basis of the diagnosis given by the
Using a prospectively collected transplant data- evaluating mental health clinician, using established
base (ie, Organ Transplant Tracking Record, OTTR), criteria in the Diagnostic and Statistical Manual of
we conducted a retrospective review of all adult pri- Mental Disorders-IV.31 Nonadherence was defined as
mary liver or liver-kidney transplants at Beth Israel persistent failure to follow the prescribed medication
Deaconess Medical Center in Boston, Massachusetts, regimen, no-show or cancellation (not due to hospital-
from May 1, 2002 to February 28, 2011. A trained cli- ization) of at least 20% of scheduled outpatient clinic
nician who was blinded to the patient’s outcome after appointments, and/or failure to make other prescribed
liver transplant reviewed the pretransplant evaluation lifestyle modifications (eg, fluid or sodium restric-
records and recorded the presence or absence of each tions) as determined by patient self-report or docu-
risk factor (see below) for all patients with any docu- mentation in the medical record. Patients for whom
mented history of alcohol abuse or dependency. To employment consequences, income or housing loss,
assess interrater reliability, the clinician ratings for a loss of marital or committed relationship, or legal prob-
subset (30/118, 25%) of the patients were compared lems (eg, violation of probation) could reasonably be
with the ratings of another clinician unaffiliated with expected with resumption of regular alcohol use were
the liver transplant service. The Committee on Clini- considered to have potentially serious behavioral con-
cal Investigation at Beth Israel Deaconess Medical sequences with relapse.
Center approved the study protocol.
Other Variables
Outcome Variable The following characteristics were recorded from
The primary outcome variable was the incidence the patient’s medical record: age at time of evaluation
of return to any alcohol consumption after liver trans- for liver transplant, sex, race, cause of primary liver
plant. Much of the prior research on alcohol relapse disease, presence of hepatocellular carcinoma, time
after transplant has focused on “harmful” drinking, (days) to alcohol relapse after liver transplant, and sur-
which has been inconsistently defined and poorly oper- vival status (alive, dead) and duration (days) at time of
ationalized. For instance, Kelly et al8 defined harmful study initiation. Additionally, for alcohol relapsers,
drinking as consuming more than 140 g of ethanol a we coded the intensity of the relapse as follows: low
week or drinking with noted medical or social harm. (≥1 slip to alcohol use that was isolated and followed
Perney et al30 reported harmful use as more than 21 by an extended period of abstinence), moderate (return
units per week for males and 14 units per week for to daily or near daily alcohol use, but at amounts less
females, and Mackie et al10 described heavy use as than levels before liver transplant), and high (return to
more than 50 units per week. Clinically, however, our alcohol use at frequency or amounts at or higher than
program uses a lifetime abstinence model of care, in levels before liver transplant). This information was
which patients are educated, counseled, and expected ascertained from clinic notes written by members of
to refrain from any alcohol consumption after trans- the transplant team, reports by caregivers to transplant
plant. For our study, any documented report or evi- providers who then documented this information in
dence of alcohol consumption after liver transplant our clinical database (ie, the OTTR), or patients’ dis-
was recorded. We reviewed all outpatient and inpa- closure on our annual transplant recipient health assess-
tient clinic notes, our clinical database (ie, the OTTR), ment in which patients are asked to describe the amount
toxicology results, and our transplant center’s health and frequency of alcohol use.

Progress in Transplantation, Vol 23, No. 4, December 2013 311

Rodrigue et al

Table 1 Hypothesized predictor (risk) variables before liver transplant

Alcohol as primary cause of disease

Presence of hepatocellular carcinoma
Abstinence duration before listing for liver transplant
<3 months
<6 months
<12 months
Grams per day of alcohol before cessation (>142 g, median value for our sample)
History of alcohol dependence
History of polysubstance abuse or dependence
Tobacco dependence at time of evaluation before liver transplant
Alcohol relapse since first attempt to quit (>1)
Continued alcohol use after initial diagnosis of liver disease
Prior unsuccessful intervention before evaluation for liver transplant (≥1 relapse within 6 months of treatment initiation)
Immediate family (origin) member(s) with history of alcohol abuse or dependence
Reluctance to participation in alcohol relapse prevention treatment as condition of transplant listing (low motivation)
Unable or unwilling to accept alcohol abuse/dependence diagnosis (poor insight)
Persistent cravings for alcohol
Unable to identify 1 or more substitute activities
Shows little or no source of hope or improved self-esteem
Poor or very limited stress management skills
Presence of psychiatric comorbid conditions
Nonadherence with management of chronic liver disease
Lack of a rehabilitation relationship (eg, sober sponsor, family member, or friend)
No, or very limited, identified social supports
Active substance abuse or dependence in spouse/partner or identified primary caregiver
Lives alone
Lacks negative, nonmedical behavioral consequences if returns to alcohol use
Continued participation in social activities in which alcohol is readily available and/or consumed

Statistical Analysis on the basis of the total score. Relapse-free survival

Continuous variables are presented as means and curves by ARRA risk category were computed by
standard deviations, and t tests were used to examine using Kaplan-Meier methods and were compared by
differences between relapsers and nonrelapsers. Either using log rank tests. Finally, a receiver operating char-
χ2 or Fisher exact tests were used to identify differ- acteristic curve and the area under the curve were cal-
ences between these 2 groups of patients on categori- culated to assess the discriminative ability of the
cal variables. Variables with P less than .05 in the ARRA score. The area under the curve is interpreted
univariate analysis were included in a stepwise logis- as the probability that a randomly selected patient who
tic regression analysis to identify predictors of relapse relapsed to alcohol use after liver transplant will have
to any alcohol use after liver transplant. Only those a higher ARRA score than a randomly chosen patient
variables with an adjusted P less than .05 were included who did not relapse. Generally, values of 0.7 and
in the final model. Cohen kappa was used to assess higher for area under the curve are considered clini-
interrater agreement on coding of both predictor and cally meaningful. PASW 17.0 (IBM Corporation) was
outcome variables. used for all statistical analyses.
A clinical score (ie, Alcohol Relapse Risk Assess-
ment, or ARRA, score) was constructed on the basis Results
of the final logistic regression model. One point was Recipient Characteristics
assigned for the presence of each predictor, or risk fac- Two hundred forty-three primary transplants (223
tor. Patients were assigned to 4 ARRA risk categories liver, 20 liver-kidney) were performed during the study

312 Progress in Transplantation, Vol 23, No. 4, December 2013

 The Alcohol Relapse Risk Assessment

Adult liver transplants between May 2002 and

February 2011, N = 243
Liver, n = 223
Liver-kidney, n = 20
Excluded: No alcohol abuse or dependence history,
n = 105

Alcohol abuse or dependence history, n = 270

Not included: Died ≤6 months after transplant or

never left rehabilitation center/hospital, n = 20

Total cases included in primary

statistical analyses, n = 118

Figure 1 Overview of patient identification and inclusion in final data analysis.

observation period (Figure 1). One hundred thirty- categorical coding of predictor and outcome variables,
eight patients (57%) had a history of alcohol abuse or which suggests substantial agreement.
dependence, although 20 patients were excluded from
analysis because they died less than 6 months after Relapse to Alcohol After Liver Transplant
transplant or never left the rehabilitation center or hos- Forty patients (34%) relapsed to any alcohol use
pital before dying. Of the remaining 118 patients, 101 (Figure 2). The time from liver transplant surgery to
(86%) were male, 99 (84%) were white, and mean age alcohol relapse ranged from 60 days to 7 years
was 54.7 (SD, 8) years. One hundred six (90%) received (median, 625 days). Relapsers did not differ sig-
a liver transplant only, 69 (58%) had hepatitis C virus nificantly (P > .05) from nonrelapsers with respect to
infection, 45 (38%) had alcohol as the primary cause sex, race, age, MELD score at the time of liver trans-
of liver disease, and 48 (41%) had hepatocellular car- plant, or survival status. However, compared with
cinoma. Mean laboratory Model for End-Stage Liver nonrelapsers, relapsers had significantly shorter
Disease (MELD) score at the time of liver transplant duration of abstinence before listing for liver trans-
was 22.0 (SD, 11; range, 7-48) and mean MELD score plant (mean [SD], 102.5 [118.0] months vs 27.7 [55.3]
with exception points at time of liver transplant was months, P < .001), were more likely to have alcohol
28.2 (SD, 7; range, 9-40). Mean duration of follow-up as their primary cause of disease (28% vs 58%, P =
after liver transplant was 55 months. .003), and were less likely to have hepatocellular car-
cinoma (54% vs 15%, P < .001). Of the 40 relapsers,
Interrater Agreement 9 (22%) had a low-intensity relapse, 19 (48%) had a
Examination of interrater agreement yielded a moderate-intensity relapse, and 12 (30%) had high-
Cohen kappa of 0.80 (95% CI, 0.72-0.97) across the intensity relapses (Figure 2).

60

50
Percentage of patients

40

30

20

10

0
Relapsers Low Moderate High
Relapse intensity

Figure 2 Percentage of patients who relapsed to any alcohol use after liver transplant and the intensity of the relapse (low, moderate, high).

Progress in Transplantation, Vol 23, No. 4, December 2013 313

Rodrigue et al

Table 2 Risk factors before liver transplant significantly associated with alcohol relapse after liver transplant in univariate analysis

No. (%) of patients

Relapsers Nonrelapsers
Predictor variable (n = 40) (n = 78) P
Alcohol as primary cause of disease 23 (58) 22 (28) .003
Presence of hepatocellular carcinoma 6 (15) 42 (54) <.001
Abstinence duration before listing for liver transplant
<6 months 18 (45) 14 (18) .004
<12 months 27 (68) 22 (28) <.001
Tobacco dependence at time of evaluation before liver transplant 24 (60) 26 (33) .006
Continued alcohol use after initial diagnosis of liver disease 25 (63) 18 (23) <.001
Prior unsuccessful intervention before evaluation for liver transplant (≥1 relapse
within 6 months of treatment initiation) 33 (83) 44 (66) .007
Reluctance to participate in alcohol relapse prevention treatment as condition of
transplant listing (low motivation) 25 (63) 17 (22) <.001
Unable or unwilling to accept alcohol abuse/dependence diagnosis (poor insight) 19 (48) 9 (12) <.001
Unable to identify 1 or more substitute activities 31 (78) 17 (22) <.001
Shows little or no source of hope or improved self-esteem 9 (23) 5 (6) .02
Poor or very limited stress management skills 35 (88) 24 (31) <.001
Nonadherence with management of chronic liver disease 10 (25) 7 (9) .03
Lack of a rehabilitation relationship (eg, sober sponsor, family member, or friend) 30 (75) 31 (40) <.001
No, or very limited, identified social supports 17 (43) 16 (21) .02
Lacks negative, nonmedical behavioral consequences if returns to alcohol use 32 (80) 9 (12) <.001
Continued participation in social activities in which alcohol is readily available
and/or consumed 37 (93) 16 (21) <.001

Univariate and Multivariate Analyses for ARRA II (43%), 4 to 6 for ARRA III (36%), and
In univariate analysis, 16 variables met the inclu- 7 to 9 for ARRA IV (14%). The ARRA score was pre-
sion criteria and were associated with a higher inci- dictive of relapse to any alcohol use after liver trans-
dence of alcohol relapse after liver transplant (Table 2). plant (log rank χ2 = 57.9, P < .001). The alcohol relapse
In the subsequent multivariate prediction model, 9 rates were 0% for the ARRA I, 8% for the ARRA II,
variables were statistically significant (Table 3), explain- 57% for the ARRA III, and 75% for the ARRA IV
ing 73% of the variance in the outcome (P < .001) and group (P < .001). Also, ARRA classification was asso-
correctly classifying the outcome of 91% of patients. ciated with relapse intensity for those who returned to
The model had a positive predictive value of 87%, alcohol use (χ2 = 15.7, P = .003). Low- and moderate-
correctly classifying 34 of 39 patients it predicted would intensity relapsers were more likely to be in ARRA III
relapse to any alcohol use after liver transplant. The classification, whereas high-intensity relapsers were
negative predictive value was 92%, where the model more likely to be in the ARRA IV group (Figure 3).
correctly classified 73 out of 79 patients it predicted The area under the receiver operating characteristic
would have no relapse to alcohol use. curve was 0.892 (95% CI, 0.833-0.950) for the simpli-
fied ARRA total score.
ARRA Prediction Score
An ARRA prediction score was calculated by Discussion
adding points assigned to each of 9 statistically signif- In this study, we present a new scoring system
icant predictor variables in the final regression model. (ARRA) based on 9 pretransplant clinical parameters,
For the purpose of simplicity, and pending validation which may help to identify before liver transplant those
of these study findings, we assigned 1 point to each patients who are at high risk of relapsing to alcohol
variable despite the difference in regression coeffi- use after transplant. The ARRA score allowed patients
cients. The resulting ARRA score ranged from 0 to 9. to be classified into 4 groups with minimal (0/9, 0%),
Four groups of patients were defined on the basis of low (4/51, 8%), moderate (24/42, 57%), and high
the ARRA score: score of 0 for ARRA I (8%), 1 to 3 (12/16, 75%) risk of relapse to any alcohol use after

314 Progress in Transplantation, Vol 23, No. 4, December 2013

 The Alcohol Relapse Risk Assessment

Table 3 Multivariate logistic regression model for prediction of relapse to any alcohol use after liver transplant
Predictor variable Regression coefficient Odds ratio (95% CI) P
Presence of hepatocellular carcinoma -1.89 0.15 (0.06, 0.40) <.001
Tobacco dependence 2.24 2.46 (1.18, 10.65) .01
Continued alcohol use after liver disease diagnosis 1.80 1.79 (1.13, 3.27) <.001
Low motivation for relapse prevention treatment 1.62 1.59 (1.06, 2.41) .02
Poor stress management skills 1.31 3.61 (1.09, 14.12) .049
Lack of a rehabilitation relationship 1.67 2.09 (1.13, 4.65) .04
Limited social supports 1.59 3.02 (1.72, 10.19) .03
Lacks nonmedical behavioral consequences 1.89 6.15 (1.23, 18.42) .01
Continued engagement in social activities with alcohol present 2.31 8.77 (2.01, 42.17) .004

liver transplant. Supporting the clinical utility of the abuse patterns, the acquisition of more effective cop-
ARRA score is an area under the curve of 0.892, which ing and stress management skill, the identification of
suggests that if a relapser and nonrelapser were both high-risk social situations (eg, social activities involv-
drawn at random, the probability that the patient who ing alcohol), and strategies to avoid or manage such
relapsed to alcohol use would have a higher ARRA situations effectively. Additionally, treatment efforts
score is 89.2%. focused on identifying the relative benefits and rewards
The ARRA score is based on 9 parameters, includ- of continued alcohol abstinence after liver transplant
ing some that have previously been identified as may improve the patient’s motivation for sustained
correlates or predictors of alcohol relapse after trans- lifestyle change after transplant. Of particular note, cig-
plant.4,5,8,11,12,19-22,27,32,33 Importantly, these predictors seem arette smoking commonly co-occurs with alcohol abuse
logical and clinically meaningful. All but 2 of the pre- or dependence,35 and smokers appear to be at higher
dictors (presence/absence of hepatocellular carcinoma, risk of relapse after treatment for alcohol depend-
continued alcohol use after liver disease diagnosis) are ence.8,36,37 For exsmokers, the development of effective
modifiable, which suggests that risk factors can be coping skills to prevent relapse and the positive rein-
attenuated or reduced with appropriate intervention forcement associated with smoking cessation may be
before liver transplant. Many active alcohol relapse beneficial in facilitating longer-term abstinence from
prevention programs (eg, Motivational Enhancement alcohol. In contrast, continued smoking may, over
Therapy)34 target greater insight into one’s substance time, contribute to cue-induced craving for alcohol,

100
90
80
Percentage of patients

70
60
50
40
30
20
10
0
Low Moderate High
Intensity

ARRA II ARRA III ARRA IV

Figure 3 The association between Alcohol Relapse Risk Assessment (ARRA) classification and alcohol relapse intensity. None of
the patients in the ARRA I classification had an alcohol relapse. ARRA II = 1-3 points, ARRA III = 4-6 points, ARRA IV = 7-9 points.

Progress in Transplantation, Vol 23, No. 4, December 2013 315

Rodrigue et al

especially in the context of psychological distress. Many other variables, a finding consistent with several
liver transplant programs strongly advise or require other studies.8,10,12,14,22,45,46 The recent study by Mathurin
patients to quit smoking before transplant surgery to et al47 showing that early liver transplant for those
decrease the risk of adverse health risks.38-40 Our data with acute alcoholic hepatitis yields favorable out-
suggest that patients with an alcohol history should be comes and low short-term relapse rates further calls
offered interventions to facilitate smoking cessation into the question the need for a specified abstinence
(eg, online programs, self-help groups, behavioral and period before transplant.
pharmacological interventions), as this may also atten- Findings should be evaluated within the context
uate the risk of alcohol relapse after liver transplant. of several important study limitations. First, this is a
One interesting finding is that having hepatocel- single-center study, and findings cannot be general-
lular carcinoma was associated with a lower likeli- ized beyond the characteristics of our particular
hood of relapse. Research has shown that adults who sample (eg, predominantly male, white, and having
believe an unhealthy lifestyle contributed to a cancer infection with hepatitis C virus as the primary diagno-
diagnosis are more likely to make and maintain lifestyle sis). The ARRA scoring system requires external val-
changes.41-44 It is possible that the heightened sense of idation by researchers in other transplant programs
urgency, higher risk of mortality, and added uncer- and with different populations of patients. Such vali-
tainty of cancer progression and possible removal from dation is especially important given the inherent selec-
the transplant list before liver transplant contribute to tion bias in a study of this nature, where it is possible
a greater commitment to maintaining an alcohol-free that our program’s selection criteria for liver trans-
lifestyle after transplant. Based on their understanding plant most likely excluded patients who were thought
or representation of hepatocellular carcinoma, some to be at the highest risk of alcohol relapse after liver
liver transplant recipients may maintain alcohol absti- transplant. Second, it is possible that our alcohol relapse
nence because they believe doing so will prevent can- rates are higher than reported in this study, because
cer recurrence. Further study of cancer-specific beliefs our data collection processes may not have detected
and their association with other lifestyle changes, in some patients who returned to drinking. Similarly, the
addition to alcohol abstinence, is warranted. length of time from liver transplant to return to drink-
Several variables were associated with alcohol ing may have been shorter than described because
relapse in the univariate analysis but were not signifi- some patients may have successfully concealed their
cant predictors in the multivariate analysis, including drinking from members of the transplant team, other
cause of primary disease, abstinence duration before health providers, and family members. Third, the
transplant listing, prior unsuccessful alcohol treatment, number of predictor variables is large and the number
poor insight, and nonadherence with other aspects of of events (ie, relapse) is small, possibly resulting in
liver disease management. Many other studies have overfitting of the model. Also, we may have overlooked
focused exclusively on evaluating relapse rates in other variables that may be important to consider in
patients with alcoholic liver disease as the primary assessing risk to alcohol relapse after transplant, includ-
indication for liver transplant. Indeed, we found that ing genetic markers.48
those with alcoholic liver disease had a higher rate of The aforementioned limitations notwithstanding,
alcohol relapse and a shorter duration of pretransplant this study has several notable strengths. For instance,
abstinence compared with those with other primary it includes all patients with a history of alcohol abuse
indicators for transplant. However, patients with any or dependence, not only those with alcoholic liver dis-
diagnosis may have a history of alcohol abuse or ease. This factor is clinically important because all
dependency, and decisions about substance abuse pol- patients with a known history of alcohol abuse, regard-
icy or practices should not be made on the basis of dis- less of the cause of their liver disease, are advised to
ease etiology alone. Indeed, those for whom alcohol adopt a lifetime abstinence model to reduce the risk of
was not the primary indicator for liver transplant con- liver graft injury, to ensure adherence to the immuno-
sumed as much alcohol per day (224 g vs 205 g, P = suppression regimen, and to optimize the patient’s over-
.56) as those with alcoholic liver disease. In our own all health and well-being. Moreover, including patients
program, we require a minimum of 3 months of out-of- with all indications for liver transplant broadens the
hospital abstinence before a patient, including those potential utility of the ARRA scoring system. Another
without alcoholic liver disease, with any history of sub- relative strength is that our primary outcome was any
stance abuse can be listed for transplant. Most patients alcohol use after liver transplant, regardless of amount
(89%) in our study met this criterion; however, 13 or level of harm potential as has been previously
patients were abstinent for less than 3 months and 6 reported.8 The ARRA classification could be used to
patients (46%) relapsed to alcohol use after transplant. distinguish those patients with high-intensity relapses
Regardless, it appears that abstinence duration is a less from those who had an isolated slip, thus suggest-
salient indicator of relapse after liver transplant than ing that the scoring system may have some utility in

316 Progress in Transplantation, Vol 23, No. 4, December 2013

 The Alcohol Relapse Risk Assessment

identifying patients at risk of more harmful levels of 2. McCullough AJ. Alcoholic liver disease. In: Schiff ER, Sorrell
MF, Maddrey WC, eds. Schiff’s Diseases of the Liver. 8th ed.
drinking. This finding warrants replication in a larger Philadelphia, PA: Lippincott Williams & Wilkins; 1999:941-971.
cohort of patients because of the small cell sizes in the 3. Bhagat V, Mindikoglu AL, Nudo CG, Schiff ER, Tzakis A,
present study. Finally, we used multiple strategies to Regev A. Outcomes of liver transplantation in patients with
cirrhosis due to nonalcoholic steatohepatitis versus patients with
identify patients who had returned to any alcohol use, cirrhosis due to alcoholic liver disease. Liver Transpl. 2009;
including self-reports, caregiver reports, confirmed 15(12):1814-1820.
clinical suspicion, medical records and hospital notes, 4. Biselli M, Gramenzi A, Del Gaudio M, et al. Long-term fol-
low-up and outcome of liver transplantation for alcoholic liver
and toxicology results. disease: a single center case-control study. J Clin Gastroenterol.
Clinically, we recommend a strategy for risk strat- 2010;44(1):52-57.
ification of patients with a history of alcohol abuse or 5. Lucey MR, Merion RM, Henley KS, et al. Selection for and
outcome of liver transplantation in alcoholic liver disease. Gas-
dependence being evaluated for liver transplant and a troenterology. 1992;102(5):1736-1741.
tailoring of alcohol treatment on the basis of risk level. 6. Watt KD, McCashland TM. Transplantation in the alcoholic
Our relapse rate of 34%, while consistent with the rates patient. Semin Liver Dis. 2004;24(3):249-255.
7. Björnsson E, Olsson J, Rydell A, et al. Long-term follow-up
reported in other studies,8 is higher than we expected. of patients with alcoholic liver disease after liver transplanta-
Historically in our program, all patients with a sub- tion in Sweden: impact of structured management on recidivism.
stance abuse history have been referred for relapse Scand J Gastroenterol. 2005;40(2):206-216.
8. Kelly M, Chick J, Gribble R, et al. Predictors of relapse to
prevention treatment, but we have not closely moni- harmful alcohol after orthotopic liver transplantation. Alcohol
tored the frequency and intensity of treatment and have Alcohol. 2006;41(3):278-283.
not distinguished between those at low versus high risk 9. Pageaux GP, Bismuth M, Perney P, et al. Alcohol relapse after
liver transplantation for alcoholic liver disease: does it matter?
of relapse. The ARRA risk stratification offers the pos- J Hepatol. 2003;38(5):629-634.
sibility of stratifying patients on the basis of risk level 10. Mackie J, Groves K, Hoyle A, et al. Orthotopic liver trans-
for relapse and prescribing relapse prevention treat- plantation for alcoholic liver disease: a retrospective analysis
of survival, recidivism, and risk factors predisposing to recidi-
ment on the basis of such risk. Additionally, because vism. Liver Transpl. 2001;7(5):418-427.
two-thirds of alcohol relapses occurred more than 1 11. DiMartini A, Day N, Dew MA, et al. Alcohol consumption
year after liver transplant, programs should consider patterns and predictors of use following liver transplantation
for alcoholic liver disease. Liver Transpl. 2006;12(5):813-820.
recommending booster sessions for those who are no 12. Lucey MR, Carr K, Beresford TP, et al. Alcohol use after liver
longer in treatment after transplant. transplantation in alcoholics: a clinical cohort follow-up study.
Predicting relapse to alcohol use after liver trans- Hepatology. 1997;25(5):1223-1227.
13. Bird GL, O’Grady JG, Harvey FA, Calne RY, Williams R.
plant is clinically important and represents a ripe area Liver transplantation in patients with alcoholic cirrhosis: selec-
for scientific inquiry. Future studies should validate tion criteria and rates of survival and relapse. BMJ. 1990;
the predictive utility of the ARRA model with other 301(6742):15-17.
14. Gish RG, Lee A, Brooks L, Leung J, Lau JY, Moore DH 2nd.
patient populations at other transplant centers. Other Long-term follow-up of patients diagnosed with alcohol depend-
recommended areas of study include assessing inter- ence or alcohol abuse who were evaluated for liver transplan-
rater agreement among liver transplant team members tation. Liver Transpl. 2001;7(7):581-587.
15. Beresford TP. Psychiatric follow-up care of alcohol-dependent
in classifying patients based on ARRA risk score, liver graft recipients. In: MR Lucey, RM Merion, TP Beres-
assessing the utility of the ARRA score in predicting ford, eds. Liver Transplantation and the Alcoholic Patient:
relapse to other substances of abuse, and determining Medical, Surgical and Psychosocial Issues. New York, NY:
Cambridge University Press; 1994:96-112.
whether specific interventions can mitigate the risk 16. Anantharaju A, Van Thiel DH. Liver transplantation for alco-
level of patients and reduce the incidence of alcohol holic liver disease. Alcohol Res Health. 2003;27(3):257-268.
17. Neuberger J, Webb K. Liver transplantation for alcoholic liver
relapse after transplant. disease: knowing the future informs the present. Am J Trans-
plant. 2010;10(10):2195-2196.
Acknowledgments 18. Lucey MR, Brown KA, Everson GT, et al. Minimal criteria for
placement of adults on the liver transplant waiting list: a report
The authors thank Timothy Antonellis, Ariel Hodara, of a national conference organized by the American Society of
Richard McCartney, Matthew Paek, Stacey Senat, and Transplant Physicians and the American Association for the
Hongying Tang for the data collection and entry assistance. Study of Liver Diseases. Liver Transpl Surg. 1997;3(6):628-637.
19. DiMartini A, Dew MA, Day N, et al. Trajectories of alcohol
consumption following liver transplantation. Am J Transplant.
Financial Disclosures 2010;10(10):2305-2312.
This study was funded, in part, by the Julie Henry 20. Karim Z, Intaraprasong P, Scudamore CH, et al. Predictors of
Research Fund and the Center for Transplant Outcomes and relapse to significant alcohol drinking after liver transplanta-
tion. Can J Gastroenterol. 2010;24(4):245-250.
Quality Improvement, The Transplant Institute, Beth Israel 21. Platz KP, Mueller AR, Spree E, et al. Liver transplantation for
Deaconess Medical Center, Boston, Massachusetts. alcoholic cirrhosis. Transpl Int. 2000;13(suppl 1):S127-S130.
22. Jauhar S, Talwalkar JA, Schneekloth T, Jowsey S, Wiesner RH,
Menon KV. Analysis of factors that predict alcohol relapse
References following liver transplantation. Liver Transpl. 2004(3);10:
1. Durning PE, Perri MG. Substance abuse and transplantation. 408-411.
In: Rodrigue JR, ed. Biopsychosocial Perspectives on Trans- 23. Foster PF, Fabrega F, Karademir S, Sankary HN, Mital D,
plantation. New York, NY: Kluwer; 2001:125-149. Williams JW. Prediction of abstinence from ethanol in alcoholic

Progress in Transplantation, Vol 23, No. 4, December 2013 317

Rodrigue et al

recipients following liver transplantation. Hepatology. 1997; evidence for beneficial carry-over effects of abstinence from
25(6):1469-1477. smoking on alcohol use and vice versa. Addict Behav. 2007;
24. Brown RS Jr. Transplantation for alcoholic hepatitis--time to 32(12):3093-3100.
rethink the 6-month “rule.” N Engl J Med. 2011;365(19):1836- 37. Kohn CS, Tsoh JY, Weisner CM. Changes in smoking status
1838. among substance abusers: baseline characteristics and absti-
25. Shawcross DL, O’Grady JG. The 6-month abstinence rule in nence from alcohol and drugs at 12-month follow-up. Drug
liver transplantation. Lancet. 2010;376(9737):216-217. Alcohol Depend. 2003;69(1):61-71.
26. Beresford TP, Everson GT. Liver transplantation for alcoholic 38. Herrero JI, Pardo F, D’Avola D, et al. Risk factors of lung,
liver disease: bias, beliefs, 6-month rule, and relapse—but head and neck, esophageal, and kidney and urinary tract car-
where are the data? Liver Transpl. 2000;6(6):777-778. cinomas after liver transplantation: the effect of smoking with-
27. De Gottardi A, Spahr L, Gelez P, et al. A simple score for pre- drawal. Liver Transpl. 2011;17(4):402-408.
dicting alcohol relapse after liver transplantation: results from 39. Mathur AK, Ranney DN, Patel SP, et al. The effect of smoking
387 patients over 15 years. Arch Intern Med. 2007;167(11): on biliary complications following liver transplantation. Transpl
1183-1188. Int. 2011;24(1):58-66.
28. Beresford TP. Alcohol abuse and liver transplantation. In: J 40. Leithead JA, Ferguson JW, Hayes PC. Smoking-related morbid-
Craven, GM Rodin, eds. Psychiatric Aspects of Organ Trans- ity and mortality following liver transplantation. Liver Transpl.
plantation. New York, NY: Oxford University Press; 1992: 2008;14(8):1159-1164.
22-32. 41. Rabin C, Pinto B. Cancer-related beliefs and health behavior
29. Yates WR, Martin M, LaBrecque D, Hillebrand D, Voigt M, change among breast cancer survivors and their first-degree
Pfab D. A model to examine the validity of the 6-month relatives. Psychooncology. 2006;15(8):701-712.
abstinence criterion for liver transplantation. Alcohol Clin Exp 42. Pinto BM, Eakin E, Maruyama NC. Health behavior changes
Res. 1998;22(2):513-517. after a cancer diagnosis: what do we know and where do we
30. Perney P, Bismuth M, Sigaud H, et al. Are preoperative pat- go from here? Ann Behav Med. 2000;22(1):38-52.
terns of alcohol consumption predictive of relapse after liver 43. Satia JA, Campbell MK, Galanko JA, James A, Carr C, San-
transplantation for alcoholic liver disease? Transpl Int. 2005; dler RS. Longitudinal changes in lifestyle behaviors and health
18(11):1292-1297. status in colon cancer survivors. Cancer Epidemiol Biomark-
31. American Psychiatric Association. Diagnostic and Statistical ers Prev. 2004;13(6):1022-1031.
Manual of Mental Disorders. 4th ed. Washington, DC: Amer- 44. Blanchard CM, Denniston MM, Baker F, et al. Do adults change
ican Psychiatric Association; 1994. their lifestyle behaviors after a cancer diagnosis? Am J Health
32. Hartl J, Scherer MN, Loss M, et al. Strong predictors for alco- Behav. 2003;27(3):246-256.
hol recidivism after liver transplantation: non-acceptance of the 45. Weinrieb RM, Van Horn DH, McLellan AT, Lucey MR.
alcohol problem and abstinence of <3 months. Scand J Gas- Interpreting the significance of drinking by alcohol-dependent
troenterol. 2011;46(10):1257-1266. liver transplant patients: fostering candor is the key to recovery.
33. McCallum S, Masterton G. Liver transplantation for alcoholic Liver Transpl. 2000;6(6):769-776.
liver disease: a systematic review of psychosocial selection cri- 46. Beresford TP. Predictive factors for alcoholic relapse in the
teria review. Alcohol Alcohol. 2006;41(4):358-363. selection of alcohol-dependent persons for hepatic transplant.
34. Weinrieb RM, Van Horn DH, Lynch KG, Lucey MR. A ran- Liver Transpl Surg. 1997;3(3):280-291.
domized, controlled study of treatment for alcohol dependence 47. Mathurin P, Moreno C, Samuel D, et al. Early liver transplanta-
in patients awaiting liver transplantation. Liver Transpl. 2011; tion for severe alcoholic hepatitis. N Engl J Med. 2011;365(19):
17(5):539-547. 1790-1800.
35. Little HJ. Behavioral mechanisms underlying the link between 48. Dahlgren A, Wargelius HL, Berglund KJ, et al. Do alcohol-
drinking smoking and drinking. Alcohol Res Health. 2000; dependent individuals with DRD2 A1 allele have an increased
24(4):215-224. risk of relapse? A pilot study. Alcohol Alcohol. 2011;46(5):
36. Hintz T, Mann K. Long-term behavior in treated alcoholism: 509-513.

318 Progress in Transplantation, Vol 23, No. 4, December 2013