Alcohol recidivism impairs long-term patient survival after orthotopic liver transplantation for alcoholic liver disease


  • Antonio Cuadrado,

    1. Gastroenterology and Hepatology Unit, University Hospital “Marqués de Valdecilla,” Faculty of Medicine, Santander, Spain
    Search for more papers by this author
  • Emilio Fábrega,

    Corresponding author
    1. Gastroenterology and Hepatology Unit, University Hospital “Marqués de Valdecilla,” Faculty of Medicine, Santander, Spain
    • Gastroenterology and Hepatology Unit, University Hospital “Marqués de Valdecilla,” Faculty of Medicine, Avenida Valdecilla s/n, Santander 39008 Spain
    Search for more papers by this author
    • Telephone: 073442202544; FAX: 073442202544

  • Fernando Casafont,

    1. Gastroenterology and Hepatology Unit, University Hospital “Marqués de Valdecilla,” Faculty of Medicine, Santander, Spain
    Search for more papers by this author
  • Fernando Pons-Romero

    1. Gastroenterology and Hepatology Unit, University Hospital “Marqués de Valdecilla,” Faculty of Medicine, Santander, Spain
    Search for more papers by this author


The aim of this study was to evaluate the rate of alcohol recidivism after orthotopic liver transplantation (OLT) for alcoholic liver disease (ALD) and its influence on the allograft and patient survival, as well as the development of comorbidities and de novo cancers. The study was performed on 54 subjects previously analyzed and transplanted in our center for ALD, whose follow-up was prolonged to a mean of 99.2 (SD 31.7) months (range, 14–155). Medical records were reviewed, and data on alcohol consumption, therapeutic compliance, graft evolution, rejection, infections, comorbidities, rates of de novo malignancies and other clinical events, and survival were collected. Comparisons between groups were performed by the Fisher's exact test, and survival was assessed by the Kaplan-Meier method. Survival curves were compared using the Mantel-Cox statistic. The risk of death resulting from alcohol recidivism was analyzed with a Cox proportional hazards model. Fourteen patients who underwent transplantation for ALD (25.9%) returned to alcohol use between 5.0 and 86.9 months after OLT (median, 47.5). There was no significant association between the presence or absence of alcohol recidivism and the occurrence of graft rejection, infections, associated comorbidities after OLT, or compliance. The 5- and 10-year survival rates for patients with alcohol recidivism were 92.9% and 45.1%, respectively, compared with 92.4% and 85.5%, respectively, for patients without alcohol recidivism. These figures show significantly lower survival rates in recidivistic patients after 10 years (P < 0.01, Mantel-Cox). The fact that patients who resumed alcohol consumption have a worse 10-year survival rate might be attributed to a higher frequency of deaths, primarily from cancer and cardiovascular events. (Liver Transpl 2005;11:420–426.)

Alcoholic liver disease (ALD) is the leading cause of cirrhosis in Western countries and represents the second most common indication for orthotopic liver transplantation (OLT) in the United States and Europe.1–3 Reported survival rates after liver transplantation appear to be similar among alcoholics and nonalcoholics, reaching up to 60% 7 years after OLT.1, 4 In view of patient and graft survival, rejection rates, and infection rates, ALD appears to be a good indication for OLT.5

Nonetheless, this indication still remains controversial, based partly on concerns regarding alcohol relapse and functional outcome after transplantation in an era of donor organ shortage and priority setting.1 In fact, it is common practice to establish a period of abstinence from alcohol (usually 6 months) before listing a patient with ALD for OLT based on (1) patients in whom stabilization or improvement of liver function may occur, obviating the need for transplantation,3 and (2) higher reported rates of alcohol relapse in patients who had been abstinent from alcohol for less than 6 months before transplantation.6

Several previous studies analyzed the rate of alcohol relapse after OLT for ALD and its influence on patient and graft survival.2, 4, 7–18 Thus, reported rates of alcohol relapse range from 11.5% to 49%, although this fact was rarely considered a reason for graft failure in recipients with ALD.2 Patients who resume alcohol consumption after OLT for ALD may damage their graft because of poor compliance with immunosuppressive drugs and alcohol-related liver injury. Graft dysfunction related to relapse ranged from 0% to 17%, although death related to relapse ranged from 0% to 5%.2, 4, 8, 9, 11, 14, 15, 17, 18 The prognosis of liver transplantation in this setting after a more prolonged follow-up has been seldom studied.

In a previous study from our group, 44 patients who underwent an OLT for ALD and 17 control subjects were included for study and followed for a mean of 39.5 months; no deleterious effect of alcohol recidivism on therapeutic compliance to treatment or on graft function could be demonstrated.16 For the current study, the previous cohort of alcoholic patients was enlarged, and follow-up was extended to provide more information about the long-term effects of alcohol recidivism on patient and graft survival and on comorbidities, as well as on the rate of occurrence of de novo malignancies and other clinical events after OLT for ALD.


OLT, orthotopic liver transplantation; ALD, alcoholic liver disease; 95% CI, 95% confidence interval.

Patients and Methods

Patient Selection

This retrospective cohort study included the long-term follow-up of a previously analyzed cohort of 44 consecutive patients who underwent transplantation for ALD in our center before 1997.16 This group was enlarged to 54 patients to include the complete cohort of ALD patients who underwent transplantation up to the date of publication of our previous study. The diagnosis of ALD in each patient was based on a history of habitual and excessive alcohol consumption in the absence of other causes of liver cirrhosis, compatible clinical and laboratory findings, and typical histology in the explanted liver. The criteria for the selection of patients for transplantation included (1) a diagnosis of end-stage liver disease, in which transplantation was the only option for survival; (2) absence of medical contraindications; and (3) a standard psychiatric evaluation to predict long-term abstinence from alcohol. Before transplantation was considered, a period of abstinence from alcohol consumption of 6 months was required to ensure that there was no improvement in liver function without alcohol intake.

Patient Follow-up

All patients received blood group compatible grafts and were treated with the same immunosuppressive regimen, which combined cyclosporin A, steroids, and azathioprine. By the end of the third postoperative month, azathioprine was stopped, and prednisone was progressively tapered over the first year. Routine biochemical examinations were performed by a member of the transplant team on an outpatient basis, and patients were instructed to continue abstinence after OLT. The outpatient follow-up visits were usually conducted once a week during the first month after leaving hospital, twice a month during the second and third months, monthly during the first year, and every 3 or 4 months thereafter, regardless of the length of the observation period after transplantation. Additional visits also were necessary because of particular problems. Patients were personally questioned about alcohol consumption, compliance with immunosuppressive regimen, and health status. A complete laboratory investigation, including hematology, liver parameters, coagulation, electrolytes, total protein, renal parameters, electrophoresis, and lipid profile, and cyclosporin A whole blood level was conducted at each visit. As part of our routine management of all patients receiving transplants, we performed a liver biopsy during the second week after transplantation, at 3 months, at 1 year, and when clinically indicated. Only biopsies performed at least 2 weeks after OLT were considered to exclude early postoperative complications. Histologic abnormalities were grouped into the following pathologic categories: isolated fatty change, cholestatic changes, pericellular or perivenular fibrosis, acute alcoholic hepatitis, cirrhosis, acute rejection, chronic rejection, and miscellaneous.

Alcohol Recidivism

Alcohol relapse was defined as a daily alcohol intake of more than 30 g/day. The diagnosis was mainly based on the information collected from patients and family members during protocol interviews conducted at each outpatient visit by a transplant hepatologist and from a review of alcohol and non-alcohol-related hospitalizations. Moreover, biochemical liver enzyme tests were useful in confirming alcohol consumption. A formal prospective longitudinal protocol of urine alcohol measurements was not used throughout the duration of the study.

The consequences of alcohol relapse were analyzed as follows: patient survival, graft evolution as favorable/unfavorable, acute and chronic rejection, number of infections, medical complications (defined as arterial hypertension, renal failure, or metabolic disturbances such as diabetes as a whole), hyperlipidemia, or hyperuricemia. Other clinical events such as cancer occurrence, cardiovascular events (e.g., acute myocardial infarction, cerebrovascular accidents), or any other clinical condition requiring medical treatment after OLT whatever its severity, tobacco consumption, and compliance with the immunosuppressive regimen also were recorded.

All data analyzed in this study were retrospectively collected from medical records. The variables used, including alcohol relapse, have been recorded in accordance with a standard protocol used in our center for transplanted patients during follow-up.

Statistical Analysis

Comparisons between qualitative variables were performed using the chi-square test or the Fisher's exact test. Survival was assessed by the Kaplan-Meier method; survival curves were compared using the Mantel-Cox statistic. A Cox proportional hazards model was used to evaluate the risk of death resulting from recidivism in a multivariate model (P < 0.05 was considered significant). Analyses were performed using SPSS 8.0 (Statistical Package for Social Sciences software for Windows; SPSS, Chicago, IL).


Table 1 shows the clinical characteristics of the patients studied. Of the 54 patients who underwent OLT for end-stage ALD, 43 were men and 11 were women. Their ages ranged from 29 to 62 years, and the median age was 48.9 years. ALD was the only etiologic factor in 90% of patients (49 patients), although the remaining had concomitant hepatocellular carcinoma (5 patients). Follow-up periods ranged from 14 to 155 months, with a mean of 99.2 months for the entire population examined (SD = 31.8).

Table 1. Clinical Characteristics of Patients Who Underwent Transplantation for Alcoholic Liver Disease and Their Alcohol Habit After Transplantation
CharacteristicAbstinentAlcohol relapse
  1. Abbreviations: SD, standard deviation.

Number of patients4014
Gender (male/female)32/811/3
Age, yrs (SD)49.3 (8.5)47.6 (7.7)
Hepatocellular carcinoma41
Abstinence >6 mo pretransplant4014
Follow-up, mo (range)99.8 (30–151)97.5 (14–155)

Alcohol Relapse

After OLT, 14 patients (25.9%) resumed alcohol consumption; the incidence of relapse was 3.01/1000 person-months (1.64 to 5.04). The mean interval from transplantation to alcohol relapse was 47.5 months, with a range of 5.0 to 86.9 months. Five years after transplantation, the percentage of patients free from relapse reached 78.69% (67.48%–89.90%) (Fig. 1).

Figure 1.

Cumulative proportion curve for the risk of alcohol relapse during follow-up.

Influence of Alcohol Relapse on Graft Evolution and Comorbidities

None of the patients who returned to alcohol use had serious medical complications requiring admission to hospital. No significant differences were found when comparing patients who resumed alcohol consumption with those who remained abstinent in the ALD group in terms of graft function (85.7% vs. 80.0% were favorable, respectively), rates of infection after OLT (64.3% vs. 70.0% had 1 or more documented infections of different degrees of severity, respectively), or medical complications as defined previously (100% vs. 97.5% had 1 or more medical complications of different degrees of severity, respectively) (Fig. 2). Although we could not demonstrate significant differences when comparing the rate of acute rejection between both groups, there was a trend toward a lower rate of rejections in the group who relapsed, limiting the statistical significance (28.6% vs. 60.0% had 1 or more documented rejections of different degrees of severity, respectively; P = 0.06 by Fisher's exact test). None of the patients had acute rejection attributable to poor compliance with immunosuppressive drugs, as therapeutic compliance assessed by clinical interview and therapeutic drug monitoring of cyclosporine reached 100% in both groups.

Figure 2.

Influence of alcohol relapse on graft evolution and comorbidities. Medical complications included arterial hypertension, renal failure, or metabolic disturbances such as diabetes, hyperlipidemia or hyperuricemia. Clinical events correspond to other clinical circumstances (cancer occurrence, cardiovascular events), which required medical treatment after liver transplantation regardless of their severity. No statistical significance was found between patients who relapsed and those who rested abstinents.

Moreover, there were no differences in the rate of occurrence of other medical events (71.4% vs. 72.5% had 1 or more documented medical events, respectively). This category includes any clinical event requiring medical treatment after OLT, regardless of its severity.

Histopathologic Features

One hundred twenty liver biopsies were performed throughout the study: 86 (71.6%) in the abstinent group and 34 (28.4%) in patients who resumed alcohol consumption. Data shown in Table 2 correspond to the most abnormal biopsy in the biopsy series from each patient. Fatty change was more frequently observed in biopsies from patients who consumed alcohol than in those from patients who remained abstinent. No differences were observed between abstinent patients and those who relapsed regarding metabolic complications causing fatty changes. In a heavy drinker, a pattern of macrovesicular diffuse fatty change and mild fibrosis could be described, but no progression could be demonstrated thereafter. There was a similar frequency of biopsy-proven acute rejection in both groups. No patient had typical features of liver cirrhosis, acute alcoholic hepatitis, or chronic rejection.

Table 2. Histological Changes Found Throughout the Study
 Alcohol recidivism
Minimum or normal changes4015
Fatty changes45
Cholestatic changes96
Acute rejection285
Chronic rejection00
Acute alcoholic hepatitis00
Liver cirrhosis00
Hepatocellular carcinoma10
Other changes43

Influence of Alcohol Relapse on Survival and Causes of Death

The mortality rate throughout the study was 2.25/1000 person-months (1.16 to 3.93). Twelve patients died during follow-up, and alcohol relapse accounted for most of the deaths (7 patients, 58.3%). Malignant tumors and cardiovascular events are the most important causes of death in this group (Table 3). The mean latency period from alcohol relapse to the diagnosis of the clinical event that ultimately resulted in death (Table 3) was 39.6 months (SD, 18.8). When we analyzed tobacco consumption, we observed that all patients with alcohol recidivism (14 of 14 patients) maintained their smoking habit, whereas only 10 of the 40 abstinent patients continued smoking after liver transplantation. A surprising finding of our study was that none of our patients had posttransplant lymphoproliferative disorders.

Table 3. Causes of Death in the Entire Population Studied
PatientTransplant indicationAlcohol relapseCause of deathSurvival (mo)*Months after alcohol relapse
  • Abbreviations: ALD, alcoholic liver disease; HCC, hepatocellular carcinoma.

  • *

    Time in months from alcohol relapse to the diagnosis of the clinical event that caused death.

1ALDYesMetastatic dissemination of laryngeal epidermoid carcinoma7225
2ALDYesSudden death9859
3ALDNoHepatic metastases of pancreatic adenocarcinoma102
4ALDYesBone metastases of squamous pharyngeal carcinoma9536
5ALDNoInvasive bladder carcinoma31
6ALDYesAcute myocardial infarction9348
7ALDYesBiliary sepsis7263
8ALDYesAcute stroke9136
9ALDNoBiliary sepsis70
10ALDNoAcute myocardial infarction90
11ALD and HCCYesCerebral hemorrhage1410
12ALD and HCCNoRelapse of HCC with metastatic dissemination57

The 1-, 5-, and 10-year survival rates were 100%, 92.9% (95% confidence interval [95% CI], 59.1%–99.0%), and 45.1% (95% CI, 17.9%–69.3%), respectively in the group who relapsed and 100%, 92.4% (95% CI, 78.1%–97.5%), and 85.5% (95% CI, 67.9%–93.9%) in the group of abstinent patients (P = 0.006 when comparing survival curves with the Mantel-Cox statistic). The risk of death resulting from alcohol recidivism was represented by a crude hazard ratio of 4.36 (95% CI, 1.38–13.83) (P < 0.05), using a Cox proportional hazards model; when adjusted for age and sex, the hazard ratio was 5.65 (95% CI, 1.67–19.17) (P < 0.01). Figure 3 shows the survival curves of both groups assessed by the Kaplan-Meier method.

Figure 3.

Kaplan-Meier survival curves from patients with alcoholic liver disease, with or without alcohol recidivism.


We evaluated the long-term results of OLT for ALD, focusing on the occurrence of alcohol recidivism and its repercussion on the allograft, comorbidities, de novo malignancies, and survival. Interestingly, we observed excellent and higher-than-expected 1- and 5-year patient survival rates in patients who underwent transplantation for ALD. Although these rates were similar in patients who resumed alcohol consumption (100% and 92.9%, respectively) and in those who remained abstinent (100% and 92.4%, respectively), there was an abrupt and significant fall in the 10-year survival rate of the first group (45.1%) compared with that of the abstinent group (85.5%). This unexpected finding could be explained by the higher mortality rate observed in the first group, mainly because of cancer and different cardiovascular events, which are discussed in the following paragraphs.

Our study revealed a 25.9% alcohol relapse rate after OLT, which falls within the range of relapse rates reported in the literature and is consistent with that previously reported by our group.2, 4, 7, 8, 16, 19 Thus, it has been estimated that between 20% and 50% of patients who receive a liver transplant for end-stage ALD acknowledge some alcohol use in the first 5 years after liver transplantation, and 10% to 15% will resume heavy drinking.3, 20, 21 The alcoholic behavior is difficult to determine; several methods are used for this purpose: urine alcohol measurements, carbohydrate-deficient transferrin, and information obtained from patients and relatives about the resumption of alcohol use.22 The combination of different methods, a solid physician-patient relationship, and a prolonged follow-up would increase the sensitivity to detect sustained alcohol use in these patients.23, 24

As previously reported by our group, no differences were observed between abstinent patients and those who resumed alcohol consumption—as indicated by patients' reports and monitoring of drug concentrations—concerning compliance with immunosuppressive drugs, which agrees with the limited recorded evidence that suggests that alcohol relapse leads to noncompliance in liver transplant recipients.1, 25 We found no differences between abstinent patients and those who relapsed regarding the incidence of infections, which could be expected taking into consideration previous reports that were not able to demonstrate any increased incidence of infections in patients transplanted for ALD.15, 17 Conversely, although we observed a lower incidence of acute rejection in patients with alcohol recidivism compared with those who remained abstinent, this fact does not reach statistical significance and cannot be explained by poorer compliance with therapeutic regimens. Certainly, our findings are in accordance with previous reports, which showed that ALD recipients had a reduced risk of acute rejection compared with other indications, and appear to support the hypothesis suggested by these studies regarding an inhibitory effect of alcohol on various aspects of the immune response.15, 17, 26

In accordance with the results of previous studies, no differences concerning graft evolution or metabolic disturbances were observed in abstinent patients compared with those who relapsed.2 However, our histologic study revealed only mild hepatic changes directly attributable to alcohol in a few patients, which agrees with our previous results in this cohort after a shorter period of follow-up. As reported by Pageaux et al.,2 fatty changes represented the most frequent finding related to alcohol injury in patients who resumed alcohol consumption. However, we could not demonstrate any statistical difference with regard to patients who remained abstinent. Finally, no differences in the incidence of other medical conditions, such as arterial hypertension or renal failure, were found when analyzed together with metabolic disturbances.

The group of alcohol relapsers presented with a higher incidence of de novo malignancies which, in addition to an increased occurrence of cardiovascular events, appear to have accounted for a significant decline in survival at 10 years for these patients. Saigal et al. found that patients who underwent liver transplantation for ALD appeared to have an increased risk of developing posttransplantation malignancies compared with other liver diseases.27 They hypothesized that a tumorigenic action mediated by an immunosuppressive effect of alcohol on natural killer cells could explain this effect.27 In fact, in the nonimmunosuppressed population, alcoholism is associated with an increased risk for several malignancies, including liver and alimentary tract tumors.28, 29 Moreover, Duvoux et al. observed a higher incidence of oropharyngeal squamous cell carcinomas in patients receiving a transplant for alcoholic cirrhosis, suggesting an additional effect of posttransplantation immunosuppression in a group of patients “overexposed” to alcohol and tobacco before transplant.30 Moreover, we detected increased tobacco consumption in alcohol relapsers, which could explain some of the de novo malignancies observed in our study. Contrary to previous studies, we found a lower 10-year survival rate in patients transplanted for alcoholic cirrhosis when they resumed alcohol consumption. Reasonably, this fact appears to be related to the longer follow-up period in our study; thus, it is conceivable that Pageaux et al. were not able to find differences in survival rates after a mean follow-up of less than 60 months.17 Nonetheless, our findings are in agreement with those described by Jain et al., who found a decreased overall patient survival rate 5 years after liver transplantation for patients with ALD compared with patients without ALD, which appeared to be related to a higher rate of upper airway malignancies.31

Although other investigators have not observed significant differences between abstinent patients transplanted for ALD and those who resumed alcohol consumption with respect to de novo malignancies, their reported incidences of cancer were 15.9% in abstinent patients and 22.2% in drinkers within the ALD group, which are higher than the 4.5% to 12.5% previously reported incidences of de novo malignancies for liver transplantation in general.2, 27, 32–37 Thus, occurrence of de novo cancer in the group of alcohol recidivism (14.3%) was found to be significantly higher than that observed in abstinent patients (7.5%), which supports the theories outlined by Saigal et al. and Duvoux et al. Furthermore, smoking habits appear to underlie the higher incidence of adverse cardiovascular events observed in this group, which could partly explain the higher mortality in these patients.

Remarkably, in spite of the higher incidence of de novo malignancies found in our study, there were no cases of lymphoproliferative disorders, such as lymphomas, which have previously been reported to occur more frequently in liver allograft recipients.37–39 In fact, Duvoux et al. found alcoholic cirrhosis to be an independent risk factor for developing a posttransplant lymphoproliferative disorder, which could be explained by different factors, such as an overimmunosuppressed condition favoring B-cell proliferation or an imbalance in the cytokine profile, which could promote B-cell proliferation.39 The absence of posttransplant lymphoproliferative disorders in our study could be explained by the antiviral prophylaxis policy that we follow in our transplant program during the first 3 months after transplantation, which consists of universal treatment with acyclovir, regardless of the state of cytomegalovirus.40 Finally, our study reaffirms the good graft and patient survival of those who underwent transplantation for ALD without alcohol recidivism.17, 41

In conclusion, we observed a significant decline in 10-year survival rates in patients who underwent transplantation for ALD who relapsed into alcohol use compared with those who remained abstinent. These findings do not appear to be related to the graft evolution, rejections, infections, or metabolic disturbances, but a higher incidence of malignancies and cardiovascular events appears to underlie this higher mortality. More studies including longer follow-up periods in a larger number of patients are needed to corroborate these results. Meanwhile, these patients should be followed-up more closely. Furthermore, simple risk-reducing measures, such as cessation of smoking and alcohol consumption should be encouraged.


The authors thank Dr. Javier Llorca for excellent statistical analysis assistance.