Smoking increases recurrent viral hepatitis after liver transplantation



Smoking is a common behavior among transplant candidates. The aim of this study was to evaluate the effects of smoking on a range of complications after liver transplantation. We reviewed data about patient demographics and various complications after liver transplantation that were recorded in the McGill University Health Centre liver transplant database over a 14-year period. χ2 and multivariate analyses were performed. Four hundred forty-four liver transplants were performed from 1990 to 2004, and 63 were repeat transplants. Only primary liver transplant recipients were included in our analysis. Smokers (ie, active or former smokers) were more likely to be male (77.9% versus 62.7%, P = 0.009) and Caucasian (88.4% versus 78.0%, P = 0.03). The median survival time was 13.23 years for smokers and was not estimable for nonsmokers because of censoring. The median recurrent viral hepatitis–free survival time was 0.87 years for smokers and 4.10 years for nonsmokers (P = 0.03). The following variables were not found to be associated with the smoking status: patient survival (P = 0.78), time to biliary complications after liver transplantation (P = 0.67), time to the first rejection episode after liver transplantation (P = 0.61), and time to depression after liver transplantation (P = 0.67). A Cox proportional hazards regression showed that recurrent viral hepatitis–free survival was still strongly associated with smoking [HR = 2.04, 95% confidence interval (CI) = 1.13-3.68, P = 0.018] and was marginally associated with East Asian race (HR = 0.26, 95% CI = 0.06-1.06, P = 0.06) and male sex (HR = 0.59, 95% CI = 0.34-1.02, P = 0.06). In conclusion, recurrent viral hepatitis–free survival was decreased for smokers after liver transplantation, likely because of the adverse effects of tobacco on immunological host defenses. Overall, the biliary complication–free, depression-free, and rejection-free survival rates were similar for smokers and nonsmokers. These findings suggest that smoking cessation should be encouraged, particularly in recipients undergoing transplantation for viral hepatitis. Liver Transpl, 2012. © 2012 AASLD.

The success of liver transplantation has dramatically improved over the last few decades with the advent of effective and tailored immunosuppression and better perioperative care.1 As overall survival increases, long-term outcomes and quality-of-life issues are becoming important.

Postoperative complications are mainly infectious and anatomic in the immediate postoperative setting. In the longer term, diabetes, hypertension, and malignancies are more frequent among liver transplant patients versus the general population.

Smoking is a behavior that tends to be common among liver transplant recipients, with 33.6% of these patients reporting themselves to be active or former tobacco users.2 Most transplant programs across North America do not require a patient to discontinue smoking in order to be eligible for transplantation.3 Despite the known adverse physiological effects of smoking, few studies have been conducted to assess whether smoking affects post–liver transplant outcomes. An increased incidence of biliary and vascular complications has been reported in this population in the short term.4-6 In the longer term, all-cause mortality, cardiovascular mortality, and sepsis-related mortality are higher,7 and emerging evidence suggests that these patients are more susceptible to developing non–skin-related malignancies.8 Because of the scarcity of organs available for transplantation and the amount of resources that go into the process of liver transplantation, it is incumbent on those regulating access to this procedure to maximize the chances of a positive outcome. The goal of this study was to assess whether the smoking status has an impact on the incidence of a diverse range of complications after liver transplantation (including recurrent viral hepatitis) in the short term and in the longer term.


BMI, body mass index; CI, confidence interval; HR, hazard ratio; MELD, Model for End-Stage Liver Disease.


Patient Cohort

The McGill University Health Centre liver transplant database and the patient records of all adult patients who underwent liver transplantation at the McGill University Health Centre between June 6, 1990 and January 19, 2004 were reviewed retrospectively. The McGill University Health Centre liver transplant database prospectively collects demographic and clinical characteristics for all patients undergoing liver transplantation, and it is regularly updated by a full-time transplant coordinator. Only those patients who underwent primary transplantation were included in our analysis; those who underwent retransplantation were censored at the time of retransplantation. The smoking status was obtained through a review of preoperative evaluation records and pulmonary function testing reports. Patients were not asked to discontinue smoking unless they had respiratory complications. An adequate smoking history was documented for all patients. Demographic characteristics such as the age, sex, race, body mass index (BMI), smoking status, liver disease etiology, Child-Pugh score, and Model for End-Stage Liver Disease (MELD) score at the time of transplantation were recorded. Smoking behavior was self-reported by the patient during the standard preoperative evaluation, and it was defined as follows: an active smoker was a liver transplant recipient who had been smoking cigarettes within 3 months of liver transplantation, a former smoker was a patient who had a history of smoking but had stopped more than 3 months before liver transplantation, and a nonsmoker was a patient who had never smoked in his or her lifetime. The smoking status was verified in the records of pulmonary function tests, which were performed routinely for all transplant recipients. The database did not contain information on the continuation of smoking after the operation. The protocol for this study received ethics approval by the institutional review board of the McGill University Health Centre.

Definition of Outcomes

Short-term postoperative complications such as acute rejection, hepatic artery thrombosis, and biliary complications (anastomotic strictures and cholangitis) were recorded. Acute rejection was diagnosed 5 to 30 days after transplantation on the basis of typical histological findings in accordance with the rejection activity index of the Banff schema.

Longer term complications such as recurrent viral hepatitis and depression were also recorded. Recurrent viral hepatitis was determined on the basis of histological examinations of liver biopsy samples, and viral counts were checked to corroborate the histological findings. Depression was diagnosed according to the criteria of Diagnostic and Statistical Manual of Mental Disorders, 4th ed.,9 and on the basis of a psychiatric diagnostic interview.

These outcomes were assessed for patients who underwent transplantation between June 1990 and March 2004 and were followed until March 2007. Specifically, overall survival was calculated in days for all subjects from the time of liver transplantation to the date of death. Subjects who underwent retransplantation were censored at the date of retransplantation, whereas subjects who did not undergo retransplantation and were alive at the end of the study were censored at the end of study (March 8, 2007). Rejection-free survival, time to depression, time to biliary complications, and time to recurrent hepatitis were calculated in days for all subjects from the time of liver transplantation to the date of the first complication. Subjects without complications were censored at the date of graft failure, the date of retransplantation, the date of death, or the end of the study (March 8, 2007)—whichever occurred first. Recurrent viral hepatitis–free survival was assessed specifically and only for patients who had undergone transplantation for hepatitis B or C (121 patients in all) and, therefore, only for those patients at risk of developing recurrent hepatitis after transplantation.

Statistical Analysis

The independent exposure variable was smoking behavior, and the patients were categorized as smokers (active or former smokers) or nonsmokers. Demographic and clinical characteristics were compared with the χ2 analysis method on the basis of smoking behavior; active and former smokers were grouped together, and lifetime nonsmokers were a separate category.

Categorical demographic and clinical characteristics were tabulated by groups (smokers versus nonsmokers), whereas continuous variables (eg, age) were presented with summary statistics (means and standard deviations) by groups and for all subjects. Fisher's exact test (for categorical variables) and the Wilcoxon rank-sum test (for continuous variables) were performed to assess the homogeneity of demographic and baseline clinical characteristics between smokers and nonsmokers (Table 1).

Table 1. Demographic and Clinical Characteristics of Liver Transplant Recipients by Their Smoking Behavior
CharacteristicActive or Former Smokers (n = 86)Nonsmokers (n = 295)
  • *

    P = 0.009 (χ2).

  • P = 0.03 (χ2).

Age (years)52.9 ± 10.455.6 ± 11.6
Sex: male/female (n/n)67/19*185/110*
BMI (kg/m2)26.9 ± 7.026.3 ± 5.6
MELD score22.1 ± 9.624.5 ± 10.0
Child-Pugh score10.8 ± 1.911.2 ± 1.8
Race (%)  
 East Asian2.37.8
 South Asian2.37.5

The differences in patient, rejection-free, depression-free, biliary complication–free, and recurrent hepatitis–free survival rates between smokers and nonsmokers were tested with the log-rank test. In addition, for each time-to-event outcome, a Cox regression model was applied (with adjustments for some other prognostic factors that were unbalanced or were correlated with the outcome in univariate analyses) to verify the impact of the smoking behavior on the survival outcome and to identify factors significantly related to the survival outcome. This Cox model included all relevant baseline variables (ie, the smoking status, sex, age, race, BMI, diagnosis, MELD score, and Child-Pugh score) as potential covariates and was determined with a stepwise model building procedure; a significance level of 0.2 was used for the entry or removal of variables. A P value less than or equal to 0.05 was considered to be statistically significant. SAS 9.2 (SAS Institute, Cary, NC) was used for all analyses.


Baseline Characteristics

Four hundred forty-four patients received liver transplants from 1990 to 2004, and 63 were repeat transplants. Only primary liver transplant recipients were included in our analysis (mean age = 55.0 ± 11.3 years, proportion of males = 66.1%, mean BMI = 26.5 ± 6.0 kg/m2, mean MELD score = 24.0 ± 9.9, mean Child-Pugh score = 11.1 ± 1.8, proportion of active smokers and former smokers = 22.6%, proportion of nonsmokers = 77.4%). Smokers were more likely to be male (77.9% versus 62.7%, P = 0.009) and Caucasian (88.4% versus 78.0%, P = 0.03; Table 1). The etiology of liver disease was more likely to be alcohol for active smokers and former smokers versus nonsmokers (29.1% versus 15.9%, P = 0.009; Table 2).

Table 2. Etiology of Liver Disease in Liver Transplant Recipients by Their Smoking Behavior
EtiologySmokers (%)Nonsmokers (%)
  • *

    P = 0.009 (χ2).

Hepatitis C virus24.423.7
Hepatitis B virus3.59.2
Nonalcoholic steatohepatitis3.56.5
Hepatocellular carcinoma12.819.7
Primary sclerosing cholangitis4.76.1
Primary biliary cirrhosis5.88.5
Autoimmune hepatitis1.22.0

Impact of Smoking Behavior on Complications After Liver Transplantation

For nonsmokers, the patient survival rates at 1 and 5 years were 83.3% and 69.5%, respectively. For smokers, the patient survival rates at 1 and 5 years were 86.8% and 71.8%, respectively (Fig. 1). For nonsmokers, the recurrent viral hepatitis–free survival rates at 1 and 5 years were 67.9% and 47.6%, respectively. For smokers, the recurrent viral hepatitis–free survival rates at 1 and 5 years were 47.3% and 31.0%, respectively (Fig. 2).

Figure 1.

Overall survival for liver transplant recipients: smokers versus nonsmokers. [Color figure can be viewed in the online issue, which is available at]

Figure 2.

Recurrent hepatitis–free survival for patients undergoing transplantation for viral hepatitis. [Color figure can be viewed in the online issue, which is available at]

Therefore, overall survival was similar for smokers and nonsmokers (P = 0.78), and recurrent viral hepatitis–free survival was significantly lower for smokers who underwent transplantation for viral hepatitis (P = 0.03). Rejection-free survival (P = 0.61), biliary complication–free survival (P = 0.67), and depression-free survival (P = 0.67) were all similar for smokers and nonsmokers.

A multivariate analysis was performed with the covariates listed in Table 3. The only association that held was recurrent viral hepatitis–free survival, with smoking [HR = 2.04, 95% confidence interval (CI) = 1.13-3.68, P = 0.02], East Asian race (HR = 0.26, 95% CI = 0.06-1.06, P = 0.06), and male sex (HR = 0.59, 95% CI = 0.34-1.02, P = 0.06) being predictive of recurrent viral hepatitis–free survival. The smoking status was not predictive of overall, biliary complication–free, or rejection-free survival because other confounding variables were found (Table 4).

Table 3. Adjustment Covariates for the Logistic Regression Analysis of Outcomes After Liver Transplantation
Etiology of liver disease (hepatitis C virus, alcohol, or other)
Smoking status
MELD score
Child-Pugh score
Race (Caucasian, East Asian, Black, or other)
Table 4. Predictors of Outcomes for Liver Transplant Recipients
Predictors of Overall SurvivalPredictors of Rejection-Free SurvivalPredictors of Biliary Complication–Free SurvivalPredictor of Depression-Free SurvivalPredictors of Recurrent Viral Hepatitis–Free Survival
Age (HR = 1.02, 95% CI = 1.00-1.04, P = 0.016)Hepatitis C virus (HR = 1.50, 95% CI = 1.02-2.20, P = 0.04)Caucasian race (HR = 0.43, 95% CI = 0.25-0.74, P = 0.002)Sex (HR = 0.40, 95% CI = 0.25-0.55, P = 0.07)Smoking (HR = 2.04, 95% CI = 1.13-3.68, P = 0.018)
MELD score (HR = 1.04, 95% CI = 1.02-1.07, P = 0.02)Age (HR = 0.98, 95% CI = 0.97-1.00, P = 0.02)Black race (HR = 0.032, 95% CI = 0.11-0.97, P = 0.04) East Asian race (HR = 0.26, 95% CI = 0.06-1.06, P = 0.06)
    Sex (HR = 0.59, 95% CI = 0.34-1.02, P = 0.06)

The duration of the intensive care unit stay (9.2 ± 10.0 versus 11.0 ± 14.9 days, P = 0.63) and the duration of the entire hospital admission (31.7 ± 30.7 versus 36.4 ± 47.0 days, P = 0.40) did not differ between smokers and nonsmokers.


Better immunosuppression and better perioperative care of liver transplant recipients have improved outcomes: the 1- and 5-year survival rates are 85% and 75%, respectively. However, recent literature has suggested that this progress could be further enhanced with preventive strategies such as smoking cessation. Our study reveals that recurrent viral hepatitis is significantly more common among smokers; thus, preoperative smoking cessation may be beneficial in this select patient population. A survey of liver transplant programs 10 years ago showed that 20% of the programs viewed tobacco use as an absolute contraindication to liver transplantation.3 Many American transplant programs now mandate that patients enroll in smoking cessation programs or stop smoking before they are placed on the transplant list, although firm statistics are not available. Because of the great shortage of organs, it is imperative to optimize outcomes after liver transplantation. In 2005, 132 adult patients died while they were waiting for a liver in Canada. The question is whether there is sufficient evidence to deny liver transplantation to active smokers because this denial also has ethical connotations.10 The findings of our study do not support such routine denial of transplantation, but they instead strongly encourage discontinuation among those patients with viral hepatitis as their reason for transplantation.

Smoking is a common habit among liver transplant patients, with increasing evidence showing that it causes adverse events after transplantation.2 Although it may not affect short-term mortality (ie, 2 years after transplantation), some differences have been noted in longer term survival.2 A 4-year series of liver transplant patients by Leithead et al.7 described an increased overall mortality rate along with higher cardiovascular- and sepsis-specific mortality rates. Smoking has been correlated with greater resource utilization and hospital charges over a 2-year period after liver transplantation.11 Biliary strictures and leaks have been reported to be more common among active smokers, with a retrospective study documenting a 92% higher rate of biliary complications among patients who smoked.4 A higher incidence of vascular complications, including hepatic artery thrombosis, has been shown after transplantation in active smokers.5 Smoking has also been independently associated with the development of de novo noncutaneous malignancies more than 10 years after liver transplantation.12, 13

In the general surgery literature, smoking behavior has been known to adversely affect outcomes.14, 15 This is due to a deranged microcirculation, which leads to slower wound healing, skin necrosis, and an increased risk of infection. The pathophysiological basis of tobacco-induced injury includes vasoconstriction, reduced blood oxygenation, endothelial dysfunction, and increased platelet adhesiveness. Tobacco-induced damage to mucoid tissues and reduced cellular immune function lead to an increased infection rate.16 A recent meta-analysis concluded that preoperative smoking cessation led to a 41% relative risk reduction in total postoperative complications.17

The presence of a higher number of smokers among those patients with alcoholic cirrhosis was expected in our study. Duvoux et al.18 determined that patients who underwent transplantation for alcoholic cirrhosis smoked 10 more pack-years on average than patients with nonalcoholic steatohepatitis. Smoking may hasten the progression of alcoholic cirrhosis, but it is unclear whether it aggravates the development of fibrosis in patients with nonalcoholic steatohepatitis. A prospective study of smoking behavior determined that patients who underwent transplantation for alcoholic liver disease tended to resume smoking behavior early after transplantation, and their use increased over time.19 This translated into 15% of active smokers resuming smoking and a relapse rate of 20% among former smokers.20

The higher rate of recurrent hepatitis among smokers undergoing transplantation for viral hepatitis is a novel finding in the transplant literature. Smoking has been independently associated with more severe liver fibrosis in patients with chronic hepatitis C (but not hepatitis B) in a prospective study,21 likely because of its proinflammatory effects. The higher rate of recurrent viral hepatitis among smokers in our study may be explained by the adverse effects of tobacco on both structural and immunological host defenses. In addition, a significant proportion of the patients in our population underwent transplantation for hepatitis C versus hepatitis B. East Asian race was negatively predictive of recurrent viral hepatitis, and this can probably be explained by the more favorable interleukin-28B genotype in this region and the presence of less treatment-resistant hepatitis C genotypes (2 and 3) among East Asians.

We did not find that other outcomes such as overall survival, rejection-free survival, biliary complication–free survival, and hepatic artery thrombosis–free survival5 differed between smokers and nonsmokers, despite what has been reported in the literature. However, our study was underpowered for hepatic artery thrombosis because of the low occurrence of this complication in our population.

We noted a trend toward decreased depression-free survival among smokers, and this may affect adherence to immunosuppressant medications.22 In addition, inadequately managed pretransplant depression has been correlated with acute rejection.23 Depression is known to be overrepresented among smokers,24 with abstinence tending to trigger depressive symptoms. The hypothesis is that nicotine may dysregulate acetylcholine receptors and thus bring about clinical depression.

Our retrospective study examined a wider spectrum of complications after liver transplantation, some of which have never been assessed in the context of smoking behavior. However, a few limitations to this study should be noted. The smoking status at the time of transplantation was obtained, but we did not have accurate information on the continuation of smoking habits after transplantation, nor did we have pack-year history data for a large majority of our patients. The determination of the smoking status for our study was based on self-reporting rather than cotinine levels. There is thus a possibility of misclassification bias. However, cotinine levels are not routinely measured in transplant patients, and none of the studies concerning the effects of smoking on complications after liver transplantation employed cotinine as their means of determining a patient's smoking status. Another limitation was the fact that we did not know whether the donor had been a smoker, and this likely had an impact on the condition of the donor vasculature and common bile duct being anastomosed. Certain complications could have arisen if patients had been followed for a longer period of time. It is also possible that our study was underpowered for certain outcomes.

In conclusion, smoking behavior has been associated with various adverse outcomes after liver transplantation. Our study in particular has demonstrated that recurrent viral hepatitis is more frequent among transplant recipients who are either active or former smokers at the time of transplantation. These findings indicate the need to identify these patients at the first transplant assessment and to strongly encourage smoking cessation through enrollment in a cessation program. Our findings make it ethically difficult to justify the absolute denial of liver transplantation. Therefore, if a transplant recipient persists in smoking, we must be more vigilant in our surveillance for complications after transplantation in order to optimize outcomes.