Long-term survival and impact of ursodeoxycholic acid treatment for recurrent primary biliary cirrhosis after liver transplantation

Authors

  • Phunchai Charatcharoenwitthaya,

    1. Miles and Shirley Fitterman Center for Digestive Diseases, Mayo Clinic and Foundation, Rochester, MN
    Search for more papers by this author
  • Sylvania Pimentel,

    1. William J. Von Liebig Liver Transplantation Center, Mayo Clinic and Foundation, Rochester, MN
    Search for more papers by this author
  • Jayant A. Talwalkar,

    Corresponding author
    1. Miles and Shirley Fitterman Center for Digestive Diseases, Mayo Clinic and Foundation, Rochester, MN
    2. William J. Von Liebig Liver Transplantation Center, Mayo Clinic and Foundation, Rochester, MN
    • Division of Gastroenterology and Hepatology, Mayo Clinic and Foundation, 200 First Street. SW Rochester, Minnesota, USA, 55905
    Search for more papers by this author
    • Telephone: (507) 284-4823; FAX: (507) 284-0538

  • Felicity T. Enders,

    1. Division of Biostatistics, Mayo Clinic and Foundation, Rochester, MN
    Search for more papers by this author
  • Keith D. Lindor,

    1. Miles and Shirley Fitterman Center for Digestive Diseases, Mayo Clinic and Foundation, Rochester, MN
    Search for more papers by this author
  • Ruud A.F. Krom,

    1. William J. Von Liebig Liver Transplantation Center, Mayo Clinic and Foundation, Rochester, MN
    Search for more papers by this author
  • Russell H. Wiesner

    1. Miles and Shirley Fitterman Center for Digestive Diseases, Mayo Clinic and Foundation, Rochester, MN
    2. William J. Von Liebig Liver Transplantation Center, Mayo Clinic and Foundation, Rochester, MN
    Search for more papers by this author

  • See Editorial on Page 1214

Abstract

The recurrence of primary biliary cirrhosis (PBC) in the hepatic allograft may impact patient and graft survival with long-term follow-up. The efficacy of ursodeoxycholic acid (UDCA) for treatment of recurrent PBC after liver transplantation (LT) remains less well known. The aims of this study were as follows: 1) to determine the significance of recurrent PBC on overall survival among PBC patients who underwent LT, and 2) to determine the efficacy of UDCA treatment after LT in patients with recurrent PBC. A retrospective cohort study was conducted of 154 PBC patients who underwent LT with at least 1 yr of follow-up after transplantation from 1985 through 2005. A total of 52 patients with recurrent PBC were identified. After adjusting for age and gender, recurrent PBC was not associated with death or liver retransplantation (hazard ratio, 0.97, 95% confidence interval, 0.41–2.31; P = 0.9). A total of 38 patients with recurrent PBC received UDCA at an average dose of 12 mg/kg/day for a mean duration of 55 months. Over a 36-month period, an estimated 52% of UDCA-treated patients experienced normalization of serum alkaline phosphatase and alanine aminotransferase compared to 22% of untreated patients. There was no significant difference in the rate of histological progression between subgroups. UDCA did not influence patient and graft survival. In conclusion, the development of recurrent PBC has little impact on long-term survival or need for retransplantation. While UDCA therapy is associated with biochemical improvement, its role in delaying histologic progression remains unknown. In this short period of treatment, UDCA was not associated with improved patient and graft survival compared to untreated patients. Liver Transpl 13:1236–1245, 2007. © 2007 AASLD.

Primary biliary cirrhosis (PBC) is a chronic, nonsuppurative cholangitic liver disease, with progressive loss of interlobular and septal bile ducts.1 PBC frequently progress to cirrhosis, which may lead to death or liver failure. Liver transplantation (LT) has emerged as an attractive therapeutic option for advanced PBC, with good survival and substantial improvement in quality of life. Currently, PBC represents the sixth leading indication for LT in the United States. It is now well established that PBC may recur in the allograft in some patients.2–8 However, there are no clear factors or clinical features that identify those at risk of recurrence. With the increasing success of LT and the greater number of patients with long-term survival, recurrent disease may become a greater problem and has impact on patient survival and graft function with long-term follow-up of these patients.

Ursodeoxycholic acid (UDCA) is currently the only drug approved for the treatment of patients with PBC.9 Several randomized trials and combined analyses have shown that this agent not only improves biochemical liver indices10–13 but also delays histologic progression14–17 and development of gastroesophageal varices as well as improves long-term survival in PBC patients.18, 19 The efficacy of UDCA for treatment of recurrent PBC after LT, however, remains less well known.

The aims of the present study were as follows: 1) to determine the clinical significance of recurrent PBC on overall patient and graft survival among PBC patients who undergo LT; and 2) to determine the efficacy of UDCA treatment after LT in a large group of patients with recurrent PBC.

Abbreviations

PBC, primary biliary cirrhosis; UDCA, ursodeoxycholic acid; LT, liver transplantation; OLT, orthotopic liver transplantation; ALP, alkaline phosphatase; CI, confidence interval.

PATIENTS AND METHODS

Study Population

The study group is comprised of all patients with an established diagnosis of PBC who underwent orthotopic LT (OLT) at the Mayo Clinic in Rochester, MN between 1985 and 2002. Follow-up after LT was extended to December 31, 2005. All patients with at least 1 yr of posttransplantation follow-up were included in this study. Data were collected from the Mayo Clinic Liver Transplantation Database and from medical records after receiving approval for the study from the Institutional Review Board. The pretransplantation diagnosis of PBC was established in all patients based on clinical, serological, and histological findings. In all instances, histopathologic study of the resected native liver confirmed the original diagnosis.

Patient Information

The complete medical dossiers of patients who have PBC recurrence and those who did not were reviewed to abstract information about recipient age and gender, donor age and gender, calendar year of OLT, human leukocyte antigen (A, B, and DR) matches, the type of primary immunosuppression (calcineurin inhibitors) and use of corticosteroids, history of rejection, calendar year of the first diagnosis of recurrent PBC, laboratory data including liver biochemical parameters (serum alkaline phosphatase [ALP], aspartate aminotransferase, alanine aminotransferase, total bilirubin, albumin, and prothrombin time), antimitochondrial antibody, and liver histologic data. The Model for End-Stage Liver Disease scores were calculated using the formula.20 Model for End-stage Liver Disease = 3.8 [loge serum bilirubin (mg/dL)] + 11.2 [loge international normalized ratio of prothrombin] + 9.6 [loge serum creatinine (mg/dL)] + 6.4. The PBC Mayo risk scores were determined using the formula described in the original model21: risk = 0.871 [loge bilirubin (mg/dL)] − 2.53 [loge albumin (g/dL)] + 0.039 [age (yr)] + 2.83 [loge prothrombin time (second)] + 0.859 edema. The follow-up of patients was obtained through an existing computerized database that provides the date of last contact with the record linkage system, as well as the status at last contact.

Posttransplantation Treatment

LT had been performed with standard technique in all patients, as described elsewhere.22 Patients were treated with triple-drug immunosuppression. The protocol used for immunosuppressive therapy after LT included cyclosporine A as the primary agent between 1985 and 1986. Tapering doses of prednisone were also included. After 1986, the primary immunosuppression regimen was changed to cyclosporine A, azathioprine, and prednisone. Beginning in 1993, tacrolimus was introduced into the immunosuppression protocol in combination with azathioprine and prednisone. Over time, tacrolimus became the preferred calcineurin inhibitor in all patients undergoing LT. In 1999, mycophenolate mofetil was used in place of azathioprine for all patients. Oral prednisone was tapered to discontinuation between 4 and 6 months from the year 1999 on. Within the first year after LT, a clinical algorithm was employed that allowed for the discontinuation of mycophenolate mofetil. If an individual patient experienced no episodes of acute cellular rejection within 2 months after LT, mycophenolate mofetil could be discontinued. If 1 episode of acute cellular rejection occurred within the first 2 months, an additional 2 months of mycophenolate mofetil was required. If 2 episodes of acute cellular rejection or at least 1 episode of steroid resistant rejection occurred within the first 2 months, patients were required to stay on mycophenolate mofetil for a duration of 1 yr before considering tapering to discontinuation.

Definition of Recurrent PBC Posttransplantation

Recurrent PBC was defined by the following criteria: 1) original diagnosis of PBC before transplant, 2) no history of postoperative ischemic cholangiopathy or chronic ductopenic rejection, and 3) liver histology diagnostic or compatible with PBC. Post-OLT protocol liver biopsies were obtained at 1 week, 3 weeks (1985 to 1996 only), 4 months, 1 yr, and 1- to 2-yr intervals thereafter (after 2004, biopsies have been done at 1, 3, and 5 yr intervals) as well as upon the occurrence of undefined hepatic graft dysfunction. The histologic feature of recurrent PBC is the presence of a florid duct lesion (granulomatous bile duct destruction) or destructive lymphocytic cholangitis with a dense portal infiltrate (in the absence of features of acute cellular rejection). A total of 2 patients who have other causes of histologic features of bile duct damage in the hepatic allograft, such as drug-induced cholestatic hepatitis (n = 1) and chronic viral hepatitis post-OLT (n = 2) were excluded. The onset of recurrent PBC was defined as the date on which the histologic features of disease was first reported by the pathologist.

Treatment of Recurrent PBC

Recurrent PBC patients who were symptomatic, or had persistent elevation of serum ALP level or hepatic fibrosis, were offered treatment with UDCA. The average daily dosage of UDCA prescribed was 900 mg (range, 500–1,200). In patients who were successfully withdrawn from steroids at any time before the diagnosis of PBC recurrence, steroids were not reintroduced for treatment of recurrent disease.

To determine the efficacy of UDCA therapy for recurrent PBC after LT, patients with recurrent PBC who were on UDCA therapy were identified. The comparative group was made up of those patients with recurrent PBC who did not receive UDCA treatment post-LT. Data on UDCA, including administration, dose, and duration of therapy, were collected. Liver biochemical parameters after diagnosis of recurrent PBC were collected every 12 months.

The progression of recurrent disease was assessed by histologic examination of liver-biopsy specimens. We used as the baseline the liver biopsy that first demonstrated histologic evidence of recurrent disease, and the follow-up biopsy as the last liver biopsy performed while on treatment. Histological stage was recorded on a 0 to 4 scale as previously reported by Ludwig et al.23 Briefly, in stage 1 disease, inflammation and/or abnormal connective tissue is confined to portal areas; in stage 2, the inflammation and/or fibrosis is confined to portal and periportal areas; in stage 3, there is bridging fibrosis; and in stage 4, there are regenerative nodules. In intermediate stages (e.g., stage 1–2 or 2–3), the biopsy was assigned the higher stage.

Statistical Analysis

The results are displayed in tables, with categorical variables presented as numbers and percentage and continuous variables presented as mean and standard deviation or median (range), when appropriate. Continuous data were analyzed using the Student's t-test if normally distributed, or the nonparametric Wilcoxon rank sums test if not normally distributed. The chi-squared test or Fisher's exact test, was used for comparison of frequency data, when appropriate. Logistic regression analysis was used to determine the factors associated with recurrence of PBC. We estimated the cumulative incidence of recurrent PBC using a competing risks method to account for the risk of death or retransplantation without recurrent PBC. The overall patient and graft survival was determined for all PBC patients using the Kaplan-Meier method. Survival was analyzed using Cox proportional hazards from the time of initial LT to the date of death or retransplantation, with PBC recurrence treated as a time-dependent covariate. In addition, Cox regression analysis was used to identify variables associated with overall death and retransplantation for PBC patients who underwent OLT. Subjects who remained free of graft failure or were alive through the follow-up were censored at the date of last contact in medical records, or December 31, 2005 (whichever came first).

Efficacy of UDCA for treatment of recurrent PBC after LT was measured in terms of: 1) changes in liver biochemical parameter values, 2) the rate of histological progression over time, and 3) the probability of death or retransplantation. The Kruskal-Wallis tests were used to examine differences within each treatment group with respect to changes from baseline in biochemical values at years 1 through 5. Histologic progression was defined as the time interval between date of first diagnosis of recurrent PBC and the date of last follow-up liver biopsy. Histologic progression was analyzed according to change in the stage and was deemed to be improving (decrease by 1 or more stages), unchanged, or worsening (increase by 1 or more stages). The rates of histologic progression for recurrent PBC patients who were treated with UDCA and those who did not receive UDCA therapy were estimated from survival curves based on the Kaplan-Meier method, with worsening in histological fibrosis stage as the event of interest, and compared by the Wilcoxon test. Probability of death from any cause or liver retransplantation was estimated by the Kaplan-Meier method for these subgroups and compared using the log-rank test. All statistical testing was done at the conventional 2-tailed α level of 0.05.

RESULTS

Study Population

Between 1985 and 2002, there were 164 consecutive patients who underwent LT for PBC and were prospectively followed-up at the Mayo Clinic in Rochester. Individuals undergoing LT during the study period received whole livers from deceased donors except for 1 patient who received a split-liver transplant using a right liver lobe. The mean age at time of LT for the cohort was 53.2 ± 8.5 yr (range, 34.4-71.2 yr) and 87% of patients were female. A total of 92% of individuals had positive serum antimitochondrial antibody titers before LT. At the time of LT; the mean Model for End-stage Liver Disease score was 16.2 ± 8.3 and the mean PBC Mayo risk score was 7.5 ± 1.2. Sepsis (2), spontaneous hollow viscous perforation with sepsis (2), disseminated intravascular coagulation (1), restrictive lung disease (1), cardiogenic shock (1), intracerebral hemorrhage (1), and graft vs. host disease (1) were the causes of early deaths within 6 months after LT. Of all PBC patients undergoing OLT, the actuarial overall patient survival at 1, 5, 10, 15, and 20 yr was 93%, 90%, 79%, 66%, and 49%, respectively. The actuarial overall graft survival at 1, 5, 10, 15, and 20 yr was 85%, 82%, 72%, 59%, and 42%, respectively (Fig. 1).

Figure 1.

The actuarial overall patient (upper line) and graft (lower line) survival of 164 PBC patients follow-up after LT from 1985 through 2005.

Recurrent PBC

A total of 52 patients were identified with histologic evidence of PBC recurrence. However, 10 patients in the nonrecurrent PBC group were excluded from the study due to 9 deaths within 1 yr after LT and 1 loss to follow-up at 4 months posttransplantation. Therefore, our analyses were based on a total of 52 patients with recurrent PBC and 102 nonrecurrent PBC patients.

The clinical and biochemical characteristics of the patients with and without recurrent PBC are summarized in Table 1. At the time of the first diagnostic lesion of recurrent disease, the ALP was elevated in 46% of these patients, with values ranging from 114 to 1,274 U/L (reference range, 90–234 U/L). Patients with recurrent PBC were significantly older, more often male, received grafts from older donors, were more likely to receive tacrolimus based-immunosuppression, and had a shorter duration of maintenance corticosteroids therapy when compared to those without disease recurrence. Patients receiving tacrolimus-based immunosuppression were withdrawn from corticosteroids at a median time of 3.0 yr after transplantation (range, 0.3–17.1 yr), compared to a median time of 10.1 yr (range, 1.5–20.1 yr) (P < 0.0001) among patients treated with cyclosporine-based immunosuppression. Furthermore, patients receiving tacrolimus developed recurrent PBC at a median time of 3 yr (range, 0.4–9.6 yr), whereas patients treated with cyclosporine developed disease recurrence after a median time of 6.7 yr (range, 0.3–18.1 yr) (P = 0.005).

Table 1. Comparison of Baseline Characteristic of Patients Who Had Recurrent PBC and Those Who Did Not Have Disease Recurrence
 Recurrent PBCNonrecurrent PBCP value
  • NOTE: Data represent mean ± SD and percentage. For alkaline phosphatase, 90–234 U/L; for AST the normal range is 12–31 U/L; for ALT, 9–29 U/L; for total bilirubin, 0.1–1.1 mg/dL; for albumin, 3.5–5 gm/dL; for prothrombin time, 8.4–12 seconds.

  • Abbreviations: MELD, Model for End-Stage Liver Disease; HLA, human leukocyte antigen; AMA, antimitochondrial antibodies; AST, aspartate aminotransferase; ALT, alanine aminotransferase.

  • *

    Pretransplantation data were collected at time of liver transplantation.

  • AMA and liver biochemical indices were collected at the time of first diagnosis of PBC recurrence in the recurrent PBC group, and at the time of performing the last liver biopsies in the nonrecurrent PBC group.

Number of patients52102 
Recipient age (yr)*55.7 ± 7.851.7 ± 8.40.005
Recipient gender (male)12 (23%)9 (9%)0.01
Donor age (yr)41.3 ± 17.834.6 ± 16.40.03
Donor gender (male)28 (54)53 (52)0.8
Gender match32 (62)54 (53)0.3
MELD*17.0 ± 9.516.2 ± 7.80.6
Mayo risk score*7.4 ± 1.27.5 ± 1.30.5
HLA match ≥162%51%0.2
Primary immunosuppression   
 Tacrolimus60%35%0.004
 Cyclosporine40%65% 
Duration of corticosteroids (yr)6.0 ± 5.08.4 ± 5.70.02
Acute cellular rejection56%54%0.8
AMA titer ≥1:4063%59%0.7
Liver biochemical indices   
 Alkaline phosphatase (U/L)394 ± 255469 ± 8700.5
 AST (U/L)52 ± 3746 ± 430.4
 ALT (U/L)67 ± 6358 ± 950.5
 Serum bilirubin (mg/dL)0.8 ± 0.31.5 ± 2.60.051
 Serum albumin (gm/dL)4.0 ± 0.33.9 ± 0.40.2

Peritransplant characteristics associated with higher probability of developing recurrent PBC were tacrolimus-based immunosuppression, duration of maintenance corticosteroid therapy, recipient age, recipient gender, and donor age from univariate logistic regression analysis. In the multiple logistic regression model with mixed elimination, the significant predictors of developing recurrent PBC were tacrolimus-based immunosuppression (P = 0.005) and male recipient (P = 0.02).

Histologic Characteristics of Recurrent PBC

Florid duct lesions were identified in 33 of 52 patients with recurrent PBC (63%) on initial diagnostic liver biopsy. Destructive lymphocytic cholangitis within dense portal infiltrates (in the absence of features of acute cellular rejection) was identified in the remaining 19 patients (37%). Among the PBC patients who met the criteria of recurrent disease, the great majority (45 out of the 52, 86%) of patients had developed portal fibrosis (stage 1 fibrosis) on the initial diagnostic liver biopsy. A total of 3 patients did not have fibrosis (stage 0 fibrosis), 3 patients had periportal fibrosis (stage 2 fibrosis), and 1 patient had septal fibrosis (stage 3 fibrosis). No evidence of septal fibrosis with regenerative nodules or histological cirrhosis at diagnosis of recurrent PBC was seen in this population.

Cumulative Incidence of Recurrent PBC

The mean follow-up after LT in the group of 154 patients was 10.8 ± 4.8 yr (range, 1.0–20.7 yr). A total of 52 of 154 patients (34%) had histologic evidence of disease recurrence at a median time after transplantation of 3.5 yr (range, 0.3–18.1 yr). Histologic features of 2 patients were suggestive of recurrent PBC less than 1 yr after transplantation. The cumulative incidence of developing recurrent PBC in overall patients undergoing LT was 0.22 (95% confidence interval [CI], 0.15–0.29) at 5 yr, 0.37 (95% CI, 0.27–0.45) at 10 yr, and 0.43 (95% CI, 0.33–0.52) at 15 yr following primary LT, respectively (Fig. 2). Furthermore, the cumulative incidence of recurrent PBC in patients who underwent LT according to date of transplantation (Fig. 3) showed that patients undergoing primary OLT between 1994 and 2004 were more likely to develop PBC recurrence in the liver allograft than those undergoing primary OLT between 1985 and 1993 (P < 0.001). Given tacrolimus became the preferred immunosuppressant in most of patients undergoing OLT after 1993; we also demonstrated the impact of calcineurin inhibitors on the incidence of recurrent PBC after LT (Fig. 4). The cumulative incidence of recurrent PBC among patients receiving a tacrolimus-based regimen were higher than the incidence observed in those receiving a cyclosporin-based regimen (P < 0.001).

Figure 2.

Probability of developing recurrent PBC in overall patients undergoing LT for PBC. The cumulative incidence of recurrent PBC (solid line) is shown with 95% CIs (broken lines).

Figure 3.

Probability of developing recurrent PBC in patients undergoing LT for PBC according to date of transplantation. The cumulative incidence of recurrent PBC in patients undergoing LT during 1994 to 2004 (broken lines) is significantly higher than the incidence observed in those undergoing liver transplantation during 1985 to 1993 (solid line) (log-rank test, P < 0.001).

Figure 4.

Probability of developing recurrent PBC in patients undergoing LT for PBC according to primary type of calcineurin inhibitors–based immunosuppression. The cumulative incidence of recurrent PBC in patients receiving tacrolimus-based immunosuppression (broken lines) is higher than the incidence observed in those receiving cyclosporin-based immunosuppression (solid line) (log rank test, P < 0.001).

Long-Term Survival of Patients With Recurrent PBC

The mean follow-up after LT was 10.1 ± 4.3 yr (range, 1.0–20.7 yr) for patients with recurrent PBC and 11.2 ± 5.0 yr (range, 1.0–20.6 yr) for nonrecurrent PBC patients. Overall, 6 patients with PBC recurrence and 30 nonrecurrent PBC patients died from any cause or were retransplanted during follow-up.

A total of 6 deaths and 2 retransplantations occurred in the recurrent PBC group. Of the deaths, 2 patients died of pneumonia, 1 patient died of drug induced hepatotoxicity, 1 patients died because of respiratory failure from myopathy of undetermined etiology, and another died of acute coronary syndrome. The cause of death was unknown in the remaining 1 patient. Two patients required retransplantation due to end-stage liver disease from recurrent disease in the hepatic allograft at 8 and 10 yr after first LT, respectively.

Of the patients without histologic evidence of recurrent PBC, there have been 29 deaths and 14 retransplantations. De novo malignancy (4), chronic ductopenic rejection (3), hepatic failure from vascular thrombosis (3), cardiac events (3), renal failure (2), acute stroke (2), sepsis (1), idiopathic pulmonary hypertension (1), and disseminated intravascular coagulation (1) were the causes of death in the nonrecurrent PBC group. The causes of death were unknown in the remaining 9 nonrecurrent PBC patients. Of the malignant deaths, 2 patients developed lung cancer (1 patient with non-small-cell carcinoma, and another with squamous cell carcinoma), 2 developed posttransplantation lymphoproliferative disorder (1 patient with central nervous system lymphoma, Epstein-Barr virus –related lymphoma in another). Hepatic artery thrombosis (2), primary graft failure (1), outflow obstruction (1), and chronic rejection (1) were the causes of early LT (defined as within 3 months from primary LT), while ductopenic rejection (5), hepatic artery thrombosis (2), and biliary duct strictures (2) were the causes of late LT.

The presence of recurrent PBC was not associated with overall death or retransplantation (hazard ratio 1.04; 95% CI, 0.45–2.43; P = 0.9), even after adjusting for age and gender (hazard ratio 0.97; 95% CI, 0.41–2.31; P = 0.9). Furthermore, cyclosporine- or tacrolimus-based immunosuppression showed no significant influence on long-term survival for PBC patients after adjusting for age and gender (hazard ratio 0.86; 95% CI, 0.42–1.68; P = 0.7). The prognostic factor associated with all causes of mortality or liver retransplantation for all PBC patients was duration of maintenance corticosteroids (yr) after adjusting for age and gender (hazard ratio 0.90; 95% CI, 0.83–0.97; P = 0.008).

UDCA Treatment in Patients With Recurrent PBC

A total of 38 (80%) patients with recurrent PBC received UDCA while 14 patients (20%) with recurrent PBC did not receive UDCA. The mean dosage of UDCA prescribed was 12 mg/kg/day (range, 6.7–20 mg/kg/day). The average time period for UDCA therapy was 55 months (range, 4–121 months). A total of 33 patients (87%) were treated with UDCA for a duration of over 24 months. Of 38 patients (39%), 15 were treated with UDCA dosage ≥13 mg/kg/d for 62.3 ± 37.3 months (range, 5–121 months).

Baseline characteristics, including age at the initial transplantation, gender, serum aminotransferase level, total bilirubin level, Mayo risk score, antimitochondrial antibody status, type of calcineurin inhibitor used, and maintenance corticosteroid therapy, were similar between treated and untreated patients (Table 2). Patients receiving UDCA had a higher baseline serum ALP level compared to untreated patients (P = 0.004). Six patients with recurrent PBC (12%) had experienced disease-related symptoms before treatment with UDCA. Of these patients, 2 had pruritus, 2 had fatigue, 1 had pruritus and Raynaud's phenomenon, and 1 had fatigue and Sjögren's syndrome.

Table 2. Comparison of Baseline Characteristics of Both Treatment Groups for Recurrent PBC
 UDCA treatmentNo UDCA treatmentP value
  • NOTE: Data represent mean ± SD and number (percentage).

  • Abbreviations: HLA, human leukocyte antigen; AST, aspartate aminotransferase; ALT, alanine aminotransferase; AMA, antimitochondrial antibodies.

  • *

    Corticosteroid withdrawal data, Mayo risk score, liver biochemical indices, and AMA titers were obtained at the time of first diagnosis of PBC recurrence.

Number of patients3814 
Age at time of recurrence (yr)61.1 ± 7.959.1 ± 9.60.5
Gender (female)30 (79%)10 (71%)0.6
Race (Caucasian)37 (97%)13 (93%)0.5
HLA match ≥122 (65%)7 (54%)0.5
Primary immunosuppression   
 Tacrolimus22 (58%)9 (64%)0.8
 Cyclosporine16 (42%)5 (36%) 
Corticosteroids withdrawal*19 (50%)6 (43%)0.6
Mayo risk score*4.3 ± 0.64.0 ± 0.60.09
Liver biochemical indices*   
 Alkaline phosphatase (U/L)446 ± 270245 ± 1250.004
 AST (U/L)52 ± 3451 ± 450.2
 ALT (U/L)70 ± 6559 ± 560.3
 Serum bilirubin (mg/dL)0.9 ± 0.40.7 ± 0.30.3
 Serum albumin (gm/dL)4.0 ± 0.34.1±0.40.1
 Prothrombin time (seconds)9.7 ± 0.810.1 ± 0.80.1
AMA titer ≥1:40*17 (59%)8 (73%)0.4
Histologic stage   
 Stage 02 (5%)1 (7%) 
 Stage 132 (84%)13 (93%) 
 Stage 23 (8%)0 (0%) 
 Stage 31 (3%)0 (0%) 

In the group receiving UDCA therapy, serum ALP levels showed improvement at 3 yr (P = 0.02), as did levels of serum aspartate aminotransferase (P = 0.004) and serum alanine aminotransferase (P = 0.001). No significant changes in levels of serum total bilirubin, albumin, and prothrombin time were apparent with UDCA treatment. Following a mean duration of 3 yr, an estimated 52% of UDCA-treated patients experienced normalization of serum ALP and alanine aminotransferase levels with minor fluctuations above normal range, compared to 22% of untreated patients.

Serial liver histology to compare progression rates was available in 18 UDCA-treated patients and 8 untreated patients. The median time interval between diagnostic and last follow-up liver biopsies was 48 months (range, 12–131 months) in the UDCA-treated group and 27.5 months (range, 6–108 months) in the untreated group. Progression in histological stage occurred in 9 UDCA-treated patients (5 by 1 stage, 3 by 2 stages, and 1 by 3 stages) and 3 untreated patients (1 by 1 stage, 1 by 2 stages, and 1 by 3 stages), while no change in histological stage was noted in the other 7 UDCA-treated patients and 4 untreated patients. Regression in histological stage occurred in 2 treated patients by 1 and 2 stages, respectively, and 1 untreated patient by 1 stage. Overall, there was no significant difference in histologic stage progression rate between the UDCA-treated and untreated groups (P = 0.9) over a median follow-up of 40 months (range, 6–131 months) as shown in Figure 5.

Figure 5.

Probability of fibrosis progression of recurrent PBC in patients treated with UDCA (solid line) and nontreated patients (broken line) (P = 0.9).

The effect of UDCA on histologic progression in a subgroup of patients based on biochemical response was also calculated. Of the 26 patients with comparative histology, 58% (11/18 UDCA-treated patients and 4/8 untreated patients) had evidence of biochemical response as defined by a decrease in ALP greater than 40% of baseline or normal levels after 1 yr. A total of 3 (27%) of 11 UDCA-treated patients with biochemical response had histologic progression, while 1 (25%) untreated patient had evidence of fibrosis progression during follow-up. Of the 11 patients without biochemical response, progression in histological stage occurred in 6 UDCA-treated patients and 2 untreated patients. Fortunately, the proportion of individuals with histologic progression in the setting of a biochemical response (27%) was significantly lower compared to patients without this response (73%), regardless of treatment group (P = 0.02).

One patient had fibrosis progression from stage 1 to cirrhosis and required retransplantation approximately 4 yr after receiving UDCA treatment, whereas another untreated patient had histologic progression from fibrosis stage 1 to cirrhosis within 6.3 yr and required retransplantation. Overall, 5 of 38 patients receiving UDCA and 1 of 14 untreated patients died within the study period. Probability of death or liver retransplantation for the UDCA-treated group was not significantly different from those who were not treated with UDCA (P = 0.7), as shown in Figure 6.

Figure 6.

Probability of death or retransplant for recurrent PBC in patients treated with UDCA (solid line) and nontreated patients (broken line) (P = 0.7).

DISCUSSION

This study reports the long-term follow-up of PBC patients who underwent OLT and the impact of recurrent PBC as well as treatment with UDCA from a single center. Our results show that recurrent PBC is not likely to affect long-term patient or graft survival after primary LT. While UDCA appears to be the most effective drug available for PBC in the native liver, the efficacy of UDCA for recurrent PBC remains unknown. From this experience, it appears that UDCA may have benefits in patients with recurrent PBC who develop elevated serum liver biochemistries. Furthermore, patients with recurrent PBC who experience a biochemical response at least 1 yr after treatment could also have a delay in histologic progression, although extended follow-up is required to verify this observation.

Overall cross-sectional rates of recurrent PBC in previous studies increase with time, ranging from 8 to 18% at 3–5 yr after transplantation and 23 to 32% at 10 yr.3–5, 24–30 Few investigations, however, have reported the cumulative incidence rate,3, 5 which provides better information on the overall risk for recurrent disease. In our experience, the cumulative incidence rate of recurrent PBC is 22%, 37%, and 43% at 5, 10, and 15 yr, respectively. There are several factors that can influence the identification of disease recurrence, including: 1) the variable use of protocol biopsies; 2) variably stringent criteria for recurrent disease; 3) no consensus as to the minimum number of portal tracts required for diagnosis; 4) the immunosuppressive regimen used; and 5) the length of follow-up. The most important factor relates to the use and timing of protocol biopsies in follow-up. Because histological features of recurrent PBC are often unrelated to biochemical indices as defined in this study, the performance of liver biopsy for clinical indications (i.e., symptoms, abnormal serum liver biochemistries) alone will underestimate the extent of this problem. Thus, our center continues to perform liver biopsies in a protocolized fashion to identify patients whose risk for developing recurrent PBC may increase over time.

The pathogenesis of PBC remains unclear; however, several studies suggest an autoimmunopathogenesis.31 Recurrence of disease in the liver allograft suggests that the mechanisms that lead to the initial development of PBC persist after transplantation. Moreover, defining the risk factors associated with recurrent PBC may give further insight into the pathogenesis of this disease. The absence of corticosteroid therapy may be an important factor in developing recurrent PBC.32 Our data illustrate that patients without evidence for recurrent disease received, on average, a longer duration of maintenance corticosteroids therapy compared to individuals with recurrent PBC. However, this observation did not remain significant in multivariate analysis, in which type of calcineurin-based immunosuppression used and recipient gender emerged as more important associated factors. It appears that duration of corticosteroid use is a consequence of immunosuppression strategy since tapering of corticosteroids became standard practice with the widespread use of tacrolimus. Unfortunately, the relationship between systemic corticosteroid use and recurrent PBC remains uncertain.

The mechanism of calcineurin inhibition as it relates to the pathogenesis of recurrent PBC remains unclear.33, 34 The current study confirms data from several centers showing a significant reduction in the time to PBC recurrence with tacrolimus as compared to cyclosporine.3, 27, 28 Tacrolimus is also an independent predictor of overall risk for recurrent PBC.5, 28–30 In contrast; a number of studies have not observed a relationship between times to recurrence based on type of calcineurin inhibitor.25, 27, 30 Smaller patient numbers in these studies may in part be responsible for these results. Sex hormones are believed to influence the onset and severity of autoimmune disease by modulating lymphocytes at various stages in life.35 A recent study described the ability of the hypothalamic-pituitary-gonadal hormone axis and sex steroid metabolism to resume physiological function following OLT in men, despite significant alterations prior to OLT.36 The majority of patients receiving LT for PBC are female, but there is no clear reason for increased susceptibility in male recipients. Moreover, calcineurin inhibitors have favorable effects on sexual hormone levels of male transplant recipients and there is no difference in baseline hormone levels between cyclosporine- and tacrolimus-treated male patients.37

Previous studies suggested that recurrent PBC may not significantly impact posttransplantation survival rate up to 10 yr3, 30; however, no larger studies have examined patient survival after this time period. Thus, we extended follow-up after transplantation to 20 yr and found that there was no significant difference in patient or graft survival for patients with and without recurrent PBC. In our series, no patient with recurrent PBC died because of this condition, but there are 2 patients with recurrent PBC requiring retransplantation for end-stage liver disease. Furthermore, our data confirm that either cyclosporine- or tacrolimus-based immunosuppression provide excellent long-term survival for PBC patients who undergo LT.38 However, using Cox regression analysis, we were able to demonstrate that long-term use of corticosteroids had a significant impact on decreased patient and graft survival of PBC patients who underwent OLT. Specifically, both cardiovascular death and infection were observed as the 2 leading causes of death in patients maintained on systemic corticosteroid therapy, which might be explained by higher rates of metabolic complications such as hypertension, diabetes mellitus, and dyslipidemia.39, 40

Few studies have investigated the potential impact of donor age on the risk for recurrent PBC as compared to chronic hepatitis C.41 Given the cadaveric donor organ shortage in the United States and the increasing use of older donor livers, this observation may became important in the future. In our experience, patients with recurrent PBC were more likely to have received grafts from older donors when compared to those without disease recurrence. However, this observation did not remain a significant risk factor for developing recurrent PBC in multivariate analysis.

No standard approach exists for the treatment of recurrent PBC. Given the near universal presence of early-stage disease at diagnosis, a potential role for UDCA therapy exists. Several centers reported their anecdotal experience with UDCA.3, 27, 29 The present study demonstrates that most asymptomatic patients with recurrent PBC have early-stage histological disease. Furthermore, a majority of these patients will have an improvement in serum liver biochemistries with UDCA treatment. Nevertheless, the lack of effect on delaying histological progression appears to correlate with the drug's apparent ineffectiveness in reducing mortality or the need for retransplantation. However, the reduction or normalization of the serum ALP level induced by UDCA treatment occurs early after the beginning of therapy. Therefore, this marker could be useful in identifying patients who are likely to respond to UDCA treatment. In our study, a subset of patients with evidence of biochemical response (defined by a decrease in serum ALP greater than 40% of baseline or normal levels after 1 yr of treatment) were observed to have a trend toward delayed histologic progression. Recently, a study of UDCA treatment for PBC patients prior to LT showed that the survival of patients with this biochemical response was better than that predicted by the Mayo model and similar to that estimated for the control population.42 However, conclusive evidence will require a greater number of patients who are followed-up for a longer period of time.

The effect of UDCA therapy on histologic progression and survival of recurrent PBC in this study should be interpreted in the context of study limitations. First, the sample size of this study was too small to detect any clinically important difference, if there was any. Second, patients receiving UDCA therapy are more likely to have features of liver disease than untreated patients, such as higher levels of baseline serum ALP, and more advanced fibrosis (stage 2 or greater). Third, the most important requirement for a reliable assessment of any drug effect on liver histology is comparability of serial liver specimens. Various stages of disease can be present simultaneously in a single liver, and histological changes in the liver evolve at different rates. Thus, histological changes as the result of sampling variation cannot be excluded. Furthermore, not all patients underwent follow-up biopsy, nor was the interval between specimens the same. As a result, information on the histological change occurring during therapy was incomplete. Fourth, this study did not provide estimates of efficacy for UDCA therapy against a placebo. Indeed, it was evaluated between immunosuppressive drugs and the combination of UDCA and immunosuppressive drugs. Immunosuppression as part of post-OLT therapy not only controls activation of allogeneic immune-mediated responses but it also could potentially modify recurrent PBC, as it is a chronic immune-mediated disease. Fifth, given that we cannot evaluate compliance with UDCA treatment, the overall efficacy may be underestimated. Sixth, several randomized trials of UDCA therapy for PBC patients in the native liver have shown that a dosage of 13–15 mg/kg/day leads to delayed disease progression and prolonged survival free of LT.14–18 Unfortunately, two-thirds of our patients received UDCA at a potentially suboptimal dose (less than 13 mg/kg/day) because some physicians prescribed the same dosage of UDCA as pre-OLT treatment without awareness of individuals' weight gain after OLT. Finally, an effect of UDCA therapy on patient survival or eliminating the need for retransplantation cannot be seen in patients with mild disease because they did not have time to progress to an advanced stage in the short period of treatment. Theoretically, demonstration of such an effect in these patients would require a clinical trial with a large number of patients and long-term follow-up. However, this may be difficult to perform based on the declining rate of primary LT for PBC.

In conclusion, it appears that recurrent PBC after LT has had little impact to date on long-term survival or need for retransplantation. Long-term immunosuppression with a cyclosporine-based regimen is associated with a reduced incidence of recurrent PBC, while the prolonged use of corticosteroids within this regimen is associated with a decrease in patient and graft survival. The effects of UDCA therapy for recurrent PBC, including normalization of liver biochemistries and a trend toward delayed histologic progression in treatment responders, suggest that UDCA may be a useful therapy for preventing disease progression. Determining whether UDCA should be used to influence the natural history of recurrent disease requires the conduct of randomized controlled trials with extended follow-up.

Acknowledgements

P.C. was supported by an overseas medical scholarship from the Faculty of Medicine, Siriraj Hospital, Mahidol University.

Ancillary