Inflammatory bowel disease after liver transplantation: the effect of different immunosuppressive regimens

Authors


Dr E. B. Haagsma, Department of Gastroenterology and Hepatology, University Hospital Groningen, Hanzeplein 1, PO Box 30.001, 9700 RB Groningen, The Netherlands.
E-mail: e.b.haagsma@int.azg.nl

Summary

Background : Seemingly conflicting results have been reported on the prevalence and severity of inflammatory bowel disease after liver transplantation. Regimens with different combinations of drugs can be used for immunosuppression after transplantation.

Aim : To study retrospectively the prevalence of inflammatory bowel disease after liver transplantation, and the possible relationship with maintenance immunosuppressive regimens.

Methods : All 78 patients with end-stage primary sclerosing cholangitis (48 patients) or autoimmune cirrhosis (30 patients), transplanted between 1979 and July 2001, and with a follow-up of at least 1 year, were eligible for this study. In addition to patient and transplant characteristics, data on inflammatory bowel disease and immunosuppression before and after transplantation were collected. The Kaplan–Meier method was used for survival analysis. Possible risk factors for inflammatory bowel disease after transplantation were analysed by Cox univariate and multivariate regression.

Results : The median follow-up after transplantation was 7.2 years (range, 1.1–22.3 years). Nine of 25 patients with pre-transplant inflammatory bowel disease experienced flare-ups after transplantation. Six of 53 patients without pre-transplant inflammatory bowel disease developed de novo inflammatory bowel disease after transplantation. The cumulative risks (standard errors in parentheses) for inflammatory bowel disease were 6% (3%), 12% (4%) and 20% (5%) at 1, 3 and 5 years after transplantation, respectively. The inflammatory bowel disease-free survival was significantly higher in patients not receiving tacrolimus vs. those receiving tacrolimus, in patients receiving azathioprine vs. those not receiving azathioprine and in patients taking the regimen prednisolone–azathioprine–ciclosporin A vs. those taking tacrolimus–prednisolone. Pre-transplant inflammatory bowel disease and the use of tacrolimus were found to be independent predictors for inflammatory bowel disease after transplantation.

Conclusions : The prevalence of inflammatory bowel disease after liver transplantation is affected by the immunosuppression used. Azathioprine seems to have a protective effect and tacrolimus a promoting effect.

Introduction

Inflammatory bowel disease is thought to result from inappropriate and ongoing activation of the mucosal immune system.1 Aminosalicylates and courses of corticosteroids are effective treatment regimens for exacerbations of the disease. In severe cases of ulcerative colitis, ciclosporin A may be tried. In addition to aminosalicylates, azathioprine has a role as maintenance therapy in preventing exacerbations. End-stage primary sclerosing cholangitis and autoimmune cirrhosis are well-known indications for orthotopic liver transplantation (OLT). Both diseases are associated with an increased incidence of inflammatory bowel disease. In patients with primary sclerosing cholangitis, 55–75% develop inflammatory bowel disease during the course of their condition and, in patients with autoimmune cirrhosis, 4% do so.2–4

After OLT, exacerbations of inflammatory bowel disease and the development of de novo inflammatory bowel disease may occur despite the continuous use of immunosuppressive drugs for the prevention of liver graft rejection. Seemingly conflicting results have been reported on the prevalence and severity of inflammatory bowel disease after OLT.5–8 It is possible that these different outcomes are related to the use of different maintenance immunosuppressive regimens after OLT. Immunosuppressive regimens, with different combinations of drugs, have been implicated in the timing of the recurrence of autoimmune disease in the transplanted liver.9 Papatheodoridis et al. reported an aggressive course of inflammatory bowel disease after OLT in patients with primary sclerosing cholangitis who were treated without long-term steroids as part of their maintenance immunosuppressive regimen.5 On the other hand, Befeler et al. reported a mild course in patients on triple therapy with prednisone, azathioprine and ciclosporin A.6

The first OLT was performed in 1963 and, in the following decades, maintenance immunosuppression consisted of a combination of prednisone and azathioprine. Ciclosporin and tacrolimus, both calcineurin inhibitors and potent anti-rejection drugs, became available in the 1980s and 1990s, respectively. Our liver transplant programme started in 1979; in recent years, we have received the impression that the prevalence of exacerbations of inflammatory bowel disease and of de novo inflammatory bowel disease has increased. Based on the above, we decided to study retrospectively the prevalence of inflammatory bowel disease after OLT, and the possible relationship between inflammatory bowel disease and maintenance immunosuppressive regimens.

Patients and methods

All adult patients with end-stage primary sclerosing cholangitis or autoimmune cirrhosis, who were transplanted between 1979 and 1 July 2001, were eligible for the present study. Patients who died within the first year after OLT were excluded. The end-of-study date was 1 July 2002, and so the minimal follow-up per patient was 1 year.

Data were collected from the medical records at our hospital. Whenever a patient is referred to our hospital for liver transplantation, copies of the medical records from the referring hospital, mainly letters from internists and gastroenterologists, are collected and added to our records. During the evaluation for OLT, as a rule, all patients with primary sclerosing cholangitis are screened for inflammatory bowel disease with at least sigmoidoscopy. After OLT, records are maintained of all events concerning the health of the patients.

Data were collected on age, gender, diagnosis of liver disease, immunosuppression used for liver disease at the time of OLT, date(s) of transplantation and date of death.

Inflammatory bowel disease before OLT

Retrospective confirmation of a recorded diagnosis of inflammatory bowel disease before OLT was based on the symptoms at presentation, endoscopic and sometimes radiological investigations of the colon and, in most cases, histology. Follow-up data on the response to therapy, exacerbations and repeat endoscopies were helpful for confirmation or rejection of the diagnosis, when in doubt. An exacerbation was defined by a combination of recurrent symptoms, endoscopic findings of colitis and the need for a significant change or re-start of therapy. The following parameters on pre-transplant inflammatory bowel disease were collected: date of diagnosis (interval to OLT), type of inflammatory bowel disease, presence of symptoms at diagnosis, maximum extent of inflammatory bowel disease, maximum therapy and last recorded exacerbation before OLT (interval to OLT). In addition, we noted the clinical degree of disease at the time of OLT and the inflammatory bowel disease-related medication at that time.

Inflammatory bowel disease after OLT

None of the patients with pre-transplant inflammatory bowel disease continued specific inflammatory bowel disease therapy, mainly aminosalicylates, after OLT, as the immunosuppression given to prevent rejection was considered to be sufficient for the prevention of inflammatory bowel disease activity. A diagnosis of recurrent or de novo inflammatory bowel disease was based on symptoms, endoscopic findings, histology and the exclusion of other possible causes of colitis, such as enteropathogenic bacteria, Clostridium difficile toxins or cytomegalovirus infection. The following parameters on post-transplant inflammatory bowel disease were collected: date of diagnosis (interval after OLT), type of inflammatory bowel disease, exacerbation or de novo inflammatory bowel disease, immunosuppression, presenting symptoms, extent as judged by endoscopy and inflammatory bowel disease therapy needed for remission. The course and outcome were also noted.

Immunosuppressive regimens

Basically, three immunosuppressive regimens have been used for long-term maintenance anti-rejection therapy since the start of our programme in 1979. First, in the earlier years, a combination of prednisone and azathioprine was used. Standard maintenance dosages were prednisolone, 10 mg/day, and azathioprine, 125 mg/day. Second, ciclosporin A was added in later years, which resulted in a triple drug regimen. Standard maintenance dosages were prednisolone, 10 mg/day, azathioprine, 125 mg/day, and ciclosporin A at a dosage aimed at low whole-blood levels as measured by high-performance liquid chromatography (∼ 100 µg/L). Third, when tacrolimus became available, a regimen with prednisolone and tacrolimus was frequently used. In addition to these three basic maintenance regimens, variations were used if needed. Variations were related to side-effects of the drugs (renal function, diabetes, osteoporosis, leucopenia) or to treatment of allograft rejection. Mycophenolate mofetil was given when more immunosuppression was needed, in cases of azathioprine intolerance or in cases of decreased renal function. A change to a regimen with tacrolimus was made when signs of chronic rejection were present. Most changes in immunosuppressive regimens took place in the first 6 months after OLT. For the present study, we noted the immunosuppressive regimen on discharge from hospital, at 1 year after OLT and at the time of the first episode of inflammatory bowel disease after OLT.

Statistical analysis

The chi-squared test was used when appropriate. The Kaplan–Meier method was used for survival analysis. The influence of possible risk factors for inflammatory bowel disease after OLT was first analysed by Cox univariate regression. Variables that achieved a significance level below 0.20 were subsequently analysed by forward and backward stepwise Cox multivariate regression. All data were analysed using the Statistical Package for Social Sciences 11.0 (SPSS Inc.). A two-tailed P value below 0.05 was considered to indicate statistical significance.

Results

The study group consisted of 78 patients with a median age at the time of OLT of 38 years (range, 16–66 years); 35 were female. The median follow-up after OLT was 7.2 years (range, 1.1–22.3 years).

Pre-transplant characteristics

The pre-transplant characteristics are listed in Table 1. The pre-transplant diagnosis was primary sclerosing cholangitis in 48 patients and autoimmune cirrhosis in 30 patients. A gender difference was noted, such that two-thirds of the patients with primary sclerosing cholangitis were male, whereas two-thirds of the patients with autoimmune cirrhosis were female. At the time of transplantation, immunosuppression was used by 6% and 87% of the patients with primary sclerosing cholangitis and autoimmune cirrhosis, respectively. The use of immunosuppression was related to liver disease in all patients with autoimmune cirrhosis, 63% of whom used a combination of prednisone and azathioprine. Three patients with primary sclerosing cholangitis used immunosuppression for ulcerative colitis, Wegener's disease and sarcoidosis, respectively.

Table 1.  Pre-transplant characteristics
 All patients (n = 78)PSC (n = 48)AIC (n = 30)
  1. AIC, autoimmune cirrhosis; OLT, orthotopic liver transplantation; PSC, primary sclerosing cholangitis.

Age (years)38 (16–66)39 (16–66)27 (18–58)
Gender (female/male)35/43 (45%/55%)15/33 (31%/69%)20/10 (67%/33%)
Diagnosis of liver disease
 PSC48 (61%)48 
 AIC30 (39%) 30
Calendar month and year of OLT (median, range)January 1995 (Mar 1980–June 2001)February 1995 (May 1987–June 2001)December 1994 (Mar 1980–June 2001)
Immunosuppression at donor call (n)
 None49 (63%)45 (94%)4 (13%)
 Prednisone8 (10%)2 (4%)6 (20%)
 Azathioprine1 (1%)01 (3%)
 Prednisone + azathioprine20 (26%)1 (2%)19 (63%)
Inflammatory bowel disease
 No of patients with inflammatory bowel disease25 (32%)23 (48%)2 (7%)
 Ulcerative colitis/indeterminate/Crohn's disease20/4/118/4/12/0/0
 Symptomatic at diagnosis (n)23212
 Interval to OLT at diagnosis (years)11.7 (0.5–21.7)11.7 (0.5–21.7)1.7–16.3
 Interval to OLT of last flare-up (years)7.9 (0.4–19.7)7.9 (0.4–19.7)1.7–8.4
 Maximum extension before OLT (n)
  Pancolitis18 (72%)16 (70%)2 (100%)
  Left4 (16%)4 (17%) 
  Ileo-colonic1 (4%)1 (4%) 
  Unknown2 (8%)2 (9%) 
 Maximum therapy before OLT (n)
  None3 (12%)3 (13%)0
  Mesalazine/salazosulfapyridine14 (56%)13 (57%)1 (50%)
  Corticosteroids3 (12%)2 (9%)1 (50%)
  Hospitalization5 (20%)5 (22%)0
  Colectomy000
 Medication at donor call (n)
  None5 (20%)4 (17%)1
  Mesalazine18 (72%)17 (74%)1
  Budesonide enema1 (4%)1 (4%)0
  Corticosteroids orally1 (4%)1 (4%)0

Inflammatory bowel disease was diagnosed in 25 (32%) of the 78 patients (48% of the patients with primary sclerosing cholangitis and 7% of those with autoimmune cirrhosis). The diagnosis was ulcerative colitis in 20 patients, indeterminate colitis in four patients and Crohn's disease in one patient. All patients were symptomatic (mostly diarrhoea) at the time of diagnosis, except for two in whom the diagnosis was made during the work-up for liver transplantation. The median intervals between the diagnosis of inflammatory bowel disease and OLT and between the last exacerbation and OLT were 11.7 and 7.9 years, respectively. A majority of 72% showed pancolitis, defined as extension beyond the splenic flexure. Most patients (56%) responded well to mesalazine, 12% needed out-patient treatment with corticosteroids and 20% needed hospitalization during treatment. None of the patients underwent colectomy. At the time of OLT, 72% of the patients with inflammatory bowel disease used mesalazine and 4% used oral corticosteroids.

Immunosuppression after OLT

On discharge from hospital, 18 patients (23%) used an immunosuppressive regimen containing tacrolimus. The regimens are listed in Table 2. Within 13 months after OLT, five patients had changed from a combination of prednisolone, azathioprine and ciclosporin A to a regimen containing tacrolimus (prednisolone–tacrolimus, three patients; prednisolone–azathioprine (or mycophenolate mofetil)–tacrolimus, two patients). Thus, at 13 months after OLT, 23 patients (29%) used a regimen containing tacrolimus.

Table 2.  Immunosuppressive regimens after orthotopic liver transplantation (OLT)
RegimenAt discharge from hospital (n = 78)At 13 months after OLT (n = 78)
Ciclosporin–prednisolone– azathioprine5448
Ciclosporin–prednisolone– mycophenolate mofetil11
Prednisolone–azathioprine55
Prednisolone–ciclosporin01
Tacrolimus–prednisolone1517
Tacrolimus–prednisolone– azathioprine02
Tacrolimus–prednisolone– mycophenolate mofetil34

The reasons for a change to tacrolimus included episodes of acute rejection related to drug intolerance (three patients) and signs of chronic rejection (two patients). None of these five patients developed signs of inflammatory bowel disease before the change to tacrolimus.

Inflammatory bowel disease after OLT

Fifteen patients (19%) experienced episodes of inflammatory bowel disease after OLT, either as flare-ups (nine patients) or as de novo inflammatory bowel disease (six patients). The cumulative risks (with standard errors in parentheses) for the development of inflammatory bowel disease after OLT were 6% (3%), 12% (4%), 20% (5%) and 23% (5%) at 1, 3, 5 and 10 years after OLT, respectively (Figure 1a). The median interval after OLT was 1.5 years (range, 0.3–6.3 years).

Figure 1.

(a) Cumulative incidence of active inflammatory bowel disease (IBD) after orthotopic liver transplantation (OLT). (b) Cumulative incidence of active IBD after OLT in subgroups with (first exacerbation; top curve) or without (de novo; bottom curve) pre-OLT IBD.

Subgroup with pre-OLT inflammatory bowel disease. Nine of the 25 patients (36%) with pre-OLT inflammatory bowel disease experienced exacerbations a median of 1.0 year (range, 0.3–4.6 years) after OLT. The cumulative risks (with standard errors in parentheses) for exacerbation were 20% (8%), 28% (9%), 39% (10%) and 39% (10%) at 1, 3, 5 and 10 years after OLT, respectively (Figure 1b). The details of these nine patients are listed in Table 3. Five of the nine patients responded well to therapy, and remained in remission afterwards on maintenance aminosalicylates. The other four patients showed a more complicated course with either a poor response to therapy or recurrent exacerbations.

Table 3.  Characteristics of patients with flare-up of inflammatory bowel disease after orthotopic liver transplantation
OLTSexAge
at
OLT
Liver
disease
IS at
discharge
Flare-up after
1st OLT
(2nd) (months)
Symptoms
and
endoscopy
IS at
diagnosis
Therapy and
outcome
  1. A, azathioprine; C, ciclosporin A; IS, immunosuppression; M, mycophenolate mofetil; OLT, orthotopic liver transplantation; P, prednisolone; PAC, prednisolone–azathioprine–ciclosporin A; PMT, prednisolone–mycophenolate mofetil–tacrolimus; PSC, primary sclerosing cholangitis; PT, prednisone–tacrolimus; T, tacrolimus.

1-73M45PSCPAC6Diarrhoea
Sigmoidoscopy:
moderate colitis
P, 11 mg
A, 150 mg
C level, 80 µg/L
Mesalazine, 4 × 1 g
P enema
Remission
Colectomy for serious
dysplasia 8 years
after OLT
2-185F51PSCPAC42Diarrhoea
Colonoscopy:
mild pancolitis
P, 10 mg
A, 125 mg
Olsalazine,
3 × 500 mg
Remission
3-284M26PSCPAC4Diarrhoea
Colonoscopy:
moderate
pancolitis
P, 20 mg
A, 125 mg
C level, 80 µg/L
P, budesonide
enemas. Remission
At least 5 flare-ups in
next 5 years
P, mesalazine
4-299M41PSCPAC55Diarrhoea, fever,
pain
Colonoscopy:
moderate colitis
of the right colon
P, 10 mg
A, 150 mg
C level, 150 µg/L
Mesalazine, 3 × 1 g
Remission
5-314M33PSCPMT19 (6)Diarrhoea
Sigmoidoscopy:
mild colitis
P, 10 mg
M, 2 g
T level, 8.5 µg/L
Mesalazine, 3 × 1 g,
budesonide enemas
Remission
Flare-ups 2 years
later with mainly
right-sided moderate
colitis
Difficult to treat
Mesalazine,
budesonide
orally and enemas
Switch from T to C
Remission
6-379M37PSCPT7Diarrhoea
Colonoscopy:
moderate pancolitis
P, 10 mg
T level, 9.5 µg/L
P, budesonide
enemas. Remission
Flare-up 1 year later
Remission on P
(allergy to
mesalazine)
7-424M35PSCPT16Diarrhoea, blood
Colonoscopy:
moderate patchy
pancolitis
T level, 8.2 µg/LP, mesalazine
Remission
8-509M52PSCPMT11Diarrhoea
Sigmoidoscopy:
mild proctitis,
rectal sparing,
upwards moderate
colitis
P, 10 mg
M, 2 g
T level, 9.1 µg/L
Mesalazine, 3 × 1 g
Remission
9-549M33PSCPMT5Diarrhoea
Colonoscopy:
moderate
pancolitis
P, 10 mg
M, 2 g
T level, 10.0 µg/L
P, mesalazine 4 × 1 g
Switch M to A
Poor response
Budesonide orally
and enemas
After 6 months,
switch T to C
Post or propter
remission

Subgroup without pre-OLT inflammatory bowel disease. Six of the 53 patients (11%) without pre-OLT inflammatory bowel disease developed de novo inflammatory bowel disease after OLT. This occurred at a median of 3.9 years (range, 1.1–6.3 years) after OLT, which is significantly later than the exacerbations (Mann–Whitney: P = 0.045). The cumulative risks (with standard errors in parentheses) for de novo inflammatory bowel disease after OLT were 0% (0%), 4% (3%), 11% (5%) and 14% (5%) at 1, 3, 5 and 10 years after OLT, respectively (Figure 1b). The details of the six patients are listed in Table 4. The type of inflammatory bowel disease was ulcerative colitis in three patients, indeterminate colitis in two patients and Crohn's colitis in one patient. Three of these six patients had autoimmune cirrhosis as original liver disease. All but two patients responded well to therapy. These two patients needed extra treatment with budesonide for exacerbations.

Table 4.  Characteristics of patients with de novo inflammatory bowel disease after orthotopic liver transplantation
OLTSexAge
at
OLT
Liver
disease

IS at
discharge
IBD after
1st OLT
(2nd) (months)
Symptoms
and
endoscopy

Diagnosis
of IBD
IS at
diagnosis
Therapy and
outcome
  1. A, azathioprine; AIC, autoimmune cirrhosis; C, ciclosporin A; IBD, inflammatory bowel disease; IS, immunosuppression; OLT, orthotopic liver transplantation; P, prednisolone; PAC, prednisolone–azathioprine–ciclosporin A; PSC, primary sclerosing cholangitis; PT, prednisolone–tacrolimus; T, tacrolimus; UC, ulcerative colitis.

1-85M32PSCPAC76Diarrhoea, blood,
mucus
Sigmoidoscopy:
diffuse colitis from
40 cm upwards
IndeterminateP, 10 mg
A, 150 mg
C level, 110 µg/L
Mesalazine, 4 × 1 g
Remission
2-326F23AICPAC54 (4)Diarrhoea, blood
Colonoscopy:
mild pancolitis with
rectal sparing
UCP, 10 mg
A, 100 mg
T level, 11 µg/L
Mesalazine, 3 × 1 g
Remission
Exacerbation after
discontinuation of
mesalazine
With mesalazine
remission
Later budesonide
enema added
for proctitis
3-383M44PSCPAC36Diarrhoea
Colonoscopy:
mild pancolitis
UCT level, 5.0 µg/LMesalazine, 3 × 1 g
Remission
4-389F38PSCPAC53 (25)Diarrhoea
Giardia infestation
Colonoscopy:
normal rectum,
colitis with small
and linear ulcers
compatible with
Crohn's colitis
Normal terminal
ileum
Crohn's colitisP, 10 mg
T level, 9.0 µg/L
P course (40 mg)
Switch to PAC
Remission
5-398F26AICPAC40Diarrhoea, blood
Colonoscopy:
normal
Pathology: colitis
UCP, 10 mg
A, 50 mg
T level, 9 µg/L
Mesalazine 3 × 1 g,
budesonide 6 mg
orally. Remission
Exacerbation after
discontinuation of
A and budesonide
Remission on
budesonide
6-499M43AICPT14Diarrhoea, blood
Colonoscopy:
moderate colitis
with distal sparing
IndeterminateP, 7.5 mg
T level, 8 µg/L
Mesalazine 3 × 1 g,
budesonide 9 mg
Remission
Exacerbation on
budesonide 3 mg
Remission on 9 mg

Immunosuppressive drugs and inflammatory bowel disease

Tacrolimus. About 1 year after OLT, as described above, 29% of patients used an immunosuppressive regimen containing tacrolimus. Survival analysis showed that the inflammatory bowel disease-free survival after OLT was significantly higher in the 55 patients who did not receive tacrolimus than in the 23 patients who received tacrolimus (P = 0.0000). At 1, 3 and 5 years after OLT, the inflammatory bowel disease-free survival rates (with standard errors in parentheses) in patients not receiving tacrolimus were 96% (3%), 96% (3%) and 90% (4%), respectively. In patients receiving tacrolimus, these values were 87% (7%), 63% (12%) and 36% (17%), respectively (Figure 2a). It is noteworthy that the group of 23 patients receiving tacrolimus included only two patients who also used azathioprine (Table 2).

Figure 2.

(a) Active inflammatory bowel disease (IBD)-free survival after orthotopic liver transplantation (OLT) in patients on an immunosuppressive regimen containing (bottom curve) or not containing (top curve) tacrolimus. P = 0.0000. (b) Active IBD-free survival after OLT in patients on an immunosuppressive regimen containing (top curve) or not containing (bottom curve) azathioprine. P = 0.0003.

The exclusion of the five patients who used mycophenolate mofetil in their immunosuppressive regimen (four on tacrolimus and one on ciclosporin A) revealed that the inflammatory bowel disease-free survival remained statistically higher in patients not receiving tacrolimus than in those receiving tacrolimus (P = 0.0004).

Azathioprine. A comparison of the 55 patients who received azathioprine with the 23 patients who did not receive azathioprine revealed a significantly higher inflammatory bowel disease-free survival in the former (P = 0.0003). Two of the 55 patients receiving azathioprine in this comparison also used tacrolimus. At 1, 3 and 5 years after OLT, the inflammatory bowel disease-free survival rates (with standard errors in parentheses) in patients receiving azathioprine were 96% (3%), 96% (3%) and 88% (4%), respectively. In patients not receiving azathioprine, these values were 87% (7%), 63% (12%) and 54% (13%), respectively (Figure 2b).

The exclusion of the five patients who used mycophenolate mofetil in their immunosuppressive regimen revealed that the inflammatory bowel disease-free survival was still significantly higher in patients receiving azathioprine (P = 0.0070).

A comparison of the 60 patients who received azathioprine or mycophenolate mofetil with the 18 patients who did not receive azathioprine or mycophenolate mofetil revealed no statistical difference (P = 0.0569). In this comparison, six of the 60 patients on azathioprine or mycophenolate mofetil also used tacrolimus.

Ciclosporin A–prednisolone–azathioprine vs. tacrolimus–prednisolone. A comparison of the subgroup of 48 patients who used the combination ciclosporin A–prednisolone–azathioprine with the subgroup of 17 patients who used tacrolimus–prednisolone (Table 2) again showed more inflammatory bowel disease in the patients receiving tacrolimus (P = 0.049). At 1, 3 and 5 years after OLT, the inflammatory bowel disease-free survival rates (with standard errors in parentheses) with the regimen ciclosporin A–prednisolone–azathioprine were 96% (3%), 96% (3%) and 89% (5%), respectively. With the regimen tacrolimus–prednisolone, these values were 94% (6%), 69% (14%) and 58% (16%), respectively.

Subgroup with pre-OLT inflammatory bowel disease. Survival without an exacerbation of pre-existing inflammatory bowel disease tended to be higher in patients not receiving tacrolimus (P = 0.0553). Five of the nine patients receiving tacrolimus (56%) experienced an exacerbation, in contrast with four of the 16 patients (25%) not receiving tacrolimus. A trend for a higher inflammatory bowel disease-free survival was also found for patients taking azathioprine (P = 0.0553). No difference was found between the regimens ciclosporin A–prednisolone–azathioprine (14 patients) and tacrolimus–prednisolone (five patients), but the numbers were small in this comparison.

Subgroup without pre-OLT inflammatory bowel disease. Survival without de novo inflammatory bowel disease was significantly higher in patients not receiving tacrolimus (P = 0.0000). Five of the 15 patients (33%) receiving tacrolimus developed de novo inflammatory bowel disease, in contrast with one of the 38 patients (3%) not receiving tacrolimus. A significantly higher inflammatory bowel disease-free survival was also found in patients receiving azathioprine (P = 0.0022) and in patients taking the regimen ciclosporin A–prednisolone–azathioprine (P = 0.0004).

Predictive factors for inflammatory bowel disease after OLT

Univariate analysis showed that male gender, the presence of pre-OLT inflammatory bowel disease, transplantation after 20 January 1995 and the use of an immunosuppressive regimen containing tacrolimus were predictive for inflammatory bowel disease after OLT (P = 0.0509, 0.0032, 0.0041 and 0.0000, respectively). Age and the diagnosis of primary sclerosing cholangitis were not predictive of inflammatory bowel disease after OLT. Subsequent multivariate analysis revealed that pre-OLT inflammatory bowel disease and the use of tacrolimus were independent risk factors for inflammatory bowel disease after OLT (P = 0.01 and 0.00, respectively).

Discussion

Seemingly conflicting results have been reported on the prevalence and severity of inflammatory bowel disease after OLT. The present study shows that different outcomes may relate to the specific drugs used for maintenance anti-rejection therapy.

The patients with pre-transplant inflammatory bowel disease in our study had experienced quiescent disease for many years before OLT, mostly on mesalazine maintenance therapy. Exacerbations occurred, however, as early as a median of 1.0 year after OLT. This observation matches the results of others that inflammatory bowel disease runs a more aggressive course after OLT, despite the use of immunosuppression.5, 7 As most patients (five of nine) responded well to mesalazine therapy, the number of patients with exacerbations might have been smaller had we not discontinued mesalazine post-operatively. In most series, however, mesalazine has been discontinued after OLT as the immunosuppression given has been considered to be sufficient for the prevention of exacerbations.

The prevalence of inflammatory bowel disease in autoimmune hepatitis has been found to be about 4%.4 In our series, it was 7% pre-transplant (two patients), with an increase to 17% (five patients) after OLT. Therefore, it may be speculated whether these patients with autoimmune cirrhosis would have developed inflammatory bowel disease if they had not undergone OLT. As five of the six patients with de novo inflammatory bowel disease received tacrolimus, the role of tacrolimus in the pathogenesis of inflammatory bowel disease after OLT may be questioned (see below).

Since the early years of liver transplantation, new immunosuppressive drugs have become available. Nowadays, different combinations can be used, including prednisolone, azathioprine or mycophenolate mofetil and ciclosporin A or tacrolimus. The patients in our study used different combinations of drugs for maintenance therapy, as the first was transplanted in 1980 and the last in 2001. This enabled us to compare inflammatory bowel disease-free survival rates between groups of patients on different drugs. The effect of prednisolone could not be tested as all but a few patients were using prednisolone at the time of their first inflammatory bowel disease episode. Likewise, ciclosporin A was not tested as all but one patient on ciclosporin A was also using azathioprine or mycophenolate mofetil. We found that the inflammatory bowel disease-free survival was increased in patients not receiving tacrolimus as part of their regimen compared with those receiving tacrolimus. In addition, the inflammatory bowel disease-free survival was increased in patients receiving azathioprine as part of their regimen compared with those not receiving azathioprine. Some overlap existed between the tacrolimus and azathioprine groups, but it seems likely that the reported differences basically reflect differences between tacrolimus use and azathioprine use. This is also illustrated by the clear difference in inflammatory bowel disease-free survival between the two most frequently used regimens, namely ciclosporin A–azathioprine–prednisolone and tacrolimus–prednisolone. These findings may explain the seemingly conflicting results in the literature so far. A comparison of the studies is hampered by differences in the study design, number of patients, length of follow-up and detail of the information. In contrast with others, for example, we used survival analysis as the most appropriate method to exclude the influence of different follow-up times. Nevertheless, it may be derived from the most detailed studies5–8 that the lowest rates of inflammatory bowel disease exacerbation (around 30%) are reported in patient groups receiving a ciclosporin A–azathioprine–prednisolone regimen6, 7 and the highest rates (around 50%) are found in patient groups in whom tacrolimus is often used8 or azathioprine is not always used.5 In this respect, the data on de novo inflammatory bowel disease cannot easily be compared due to the small numbers of patients in most studies.5–7 Riley et al. described 14 patients with de novo inflammatory bowel disease after liver or kidney transplantation, with 10 of the 14 patients using tacrolimus.10

We detected the relevance of immunosuppressive drugs after survival analysis of the whole group of patients, including those with and without pre-transplant inflammatory bowel disease. Subsequent analysis of the two smaller subgroups, with and without pre-transplant inflammatory bowel disease, revealed that the effect of tacrolimus and azathioprine was strongest in the subgroup without pre-transplant inflammatory bowel disease. In the subgroup with pre-transplant inflammatory bowel disease, the effect did not quite reach significance with P = 0.0553. Therefore, it may be concluded that the negative effect of tacrolimus or the positive effect of azathioprine is not present in patients with pre-transplant inflammatory bowel disease. However, we feel that the effect is probably there, but that the number of patients was too small to show the effect. In the whole group, we found that pre-transplant inflammatory bowel disease and the use of tacrolimus were independent predictors for inflammatory bowel disease after OLT. This finding also implies that a larger group of patients with pre-transplant inflammatory bowel disease is probably needed to find an unequivocal effect of tacrolimus (or azathioprine). Therefore, larger studies are needed in this respect.

The smoking history could be a possible confounding factor. In retrospect, we did not register the smoking habits of the patients in sufficient detail. The cessation of smoking after OLT could have increased the risk of the development of ulcerative colitis. However, all but one of the patients with active inflammatory bowel disease after OLT were non-smokers long before liver transplantation.

In the past, it was anticipated that inflammatory bowel disease would not recur after OLT because of the continued immunosuppression. However, it is now clear that inflammatory bowel disease can recur or develop de novo, that the course of the disease may be more aggressive than before OLT, and that there is a relationship with the types of drug used in the maintenance immunosuppressive regimens. It could be argued that corticosteroids and ciclosporin/tacrolimus are not effective maintenance drugs for the prevention of inflammatory bowel disease, and have been used mainly for the induction of remission.1, 11 In contrast, azathioprine is an effective drug for the prevention of inflammatory bowel disease activity.1 The observation that patients on azathioprine are not free from inflammatory bowel disease after OLT may then relate to the relatively low dosages used for anti-rejection therapy (1–2 mg/kg body weight), compared with the dosages used for the prevention of inflammatory bowel disease flares (2–2.5 mg/kg body weight). The discontinuation of mesalazine immediately after OLT, as reported in some studies, including ours, may also play a role.6

In addition, and possibly of greater importance, changes in the digestive tract induced by liver transplantation should be considered. Inflammatory bowel disease is thought to result from inappropriate and ongoing activation of the mucosal immune system driven by the presence of normal luminal flora. The aberrant response is most probably facilitated by defects in the barrier function of the intestinal epithelium and the mucosal immune system.1 In general, immunosuppressed patients are more prone to infections for which appropriate, often antibiotic, treatment is given. This may lead to episodes of changed bacterial gut flora and decreased barrier function. At these times, the mucosal immune system will be exposed to higher concentrations of different bacterial antigens. Furthermore, immunosuppressive drugs may give rise to an imbalance of the mucosal immune system. Tacrolimus is a strong inhibitor of interleukin-2 production. It is interesting to note that interleukin-2-deficient mice spontaneously develop colitis when reared with conventional gut flora.12 Deficiency in interleukin-2 can result in T-cell dysregulation, leading to the development of colitis because of the inability to mount a counter-regulatory (suppressive), transforming growth factor-β-mediated, response.13 Animal experiments have shown that the use of ciclosporin A predisposes to autoimmunity and autoimmune disease.14 Differences have been noted between tacrolimus and ciclosporin A. In renal transplant patients, the frequency of interleukin-2-expressing T cells was significantly lower with tacrolimus than with ciclosporin A.15 In addition, it has been established that the expression of transforming growth factor-β in tissues is higher with ciclosporin A treatment than with tacrolimus.16 The mucosa in ulcerative colitis contains CD4+ lymphocytes with a T-helper-2 phenotype,1 and interleukin-10 is an important T-helper-2 cytokine. In rats, interleukin-10 production by CD4+ lymphocytes was higher during tacrolimus therapy than during ciclosporin A therapy.17 In renal transplant patients, switching from ciclosporin A to tacrolimus resulted in enhanced CD4+ interleukin-10 responses and lipopolysaccharide-stimulated monocyte responses.18 Studies on the recurrence of primary biliary cirrhosis after OLT showed that patients on tacrolimus monotherapy experienced earlier recurrence of disease than those on ciclosporin A monotherapy.19, 20 All of these studies make it conceivable that patients receiving tacrolimus therapy will be more prone to inflammatory bowel disease than those receiving ciclosporin A and/or azathioprine therapy.

In conclusion, our findings confirm earlier reports on the exacerbation of inflammatory bowel disease and the occurrence of de novo inflammatory bowel disease after OLT for primary sclerosing cholangitis or autoimmune cirrhosis. The rate of inflammatory bowel disease is affected by the immunosuppression used. Azathioprine seems to have a protective effect, and tacrolimus a promoting effect. Further studies are needed to substantiate the role of tacrolimus in inflammatory bowel disease after OLT.

Ancillary