Long-term effect of tacrolimus therapy in patients with refractory ulcerative colitis


Dr H. Nakase, Department of Gastroenterology and Hepatology, Graduate School of Medicine, Kyoto University, 54 Shogoin Kawahara-cho, Sakyo-ku, Kyoto 606-8507, Japan.
E-mail: hiropy_n@kuhp.kyoto-u.ac.jp


Background  Little is known about long-term outcome of tacrolimus therapy for ulcerative colitis.

Aim  To evaluate long-term efficacy and safety of tacrolimus in Japanese patients with refractory ulcerative colitis.

Methods  Twenty-seven patients with UC refractory to conventional therapy were administered tacrolimus with trough whole-blood levels of 10–15 ng/mL to induce remission and 5–10 ng/mL to maintain remission. Median treatment duration was 11 months (1–39 months) and median follow-up duration was 17 months (2–65 months). Evaluation of the clinical response was based on a modified Truelove–Witts severity index (MTWSI).

Results  Tacrolimus produced a clinical response in 21 patients (77.8%), and remission was achieved in 19 of these 21 (70.4%) within 30 days. Overall cumulative colectomy-free survival was estimated as 62.3% at 65 months. In 18 of 19 patients treated with corticosteroids at the initiation of tacrolimus therapy, corticosteroids were discontinued or tapered. Adverse events were tremor (25.9%), renal function impairment (18.5%), infectious disease (14.8%), hot flashes (11.1%), hyperkalaemia (7.4%), headache (7.4%), epigastralgia (7.4%) and nausea (3.7%). No mortality occurred.

Conclusion  Long-term administration of tacrolimus appears to be an effective and well-tolerated treatment for Japanese patients with refractory ulcerative colitis.


Ulcerative colitis (UC) is an idiopathic chronic inflammatory disorder characterized by manifestations such as diarrhoea, rectal bleeding, abdominal pain, fever, anaemia and body weight loss.1 Immunomodulators have been integral to the treatment of refractory inflammatory bowel disease (IBD) since their introduction. The goals of treatment with these agents are to control active IBD, allow for the withdrawal of corticosteroids and maintain long-term remission in patients with IBD. Various immunomodulators, including azathioprine (AZA), mercaptopurine (MP), ciclosporin (CyA) and tacrolimus, are efficacious in patients with UC.1–3 Generally, the major limitation in the use of both AZA and MP for the management of active disease, however, is a delay in their onset of action, although response to thioprine may depend on its administered dose. Therefore, for its rapid action, CyA is widely used for steroid-refractory and severe UC. One placebo-controlled study4 and several open studies5–9 demonstrated the clinical efficacy of intravenous CyA in patients with an acute UC attack who failed to respond to intravenous corticosteroids. The disadvantages of CyA treatment include a high potential for serious side effects and prolonged hospitalization necessitated by the intravenous administration of the drug.10–12 Although a new oral microemulsion form of CyA appears to be effective for UC therapy, there are only limited data supporting its use for this indication.13

Tacrolimus is a macrolide antibiotic isolated from Streptomyces tsukubaenesis that has immunomodulatory properties. It inhibits the complexation of calcineurin with its respective cytoplasmic receptors cyclophilin and FK-binding protein 12, both of which regulate a calmodulin-dependent phosphatase, thus interrupting signal transduction pathways in T -cells.13–15 Although its mode of action is similar to that of CyA, the immunosuppressive effect is 30–100 times greater in vitro and 10–20 times greater in vivo than that of CyA and its intestinal absorption is more reliable, even in the presence of gastrointestinal disease.16 Tacrolimus is currently approved for the prevention of allograft rejection in patients undergoing liver or kidney transplantation.15, 17 Cochrane systematic reviews on the effectiveness and safety of immunosuppression with CyA vs. tacrolimus for liver18 and kidney19 transplant patients demonstrated that tacrolimus is superior to CyA in improving graft survival and in preventing acute rejection after transplantation. In bone marrow transplant recipients, the incidence of grade II–IV graft-versus.-host disease is significantly lower in patients treated with tacrolimus than in those treated with CyA.20, 21

Tacrolimus is also effective for patients with refractory UC. A randomized control study demonstrated the efficacy and safety of oral tacrolimus for inducing remission of refractory UC.22 To our knowledge, however, few studies have evaluated the medium- to long-term outcome of tacrolimus therapy in patients with UC refractory to conventional therapy.23–25 Here, we report our findings of the long-term effects and safety of tacrolimus in patients with refractory UC, with a median follow-up of 17 months.

Patients and methods


Between April 2001 and October 2007, 27 patients with UC who were resistant to or could not be treated with conventional therapy were enrolled in this retrospective, observational, single-centre study, which was reviewed and allowed by the Institutional Review Board of Kyoto University. Patients were informed about the potential risks and benefits of tacrolimus therapy and informed written consent to its use was provided by them. In all cases, the diagnosis was established according to standardized criteria by prior clinical assessment, radiology, endoscopy and histology. Patient characteristics at baseline are shown in Table 1. Fifteen men (55.6%) and 12 women (44.4%) were treated with tacrolimus. The median age at tacrolimus initiation was 31 years (range: 15–78 years). The median disease duration prior to tacrolimus treatment was 37 months (range: 2–236 months). Twenty-two of these patients (81.5%) were afflicted with pancolitis and five of these (18.5%) had left-sided colitis. No patient had proctitis. Patients were classified as steroid-resistant or steroid-dependent in accordance with the definition published previously by Ogata et al.22. Steroid resistance was defined as the lack of response to a systemic daily dose of 30 mg or more of prednisone per body over at least 2 weeks and steroid dependency was defined as either chronic active UC for more than 6 months or more than once a year or at least three times every 2 years regardless of intensive medical therapy including corticosteroids. Seven patients (25.9%) were steroid-resistant and 18 patients (66.7%) were steroid-dependent. The remaining two patients (7.4%) had not been treated with prednisone, because one patient was aged 73 years at disease onset and the other patient refused to take corticosteroids.

Table 1.   Patients’ baseline characteristics
  1. * Age and † months are shown.

  2. Numbers of patients are shown unless specified.

Age at tacrolimus initiation [median (range)]31* (15–78)
Disease duration prior tacrolimus [median (range)]37† (2–236)
Gender (%)
 Men15 (55.6)
 Women12 (44.4)
Disease site (%)
 Pancolitis22 (81.5)
 Left sided5 (18.5)
Steroid response (%)
 Steroid resistant7 (25.9)
 Steroid dependent18 (66.7)
 Naive2 (7.4)
Treatment prior to tacrolimus therapy (%)
 5-Aminosalicylic acid or salazosulfapyridine27 (100)
 Thioprines7 (25.9)
 Leucocytapheresis16 (59.3)
 Inflixmab1 (3.7)

Infliximab therapy failed to let one steroid-dependent patient into remission. Leucocytapheresis was not effective in any of the seven steroid-resistant patients, seven of 18 steroid-dependent patients and both steroid–naive patients. Thioprines had been administered to one steroid-resistant and six steroid-dependent patients. All patients had been treated with 5-aminosalicylic acid or salazosulfapyridine.

Definition of response

Disease activity was evaluated using a modified Truelove–Witts severity index (MTWSI);26 details are shown in Table 2. The potential maximum (worst) score is 21 and the potential minimum (best) score is 0. Clinical response was defined as a decrease in the MTWSI score of more than four points below baseline. Clinical remission was retrospectively defined as an estimated MTWSI score of 4 or less.

Table 2.   Modified Truelove–Witts severity index26
  1. Severe category: 12 or higher; remission category: 4 or less.

Bowel movement0–23–45–67–9≥10 
Nocturnal diarrhoeaNoYes    
Visible blood in stool (%)0<50≥50100  
Abdominal tendernessNoneMildModerateSevere  
Abdominal pain/crampingNoneMildModerateSevere  
Need for antidiarrhoealsNoYes    
General statusPerfectVery goodGoodAveragePoorTerrible
Faecal incontinenceNoYes    


All patients were hospitalized at the time of initiation of tacrolimus therapy, except one steroid-dependent woman. In 23 patients, the initial dose of oral tacrolimus was 0.1 mg/kg body weight per day. Four patients initially received tacrolimus intravenously at 0.01 mg/kg body weight per day and were later switched to the oral formulation. Dosage was then adapted to achieve trough tacrolimus whole-blood levels between 10 and 15 ng/mL for the induction of remission. The disease course was evaluated clinically with routine laboratory tests. Patients were closely monitored for side effects. Patients were hospitalized until their clinical condition was stabilized with tacrolimus levels in the therapeutic range. After patients responded to tacrolimus, corticosteroids were tapered by reducing the dose 5 to 10 mg/week (more slowly in patients with high disease activity) and completely withdrawn if possible. After inducing clinical remission, tacrolimus whole-blood trough concentrations were maintained at a lower level, between 5 and 10 ng/mL.

Patients in an outpatient setting were followed up at least every 4 weeks. Trough whole blood levels of tacrolimus were measured on each visit and the dose adjusted for each patient to achieve an appropriate trough level. Biochemical values, including liver function and serum creatinine levels were measured and patients were regularly interviewed about adverse events on each visit.

Tacrolimus was administered for a median of 11 months (range: 1–39 months). Patients were followed up for a median of 17 months (range: 2–65 months) following the initiation of tacrolimus therapy.


The primary endpoints of this study were the induction of remission and colectomy-free survival. Secondary endpoints were the percentage of patients who achieved corticosteroid withdrawal and treatment safety.


Continuous variables were expressed as the mean (s.d.) and compared by paired t-test. Clinical scores (i.e., MTWSI) were expressed as percentages and were compared by Wilcoxon signed-rank test. A P-value of <0.05 was considered statistically significant. Colectomy-free survival was assessed using the Kaplan–Meier method.


Induction of remission

Immediate outcome was assessed 30 days after starting tacrolimus, based on the MTWSI score and the serum C-reactive protein (CRP) levels. Of 27 patients, 19 (70.4%) achieved clinical remission within 30 days, and two other patients (7.4%) had an improved MTWSI score. Thus, 21 (77.8%) of 27 patients responded to tacrolimus therapy within 30 days. None of the patients underwent colectomy in the 30 days following the initiation of tacrolimus therapy. The median MTWSI score decreased from 10.5 at the initiation of tacrolimus therapy to 2.5 at 30 days after the start of treatment (Figure 1). The mean-serum CRP levels decreased from 16.7 ± 21.6 mg/L at tacrolimus initiation to 2.8 ± 3.8 mg/L at 30 days after the start of treatment (Figure 2). Both the MTWSI score (Wilcoxon signed-rank test: < 0.0001) and CRP levels (paired t-test: = 0.0024) decreased significantly.

Figure 1.

 Changes in modified Truelove–Witts severity index (MTWSI) scores during tacrolimus treatment (= 27). The MTWSI decreased significantly at 1 month after the initiation of tacrolimus treatment (Wilcoxon signed-rank test; < 0.0001) and the scores remained low during treatment. The interquartile range (IQR) is represented by the box, with the median MTWSI value indicated by the line within the box; the bars represent the most extreme values that did not qualify as outliers or 1.5 times the size of the IQR, whichever was smaller. Outliers are represented by small open circles.

Figure 2.

 Changes in the serum C-reactive protein (CRP) levels during tacrolimus treatment (= 27). The serum CRP levels decreased significantly at 1 month after the initiation of tacrolimus treatment and remained low during the treatment. Means ± S.E. are shown.

Colectomy-free survival

One patient opted to have surgery rather than using immunosuppressive drugs despite being in remission, thus colectomy-free survival was assessed only in the remaining 26 patients. Within a median follow-up of 17.5 months, seven patients (26.9%) underwent colectomy between 2 and 35 months following the initiation of tacrolimus therapy. Based on Kaplan–Meier survival analysis, overall cumulative colectomy-free survival was estimated as 62.3% at 65 months (Figure 3). Colectomy-free survival was significantly different between patients who achieved clinical remission or had an improved MTWSI score within 30 days and those who showed no clinical response to tacrolimus therapy within 30 days. The test statistics for the equality of survival distributions were = 0.0135 (log-rank), = 0.0074 (Breslow) and = 0.0087 (Tarone–Ware; Figure 4).

Figure 3.

 Proportion of patients without colectomy (= 26). The overall cumulative colectomy-free survival was estimated as 62.3% at 65 months, based on Kaplan–Meier survival analysis.

Figure 4.

 Colectomy-free time interval of patients who responded to tacrolimus therapy within 30 days (= 20, 76.9%) and of those who did not respond (= 6, 23.1%). Colectomy-free survival of patients who achieved clinical remission or had an improved modified Truelove–Witts severity index score within 30 days following tacrolimus initiation is depicted by the solid line and that of patients who did not respond to the therapy within 30 days is represented by the dashed line. Kaplan–Meier analysis was performed; log-rank test: = 0.0135; Breslow test: = 0.0074; and Tarone–Ware test: = 0.0087.

Steroid-sparing effect of tacrolimus

Eight patients (29.6%) were not on corticosteroids at the time of tacrolimus initiation and in all of them clinical remission was maintained for 30 days following the initiation of tacrolimus therapy. In the remaining 19 patients, the mean prednisolone dose at tacrolimus initiation was 23.2 mg/day. Steroid therapy was completely withdrawn in 16 of 19 patients and could be reduced in two other patients within 3 months after starting tacrolimus. In the remaining patient, tacrolimus was discontinued before the withdrawal of steroid therapy because of Pneumocystis carinii infection and this patient eventually underwent colectomy. Of 16 patients in whom corticosteroids could be withdrawn after starting tacrolimus, 14 did not require corticosteroid re-administration, but the remaining two patients required re-administration because of an UC flare-up. Of eight patients who eventually underwent colectomy, seven were no longer on corticosteroids when the colectomy was performed.

Adverse effects

The frequency of adverse events is shown in Table 3. All adverse events were controlled by conservative therapies. Tacrolimus withdrawal was necessary in five patients (18.5%). In four patients (14.8%), infection occurred, and one patient (3.7%) developed renal dysfunction. Among the four patients with infection, one developed infection with P. carinii, one developed gingival infection and the other two had bacterial pneumonia. In all cases with bacterial infection, tacrolimus was discontinued temporarily and the patients recovered with antibiotic administration. In one patient who developed renal dysfunction (his serum creatinine level increased from 0.8 to 2.5 mg/dL), diarrhoea and resulting dehydration caused by UC exacerbation may have been contributing factors. Tacrolimus therapy was discontinued and his renal function normalized. In four other cases (14.8%), serum creatinine levels increased above 1.3 mg/dL, but normalized in all cases after the dose of tacrolimus was reduced. Adverse events included tremor (25.9%, = 7), hot flashes (11.1%, = 3), hyperkalaemia (7.4%, = 2), headache (7.4%, = 2), epigastralgia (7.4%, = 2) and nausea (3.7%, = 1). In 14 patients (51.9%), hypomagnesaemia occurred with no symptoms and drug withdrawal was not necessary in any case. No mortality occurred.

Table 3.   Adverse events that developed during tacrolimus treatment
Adverse eventsCases (%)
  1. * In one case, tacrolimus withdrawal was necessary.

  2. † Infectious disease consisted of Pneumocystis carinii, gingival infection and bacterial pneumonia.

Hypomagnesaemia14 (51.9)
Tremor7 (25.9)
Renal function impairment (rise in creatinine above 1.3 mg/dL)*5 (18.5)
Infectious disease†4 (14.8)
Hot flashes3 (11.1)
Hyperkalaemia2 (7.4)
Epigastralgia2 (7.4)
Headache2 (7.4)
Nausea1 (3.7)


One placebo-controlled study22 and several uncontrolled studies23–25, 27, 28 have demonstrated short-term efficacy of tacrolimus treatment in patients with UC, but there are few reports of its long-term effects.23–25 To our knowledge, the clinical trial reported here is the first to demonstrate the long-term effects and safety of tacrolimus therapy in Japanese patients with UC.

First, we evaluated the effect of tacrolimus therapy on the induction of remission. Of 24 patients who were in an active disease phase at the initiation of tacrolimus therapy, 16 achieved clinical remission and two were clinically improved within 30 days after initiating tacrolimus therapy. Consistent with previous reports,22, 23, 27, 28 our results suggested that tacrolimus effectively induced a rapid remission in patients with refractory UC. In 16 of 19 patients who were treated with corticosteroids, administration was completely tapered off, and in two others, the corticosteroid dose could be reduced within 3 months after starting tacrolimus. These data suggested that tacrolimus is useful for inducing rapid improvement in patients with refractory UC.

Next, we evaluated whether a clinical response within 30 days after initiating tacrolimus therapy affected long-term clinical outcome, namely colectomy-free survival. Colectomy-free survival rates were significantly higher in patients who responded to tacrolimus therapy within 30 days than in those who did not. Indeed, six patients did not respond to tacrolimus therapy within 30 days and four of these patients underwent colectomy despite additional therapy, such as leucocytapheresis or infliximab, after tacrolimus administration. Baumgart et al.23 also compared colectomy-free survival rates based on the immediate outcome at 4 weeks after the initiation of tacrolimus therapy and demonstrated a statistically significant difference between patients who initially responded to tacrolimus and those who did not. Therefore, evaluation of the effect at approximately 4 weeks after starting tacrolimus may be a good strategy for predicting the clinical outcome of patients with refractory UC treated with tacrolimus.

The most important clinical outcome in this case series is the long-term efficacy and safety of tacrolimus for patients with refractory UC. Baumgart et al.23 reported that tacrolimus treatment prevented the need for colectomy in 31 (77.5%) of 40 UC patients for up to a mean follow-up period of 39 months. Fellermann et al.24 demonstrated an overall colectomy-free rate of 66% in 38 UC patients treated with tacrolimus with a mean observation period of 16.2 months. In the present series, 19 (73.1%) of 26 patients avoided colectomy during a median follow-up period of 17.5 months. Thus, our data of Japanese patients with refractory UC are comparable with previously reported Western data. Moreover, the adverse effects observed in the present series were similar to those in previous reports and all adverse effects were controlled by optimal therapies during long-term administration of tacrolimus. Taken together, these findings indicate that long-term administration of tacrolimus is a well-tolerated and promising therapy for patients with refractory UC.

A critical issue in tacrolimus treatment for patients with IBD is determining the optimal trough levels for inducing, as well as maintaining, remission without causing adverse effects. As for tacrolimus trough levels for inducing remission, Ogata et al.22 recently reported in a placebo-controlled, double-blind, randomized trial that a high trough concentration (10–15 ng/mL) significantly affects the induction of clinical remission in patients with refractory UC compared to a placebo group. In this series, we tried to adjust the trough level of tacrolimus to 10–15 ng/mL to induce remission, and demonstrated that 70.4% of the patients with refractory UC achieved clinical remission within 30 days after the initiation of tacrolimus therapy.

Regarding the long-term administration of tacrolimus, several reports indicate that although a high trough level of tacrolimus is required to control the intestinal immune response at the initial period after transplantation, the dose can be reduced successfully to 5 and 10 ng/mL by the end of the first post-transplant year in small bowel transplantation.29 Moreover, trough levels of 5–10 ng/mL are recommended for long-term administration of tacrolimus to avoid rejection in patients with liver, renal and small bowel transplantation.30 Accordingly, we also tried to taper the trough levels of tacrolimus to 5–10 ng/mL in our patients after achieving remission and with this dose 73.1% of our patients avoided colectomy during a median follow-up period of 17.5 months after initiating tacrolimus treatment. Thus, tacrolimus trough levels of 5–10 ng/mL may be optimal for maintaining remission in patients with refractory UC, although further investigation is needed.

Thioprines are the most widely used immunomodulators for maintenance therapy in refractory UC patients. Several open studies have demonstrated that thioprines are effective for treating patients with steroid-dependent UC,31–33 but the efficacy of thioprines in UC remains controversial.31 In our series, 17 of 21 tacrolimus responders (81%) avoided colectomy for a median follow-up period of 17 months after initiating tacrolimus treatment. This long-term effect of tacrolimus as maintenance therapy is comparable to the efficacy of thioprines in steroid-dependent UC.31 In this regard, the long-term administration of tacrolimus may be an alternative therapy for patients with UC refractory to thioprines or for those who cannot use thioprines because of the side effects.

Different medical options for induction and maintenance for patients with refractory UC have been considered. Both CyA and tacrolimus is potent drugs with rapid onset of action in patients with severe refractory UC before thioprines become effective. Fernandez-Banares et al.34 reported that the addition of thioprines improves long-term remission rates in patients with refractory UC even after discontinuation of CyA. Baumgart et al.23 reported the duration of response was from 9 months to more than 4 years after tacrolimus had been discontinued in patients who successfully responded to tacrolimus and could be maintained on long-term administration with AZA. Thus, as a bridge to thioprine therapy, CyA and tacrolimus can be efficient treatments for patients who are well-tolerated to thioprines.

The Active UC Trials 1 and 2 were large multicentre clinical studies that demonstrated the efficacy of infliximab therapy for inducing and maintaining clinical response, disease remission and mucosal healing in patients with moderate-to-severe active UC.35 These studies indicated that maintenance treatment with infliximab (by 54 weeks) could be a promising one for refractory UC. However, long-term efficacy and safety of infliximab treatment on patients with UC for 3 or 4 years have not yet been evaluated. Moreover, administration of infliximab is not the solution for all patients with UC. Thus, we should evaluate therapeutic efficacy and safety associated with maintenance therapy of infliximab for refractory UC in the future.

In addition, initially starting both infliximab and thioprines may be one of alternative options for patients with UC refractory to calcineurin inhibitors, although further investigation is necessary to determine whether or not this combination therapy exerts a more prolonged effect.

Similar to previous reports,23–25 severe adverse events rarely occurred in our patients. All the events except infectious disease were controlled by conservative therapies and all four patients who suffered from infectious disease during tacrolimus therapy were successfully treated with antibiotic therapy. Among them, one elderly patient who was administered high-dose corticosteroid with tacrolimus concomitantly suffered from P. carinii pneumonia (PCP). After our experience with this case, we prophylactically administered cotrimoxazole (960 mg b.d.; 160 mg trimethoprim and 800 mg sulfamethoxazole) twice a week when starting tacrolimus to elderly patients or patients treated with high dose of corticosteroids who might be at a relatively high risk of opportunistic infection. As a result, we have never experienced cases of UC patients with PCP since we started PCP prophylaxis.

Renal function is one of the critical side effects during long-term treatment with tacrolimus, because five of 27 patients enrolled in this series experienced elevations of serum creatinine levels. In general, calcineurin inhibitors are implicated in direct and indirect nephrotoxicity, and direct one can be divided into acute and chronic forms.36 Acute renal injury is primarily functional, while chronic kidney damage, resulting from vascular injury, is irreversible.36 In our case series, however, renal function was normalized by dose reduction or withdrawal of tacrolimus in all cases with renal dysfunction. Thus, careful monitoring of renal function should be required to avoid chronic tacrolimus-induced kidney damage during long-term administration of tacrolimus.

Both hyperglycaemia and hypertension are common adverse effects of tacrolimus therapy in liver18 and renal19 transplantations. However, these adverse effects were not observed in our series, as previous reports on long-term tacrolimus therapy for UC.23, 25

In conclusion, long-term administration of tacrolimus in patients with refractory UC is effective and well-tolerated, although we should keep in mind that severe adverse events including infectious disease may occur during tacrolimus therapy. Further studies are required to determine the optimal trough levels of tacrolimus during maintenance therapy for patients with refractory UC.


Declaration of personal and funding interests: None.