Efficacy of methotrexate in Crohn’s disease and ulcerative colitis patients unresponsive or intolerant to azathioprine /mercaptopurine


Dr S. L. Bloom, Department of Gastroenterology, University College London Hospitals, 2nd Floor, Maple House, Rosenheim Wing, 25 Grafton Way, London WC1E 6DB, UK.
E-mail: stuart.bloom@uclh.nhs.uk


Background  Despite the wide use of azathioprine/mercaptopurine (AZA/MP) therapy in the management of both Crohn’s disease (CD) and ulcerative colitis (UC), approximately 20% of patients cannot tolerate the drugs and 30% do not respond.

Aim  To examine the efficacy and safety profile of methotrexate (MTX) in patients with CD or UC who are either intolerant or non-responsive to AZA/MP.

Methods  A total of 131 patients with IBD treated with MTX were identified. Retrospective data were obtained by case note review. Clinical response (defined as steroid withdrawal, normalization of previously raised CRP or physician’s clinical assessment of improvement) was assessed at 6 months.

Results  Clinical response in Crohn’s disease occurred in 18 of 29 patients (62%) refractory to AZA/MP and 42 of 70 patients (60%) intolerant to AZA/MP, with no difference between the groups (P = 1.0). In UC, clinical response was seen in 7 of 9 (78%) patients refractory to AZA/MP and 15 of 23 (65%) intolerant to thiopurines. MTX was well tolerated in a majority of individuals.

Conclusions  Methotrexate appears effective in both CD and UC patients who fail to respond to or are intolerant to AZA/MP therapy.


Azathioprine therapy is widely used in the management of Crohn’s disease (CD), with remission rates of 42% seen at 15 months.1 Use of thiopurine therapy in the management of ulcerative colitis (UC) is also widespread. In an azathioprine withdrawal study of 67 UC patients in remission, at 1 year, the rate of relapse was 36% for patients continuing azathioprine and 59% for those taking placebo.2 However, despite the wide use of thiopurines in the management of steroid refractory inflammatory bowel disease (IBD), approximately 20% of patients cannot tolerate the drugs and 30% do not respond.3 Early azathioprine (AZA) intolerance appears to be dependent on thiopurine methyltransferase activity (TPMT) with the heterozygous TPMT genotype predicting adverse effects.4 This intolerance maybe also in part because of the imidazole derivative cleaved when AZA is converted to mercaptopurine (MP).5 Studies on azathioprine metabolite concentrations in association with remission in IBD patients have reported conflicting findings.6–8 In both conditions, patients who fail to respond or are intolerant of thiopurines, pose a management challenge.

In Crohn’s disease, methotrexate (MTX) in a dose of 25 mg once weekly is effective at inducing and a dose of 15 mg once weekly maintaining remission in thiopurine-naive patients.9, 10 However, its benefit in patients who have failed thiopurine therapy is not clear. Data on the use of MTX in UC are limited to a single placebo controlled trial in which 67 patients were randomized to receive 12.5 mg oral methotrexate or placebo. There was no significant difference in induction or maintenance of remission between the groups, although this may reflect the low dose of MTX used.11 A number of small open labelled studies have reported clinical response in up to 60–70% of patients.11–13

There are some data to support the use of MTX in patients with IBD who have failed thiopurine therapy,14–16 although differences in efficacy have been reported between those who were refractory to AZA/MP and those who were intolerant: a study of 42 patients with ulcerative colitis treated with oral methotrexate14 reported a significantly higher response in the 31 patients who had been intolerant to prior thiopurine therapy than in the 11 patients who were thiopurine refractory (58% vs. 27% respectively). Conversely, in a paediatric Crohn’s disease study, subcutaneous MTX was effective in 9 (65%) of the 14 patients, which included 6 of 11 patients who were refractory to previous MP and all three patients intolerant to mercaptopurine.17

The aim of this study was to examine the efficacy of MTX in a large cohort of patients with CD and UC who were intolerant or nonresponsive to AZA/MP. In addition, we wished to examine the safety profile of MTX in this patient group.

Materials and methods


A total of 131 patients with a diagnosis of IBD (99 CD, 32 UC) treated with MTX between 2001 and 2007 were identified using the IBD databases at University College London NHS Trust and Barts and the London NHS Trust. The diagnosis of IBD had been confirmed by endoscopic, histological and radiological features; baseline demographic data are shown in Table 1. All patients with IBD who were treated with MTX in the two trusts involved were included in the study and all had previously received thiopurine therapy.

Table 1.   Baseline demographics of patients
CharacteristicCrohn’s disease (n = 99)Ulcerative colitis (n = 32)
Age [median (range)]40 years (16–86)43 (22–88)
Duration of disease [median (range)]72 months (4–564)72 months (12–290)
Crohn’s disease
 Upper gastrointestinal4
Ulcerative colitis

The primary outcome measures assessed were clinical response at 6 months and side effects. Clinical response was defined as any one of the following: steroid withdrawal by 6 months (off steroids for a least 3 months) in patients receiving steroids at baseline, normalization of a raised C-reactive protein (CRP) or physician’s clinical assessment of improvement as recorded in the notes. Thiopurine intolerance was defined as a requirement to stop thiopurine therapy because of side effects. Thiopurine refractory patients were defined as having shown no clinical improvement to a trial of at least 12 weeks therapy. Side effects were defined by symptoms reported to the physician as documented in the patient notes and review of haematological and biochemical investigations.

Statistical analysis

Analysis was performed using Prism statistical software. Statistical differences were sought using the 2 tailed Fisher exact test for categorical data and for continuous nonparametric data the Mann–Whitney U-test.

Ethical considerations

This study was designed to assess clinical service and did not require formal ethical approval, according to UK National Research Ethics Service (NRES) guidelines.18


Crohn’s disease

A total of 99 patients with CD were treated with methotrexate. The duration of disease [median (range)] prior to MTX was 72 (4–564) months. All patients had previously been on a thiopurine (azathioprine or mercaptopurine) [median (range)] for 48 (1–864) weeks, on a dose deemed adequate by their physicians depending on weight and TPMT levels (if available), discontinued because of lack of response in 29 (29%) and intolerance in 70 (71%). Patients were intolerant to thiopurine therapy due to bone marrow suppression in 16 (22.9%), upper gastrointestinal symptoms in 13 (18.5%), pancreatitis in 5 (7.1%), abnormal liver function tests in 2 (2.9%) and nonspecific symptoms in 34 (48.6%) (including joint aches, hair loss, skin rash and flu like illness). Baseline concomitant medication is shown in Table 2.

Table 2.   Baseline medication in CD and UC patients who were nonresponsive or intolerant to azathioprine/mercaptopurine (AZA/MP)
Concomitant medicationCrohn’s diseaseUlcerative colitis
Nonresponsive to thiopurine (n = 29) (%)Intolerant to thiopurine (n = 70)* (%)Nonresponsive to thiopurine (n = 9) (%)Intolerant to thiopurine† ( n = 23) (%)
  1. * No significant difference in baseline medication between groups in patients with CD.

  2. † No significant difference in baseline medication between groups in patients with UC.

5-Aminosalicylate15 (51.7)39 (55.7)8 (88.9)20 (86.9)
Oral steroids6 (20.6)19 (27.1)5 (55.65)14 (60.8)
Anti-TNFα6 (20.7)11 (15.7)1 (11.1)2 (22.2)
Antibiotics4 (13.8)11 (15.7)1 (11.1)0 (0)

The induction dose of MTX [median (range)] used was 25 mg/week (7.5–25) followed by a maintenance dose of 15 mg/week (15–25) for a duration of [median (range)] 72 weeks (7–208) in thiopurine nonresponders and 15 mg/week (5–25) for 96 weeks (12–378) in the thiopurine intolerant group (= N.S.). During induction, the route of administration was parenteral (intramusculary) in 22 and 39 patients, in the thiopurine nonresponders and intolerant groups respectively. All patients received concomitant folic acid supplementation, with either a dose of 5 mg daily for 5 days or 5 mg once a week, 48 h after MTX.

Clinical response at 6 months occurred in 18 out of 29 patients (62%) in the thiopurine refractory group and in 42 of 70 patients (60%) of the intolerant group, with no significant difference between the groups (P = 1.0; 2 tailed Fisher exact). There was no difference in clinical response when stratified by mode of administration (oral vs. intramuscularly) (data not shown) in either group. Complete steroid withdrawal was possible in 6 of 6 in the refractory group and 8 of 19 patients in the intolerant group (P = 0.15), although 16 managed a mean dose reduction of 18 mg. A normalization of previously raised C-reactive protein was seen in 4 of 15 patients in the refractory patients and 5 of 24 in the intolerant group (P = 1.0). Overall, for the whole group, the mean drop in CRP was 6.2 mg/L in those refractory and 17.6 mg/L in those intolerant. Four (13.7%) thiopurine nonresponders went onto receive infliximab, as did 11 (15.7%) thiopurine-intolerant patients.

In the 18 of 29 thiopurine nonresponders who improved with MTX, subscores showed improvement in the physician’s global score in 13 of 18, steroid withdrawal in 6 (of 6), and normalization of raised CRP in 5 of 15. The number of patients achieving one end-point was 18 of 29 (62%), two end-points 5 (17.2%) and three end = points 0 (0%).

Of the 42 overall responders in the thiopurine-intolerant group, 39 showed improvement in the physicians global score, 8 (of 19) steroid withdrawal and 5 (of 24) normalization of previously raised CRP. The number of patients achieving one end-point was 42 of 70 (60%), two end-points 13 (18.6%) and three end = points 0 (0%).

Ulcerative colitis

Methotrexate was used to treat 32 patients with ulcerative colitis, who had all previously been managed with thiopurine therapy for [median (range)] 28 weeks (2–468). Thiopurine therapy was discontinued because of lack of response in 9 (28%) patients and intolerance in 23 (72%). Similar reasons for thiopurine intolerance were seen in the ulcerative colitis patients compared with the CD population: upper gastrointestinal symptoms in 7 (30.4%), abnormal liver function tests in 4 (17.4%), bone marrow suppression in 3 (13.0%), pancreatitis in 2 (8.7%) and various nonspecific symptoms in 7 (30.4%).

The induction dose of MTX, with a combination of oral (20 patients) and parenteral (intramuscularly) (12 patients) administration, was 25 (10–25) mg/week, with a maintenance dose of 20 (10–25) mg/week. MTX was continued for 55 (2–290) weeks.

Overall, clinical response was achieved in 22 out 32 patients (68%) with ulcerative colitis; this included 7 of 9 (78%) in the thiopurine refractory group and 15 of 23 (65%) in the intolerant group. The number of patients achieving one end-point was 7 (78%), two end-points 3 (33.3%) and three end = points 0 (0%) in those thiopurine refractory, and 15 (65%), 7 (30.4%) and 0% in those intolerant respectively. As seen in CD, there was no difference in clinical response in those patients receiving oral compared with parenteral administration of MTX (data not shown).Overall, in the patients with UC, steroid withdrawal was possible in 7 of 19 patients who were on steroids at baseline, and a normalization of previously raised CRP in 4 of 15 patients (mean drop for the whole group in CRP was 2.7 mg/L).

Side effects

In the entire study population of both patients with Crohn’s disease and ulcerative colitis, side effects to MTX therapy were seen in 23 (17.4%) requiring discontinuation in 11 (8.3%) patients. These findings are summarized in Table 3.

Table 3.   Summary of side effects of the total 132 patients (CD and UC) receiving methotrexate. The number of patients who developed each adverse effect is shown below and the number of patients in whom it was necessary to discontinue MTX
Adverse effectNo of patients N = 132 (%)MTX continued/ discontinued (no. of patients)
  1. * In four of those who developed abnormal LFTS, MTX was continued: one was diagnosed with nonalcoholic fatty liver disease, with increased monitoring in the others after excluding other liver pathology.

Abnormal liver function tests8 (6.1)Continued4*
Respiratory5 (3.8)Continued4
Bone marrow suppression4 (3.0)Continued3
Nausea and vomiting3 (2.3)Continued1
Hair loss2 (1.5)Continued1
Nonspecific (arthralgia/malaise)3 (2.3)Continued0

Four patients developed neutropenia, which, in one case, was associated with osteomyelitis; one required dose reduction and the others increased monitoring. The patient with osteomyelitis recovered fully following a complete course of intravenous antibiotics and discontinuation of MTX. Eight patients developed abnormal liver function tests; cumulative mean (range) dose of MTX in this group was 278.75 mg (175–520 mg).

Five patients developed shortness of breath. In one case, MTX was stopped by patient despite normal chest radiograph, spirometry and echocardiogram, prior to respiratory review. The others were reviewed/monitored by the respiratory physicians; in those patients MTX was continued. Two of the three patients experiencing nausea and vomiting received parenteral administration of MTX; in one of them it was necessary to stop MTX.

Twenty-seven patients had pre-existing conditions prior to commencement of MTX that were considered risk factors for MTX toxicity. These included diabetes mellitus (five patients); respiratory disease; chronic obstructive airways disease,6 interstitial lung disease;2 increased body mass index (BMI);5 NSAID use;9 renal impairment (serum creatinine >150 μmol per litre);6 and pre-existing liver disease2 (some patients may have had more than 1 risk factor). However, the side effects observed only related to the risk factors in two patients with increased BMI who developed abnormal LFTs.


Efficacy in Crohn’s disease

Our results show that 60% patients with Crohn’s disease who previously failed thiopurine therapy show a clinical response to MTX. This is the first study which shows that benefit is similar in thiopurine intolerant and refractory patients, in contrast to a previous report.17 It confirms that methotrexate is an effective and well tolerated treatment for the management of inflammatory bowel disease. In addition, MTX appears to provide some steroid sparing effect.

The current report was intended to evaluate the efficacy of MTX in inducing clinical response rather than as a maintenance strategy. In our series, 15 of the 99 patients with CD were receiving concurrent Infliximab. A possible synergistic effect must be acknowledged in light of the recent COMIT data.19 The induction dose of MTX 25 mg/week (combination of parenteral and oral administration) followed by a maintenance dose of MTX 15 mg/week may, in part, explain the response rates achieved in our study. Our results are consistent with findings in a recent retrospective study of 39 CD patients in whom remission was achieved in 71% at 16 weeks.20 Also, consistent with our findings, a paediatric study of 61 CD patients previously on thiopurine therapy reported a clinical response rates of 68.9% at 6 months, although in contrast to our findings, the reason for stopping thiopurine therapy was predominantly because of nonresponse (68.9%).21 Furthermore, response rates of at 6 months were seen in 10 of 20 patients either azathioprine resistant or intolerant treated with parenteral MTX.22

Efficacy in ulcerative colitis

In patients with ulcerative colitis, we observed an overall clinical response in 22 of 32 patients (68%). In addition, MTX also appeared to provide a steroid sparing effect. Similar high response rates have been reported in published studies; a retrospective study of 50 patients either intolerant or nonresponsive to azathioprine using oral MTX (20 mg/week) reported a response rate of 72% and a remission rate of 42%.14 Thus, our series suggests that MTX is effective in patients who have failed thiopurine therapy for lack of response as well as intolerance. This contrasts with a recent series from Oxford;14 but, in both studies, the numbers of patients with UC is relatively small (nine thiopurine intolerant patients in the current study and 11 patients respectively in the Oxford cohort) and hence the data must be interpreted with caution.

The mechanism by which MTX improves inflammatory bowel disease remains uncertain, despite an understanding of its cellular effects. Putative mechanisms include enhanced concentrations of adenosine,23 the inhibition of cellular proliferation, a decrease in the production of inflammatory mediators such as interleukins and eicosanoids and the induction of apoptosis.24, 25 At high doses, MTX antagonizes dihydrofolate reductase leading to inhibition of DNA, RNA and protein synthesis. At lower doses, inhibition of other folate dependent enzymes may be responsible for the action of MTX.26 In addition, the onset of peak activity of MTX is achieved after 12 weeks of treatment.9

Side effects

Data from our study suggest that MTX therapy is well tolerated and safe with appropriate monitoring. The incidence of side effects was 17.4%; withdrawal of MTX was required in 8.3%, slightly less than reported in previous studies, in which withdrawal of treatment rates range between 10 and 18% of patients.27, 28 The most common minor side effects of methotrexate are gastrointestinal, including nausea, vomiting and diarrhoea. This was seen only in three patients in our study which may relate to the use of concomitant folic acid supplementation29 and parenteral administration of MTX during induction in a large proportion of patients. Interestingly, the incidence of side effects seen with MTX was considerably lower than in a recent study of 217 patients treated with azathioprine, in which 39% withdrew because of adverse effects with 15.7% suffering gastric intolerance.4

In a controlled study of MTX for the induction of remission in CD,9 patients were excluded for a number of pre-existing conditions including hepatic disease (biopsy proven cirrhosis, chronic active hepatitis, serum aspartate aminotransferase, bilirubin, or alkaline phosphatise at least twice upper limit of normal), renal dysfunction (serum creatinine concentration greater than 1.7 mg per decilitre (150 μmol per liter), clinical important lung disease, systemic infection, pregnancy, previous cancer, high alcohol consumption, hypersensitivity to MTX, erythrocyte macrocytosis, body weight 40 percent higher than normal, diabetes mellitus or NSAID use. Data relating to alcohol consumption were not obtained. In our series, 27 patients (20.4%) had co-morbidities that would have excluded them from the published controlled trial.9 Although, we did not find that patient’s pre-existing conditions influenced side effects, these factors should be considered before starting MTX.

With the doses of MTX used in inflammatory bowel disease, many serious side effects such as bone marrow suppression are seldom seen. We found only one case of osteomyelitis as consequence of bone marrow suppression.

In our study, 4 of 8 patients who developed abnormal liver function tests had to stop MTX. Hepatotoxicity is an important consideration for patients on long-term MTX because of the risk of liver fibrosis and, in the past, guidelines suggested routine liver biopsy after a cumulative dose of Methotrexate. Kozarek et al. performed liver biopsies in six patients who had been treated with MTX for a mean of 34 months with a mean cumulative dose of 1733 mg. Normal histology was seen in all specimens except in one patient who had received a cumulative dose of methotrexate of 2500 mg and had minimal periportal fibrosis.30 In a case series of 20 patients with IBD treated with MTX after receiving a cumulative dose of 1.5–5.4 g over 1–5 years, liver histology showed hepatic fibrosis in one patient in the absence of abnormal liver function tests.31 Hepatoxicity of MTX may vary according to the underlying condition. In psoriasis, the risk of cirrhosis is estimated at 3%, with recommendations for liver biopsy at 1–1.5 g.32 At present, it would appear reasonable to monitor liver function tests in patients on Methotrexate every 2–3 months.33

There are a number of limitations to our study being retrospective, most importantly in the definition of clinical response and the heterogeneity of the population.34 Clinical response included not only the physician’s assessment of improvement at 6 months but also steroid withdrawal and normalization of CRP. However, given that not all patients were on corticosteroids at baseline and similarly not all patients had elevated CRP at baseline, a combination of the end-points was used. Normalization of CRP was seen in a small proportion of subjects in comparison with physician’s assessment of improvement, which, in part, may be due to the study design and in part reflect that CRP does not always correlate with disease activity scores. The response for each end-point has been shown and in addition, the numbers of patients achieving one, two and three end-points. We acknowledge the need for caution in assessing results, although retrospective assessment of clinical response has been well established in other studies of immunosuppressive therapy in inflammatory bowel disease.14, 16, 27

In conclusion, MTX appears effective in both Crohn’s disease and ulcerative colitis in patients who fail to respond to or are intolerant to thiopurine therapy with no difference between the two groups. MTX is well tolerated in the majority of individuals.


Declaration of personal and funding interest: None.