- Top of page
- Supporting Information
Crohn’s disease (CD) is a chronic inflammatory disorder of the gastrointestinal tract, of unknown aetiology. The condition is characterised by a relapsing and remitting course with patients experiencing flares of disease activity. These are often treated with glucocorticosteroids to bring about remission. Although these drugs may be effective in this situation,1 a number of patients become dependent on, or resistant to, their beneficial effects,2, 3 and their long-term use is associated with significant adverse effects, including osteoporosis, diabetes and weight gain. Immunomodulator therapies, such as the thiopurine analogues azathioprine and mercaptopurine, are useful for preventing relapse of disease activity once remission has been achieved in this subgroup of patients.4 There is also recent evidence to suggest that these drugs alter the natural history of CD, if used earlier in the disease course.5 However, up to one-fifth of patients do not respond to, or are intolerant of, therapy.6
Methotrexate is an anti-inflammatory cytotoxic drug that exerts its effects via dihydrofolate reductase inhibition. It has proven efficacy in patients with rheumatoid arthritis,7 and in the last two decades has emerged as an alternative to thiopurine analogues as first-line immunosuppressant therapy in the treatment of patients with CD. In two randomised controlled trials (RCTs) the drug, when administered intramuscularly, was more effective than placebo for both the induction of, and maintenance of, remission.8, 9 However, there have been few data from other RCTs since these studies were published,10, 11 and other investigators have reported variable efficacy of methotrexate in CD in nonrandomised studies.12–17
Part of the explanation for these disparate results may relate to the route of administration used. Although the parenteral route of administration of methotrexate was used in the aforementioned RCTs,8, 9 this carries logistical complexities in practice, as patients have to be trained to self administer the drug, and be provided with appropriate facilities for its disposal due to its cytotoxicity. However, when methotrexate is administered orally reports on its efficacy are conflicting,11, 15 and it has been shown that the oral bioavailability is reduced in those with severe intestinal disease.18
Another potential explanation may be that the efficacy is lower in patients who receive the drug as a second-line immunosuppressant, following failure with, or intolerance of, thiopurines. In the trials by Feagan et al. the majority of patients recruited were naive to immunosuppressants.8, 9 We have previously reported the efficacy of mycophenolate mofetil used as a second-line immunosuppressant in CD.19, 20 We now present our experience of the efficacy of methotrexate in terms of both inducing remission of active CD, and preventing relapse of quiescent CD, in a large cohort of patients treated in a tertiary referral centre, the majority of whom received parenteral methotrexate, and had previously failed, or were intolerant of, thiopurine therapy.
- Top of page
- Supporting Information
A total of 66 CD patients received methotrexate therapy between 2001 and 2010, and were therefore eligible for inclusion in the study. Mean age at diagnosis of CD was 29.4 years and 38 (57.6%) were female patients. There were 48 patients (72.7%) who commenced methotrexate from 2005 onwards, and who therefore received standardised induction and maintenance therapy initially. Other demographic data and characteristics of disease of included patients are detailed in Table 1. The mean duration of disease prior to initiation of methotrexate was 139.2 months. A total of 61 patients (92.4%) had been treated with another immunomodulator drug prior to using methotrexate, 60 of whom had been prescribed either azathioprine, at a dose of between 2.0 and 2.5 mg/kg, or mercaptopurine, at a dose of 1.0 mg/kg. Of these, 43 had failed therapy, and 18 were intolerant. Eighteen patients (27.3%) had previously received a biological therapy. There were a total of 31 (47.0%) patients receiving concomitant corticosteroids at the time methotrexate was commenced. Of the 66 patients, 61 (92.4%) were given methotrexate subcutaneously, and 28 (42.4%) were commenced on an initial dose of 25 mg weekly, with a further 30 (45.5%) patients started on 15 mg weekly, one patient receiving 12.5 mg weekly, and two patients 7.5 mg weekly. C-reactive protein levels prior to commencement of methotrexate were available in 61 of the 66 patients, and were 5 mg/L or less in 24 (36.4%) patients (mean 23.7 mg/L, range 5 mg/L to 218 mg/L).
Table 1. Baseline characteristics and demographics of 66 patients receiving methotrexate
|Characteristic||All patients (n = 66)|
|Mean age at diagnosis in years (s.d.)||29.4 (11.6)|
|Female gender (%)||38 (58)|
|Montreal age in years (%)|
| A1 (<16)||6 (9)|
| A2 (17–40)||50 (76)|
| A3 (>40)||10 (15)|
|Montreal location (%)|
| L1 (ileal)||4 (6)|
| L2 (colonic)||27 (41)|
| L3 (ileo-colonic)||35 (53)|
|Montreal behaviour (%)|
| B1 (nonstricturing, nonpenetrating)||41 (62)|
| B2 (stricturing)||12 (18)|
| B3 (penetrating)||13 (20)|
| Perianal disease modifier||19 (29)|
|Current or previous smoker (%)||27 (41)|
|Previous intestinal resection (%)||35 (53)|
|Extra-intestinal manifestations (%)|
| Any||25 (38)|
| Arthropathy (including sacroileitis)||21 (32)|
| Erythema nodosum||5 (8)|
|Mean disease duration in months prior to commencement of methotrexate (s.d.)||139.2 (116)|
|Previous immunomodulator used (%)||61 (92)|
|Previous biological therapy used (%)||18 (27)|
|Receiving corticosteroids at time of commencement of methotrexate (%)||31 (47)|
|C-reactive protein ≤5 mg/L at time of commencement of methotrexate (%)||24 (36)|
|Subcutaneous methotrexate (%)||61 (92)|
Response to therapy at 4 months
Response to therapy was achieved in 54 (81.8%) of the 66 patients at 4 months. Among the 19 patients with perianal disease, 17 (89.5%) responded to therapy. Of those who did not respond at this time point, reasons included failure of methotrexate therapy in seven patients, and adverse events in five. Mean age at diagnosis in responders was 30.1 years compared with 26.3 years in nonresponders (P = 0.16), and mean disease duration prior to commencement of methotrexate was 118.4 months in responders compared with 229.0 months in nonresponders (P = 0.002). Mean C-reactive protein levels prior to commencement of methotrexate were higher in responders (24.9 mg/L) than those in nonresponders (17.5 mg/L), but this difference was not statistically significant (P = 0.48). Of the 48 patients who commenced methotrexate from 2005 onwards, and who therefore received standardised induction and maintenance therapy initially, 40 (83.3%) responded to therapy at 4 months, compared with 14 (77.8%) of 18 who received nonstandardised induction and maintenance therapy (P = 0.72).
Response rates were generally higher in male patients, those with isolated colonic disease, previous or current smokers, those with no prior history of an intestinal resection, those with extra-intestinal manifestations of CD and those receiving methotrexate via the subcutaneous route (Table S1). There were no obvious predictors of response identified following univariate or multivariate analysis, although there were trends towards a higher rate of response in current or previous smokers (OR 5.28; 95% CI 0.85–32.98), and those with extra-intestinal manifestations of CD (OR 6.86; 95% CI 0.91–51.81) following multivariate analysis.
Sustained clinical benefit at last point of follow-up
At the last point of follow-up (March 2011), 22 (33.3%) of 66 CD patients were still receiving methotrexate, with a median duration of therapy in all individuals of 27.5 months (interquartile range 12–48 months). However, nine of these 22 patients were requiring concurrent biological therapy to maintain remission, with methotrexate continued to minimise the likelihood of antibody formation to biologicals. Of these, five were using infliximab and four were receiving adalimumab. A further six patients (9.1%) had discontinued methotrexate for reasons relating to a wish to start a family. Therefore, assuming these six individuals were well and in remission, of the original 66 CD patients only 19 (28.8%) experienced sustained clinical benefit with methotrexate as monotherapy, with a median duration of treatment of 37.5 months (interquartile range 24.8–60.8 months).
Among the 19 patients with perianal disease, six (31.6%) achieved sustained clinical benefit. The maintenance dose of methotrexate was 15 mg weekly in eight (61.5%) of the 13 who experienced sustained clinical benefit and continued the drug, 25 mg per week in two, 7.5 mg weekly in two and 12.5 mg per week in the remaining patient. All 19 who achieved sustained clinical benefit were receiving subcutaneous treatment. Therefore, of a total of 61 patients receiving parenteral methotrexate, 19 (31.1%) achieved sustained clinical benefit from the drug. A Kaplan-Meier analysis of the proportion of the 54 individuals deemed as having responded to methotrexate therapy at 4 months who were classified as having achieved sustained clinical benefit out to 36 months of therapy is provided in Figure 1, with a median time to treatment failure of 32 months.
Figure 1. Kaplan–Meier survival plot showing proportion of patients with sustained clinical benefit at 36 months post response to methotrexate therapy (assessed at 4 months).
Download figure to PowerPoint
Of the 47 patients who did not achieve sustained clinical benefit with methotrexate, failure of therapy occurred in 29 (61.7%), adverse events in 15 (31.9%), noncompliance in one patient (2.1%) and the drug was discontinued for reasons which were not clear in two (4.2%). Of the 29 treatment failures, outcomes were as follows: nine were receiving concomitant biological therapy as detailed above; 10 were receiving either infliximab or adalimumab monotherapy as a first-line biological; five had failed two subsequent biological therapies and had undergone surgery (subtotal colectomy with end ileostomy in all five, with or without completion proctectomy); four had failed a first-line biological and were receiving either infliximab or adalimumab as second-line therapy; and one had undergone surgery immediately after failure of methotrexate (small bowel resection).
Mean age at diagnosis in those achieving sustained clinical benefit was 31.2 years, compared with 28.7 years in those who did not (P = 0.50), and mean disease duration prior to commencement of methotrexate was 145.4 months in those achieving sustained clinical benefit compared with 136.7 months in those who did not (P = 0.81). Mean C-reactive protein levels prior to commencement of methotrexate were lower in those achieving sustained clinical benefit (15.4 mg/L) compared with those who did not (27.4 mg/L), but this difference was not statistically significant (P = 0.11). Of the 48 patients who commenced methotrexate from 2005 onwards, and who therefore received standardised induction and maintenance therapy initially, 17 (35.4%) achieved a sustained clinical benefit, compared with two (11.1%) of 18 who received nonstandardised induction and maintenance therapy (P = 0.07).
Rates of sustained clinical benefit were generally higher in female patients, those aged 40 years or more at the time of diagnosis of CD, those with colonic or ileo-colonic disease, those with stricturing disease, previous or current smokers, those with a prior history of an intestinal resection and in those without extra-intestinal manifestations (Table S2). Again, there were no obvious predictors of achievement of sustained clinical benefit identified following univariate or multivariate analysis, although no patients with isolated ileal disease experienced sustained clinical benefit, albeit among only four patients.
Tolerability of methotrexate
A total of 20 patients (30.3%) discontinued methotrexate due to adverse events. Five patients developed adverse events during initiation of therapy. Of these, two reported respiratory symptoms including exertional dyspnoea and lower respiratory tract infection, although a formal diagnosis of pneumonitis or pulmonary fibrosis was not made in either case, one developed hypersensitivity with angio-oedema on the first dose of methotrexate administered, one experienced nausea and in the remaining patient the reason for discontinuation was not clear. During maintenance therapy (median duration 12 months, interquartile range 4–20 months) a further 15 patients developed adverse events, which include development of deranged serum transaminases more than two times the upper limit of normal in six, although liver biopsy was not carried out routinely in these individuals to ascertain whether or not methotrexate was the cause, widespread lymphadenopathy in one patient, and a further eight patients withdrew treatment as a result of nonspecific side effects including lethargy, myalgia, fatigue and reduced appetite.
- Top of page
- Supporting Information
In this study, we have demonstrated that methotrexate administered parenterally or orally is efficacious in the initial treatment of patients with active CD, with a response rate after 4 months of treatment of over 80%. Methotrexate was used as a second-line immunomodulator in the majority of our patients, following failure with thiopurine analogues in more than 90%. However, its efficacy in terms of providing sustained clinical benefit in this setting was disappointing, with approximately 30% of patients achieving this according to the criteria we used, even when patients receiving oral methotrexate were excluded from the analysis. Among those patients treated with standardised induction and maintenance regimens after 2005, rates of sustained clinical benefit were slightly higher at 35%. Rates of sustained clinical benefit were generally higher in female patients, those aged 40 years or more at the time of diagnosis of CD, those with colonic or ileo-colonic disease, those with stricturing disease, previous or current smokers, those with a prior history of an intestinal resection and in those without extra-intestinal manifestations. However, none of these differences were statistically significant, so no definite conclusions can be drawn. In addition, 30% of patients had to discontinue therapy due to adverse events, with approximately 10% developing deranged serum transaminases whilst on the drug.
Strengths of this study include the fact that we included data from all patients commenced on methotrexate for CD from a single tertiary care referral centre, making this one of the largest nonrandomised studies to examine this issue to date. Complete data were collected from initiation of therapy to either cessation, or last point of follow-up, for all patients. Limitations of the study include the retrospective nature of data collection, meaning that the outcomes of response to therapy were defined using a physician’s assessment of improvement or remission, rather than a validated endpoint such as the Crohn’s disease activity index, which may have allowed a more precise measurement and stringent definition of response to therapy at 4 months, and remission at last point of follow-up. The accuracy of data collection was also dependent on the quality and comprehensiveness of medical records. Finally, the study was conducted entirely in a tertiary care population, meaning that the results may not be generalisable to patients receiving methotrexate therapy in the secondary care setting.
The response and remission rates observed in this cohort of patients are broadly corroborative of the results of previous studies. In a retrospective study conducted in 20 azathioprine-resistant or intolerant CD patients who all received parenteral methotrexate, 70% achieved a response at 3 months, but only 15% were in remission at 1 year.14 Two larger studies, conducted in France and Scotland,16, 17 both reported response rates to therapy in the order of 70%, but approximately 40% were in remission at 12 months in the French cohort,16 compared with only 20% in the Scottish study.17 The higher remission rates in the French study may relate to the fact that the patients selected for inclusion were only those who exhibited a good initial response to methotrexate, with those with early relapse or intolerance to therapy excluded.16 In another study, which administered 25 mg of intramuscular methotrexate to 14 patients with refractory CD for 12 weeks,13 36% and 29% of patients achieved endoscopic and histological remission, respectively. Although these response rates are substantially lower, it should be noted that these outcome measures were appreciably more stringent than the clinical assessments conducted in most of the other studies, and in fact clinical response was achieved in approximately 80% of these patients at 3 months.
Reasons for the low remission rates in this study, lower than those observed with azathioprine for instance,22, 23 presumably relate to the refractory population under study. The mean disease duration, prior to commencement of the drug, was in excess of 10 years. In addition, over 90% of individuals had already experienced either treatment failure with, or intolerance to, thiopurine therapy, almost 50% were receiving corticosteroids at the time methotrexate was commenced, and some had already failed biological therapies. The majority of patients received parenteral methotrexate, but none received the drug intramuscularly, which was the route of administration in the RCTs conducted by Feagan et al.,8, 9 although the bioavailability of subcutaneous and intramuscular methotrexate are thought to be equivalent.24, 25 Since patients were self administering the drug after they were judged competent to do so, there remains the possibility of noncompliance with therapy, which we were not able to assess. Nor did we use dose escalation to recapture response, or split dosing to minimise adverse events.
The side effect profile of methotrexate has been an issue of concern for some physicians. In this study, adverse events were experienced in 30% of patients. These included the development of deranged serum transaminases, nonspecific respiratory symptoms, fatigue and lethargy. This figure is much higher than that in the maintenance of remission RCT conducted by Feagan et al. over 40 weeks,9 where only one of 40 participants discontinued therapy due to adverse events, but is consistent with the Edinburgh data,17 which saw methotrexate discontinued in 29% of participants. Drug-induced hepatotoxicity is known to occur with long-term use of methotrexate, but it was not confirmed as the cause of serum transaminase derangement by liver biopsy in the six patients in our cohort, and discontinuation of therapy may not have been necessary. Similarly, the two patients who had suspected drug-induced pneumonitis discontinued treatment without high-resolution computed tomography of the thorax. In some of the patients, the adverse events experienced were somewhat nonspecific, and may have related to the disease itself. In addition, if the drug were working, one would expect the patient to report fewer symptoms that may relate to their CD, rather than more.
The use of methotrexate is often considered as a ‘sideways step’ for the maintenance of remission of CD following failure with, or intolerance of, first-line immunomodulators such as thiopurines. However, our data suggest that methotrexate has only limited efficacy in the long term in this group, even when given parenterally in the majority of patients. It may therefore be preferable to immediately ‘step-up’ to biological therapies in patients failing or intolerant of thiopurines, which have more consistent evidence for their efficacy in this situation.26–29 There may, however, still be a role for methotrexate in patients with CD and co-existent inflammatory arthropathies, as well as in reducing the formation of antibodies to infliximab in those receiving this drug, where it has demonstrated similar efficacy to azathioprine.30
In conclusion, we have demonstrated that methotrexate is effective for inducing an initial response in CD patients who have failed, or are intolerant of, thiopurines, but its effects may not be sustained in the long term. In this patient group, lasting benefit is more likely to be gained from commencement of a biological therapy. However, methotrexate still has an important role as a first-line immunomodulator in Crohn’s disease.