Objective To evaluate toxicity profiles in patients with rheumatoid arthritis (RA) treated either according to an intensive or a conventional treatment strategy approach with methotrexate (MTX) and to study factors associated with MTX-related toxicity.
Methods Data were used from the Computer-Assisted Management in Early Rheumatoid Arthritis (CAMERA) study, in which clinical efficacy of an intensive treatment strategy with MTX was more beneficial than a conventional treatment strategy approach. In this study, data on adverse events (AEs) were compared between the two strategy groups. Logistic regression analyses were used to identify possible associations between factors assessed at baseline and withdrawal due to MTX-related AEs or liver toxicity at follow-up.
Results Although significantly more patients in the intensive strategy group experienced MTX-related AEs than in the conventional strategy group, all recorded AEs were relatively mild. A higher body mass index (BMI) was significantly associated with withdrawal due to MTX-related AEs in the multiple regression analyses (odds ratio=1.207, 95% confidence interval 1.02 to 1.44, p=0.033). There was a trend towards an association between diminished creatinine clearance and MTX withdrawal. For liver toxicity, increased serum liver enzymes at baseline were associated with liver toxicity during follow-up.
Conclusion Although the occurrence of AEs in the intensive strategy group was higher than in the conventional strategy group, the previously observed clinical efficacy of an intensive treatment strategy seems to outweigh the observed toxicity profiles. When starting MTX, attention should be given to patients with a high BMI and those with increased levels of liver enzymes and decreased renal function.
Statistics from Altmetric.com
Methotrexate (MTX) is usually selected as initial treatment and as anchor drug in combination therapies for rheumatoid arthritis (RA).1,–,6 Discontinuation of MTX is more likely to be due to manifestations of adverse events (AEs) than because of inefficacy.7 8 Frequently observed AEs of MTX include hepatotoxicity and gastrointestinal (GI) problems; pulmonary toxicity and pancytopenia rarely occur.
In studies, older age,9 younger age,10 duration of MTX treatment,9 dose of MTX,11 female gender,10 11 high body mass index (BMI),10,–,12 renal impairment,13 increased liver enzyme levels,11 functional disability14 and no concomitant use of folic acid10 12 were identified as possible predictors of liver toxicity or other AEs. However, the results of these randomised clinical trials, often lacking specific data on AEs, and retrospective observational studies, were not always consistent. Retrospective observational studies have the advantage of being representative for routine practice, but have the drawback that desirable data such as possible predictors and description of AEs may be missing.
In the 2-year Computer Assisted Management in Early Rheumatoid Arthritis (CAMERA) study,15 more patients in the intensive MTX strategy group (I-group) than in the conventional MTX strategy group (C-group) achieved at least one period of remission during follow-up. However, to compare the value of the two strategies, it is important to weigh both beneficial effects and adverse effects.
The aim of this study was to compare toxicity profiles between both MTX treatment strategies and to study possible associations between baseline characteristics with MTX withdrawal and liver toxicity during follow-up.
Patients and methods
Patients in the CAMERA study were seen at one of the hospitals collaborating in the Utrecht Rheumatoid Arthritis Cohort study group. Exclusion criteria included a creatinine clearance of <75 ml/min, a serum aspartate aminotransferase (AST) or serum alanine aminotransferase (ALT) more than twice the upper limit of normal (ULN) and alcohol intake of >2 units a day. The study population in this study included all patients who used MTX for at least 1 week. Medical ethical committees of all participating hospitals approved this study, and all patients gave written informed consent before entering the study.
Patients in the C-group came to the outpatient clinic once every 3 months and patients in the I-group once every 4 weeks. In both groups, the dose of MTX could be increased until 30 mg/wk when the patient had not responded sufficiently, and after subcutaneous administration of MTX, ciclosporin was added to the MTX treatment.15 Folic acid (0.5 mg/day, except on the day MTX was taken) was prescribed to every patient. Criteria for dose adjustments, discontinuation of MTX and starting of ciclosporin as a consequence of AEs were dictated by the study protocol (see online supplementary file).
Assessments and definitions
In both strategy groups, AEs were recorded using a predefined form, categorised into GI, mucocutaneous, central nervous system, liver toxicity, renal, haematological, pulmonary disorders, infections, general AEs, post-dosing reactions and ‘other’ AEs. Creatinine increase was defined as an increase of >20 µmol/l compared with the previous protocol visit and haematological AEs included anaemia (haemoglobin <6.5 mmol/l), leucopenia (<3.5×109/l), thrombocytopenia (<150×109/l), or pancytopenia (two out of three of these criteria). Patients with normal blood pressure values at baseline were defined as hypertensive at follow-up using World Health Organization (WHO) thresholds.16 Based on normal values of the individual laboratories of participating centres, liver toxicity was defined as a value for AST or ALT greater than the ULN.
Logistic regression analysis was used to determine the association between strategy groups and withdrawal due to MTX- and/or ciclosporin-related AEs. In the total study population, possible baseline predictors of withdrawal due to AEs were identified applying simple and multiple logistic regression analyses. Both unadjusted and adjusted (for the strategy arm and baseline non-steroidal anti-inflammatory drug (NSAID) use) odds ratios (ORs) with 95% CIs are presented. In the multiple regression analyses, the variables which were significant in simple analyses and known possible predictors for MTX-related AEs (BMI, Health Assessment Questionnaire, creatinine clearance, ALT) were included in the model.
For the I-group and C-group separately, the number of patients with an AE and the total number of AEs are described. Comparison of percentages of patients with an AE was carried out using the Fisher exact test. We also expressed the number of events per number of evaluations because of differences in frequency of study visits between the two strategy groups.
Changes in creatinine clearance over 2 years between the two strategy groups, adjusted for NSAID and ciclosporin use during follow-up, were tested using multiple linear regression. For the seven main categories of AEs, the cumulative dose of MTX and dose of MTX in the month before the first recorded AE were compared between the two strategy groups using t tests or Mann–Whitney U tests, depending on the distribution of the data.
Within each treatment strategy group, the cumulative dose of MTX over 2 years was compared between completers with liver toxicity versus those without liver toxicity. For the I-group and C-group separately, multiple logistic regression analyses were used to assess the association between possible baseline predictors and liver toxicity (ie, AST and/or ALT >ULN) at follow-up. Since patients with liver enzyme levels up to twice the ULN were included in the study, we subsequently performed the same analyses including only those patients with liver enzyme levels <ULN at baseline. ORs were adjusted for NSAID use at baseline. All analyses were performed using SPSS version 15.0.
Table 1 shows the baseline patient characteristics, reasons for withdrawal and number of patients using ciclosporin during follow-up. In total, 16/149 (11%) patients in the I-group and 8/140 (6%) patients in the C-group withdrew because of MTX-related AEs including, respectively: liver toxicity (n=6 vs n=1), increased serum creatinine (n=1 vs n=0), pulmonary disorders (n=2 vs n=2), and other AEs such as nausea or oral ulcers (n=7 vs n=5) (table 1). After the start of ciclosporin, 11 patients withdrew because of different AEs, including hypertension (n=2). There was no association between the strategy group and withdrawal due to MTX-related AEs (OR=1.985, 95% CI 0.82 to 4.80, p=0.128), but patients in the I-group were more likely to withdraw due to MTX and/or ciclosporin-related AEs (OR=3.080, 95% CI 1.39 to 6.83, p=0.006).
None of the baseline variables was associated with withdrawal due to MTX-related AEs in the simple regression analyses (table 2). Data of 167 patients were available for the multiple regression analyses. In the final model, increased BMI (OR=1.207, 95% CI 1.02 to 1.44, p=0.033) and a trend towards decreased creatinine clearance (OR=0.974, 95% CI 0.94 to 1.00, p=0.093) were associated with withdrawal due to MTX-related AEs. More pain at baseline (unadjusted OR=1.016, 95% CI 1.00 to 1.03, p=0.042 and adjusted OR=1.011 95% CI 1.00 to 1.03, p=0.149) and increased ALT levels (unadjusted OR=1.027, 95% CI 1.00 to 1.05, p=0.033 and adjusted OR=1.020 95% CI 0.99 to 1.05, p=0.134) were associated with withdrawal of patients experiencing any treatment (MTX and/or ciclosporin)-related AEs in the simple regression analyses, but not in the multiple regression analyses (data not shown).
AEs were reported for 268 patients; 142/149 (95%) patients in the I-group and 126/140 (90%) patients in the C-group. For the seven main categories of AEs, the percentages of patients with an event were respectively: GI symptoms (66% vs 54%, p=0.030); mucocutaneous (54% vs 40%, p=0.025); central nervous system (59% vs 39%, p=0.001); hepatic (55% vs 35%, p=0.001); renal (39% vs 44%, p=0.403); haematological (26% vs 11%, p=0.001) and general AE (27% vs 15%, p=0.015). The reported AEs were mostly mild; frequently reported symptoms were nausea, oral ulcers, dizziness, headache, fatigue and stomach ache (online supplementary table S1). Fifteen per cent of the patients with headaches and 19% of the patients in whom creatinine levels increased during follow-up had received ciclosporin before the AE. Of patients with a normal blood pressure at baseline, 27 patients were hypertensive at follow-up—that is, WHO hypertension grade 2 and/or grade 3. Five of these patients had received ciclosporin.
After 2 years, creatinine clearance had significantly decreased from baseline; mean decrease in the I-group (n=70) was 7.7 ml/min (95% CI 4.7 to 10.7, p<0.001) and 7.8 ml/min (95% CI 4.9 to 10.7, p<0.001) in the C-group (n=95). These changes were not statistically different between the two strategy groups (estimated difference –0.46, 95% CI –4.8 to 3.9). One patient in the I-group and three patients in the C-group had reversible pancytopenia. Seven patients, four in the I-group and three in the C-group, experienced relatively severe pulmonary disorders (eg, pleuritis, restrictive lung disease and exacerbation of chronic obstructive pulmonary disease), but none of the patients developed pneumonia or pneumonitis. Three (43%) of these patients already had lung disorders before entry into the study.
Except for haematological AEs, the dose of MTX before the first recorded event was significantly higher for the I-group than for the C-group (table 3). However, the median cumulative dose of MTX until the first recorded event was, except for gastrointestinal AEs, similar between the two strategy groups.
Seventy patients in the I-group and 40 patients in the C-group had ALT abnormalities after starting treatment with MTX, of whom 14 and three patients, respectively, had at least one recorded value >3×ULN. Of the patients with AST abnormalities (63 in the I-group and 30 in the C-group), one patient in the C-group had at least one recorded value >3×ULN. In the patients with liver enzyme elevations >3×ULN, the dose of MTX was reduced or stopped and started again after resolution for 13 patients, MTX was discontinued and not restarted for one patient and for three patients it led to withdrawal from the study.
For completers in the I-group, the average±SD cumulative dose of MTX over 2 years was 2007±691 mg for patients with liver toxicity and 1878±783 mg for patients without liver toxicity (p=0.396). Similar observations were found in the C-group, (respectively, 1662±418 mg vs 1602±599 mg, p=0.556).
In both strategy groups, no clinical or demographic factors were found to be associated either with an abnormal value of AST or with ALT (online supplementary table S2). An abnormal value of ALT was solely predicted by a high ALT level before the start of MTX in the I-group and an increased ALT and creatinine levels in the C-group. No associations between baseline values and an abnormal value of AST were found in the I-group, but in the C-group an increase in both AST and ALT at study entry was associated with an abnormal value of AST during follow-up. Also, in the group of patients with liver enzyme levels <ULN at baseline, higher baseline enzymes were associated with abnormalities during the study.
Although more patients in the I-group experienced possible MTX-related AEs than patients in the C-group, the severity of AEs was relatively mild and often reversible in both groups. For 5% of the study population ALT levels >3×ULN were recorded, but they did not result in serious clinical signs or symptoms of hepatitis or cirrhosis. No patients experienced pneumonia, in agreement with results of a large population study.17 Four patients (1.4%) had reversible pancytopenia in our study, similar to the percentage found in a review.18 Since patients were allowed to take NSAIDs, and ciclosporin could be used in the strategy regimen, the reported AEs might not exclusively have been related to MTX treatment.
The average dose of MTX in the month before the first recorded event was higher in the I-group than in the C-group, but the median cumulative dose MTX until the first event was not. An explanation might be the faster incremental rise of the MTX dose in the I-group. However, the clinical efficacy of the intensive strategy was far better than that of the conventional strategy,15 which in our opinion outweighs the observed AEs.
The cumulative dose of MTX did not differ between completers with and without liver toxicity, but for 48% of the patients in the I-group and 33% of the patients in the C-group the dose of MTX was at least on one occasion not increased during one of the protocol visits because of abnormal laboratory values, although an increase was indicated according to protocol. We did not take oral or subcutaneous administration into account in calculating the cumulative dose of MTX until each event. Since more patients in the I-group received MTX subcutaneously, the absorbed dose of MTX in the I-group may have been higher, although a discrepancy in bioavailability is mainly seen at higher doses of MTX.19 20
In this study, only a higher BMI was significantly associated with withdrawal due to MTX-related AEs in the multiple regression analyses. However, only 58% of patients could be included in these analyses owing to missing baseline data. Interestingly, increased self-reported pain, but none of the variables of the response criteria for dose adjustments, was independently associated with withdrawal due to MTX- and/or ciclosporin-related AEs. These results are partly in agreement with those of other studies,9,–,14 but we did not find any association with gender, age, and Health Assessment Questionnaire and MTX-related withdrawal.
In some previous studies no association between creatinine levels and MTX toxicity was found,21 22 whereas in one study, pooling data from 11 individual studies, patients with diminished creatinine clearance were more likely to have a higher overall toxicity score.13 In our study, there was a trend towards an association between impaired creatinine clearance and withdrawal due to MTX-related AEs. Diminished creatinine clearance was not associated with liver toxicity, but we found that increased ALT, AST and creatinine levels measured at the start of the study were significantly associated with liver toxicity at follow-up in both strategy groups.
The frequency of AEs in the I-group was higher than that in the C-group, but the severity of AEs was relatively mild. Since the clinical benefit in the I-group evidently was more pronounced, the efficacy/toxicity ratio supports intensive treatment with MTX as first or anchor disease-modifying antirheumatic drug in early RA in daily practice. When starting MTX in patients with early RA, attention should be given to a high BMI, increased levels of serum liver enzymes and decreased renal function.
We thank all the rheumatologists and research nurses of the Utrecht Rheumatoid Arthritis Cohorts study group for data collection and A Jacobs-van Bree for data entry.
Patient consent Obtained.
Ethics approval This study was conducted with the approval of the medical ethical committees of all participating hospitals.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.