Juvenile idiopathic arthritis (JIA) has a prevalence of 40–150 cases per 100 000 children and is the most frequent chronic rheumatic disease in the paediatric population. It is a heterogeneous disorder and comprehends 7 clinical forms1 based on the classification of the International League Against Rheumatism (ILAR). This classification is based on the clinical and laboratory characteristics of the disease. When the predominant symptom is peripheral arthritis, it is classified based on the number of involved joints as either oligoarticular (up to 4) or polyarticular (5 or more); the latter is further classified as rheumatoid factor (RF) positive or negative. The fourth category, the systemic form, is determined by the presence of systemic manifestations (fever, erythematous rash, serositis, generalised lymph node enlargement or organ enlargement), independently of the number of affected joints. Enthesitis-related arthritis is characterised by axial skeleton involvement and/or inflammation of the entheses. The sixth form is psoriatic arthritis, and the seventh, termed undifferentiated arthritis, includes all presentations that do not fulfil the criteria for any of the previous categories. Juvenile idiopathic arthritis can cause long-term disability.2

Methotrexate (MTX) is the disease-modifying antirheumatic drug (DMARD) of choice for most clinically significant forms of JIA3,4 and has a good safety profile.5 As monotherapy it achieves remission of the disease in 60%–70% of patients, while the combination of MTX with biologic DMARDs decreases the risk of antidrug antibody production.3,6

The most frequent adverse reactions (ARs) to MTX are not severe, although they may have an impact on adherence to treatment7 and its discontinuation.

The aim of our study was to determine the frequency and type of ARs to MTX in patients with JIA, the risk factors associated with their development and how the ARs were managed.

We conducted a retrospective, observational and descriptive cohort study that included all patients aged 0–16 years with a diagnosis of JIA managed at the Paediatric Rheumatology clinic of a tertiary care hospital and treated with MTX between 2008 and 2016.

We collected clinical and laboratory data from the electronic health record database of the hospital, including sex, age at diagnosis, clinical form of JIA based on the ILAR classification1 and presence or absence of ARs. All patients received a 5 mg dose of folic acid a week.

We classified ARs as gastrointestinal (abdominal pain, nausea/vomiting, anorexia), hepatic (hypertransaminasaemia), cutaneous/mucosal (aphthous mouth ulcers, alopecia), haematologic (cytopenia), behavioural (anticipatory behaviour with development of abdominal pain, gastrointestinal symptoms or significant anxiety before administration of the drug), infectious and other.

Given the frequency and nonspecificity of gastrointestinal symptoms, we only included gastrointestinal events associated with the administration of MTX that were not transient or whose frequency and/or intensity interfered with treatment.

We defined hypertransaminasaemia as elevation of alanine aminotransferase (ALT) and of aspartate aminotransferase (AST) greater than twice superior the upper range of normal in the reference range applied in our hospital (45 IU/mL). A complete blood count (CBC) and liver function panel was performed in all patients 1 month after initiation of treatment and every 3–4 months thereafter. If the tests detected hypertransaminasaemia, we discontinued treatment and 1 month later repeated the liver function tests, reintroducing MTX if the transaminase levels had normalised or withdrawing treatment definitively if they remained elevated.

We collected the date, route of administration and dose of MTX at treatment initiation and at the time of detection of the AR, whether it was necessary to change the dose or route of administration of MTX due to the AR, and whether or not the AR resolved as a result. We also documented whether it was necessary to discontinue MTX, the reason for discontinuation and the drug chosen to replace it (biologic versus not biologic). Last of all, in patients in who MTX was reintroduced after resolution of the AR, we recorded the date of reintroduction, dose and route of administration and whether the AR recurred. For all patients, we collected the presence or absence of active arthritis or uveitis at the end of the study and the most recent C-reactive protein (CRP) level and erythrocyte sedimentation rate (ESR) values in record. We defined remission based on the criteria for clinically inactive disease published by Wallace et al.8

We have summarised quantitative data using the mean or the median as measures of central tendency and the standard deviation (SD) or interquartile range (IQR) as measures of dispersion based on whether the distribution of the data was parametric or nonparametric, respectively. We summarised qualitative data as frequency distributions.

We assessed the association between categorical variables using the χ2 test and compared means or medians by means of the Mann–Whitney U test or the Kruskal–Wallis test as appropriate based on the shape of the distribution. We used logistic regression in the multivariate analysis. We defined statistical significance as a p-value of less than 0.05.

We investigated the temporal association between the use of MTX and the development of ARs by means of survival analysis with the Kaplan–Meier method, in which the discontinuation of MTX was the event of interest and the time-to-event was expressed in months.

The data were collected in a Microsoft Access 2003 (Redmond, Washington, USA) database and the statistical analysis performed with the software Stata version 4 (College Station, Texas, USA)

Our study was the first analysis of the frequency and type of ARs associated with treatment with MTX in Spanish patients with a diagnosis of JIA. Since the efficacy of MTX for treatment of JIA was demonstrated in 1992,9 this drug has become the systemic drug of choice for the disease.10 Methotrexate has not only proven effective in decreasing the number of inflamed joints,11 but also in significantly improving health-related quality of life in these patients.12 In spite of this, few studies have addressed the ARs to MTX in patients with JIA5,13–21 (Table 4).

The overall incidence of ARs observed in the Spanish patients in our sample was similar to that reported in other case series. Between one third and half of patients with JIA treated with MTX experience ARs.5,13–21 The types of ARs found in our study were also consistent with the previous literature, the most frequent being behavioural changes, followed by gastrointestinal manifestations and third of all hypertransaminasaemia. In our sample, gastrointestinal intolerance (abdominal pain, nausea and vomiting), developed by 20% of patients, was less frequent compared to other studies (21%–73%). These differences could be due to the methodology used, as we only considered ARs that were persistently associated with the administration of MTX or with a substantial frequency and/or intensity.

Other ARs, such as cytopenia, cutaneous or mucosal changes or infection, were much less frequent. Only 1% of the patients developed significant alopecia or neutropenia, and none had a severe infection. The cutaneous/mucosal and/or haematological complications of treatment are the best known, even though every study agrees on reporting a low frequency of these events.5,15,18 As for infection, 2 previous studies have reported a higher incidence15,17 compared to the one observed in our sample; both studies were based on German registers that categorised ARs based on the Medical Dictionary for Regulatory Activities (medDRA), so they included all infections, including mild cases. A study based on the claims data of the United States state-run Medicaid health programme established that the incidence of infections associated with MTX treatment in patients with JIA that require hospital admission is of 1.46 per 100 individuals per year.22

Different variables may be associated with the development of ARs, including age, sex, the clinical form of JIA, the dose and the route of administration of MTX, supplementation with folic acid and cultural factors. When it came to age, our results suggest that children aged more than 6 years at treatment initiation were more likely to develop intolerance compared to younger patients independently of the dose and duration of treatment with MTX, which was consistent with previous studies.5,21 Sex does not seem to contribute to the development of ARs, although 1 study found that hypertransaminasaemia was more frequent in female patients.13 The form of JIA also does not seem to play a relevant role with the exception of the higher frequency of abnormal blood test results in patients with systemic arthritis,16 which is probably attributable to the underlying disease. The role of the MTX dose is still a subject of debate in the literature. Fráňová et al.5 reported a proportion of intolerance similar to the one reported by other authors20 despite the use of 40% greater doses of MTX. We also found no association between the development of ARs and the dose used in our sample. Other authors, however, have reported a higher incidence of intolerance in patients receiving higher doses.13,21 The effect of the route of administration is also under debate. Factors such as the pain associated with subcutaneous injection do not seem to have a significant impact.23 Several authors have reported a higher frequency of intolerance with the parenteral administration of MTX,15,20,21 while others have found no differences,24 as was the case in our study, or have even found fewer ARs with the parenteral route compared to oral administration.25 In our study, we were unable to assess the impact of oral folic acid/folate supplementation because it was given to 100% of patients. Although some studies in the paediatric population have found an associated reduction in the incidence of ARs,26 others have not confirmed this finding.15,20 In any case, there is evidence from clinical practice and studies in adults that such supplementation may reduce the frequency of ARs associated with MTX, and consequently its use is recommended.3,4 Lastly, there is no question that cultural factors may have an impact on the tolerability of a chronic treatment. For instance, we ought to highlight the high frequency of behavioural changes found in studies conducted in the Netherlands, with an incidence as high as 60%,14,20 which is not observed in studies conducted in other countries, like ours in Spain, in which the incidence did not exceed 20%.5,15 A recent systematic review analysed potential predictors of the development of ARs, including several clinical, laboratory and genetic variables, and did not find any risk factors that could be applied to everyday clinical practice.27

Different strategies may be used once patients develop ARs, such as the addition of antiemetic drugs, behavioural therapy, changes in dosage or the route of administration and discontinuation of treatment. The usefulness of adding an antiemetic drug in case of nausea/vomiting, in particular ondansetron, was reviewed recently,28 and the evidence supporting its effectiveness was weak. Behavioural therapy has proven less effective than subcutaneous administration of MTX.29 None of these options was used in our hospital. The third alternative, which was changing the treatment, was used most frequently in our sample and in other studies,5,25 and achieved resolution of the AR in 65% of cases in our study. Decisions regarding the strategy to use should take into account the clinical condition of the patient (remission with treatment, nonresponse to MTX) and the severity or persistence (recurrence of hypertransaminasaemia) of the AR. Finally, ARs (most frequently hypertransaminasaemia and gastrointestinal manifestations) led to discontinuation of MTX in 35% of the sample.

In conclusion, MTX is a very useful drug in the management of JIA whose use is limited by the high frequency of associated ARs that, even if not severe, lead to its discontinuation.

The authors have no conflicts of interest to declare.