We carried out a systematic search in Medline and the Cochrane Library for human studies published in English or Spanish in the past 15 years for which an abstract was available. The keywords used in the search were “idiopathic juvenile arthritis”, “methotrexate”, “routes of administration”, “dosage”, “tapering”, “discontinuation” and other, more specific terms based on each particular clinical question. Finally, we completed the search with articles provided directly by the GDG and the summaries of product characteristics for methotrexate.

Of the resulting list of publications, we selected review articles, consensus documents, guidelines by scientific associations and articles that explicitly addressed specific aspects of the proposed questions, and proceeded to summarise the evidence.

We used the nominal group technique. The experts discussed the contents of the evidence summary in a structured fashion during a work session. Based on the conclusions reached at the session, we drafted a document with the agreed-on recommendations and determined the level of evidence (LE) and grade of recommendation (GR) for each of them, either adhering to the system employed in the source (if any had been used) or applying the grading system of the Scottish Intercollegiate Guidelines Network (SIGN).4 Lastly, the resulting document was presented to the group of experts for a final review.

We obtained 344 references in the literature search. Through a subsequent screening, we selected 67 articles, 6 clinical practice guidelines and 9 summaries of product characteristics for critical reading. After evaluation by the GDG members and the implementation of the structured and participatory consensus methodology, we validated sixteen recommendations (Table 1).

A study by Giannini et al.16 showed that an initial MTX dose of 5mg/m2 is no more efficient than placebo. A study conducted by PRINTO10 showed that patients with polyarticular JIA responded to MTX within nine months of starting treatment at standard doses of 8–12.5mg/m2 (administered by the oral, subcutaneous and intramuscular routes). Thus, we have determined the following:

• •

  Recommendation 5: the following doses are recommended8: initial dose, 10–15mg/m2/week; maximum total dose: 25mg/week (IA; Niehues8).

While additional randomised controlled trials are needed, several studies suggest that MTX is useful for the treatment of chronic uveitis in patients with JIA, preventing its onset,17 and improving disease activity in a significant percentage of patients.18 Thus, we agreed on the following initial dose:

• •

  Recommendation 6: in patients with uveitis or polyarthritis, an initial dose of 15mg/m2/week is recommended (4D; SIGN4).

The study of Niehues et al.8 reviews several studies19–22 on the bioavailability of MTX. The conclusion is that for doses of up to 15mg/m2/week, the bioavailability after parenteral and oral administration is similar; at higher doses (≥15mg/m2/week) parenteral administration is recommended for its higher bioavailability and tolerability.

• •

  Recommendation 7: for doses ≥15mg/m2/week, parenteral administration is recommended due to a higher bioavailability and tolerability8 (IIIC; Niehues8).

The criteria for making decisions regarding the initial dose and route of administration are variable and subject to the preferences of the patient and/or family members and the required doses.

• •

  Recommendation 8: for initial doses of 10mg/m2/week, a choice will be made between oral or subcutaneous administration, deciding on whichever is most suitable based on the circumstances of the patient (4D√; SIGN4).

We did not find sufficient evidence to make a recommendation. However, the data of a clinical trial23 show similar results whether MTX is discontinued six or twelve months after achieving remission. Thus, in principle, MTX could be discontinued six months after achieving remission.

The panel of experts found a wide variability in the tapering off and discontinuation of MTX doses in everyday clinical practice: from withdrawal of treatment to tapering down in steps of 2.5mg in each visit (every 6–8 weeks). Other strategies consisted of increasing the interval between doses without changing the dose itself; that is, for example, giving the same dose every 15 days as opposed to every week, and so on, until discontinuation.

In gradual tapering schemes, it is acceptable to choose to switch to oral administration once a dose of 10mg is reached, applying the previously noted criteria regarding the route of administration.

Based on the review of the summaries of product characteristics, the most significant adverse effects of MTX involve the suppression of the haematopoietic system and gastrointestinal disorders. Specifically, the most frequent adverse reactions (≥1/10) are gastrointestinal (stomatitis, dyspepsia, nausea and loss of appetite) and hepatobiliary (elevated transaminases); with less frequent effects (≥1/100, <1/10) including involvement of the gastrointestinal tract (mouth ulcers and diarrhoea), cutaneous and subcutaneous tissues (exanthema, erythema and pruritus), nervous system (headache, fatigue and sleepiness), respiratory system, thorax and mediastinum (pneumonia and interstitial alveolitis/pneumonitis, often associated with eosinophilia) and the blood and lymphatic system (leukopenia, anaemia and thrombocytopenia). Hair loss, while being rare (≥1/1000, <1/100), is usually a source of concern for the families.24

The development and degree of severity of adverse reactions depend on the dose and frequency of administration. Since severe adverse reactions may occur even at the lowest doses, it is imperative that physicians monitor these patients at regular intervals (every 3–4 months).

We recommend the subcutaneous administration of MTX, as it is well tolerated at the site of injection. Only mild local skin reactions have been observed, and these diminish over the course of treatment. Intramuscular administration has led to reports of local adverse reactions (burning feeling) or lesions (formation of sterile abscesses or destruction of fatty tissues) at the site of administration.24

The contraindications for MTX need also be considered24: hypersensitivity to MTX or any of the excipients; liver failure, alcohol abuse, or kidney failure; pre-existing blood dyscrasias such as bone marrow hypoplasia, leukopenia, thrombocytopenia or significant anaemia; severe acute or chronic infection, such as tuberculosis or HIV infection; mouth ulcers or known active gastrointestinal ulcerative disease; pregnancy and breastfeeding; and concurrent vaccination with live or live-attenuated vaccines, although the evidence-based recommendations of a panel of experts of the EULAR25 recommend the live-attenuated vaccine if the dose used is less than 15mg/m2.

A study by Ortiz-Álvarez et al.26 assessed the utility of the ACR guidelines for monitoring patients with JIA, and estimated a probability of 11% for having significantly abnormal blood tests at three months, compared to 10% probability of having an abnormal test by chance alone. Consequently, they deemed that periodical routine blood tests were unnecessary. The conclusions of a different study27 suggested that the adult standard of surveillance laboratory testing every four to eight weeks might not be necessary in children treated with MTX, especially if certain risk factors are absent. Other studies suggest the need to monitor for potential and usually reversible liver function test abnormalities, especially if the dose needs to be increased or large doses of MTX are required.28 There seems to be no question that testing should be performed to confirm normal levels one month after initiation of MTX treatment.

• •

  Recommendation 9: it is recommended that laboratory testing of baseline parameters is repeated one month after initiating MTX or one to two months after any dose increases (2+D√; SIGN4).

• •

  Recommendation 10: in patients receiving stable doses and with no previous test abnormalities, surveillance laboratory testing is recommended every 3–4 months (4D; SIGN4).

Hashkes et al.29 analysed the relationship between hepatotoxic risk factors and liver histology in patients with JIA treated with MTX. They found that biochemical abnormalities were associated with liver damage based on the Roenigk classification and the presence of liver fibrosis, which concurred with studies of patients with rheumatoid arthritis (RA). This suggests that the guidelines for monitoring MTX hepatotoxicity in AR may be applicable to JIA. These data along with further evidence extrapolated from adults with RA27–30 justified the recommendations of the ACR7:

• •

  Recommendation 11: in response to liver enzyme elevation between 1 and 2 times the upper limit of normal, either no specific action or rechecking liver enzymes at a shorter interval until the levels normalise is recommended7 (C; Oxford6).

• •

  Recommendation 12: in response to liver enzyme elevation more than twice the upper limit of normal, decreasing the dose of methotrexate or temporarily withholding methotrexate administration is recommended7 (C; Oxford6).

• •

  Recommendation 13: if liver enzymes remain at levels more than 3 times the upper limit of normal following a decrease in the methotrexate dose, discontinuation of methotrexate is recommended7 (C; Oxford6).

Other possible causes of elevated liver enzymes unrelated to the administration of MTX must be investigated, and tapering off can be considered if these levels are abnormal or elevated.

More evidence is needed to establish the indications for combination therapy with MTX, be it with conventional or biologic DMARDs, in order to simplify decision-making.

The recommendations of the ACR7 propose the addition of a TNFα inhibitor for patients receiving MTX as monotherapy and with persistent JIA activity, recommending that treatment with MTX be continued or not depending on the patient's previous response to it. Other authors recommend exhausting the time frame to observe the response to MTX therapy and achieve the maximum effective doses of MTX by the parenteral route before considering initiation of combination therapies.31,32

In recent years, several studies have demonstrated that in some patients clinical worsening and a diminished response to treatment with certain TNF inhibitors is associated with the development of anti-drug antibodies.33,34 On the other hand, other studies have shown how the concomitant use of MTX reduces the immunogenicity of these drugs.35,36 For all of the above, we recommend considering the use of MTX in patients treated with biologic agents to improve the response and/or reduce the development of antibodies to certain biologics.

• •

  Recommendation 14: we recommend considering the use of MTX in patients treated with biologic agents to improve the response and/or reduce the development of antibodies to certain biologics (4D; SIGN4).

Methotrexate reduces inflammation by a mechanism related to the metabolism of folic acid. The prescription of folic or folinic acid can improve the tolerability and safety of MTX (mouth ulcers, gastrointestinal complaints, diarrhoea, haematologic changes, elevation of transaminases, etc.). Folic or folinic acid is given at least one day apart from the administration of MTX.8,12

Some expert-consensus guidelines state the need of supplementing MTX treatment with folic acid.37

• •

  Recommendation 15: to prevent side effects such as dyspepsia and nausea, the use of either folic or folinic acid is recommended as an adjuvant to MTX therapy. Folic or folinic acid will be given in tablet form in doses of 5mg/week at least one day apart from the administration of MTX or, alternatively, in doses of 1mg/day, skipping the day that MTX is administered (LE, 4; GR, D; SIGN4).