Abstract
BACKGROUND: Methotrexate (MTX) is the cornerstone disease-modifying anti-rheumatic drug in juvenile idiopathic arthritis (JIA). In JIA, it is important to start effective treatment early to avoid long-term sequelae, such as joint damage. To accomplish this goal, it is crucial to know beforehand who is going to respond well to MTX. In addition, MTX adverse effects such as MTX intolerance occur frequently, potentially hindering its efficacy. To avoid inefficacy of an otherwise effective drug, the physician should be timely aware of these adverse events. Consequently, to optimise treatment of JIA patients with MTX, predictors for efficacy and adverse events should be used in daily clinical practice. The aim of this study was to summarise the existing knowledge about such predictors. METHODS: A systematic literature search was performed in PubMed, Embase and The Cochrane Library, and 1,331 articles were identified. These were selected based on their relevance to the topic and critically appraised according to pre-defined criteria. Predictors for MTX efficacy and adverse events were extracted from the literature and tabulated. RESULTS: Twenty articles were selected. The overall quality of the studies was good. For MTX efficacy, candidate predictors were antinuclear antibody positivity, the childhood health assessment questionnaire score, the myeloid-related protein 8/14 level, long-chain MTX polyglutamates, bilateral wrist involvement and some single nucleotide polymorphisms (SNPs) in the adenosine triphosphate binding cassette and solute carrier transporter gene families. For MTX adverse events, potential predictors were alanine aminotransferase and thrombocyte level and two SNPs in the γ-glutamyl hydrolase and methylenetetrahydrofolate reductase genes. However, validation of most predictors in independent cohorts was still lacking. CONCLUSIONS: Interesting candidate predictors were found, especially for MTX efficacy. However, most of these were not validated. This should be the goal of future efforts. A clinically relevant way to validate the predictors is by means of creating a clinical prediction model.