Background/Purpose: Methotrexate (MTX) exerts its effect in part by inhibiting 5-aminoimidazole-4-carboxamide ribonucleotide transformylase (ATIC), a key enzyme in the purine biosynthetic pathway. Published data  suggests an association between a non-synonymous single nucleotide polymorphism  (SNP) in the ATIC gene (347C>G, Thr116Ser) with MTX efficacy and toxicity in patients with rheumatoid arthritis (RA); however this association has been inconsistent. The aim of this study was to investigate the effects of the ATIC 347C>G SNP on  MTX efficacy and toxicity in patients with RA by performing a meta-analysis.

Methods: Studies examining the association  of the ATIC 347C>G  polymorphism with  MTX efficacy and toxicity in RA patients were systematically identified from the Pubmed and Ovid Medline databases (from 1999 to May 31st 2012). Studies reporting genotypic or allelic distribution of ATIC 347C>G polymorphism in relation to MTX efficacy and/or toxicity were deemed eligible for the current analysis. Meta-analyses for both MTX efficacy and toxicity were separately performed to explore the composite effect of  the ATIC 347C>G SNP on these outcomes. Fixed (Mantel-Haenszel) and random effects (DerSimonian and Laird) models were utillized based on the presence or absence of between-study heterogeneity. OpenMetaAnalyst software (beta version) was used for data entry and analysis.

Results: Five studies with sufficient data  were included in the meta-analysis of the ATIC 347C>G SNP’s association with MTX efficacy which represented 1100 RA patients. The meta-analysis did not demonstrate a significant association between the ATIC 347C>G polymorphism and MTX efficacy (Figure 1A). The pooled odds ratio (OR) using the fixed effects model was 1.194 (95% CI 0.915, 1.559; p = 0.193) with a trend towards between-study heterogeneity (p=0.091) and using the random effects model was 1.168 (95% CI 0.786, 1.736; p = 0.441). For the meta-analysis of the ATIC 347C>G SNP’s association with MTX toxicity, five studies  with sufficient data were included which represented 1055 patients. Meta-analysis using the fixed effects model demonstrated a significant association of the ATIC 347C>G SNP with MTX toxicity with an OR of 0.686 (95% CI 0.515, 0.915; p = 0.010) with no evidence of between-study heterogeneity (p=0.245) (Figure 1B).

Conclusion: Our meta-analysis suggests that the ATIC 347C>G polymorphism may be a predictor of MTX toxicity in patients with RA, with the C allele being protective and the G allele conferring a higher risk for toxicity.