The contribution of candidate genetic variants involved in azathioprine biotransformation on azathioprine efficacy and pharmacokinetics in 111 young patients with inflammatory bowel disease was evaluated. Azathioprine doses, metabolites thioguanine-nucleotides (TGN) and methylmercaptopurine-nucleotides (MMPN) and clinical effects were assessed after at least 3 months of therapy. Clinical efficacy was defined as disease activity score below 10. Candidate genetic variants (TPMT rs1142345, rs1800460, rs1800462, GSTA1 rs3957357, GSTM1, and GSTT1 deletion) were determined by polymerase chain reaction (PCR) assays and pyrosequencing. Statistical analysis was performed using linear mixed effects models for the association between the candidate variants and the pharmacological variables (azathioprine doses and metabolites). Azathioprine metabolites were measured in 257 samples (median 2 per patient, inter-quartile range IQR 1-3). Clinical efficacy at the first evaluation available resulted better in ulcerative colitis than in Crohn’s disease patients (88.0% versus 52.5% responders, p = 0.0003, linear mixed effect model, LME). TGN concentration and the ratio TGN/dose at the first evaluation were significantly higher in responder. TPMT rs1142345 variant (4.8% of patients) was associated with increased TGN (LME p = 0.0042), TGN/dose ratio (LME p < 0.0001), decreased azathioprine dose (LME p = 0.0087), and MMPN (LME p = 0.0011). GSTM1 deletion (58.1% of patients) was associated with a 18.5% decrease in TGN/dose ratio and 30% decrease in clinical efficacy. GSTA1 variant (12.8% of patients) showed a trend (p = 0.049, LME) for an association with decreased clinical efficacy; however, no significant effect on azathioprine pharmacokinetics could be detected. In conclusion, GSTs variants are associated with azathioprine efficacy and pharmacokinetics.