The evidence for testing thiopurine S-methyltransferase (TPMT) enzymatic activity or genotype before starting therapy with thiopurine-based drugs is unclear.To examine the sensitivity and specificity of TPMT genotyping for TPMT enzymatic activity, reducing harm from thiopurine by pretesting, and the association of thiopurine toxicity with TPMT status in adults and children with chronic inflammatory diseases.MEDLINE, EMBASE, the Cochrane Library, and Ovid HealthSTAR (from inception to December 2010) and BIOSIS and Genetics Abstracts (to May 2009).Two reviewers screened records and identified relevant studies in English.Data on patient characteristics, outcomes, and risk for bias were extracted by one reviewer and independently identified by another.54 observational studies and 1 randomized, controlled trial were included. Insufficient evidence addressed the effectiveness of pretesting. Genotyping sensitivity to identify patients with low and intermediate TPMT enzymatic activity ranged from 70.33% to 86.15% (lower-bound 95% CI, 54.52% to 70.88%; upper-bound CI, 78.50% to 96.33%). Sparse data precluded estimation of genotype sensitivity to identify patients with low to absent enzymatic activity. Genotyping specificity approached 100%. Compared with noncarriers, heterozygous and homozygous genotypes were both associated with leukopenia (odds ratios, 4.29 [CI, 2.67 to 6.89] and 20.84 [CI, 3.42 to 126.89], respectively). Compared with intermediate or normal activity, low TPMT enzymatic activity was significantly associated with myelotoxicity and leukopenia.Available evidence was not rigorous and was underpowered to detect a difference in outcomes.Insufficient evidence addresses the effectiveness of TPMT pretesting in patients with chronic inflammatory diseases. Estimates of the sensitivity of genotyping are imprecise. Evidence confirms the known associations of leukopenia or myelotoxicity with reduced TPMT activity or variant genotype.Agency for Healthcare Research and Quality.