RESUMO
BACKGROUND: Thiopurines' toxicity often leads to dose reduction or discontinuation. This systematic review aims to synthesize the evidence on the effect of genotype-based dosing of thiopurines on treatment efficacy and safety in inflammatory bowel disease (objective #1), and the association between genotype status and the efficacy and safety profile (objective #2). METHODS: The Cochrane Library, MEDLINE, and EMBASE were searched in August 2021. A total of 80 studies (19,859 individuals) were included. Meta-analyses for mortality, different types of adverse events (AEs), withdrawal due to AE, change in disease activity and clinical remission were performed following mainly a fixed-effects model. PROSPERO registration: CRD42020148130. RESULTS: Genotype-based dosing was associated to a significantly lower incidence of hematologic AEs (risk ratio=0.71; 95% CI: 0.56-0.90; I2 : 47%; 4 randomized controlled trials; moderate quality), which may be attributable to nudix hydrolase 15 (NUDT15) testing more than to thiopurine methyltransferase (TPMT) genotyping. No differences were found in other outcomes. Mutations in TPMT and NUDT15 genes were associated to a higher probability of serious AEs [odds ratio (OR) TPMT=4.98; OR NUDT15=11.44], hematologic AEs (OR TPMT=3.18), and serious hematologic AEs (OR TPMT=7.88; OR NUDT15=12.83). TPMT was also associated with a higher risk of withdrawals due to AEs (OR=3.38), and NUDT15 with gastrointestinal AEs (OR=2.04). Mutations in the ITPA gene did not lead to significant differences. Evidence of an association between other genes and clinical outcomes is still scarce. CONCLUSIONS: Mutations in TPMT and NUDT15 genes predispose patients to suffer thiopurine-induced toxicity, and genotype-guided treatment has been shown to contribute to the prevention of thiopurine-induced toxicity, especially in the case of NUDT15 in Asians.