Background: Thiopurine-induced leukopenia, a frequently observed and potentially life-threatening adverse event, complicates the clinical management of IBD patients. Aim: To assess risk factors for thiopurine-induced leukopenia in IBD. Methods: MEDLINE, EMBASE, BIOSIS and Cochrane library were searched for studies reporting at least one risk factor for thiopurine-induced leukopenia. Pooled odds ratio (OR) was calculated for each potential risk factor using a random effects model. Studies that were not eligible for meta-analysis were described qualitatively. Results: Seventy articles were included, 34 (11 229 patients) were included in meta-analyses. A significantly higher thiopurine-induced leukopenia risk was found for TPMT (OR 3.9, 95% [CI] 2.5-6.1) and for NUDT15 R139C (OR 6.9, 95% CI 5.2-9.1), G52A (OR 3.2, 95% CI 1.3-7.9) and 36_37ins/delGGAGTC variant carriers (OR 5.6, 95% CI 2.8-11.4). A potential association between high 6-thioguanine nucleotides (6-TGN) or 6-methylmercaptopurine (6-MMP) levels and leukopenia was observed, since most studies reported higher metabolite levels in leukopenic patients (6-TGN: 204-308 (Lennard method) and 397 (Dervieux method), 6-MMP: 4020-10 450 pmol/8 x 108 RBC) compared to controls (6-TGN: 170-212 (Lennard method) and 269 (Dervieux method), 6-MMP: 1025-4550 pmol/8 x 108 RBC). Conclusions: TPMT and NUDT15 variants predict thiopurine-induced leukopenia. High 6-TGN and 6-MMP levels might induce leukopenia, although exact cut-off values remain unclear. Potential preventive measures to reduce the risk of thiopurine-induced leukopenia include pre-treatment TPMT and NUDT15 genotyping. Routine thiopurine metabolite measurement might be efficient, yet cut-off levels must be validated in advance.