Abstract Background NUDT15 genetic variants are associated with thiopurine intolerance, but their long-term clinical implications in ulcerative colitis (UC) remain unclear. We evaluated the impact of NUDT15 genotype on disease course and the durability of thiopurine therapy in UC patients. Methods A total of 1,228 UC patients with confirmed NUDT15 genotypes—wild-type (WT, n = 961), heterozygous mutant (HT, n = 245), and homozygous mutant (HM, n = 22)—were enrolled. Clinical outcomes, including disease relapse, were compared across genotype groups. Among patients receiving thiopurines, subgroup analyses were performed to identify predictors of treatment discontinuation. Results Although HM patients had a higher rate of biologics use (45.5% vs. 22.6% in WT and 20.8% in HT; p = 0.039), cumulative risks of biologics use (p = 0.171), UC-related hospitalization (p = 0.430), and relapse (p = 0.305) did not differ among genotypes. Among thiopurine-treated patients (WT: n = 418; HT: n = 103), WT received significantly higher doses (p 0.001) with fewer leukopenia events (11.0% vs. 20.4%, p = 0.028), while other adverse event rates were similar (41.4% vs. 47.6%, p = 0.256). In WT patients, higher thiopurine doses were associated with lower treatment discontinuation rates (≤25 mg: 58.5% vs. 125–≤150 mg: 3.66%, p 0.001), whereas leukopenia incidence was not dose-dependent (p = 0.698). In HT patients, thiopurine dose was not significantly associated with treatment discontinuation or leukopenia. Multivariate analysis showed that NUDT15 genotype did not independently predict thiopurine discontinuation; high-dose therapy reduced discontinuation risk (≤25 mg, reference; 125–≤150 mg, HR 0.077, p 0.001), while prior immunomodulator-related adverse events increased it (HR 3.871, p 0.001). Conclusion NUDT15 genotype was not a significant determinant of long-term clinical outcomes in UC. In WT patients, high-dose thiopurine improved drug persistence without increasing safety risks. Therefore, titrating to an adequate dose in WT patients may enhance treatment durability. Conflict of interest: Chang, Ji Young: No conflict of interest Park, Soo Jung: No conflict of interest Park, Jae Jun: No conflict of interest Kim, Tae Il: No conflict of interest Cheon, Jae Hee: No conflict of interest Park, Jihye: No conflict of interest So, Jae Eun: No conflict of interest
Chang et al. (2026) studied this question.