Absolute risk of malignancy by treatment in patients with Crohn’s disease compared to the general population: a nationwide population-based cohort study

Background: Real-world data quantifying absolute cancer risk in Crohn’s disease (CD) by treatment status are lacking but are essential for patient counselling. Method: We linked nationwide Swedish register data and assessed incident cancers overall and by type in patients with CD 2007–2023. We estimated age-stratified incidence rate (IR) differences compared to the general population in a “once-exposed-always-exposed” design. Treatment cohorts comprised new users of thiopurine, tumor necrosis factor inhibitors (TNFi), thiopurine+TNFi, vedolizumab, ustekinumab, and patients naïve to immunomodulatory drugs. Results: We followed 38,733 patients and 360,616 comparator subjects for a median 7.3 years. The IR differences for any cancer (number of excess cancers in each treatment cohort per 1,000 person-years versus the matched population) were 2.43 (95%CI:1.92;2.94) for the naive cohort; 3.14 (95%CI:2.50;3.78) for thiopurine-treated, 2.42 (95%CI:1.63;3.22) for TNFi-treated, 2.59 (95%CI:1.53;3.64) for thiopurine +TNFi, 2.61 (95%CI:-0.08;5.30) for vedolizumab, and 1.54 (95%CI:-1.34;4.41) for ustekinumab. The excess cancer incidence was primarily due to non-melanoma skin cancer (squamous cell and basal cell carcinoma). After exclusion of non-melanoma skin cancers, the excess overall cancer incidence was 1.27 to 0.96 cases/1,000 person-years in the naïve, thiopurine, and TNFi-treated groups, but not significantly increased in the other. IR differences for overall cancer increased and HRs decreased with increasing patient age. Conclusion: Elevated overall cancer incidence was observed across all treatment cohorts, also in treatment-naïve patients. After exclusion of non-melanoma skin cancer, the excess cancer incidence in CD was around 1 extra case per 1,000 person-years due to lung-, small bowel -, hepatobiliary, and hematologic cancer.