Absolute Cancer Risk in Crohn’s Disease

September 1, 2026

Issue 17

Clinical Question

What is the absolute excess cancer risk in patients with Crohn’s disease by treatment type?

Editor’s Bottom Line

Overall, these data on rates of malignancy among patients with IBD are reassuring. Setting aside non-melanoma skin cancer, the absolute increases in cancer incidence are small and more likely top reflect residual confounding than a true treatment effect.

Reference

Everhov ÅH, Eriksson J, Söderling J, et al. Absolute Risk of Malignancy by Treatment in Patients With Crohn’s Disease Compared with the General Population: A Nationwide Population-Based Cohort Study. Am J Gastroenterol. Epub ahead of print August 26, 2026. https://doi.org/10.14309/ajg.0000000000004119

Synopsis

This nationwide population-based cohort study linked Swedish registry data to assess incident cancers in 38,733 patients with Crohn’s disease (CD) and 360,616 matched individuals in the general population between 2007 and 2023. Six treatment cohorts were defined, including immunomodulator-naïve patients, thiopurine recipients, patients receiving a tumor necrosis factor inhibitor (TNFi), combination thiopurine with a TNFi, vedolizumab, and ustekinumab.

A once-exposed-always-exposed design was used, meaning patients contributed person-time from one year after treatment initiation regardless of subsequent discontinuation, and could contribute to more than one cohort. The primary measure was the incidence rate (IR) difference, defined as excess cancer cases per 1,000 person-years in CD vs. the matched population, stratified by treatment and age.

Over a median of 7.3 years of follow-up, excess overall cancer incidence was observed across all treatment cohorts, with IR differences ranging from 1.54 for ustekinumab to 3.14 for thiopurine. The 5-year cumulative cancer incidence in CD vs. the matched general population was 5.2% vs. 4.1% for immunomodulator-naïve patients, 3.9% vs. 2.4% for thiopurine exposed patients, 3.2% vs. 2.1% for TNFi exposure, 3% vs. 1.8% for combination thiopurine + TNFi exposure, 4.2% vs. 3% for vedolizumab, and 3.4% vs. 2.5% for ustekinumab. This excess was driven almost entirely by nonmelanoma skin cancer (basal and squamous cell carcinomas).

After excluding nonmelanoma skin cancer, statistically significant excess incidence was limited to the immunomodulator-naïve subgroup (IR difference: 1.27; 95% Confidence Interval [CI], 0.88–1.65), thiopurine recipients (1.14; 95% CI, 0.65–1.64), and TNFi cohorts (0.96; 95% CI, 0.34–1.58), amounting to roughly 1 extra cancer case per 1,000 person-years, attributable to lung, small bowel, hepatobiliary and hematologic cancers.

Of note, the immunomodulator-naïve cohort was older than the treatment-defined groups (mean age 47 at follow-up start vs. 33–43 for the treatment cohorts) and had the greatest comorbidity burden. After excluding nonmelanoma skin cancer, hazard ratios approached 1 across all cohorts. No elevated IRs were found for colorectal, breast, or prostate cancers in any cohort.

Lymphoma cases were elevated among the thiopurine-exposed (Hazard Ratio [HR]: 1.54; 95% CI, 1.08-2.19) and the TNFi-exposed cohorts (HR: 2.91; 95% CI, 1.84–4.61) but not in the vedolizumab or ustekinumab cohorts. For hepatobiliary cancer, HRs were elevated across cohorts but were heavily driven by co-existing primary sclerosing cholangitis (PSC), such that among immunomodulator-naïve patients with PSC, the IR difference was 6.31 extra cases per 1,000 person-years, compared to 0.11 in those without PSC at baseline. Cancer incidence varied more by age than by treatment, with IR differences not significantly elevated in patients under 18 but increased substantially in adults, particularly among those aged 60 and older.

Details

Study Design: Nationwide population-based observational cohort study
Funding: Swedish Research Council, Swedish Cancer Society, Karolinska Institutet, Vinnova, and ERA PerMed ScandRA
Allocation: Non-randomized
Setting: Nationwide, register-based
Level of Evidence: 2b