Colorectal cancer in Crohn's disease: a Scandinavian population-based cohort study
Ola Olén,Rune Erichsen,Michael C Sachs,Lars Pedersen,Jonas Halfvarson,Johan Askling,Anders Ekbom,Henrik Toft Sørensen,Jonas F Ludvigsson
DOI: https://doi.org/10.1016/S2468-1253(20)30005-4
Abstract:Background: Crohn's disease is a risk factor for colorectal cancer (CRC). However, available studies reflect older treatment and surveillance strategies, and most have assessed risks for incident CRC without taking surveillance and lead-time bias into account. Such biases can be accounted for by assessing CRC incidence by tumour stage and CRC mortality by tumour stage. We aimed to assess rates of incident CRC and CRC mortality among patients with Crohn's disease compared with the general population. Methods: For this nationwide register-based cohort study, we used International Classification of Disease codes in national patient registers and pathology reports to identify incident cases of Crohn's disease. In Denmark we searched for incident cases between January, 1977, and December, 2011, and in Sweden between January, 1969, and December, 2017. For each patient with Crohn's disease, we identified up to ten reference individuals in national population registers and matched them by sex, age, calendar year, and place of residence. Matched reference individuals had to be alive and free of inflammatory bowel disease at the start of follow-up of index patients with Crohn's disease, and stopped contributing to reference person-years if they were diagnosed with inflammatory bowel disease. Our main outcome was death from CRC (main or contributory cause of death) as captured in the cause-of-death registers. Our secondary outcome was incident CRC, as defined by the cancer registers. We used Cox regression to estimate hazard ratios (HRs) for incident CRC and CRC mortality, taking tumour stage into account. We used a series of Cox models to estimate cause-specific HRs of the different competing outcomes (CRC diagnosis, CRC death, and other causes of death) and adjusted for tumour stage at CRC diagnosis. Findings: During the 1969-2017 study period, we identified 47 035 patients with incident Crohn's disease (13 056 in Denmark and 33 979 in Sweden) and 463 187 matched reference individuals. During follow-up, 296 (0·47 per 1000 person-years) CRC deaths occurred among individuals with Crohn's disease compared with 1968 (0·31 per 1000 person-years) in reference individuals, corresponding to an overall adjusted HR of 1·74 (1·54-1·96). 499 (0·82 per 1000 person-years) cases of incident CRC were diagnosed in patients with Crohn's disease compared with 4084 (0·64 per 1000 person-years) cases in reference individuals, corresponding to an overall adjusted HR of 1·40 (95% CI 1·27-1·53). Patients with Crohn's disease who were diagnosed with CRC were at increased risk of CRC mortality compared with reference individuals also diagnosed with CRC (HR 1·42 [1·16-1·75] when adjusted for tumour stage), and tumour stage at CRC diagnosis did not differ between groups (p=0·27). Patients with Crohn's disease who had follow-up of 8 years or longer or who were diagnosed with primary sclerosing cholangitis (PSC) and hence were potentially eligible for CRC surveillance had an increased overall risk of CRC death (HR 1·40 [1·16-1·68]) or CRC diagnosis (HR 1·12 [0·98-1·28]). However, in patients potentially eligible for CRC surveillance we only found significantly increased risks in patients diagnosed with Crohn's disease before the age of 40 years, patients with disease activity in the colon only, or patients with PSC. Interpretation: Patients with Crohn's disease are at increased risk of CRC diagnosis and CRC death. Patients with Crohn's disease who have CRC have a higher mortality than patients without Crohn's disease who are also diagnosed with CRC. CRC surveillance should likely be focused on patients diagnosed with Crohn's disease before the age of 40 years, on patients with colon inflammation, and on those who have PSC. Funding: Swedish Medical Society, Karolinska Institutet, Regional Agreement on Medical Training and Clinical Research between Stockholm County Council and Karolinska Institutet (ALF), Forte Foundation, Swedish Cancer Foundation, and Independent Research Fund Denmark.