Registry-based studies on the risk of colorectal cancer (CRC) for persons with a family history (FH) typically did not control for important covariates, such as history of colonoscopy. We aimed to quantify the association between FH and CRC risk, carefully accounting for potential confounders. We conducted a population-based case-control study in Germany. A total of 4,313 patients with a first diagnosis of CRC (cases) and 3,153 controls recruited from 2003 to 2014 were included. We used multiple logistic regression analyses to assess the association between FH and risk of CRC with odds ratios (OR) and the resulting 95% confidence intervals (95% CI). A total of 582 cases (13.5%) and 321 (10.2%) controls reported a history of CRC in a first-degree relative, which was associated with a 41% increase in risk of CRC (OR: 1.41, 95% CI 1.22-1.63) after adjustment for sex and age. The OR substantially increased to 1.73 (95% CI, 1.48-2.03) after comprehensive adjustment including previous colonoscopies. Irrespective of their FH status, persons with history of colonoscopies had a lower CRC risk compared with persons without previous colonoscopies and without family history (OR: 0.25, 95% CI, 0.22-0.28 for persons without FH and OR 0.45, 95% CI, 0.36-0.56 for persons with FH). In an era of widespread use of colonoscopy, adjusting for previous colonoscopy is therefore crucial for deriving valid estimates of FH-related CRC risk. Colonoscopy reduces the risk of CRC among those with FH far below levels of people with no FH and no colonoscopy.
