Sex, Age, and Birth Cohort Effects in Colorectal Neoplasms A Cohort Analysis

Background: Prevalence of advanced colorectal neoplasms increases with age and is higher among men than women. Cross-sectional analyses estimated that men reach an equivalent prevalence at a much younger age than women. However, cross-sectional estimates may be confounded by birth cohort effects. Objective: To estimate age and cohort effects in advanced colorectal neoplasms and to adjust risk-advancement periods for men compared with women for birth cohort effects. Design: Age-cohort analyses. Setting: German screening colonoscopy program, 2003 to 2007. Participants: 2 185 153 participants aged 55 to 75 years. Measurements: Sex-and age-specific prevalence of colorectal cancer (CRC) and advanced neoplasms (CRC or advanced adenoma) were plotted with and without stratification by birth cohort. Risk-advancement periods with 95% CI for men compared with women were estimated from log-binomial regression models with and without cross-sectional analysis adjustment for birth cohort effects. Results: Overall, 17 196 participants (0.8%) had CRC and 152 429 (7.0%) had any advanced neoplasm. Age-specific prevalence was higher in men than in women and in later birth cohorts. The apparent modest increase in prevalence by age in cross-sectional analysis was much steeper after birth cohort effects were controlled for. In cross-sectional analysis, risk-advancement periods (95% CI) for men compared with women were 8.4 years (CI, 7.7 to 9.0 years) and 16.1 years (CI, 15.8 to 16.5 years) for CRC and any advanced neoplasm, respectively, and 3.4 years (CI, 2.6 to 4.3 years) and 6.9 years (CI, 6.4 to 7.4 years) after controlling for birth cohort effects. Limitation: Information on covariates that could explain cohort effects was lacking. Conclusion: In this population, strong cohort effects reduced age gradients in advanced colorectal neoplasms and inflated risk-advancement periods for men compared with women, but major risk advancement persisted, even after birth cohort effects were controlled for.