Risk-adjusted colon and rectal cancer incidence rates in the United States.

Abstract

Population-based cancer incidence rates that adjust for multiple cancer primaries and for prevalent cases of the disease provide a better approximation of risk. This study is based on a retrospective cohort. Included in the study were 9 original Surveillance, Epidemiology and End Results registries focusing on white and black males and females from 2000 through 2007. The main outcome measured was malignant colorectal cancer. Conventional colorectal cancer incidence rates overestimate population risk by 3.6% for white males, 4.0% for black males, 3.4% for white females, and 3.3% for black females. The level of overestimation bias remained similar across the age span for white and black males. However, for white females, rates were overestimated by 2.1% for ages 30 to 39 and increased to 3.8% for ages 80 years and older. Corresponding values for black females were 1.5% and 3.8%. The trends in conventional rates were generally similar to the trends in risk-adjusted incidence rates, increasing or stable before age 50, but decreasing thereafter. The number of colorectal cancer cases in the United States is estimated from conventional incidence rates. In 2007, the number of colorectal cases was 59,599 for white males, 7,670 for black males, 58,972 for white females, and 8,786 for black females. The number of colorectal cancer cases based on prevalence-corrected incidence rates increased by 2.2% for white males, 1.5% for black males, 2.1% for white females, and 1.5% for black females. Colorectal cancer incidence rates that include second and later colorectal cancer primaries and adjust for prevalence better reflect cancer burden, whereas colorectal cancer incidence rates that only include the first diagnosed case and adjust for prevalence better reflect cancer risk.