Traditional Risk Factors Are Not Associated with the Rise of Early-Onset Colorectal Cancer: An All of Us Study.
TL;DR
Traditional environmental risk factors and established common genetic variants are not associated with the rising incidence of early-onset colorectal cancer, and drivers of early-onset disease might include novel environmental exposures or early-life factors beyond traditional risk assessment.
Abstract
Background
Colorectal cancer incidence is increasing among adults under age 50, yet the etiologic drivers remain unclear. The All of Us Research Program provides a unique opportunity to examine age-specific genetic and environmental associations with colorectal cancer risk in a diverse national cohort.
Objective
To determine whether a significant association exists between traditional environmental risk factors and established genetic variants and age-specific incidence patterns of early-onset colorectal cancer.
Design
Retrospective stratified case-control study using whole genome sequencing and electronic health records. Conditional logistic regression adjusted for genetic ancestry. SETTINGS Diverse national cohort with integrated genomic data, electronic health records, and lifestyle surveys. PATIENTS A total of 2,455 colorectal cancer patients (485 early-onset diagnosed before age 50, 1,970 late-onset diagnosed at age ≥50) matched 4:1 to 9,820 controls on birth year, sex, and race. MAIN OUTCOME MEASURES Age-specific associations between environmental factors (family history, smoking timing, alcohol use disorders, body mass index) and 283 genetic variants with colorectal cancer risk.
Results
Family history was associated with higher odds uniformly across ages (odds ratio 2.10, 95% confidence interval 1.69-2.61). Heavy smoking during ages 20-40 was associated with late-onset but not early-onset disease (interaction p = 0.006). Alcohol use disorders and body mass index showed no age-specific effects. In genetic analyses, no variants demonstrated age-specific effects after correction for multiple comparisons (283 variants tested). No genetic pathways differed between early-onset and late-onset disease. Two variants were associated with rectal-specific risk (VTI1A rs4554812, FMN2 rs2078095).
Limitations
Causal inference is limited by observational design. Self-reported exposures are subject to recall bias. Age-stratified analyses had limited power for rare genetic effects.
Conclusions
Traditional environmental risk factors and established common genetic variants are not associated with the rising incidence of early-onset colorectal cancer. Thus, drivers of early-onset disease might include novel environmental exposures or early-life factors beyond traditional risk assessment. See Video Abstract.