Polygenic risk scores (PRS) are more strongly associated with colorectal cancer (CRC) risk in more recent birth cohorts, suggesting that gene–environment interaction may be contributing to the global rise in early-onset disease, according to a pooled analysis presented at Digestive Disease Week® (DDW) 2026.
“The incidence of colorectal cancer has been rising globally among adults aged under 50 years, particularly those born after the 1950s,” study presenter Xinyu Wang, MB, told GI & Hepatology News. “While genetic factors are known to play an important role in CRC risk, the environment also matters and has changed substantially across generations. What makes this question important is that we don't yet know whether genetic risk has the same impact in different birth cohorts, since individuals born in more recent cohorts have experienced substantially different exposure profiles than those born decades earlier, including an epidemic of obesity (especially in childhood), more sedentary lifestyles, greater consumption of ultra-processed foods, and increased exposure to pollution and environmental chemicals.”
Wang and colleagues analyzed 37,313 CRC cases and 35,891 controls with available birth year information and genetically defined European ancestry. They calculated each individual's PRS as the weighted sum of risk alleles across 205 CRC-associated genetic variants identified to date and used multivariable logistic models to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for CRC risk by PRS across birth cohorts. Interaction between PRS and birth cohort was assessed using the Wald test, with subgroup analyses by age at diagnosis, endoscopic screening history, family history, and CRC molecular subtypes.
PRS showed a generally stronger association with CRC risk across successive birth cohorts. The OR for CRC per 1–standard deviation increases in PRS was 1.62 among individuals born before 1920, 1.55 for those born 1920–1929, 1.6 for 1930–1939, 1.67 for 1940–1949, 1.64 for 1950–1959, and 1.79 for those born in 1960 or later, noted Wang of the Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston.
The cohort effect was concentrated in individuals without a first-degree family history of CRC, with ORs rising from 1.62 in those born before 1920 to 1.91 in those born in 1960 or later. The formal interaction tests by age at diagnosis and by history of endoscopic screening did not reach statistical significance.
Across tumor subsites, the strengthening association was specific to distal CRC and rectal CRC, with no birth-cohort effect observed for proximal CRC. For tumor molecular subtypes, the investigators observed no interactions between PRS and birth cohort for subtypes classified by BRAF and KRAS mutations, CpG island methylator phenotype, or microsatellite instability status, although only 9.7%–12.3% of CRC cases had molecular data available, limiting statistical power.
Wang said the underlying point estimates nevertheless suggested a striking divergence by age at diagnosis. “One of the most surprising findings was how genetic effects differ across birth cohorts, and that this pattern depended strongly on age,” she said. “Among individuals diagnosed before age 55, the association between polygenic risk score and colorectal cancer became stronger in more recent birth cohorts. In contrast, among those aged 75 and older, the magnitude of association was slightly decreasing over more recent birth cohorts. Since the underlying genetic architecture is relatively stable over time, this pattern points to a growing influence of environmental exposures that may be interacting with genetic susceptibility, especially in younger generations. It also suggests that genetic risk prediction tools may be more informative in the contemporary populations, particularly when applied at younger ages, before routine screening typically begins.”
The site-specific findings tracked with broader epidemiologic patterns, Wang noted. “One additional point worth highlighting is that the birth cohort effect we observed was specific to distal colon and rectal cancers and not seen for proximal colon cancer. This is consistent with broader epidemiologic trends showing that the rise in early-onset colorectal cancer is largely driven by increases in distal and rectal tumors. It also suggests the etiologic heterogeneity across tumor sites. Distal and rectal cancers tend to be more strongly linked to metabolic risk factors, such as obesity, Western dietary patterns, and sedentary lifestyles, which have become more common in recent generations. These metabolic factors may amplify or interact with underlying genetic predisposition, thus driving our observed birth cohort effect for PRS.”
The clinical implication, Wang said, is that PRS-based stratification may be most valuable when deployed early. “Our findings suggest that genetic risk stratification may be increasingly useful in today's population, especially when applied at younger ages before routine screening typically begins. In particular, polygenic risk scores could help identify individuals at higher risk earlier in life, which may support more personalized screening strategies.”
Limitations include the use of birth cohort as a proxy for generational shifts in environmental exposure rather than direct measurement of those exposures, restriction to individuals of genetically defined European ancestry, and limited statistical power for the molecular subtype analyses. The investigators called for further work to better understand the mechanisms underlying the site-specific patterns and the shifting epidemiology of CRC, particularly the rise in early-onset disease.
Wang reported having no conflicts of interest.
DDW is AGA's annual meeting, jointly sponsored by AGA, AASLD, ASGE, and SSAT. Learn more at ddw.org.