Ultraprocessed Food Consumption and Risk of Early-Onset Colorectal Cancer Precursors Among Women.
Level 3 - non-randomized controlled study
Prospective non-randomized cohort study
PubMed 41231486 · doi:10.1001/jamaoncol.2025.4777
What was done
Researchers analyzed data from 29,105 female registered nurses in the prospective Nurses' Health Study II cohort (followed from 1991 to 2015, mean age 45.2 years) who underwent at least one lower endoscopy before age 50 and had no history of cancer, polyps, or inflammatory bowel disease at baseline. Ultraprocessed food (UPF) consumption was calculated from food-frequency questionnaires administered every 4 years and categorized into quintiles using the Nova classification. The primary outcome was incidence of early-onset colorectal cancer precursors (conventional adenomas and serrated lesions) confirmed by medical records and pathology reports, analyzed using multivariable logistic regression with generalized estimating equations.
What was found
Over 24 years of follow-up, 1,189 cases of early-onset conventional adenomas and 1,598 serrated lesions were documented. UPFs accounted for 34.8% of daily caloric intake (median 5.7 servings/day). Women in the highest versus lowest quintile of UPF intake had a significantly higher risk of early-onset conventional adenomas (adjusted odds ratio [AOR], 1.45; 95% CI, 1.19-1.77; overall P < .001). No significant association was observed for serrated lesions (AOR, 1.04; 95% CI, 0.89-1.22; P for trend = .48). The association for conventional adenomas persisted after adjusting for body mass index, type 2 diabetes, dietary factors (fiber, folate, calcium, vitamin D), and overall diet quality score.
Why it matters
Early-onset colorectal cancer is rising globally; these findings provide prospective evidence linking high ultraprocessed food consumption specifically to early-onset conventional adenoma development, highlighting dietary quality as a target for prevention.
Limits
The study included only female registered nurses, limiting generalizability to men and more racially or socioeconomically diverse populations. Dietary intake was assessed via self-reported food-frequency questionnaires every 4 years, introducing potential recall error and misclassification, and residual confounding cannot be fully ruled out despite multivariable adjustment.
Cited by
- context The average American consumes 7.5 servings of ultra-processed foods per day.