Skip to main content
NUTRITION 2026 Main banner

(OR09-07-26) Ultra-Processed Foods and Cancer Risk: Associations and Methodological Considerations in the American Cancer Society’s Cancer Prevention Study-II Nutrition Cohort

Sunday, July 26, 2026
11:42 AM - 11:54 AM ET
Location: Maryland Ballroom D
Disclosure(s):
Marissa M. Shams-White, PhD, MSTOM, MS, MPH: No relevant financial relationship(s) with ineligible companies to disclose.
Kirsten A. Herrick, MSc, PhD: No relevant financial relationship(s) with ineligible companies to disclose.
Objectives: Examine associations between ultra-processed food (UPF) intake and cancer risk, and explore associations by UPF subgroups and units of measure.

Methods: Cancer Prevention Study-II Nutrition Cohort participants with no cancer history and with food frequency questionnaire and sociodemographic data in 1999 were included (N=99,044, 57% female). Diet was cumulatively updated in 2003. UPFs (classified via NOVA) were analyzed as residual energy-adjusted g/d (main analysis) and % total g/d, % total kcal/d, and servings (svg)/d (sensitivity analyses). Sex-stratified multivariable Cox proportional hazards models estimated hazard ratios (HR) and 95% confidence intervals (CI) for UPF intake per 100g/d and quintiles (Q) in relation to cancer risk: total, 13 obesity-related (ORC, combined), and select site-specific cancers. UPF subgroups (dairy, cereals, beverages, savory foods, animal proteins, sauces, sweets) were also examined.

Results: Among participants (mean age 69y, median BMI 26 kg/m2, 98% White) over ~1.2 million person-y, 23,086 cancer cases occurred (33% ORC). Median energy-adjusted UPF intake was similar for males and females in Q1 (304 g/d; 9% and 7% total g/d, respectively) but higher for males in Q5 (961 and 800 g/d; 33% and 28% g/d). Top UPF sources (g/d) were beverages (males 41%; females 36%), cereals (19% and 20%), and sweets (12% and 13%). Continuous UPF intake (g/d) was positively associated with risk (HR [CI]) for ORC in males (1.02 [1.00, 1.04]) and colorectal cancer (CRC) in males (1.04 [1.02, 1.06]) and females (1.03 [1.01, 1.05]). Comparing Q5 to Q1, UPF intake was positively associated with CRC in females (1.22 [1.00, 1.50]) and lung cancer in males (1.22 [1.00, 1.50]); total and other site-specific associations were not statistically significant. No specific UPF subgroups drove associations. Defined as %kcal and svg/d, top subgroups were cereals, sauces, and sweets; associations shifted for ORC, CRC, lung, and pancreatic cancer, but no clear patterns emerged.

Conclusions: Greater UPF intake was associated with a modestly higher risk of ORC in males and CRC in both sexes. Higher lung cancer risk in males warrants further study. Variability in results utilizing different UPF measurement units highlight the importance of UPF unit alignment with research hypotheses and careful interpretation of findings.

Funding Sources: American Cancer Society