Associations of food processing and dietary risk factors with colorectal cancer incidence in the NIH-AARP Diet and Health Study: A cohort study

BACKGROUND: Ultra-processed foods (UPF) account for >50% of energy intake by US adults. UPF intake can include dietary components linked to lower colorectal cancer (CRC) risk, such as whole grain and dairy products, and others linked to higher CRC risk, such as processed meats, potentially explaining the heterogeneity of evidence for an association between UPF and CRC risk. Our goal was to estimate the associations of UPF and established CRC-related dietary factors, by UPF and non-UPF source, with CRC incidence in a large cohort study. METHODS AND FINDINGS: US adults, aged 50-71, who enrolled in the NIH-AARP Diet and Health Study in 1995-96, self-reported dietary intake using a food frequency questionnaire (FFQ). Median UPF intake was 293 g/1000 kcal/day or 43% of daily energy intake. We assigned disaggregated FFQ items to Nova classification and categorized UPF intake (g/1000 kcal/day) into sex-specific quintiles. Participants who were cancer-free at baseline (N = 461,682) were followed for cancer incidence through 2018, and 10,075 colorectal adenocarcinoma cases were diagnosed. We used multivariable-adjusted Cox proportional hazards regression models, adjusted for demographic, lifestyle, and health-related factors, including age, sex, race/ethnicity, education, physical activity, smoking, and alcohol drinking, to estimate associations for UPF and related dietary components with CRC incidence. UPF intake was not associated with CRC incidence (hazard ratio 5th versus 1st quintile 0.97; 95% confidence interval [0.91-1.03]; ptrend = 0.55). Whole grain, fiber, dairy, and calcium intakes from UPF sources were inversely but meat intake from UPF sources was positively associated with CRC incidence. The main limitations of the study relate to its observational design and include potential residual confounding and exposure misclassification owing to systematic or random error in self-reported dietary data. CONCLUSIONS: Total UPF intake was not associated with incident CRC in this cohort of older, US adults. This may be explained, in part, by opposing associations of some UPF with CRC risk. Our findings support current dietary guidance to consume whole grains, fiber, dairy, and calcium and avoid processed meat for CRC prevention.