Prospective investigation of ultra-processed food intake with risk of cancer and mortality in the Prostate Lung, Colorectal and Ovarian Cancer Screening Trial cohort
In the NutriNet-Santé Study, a prospective cohort of 44,551 French adults, a 10% increase in the proportion of UPF was associated with a 14% higher risk of all-cause mortality after a median of 7.1 years of follow-up [4]. More recently, a UK Biobank study found that a 10% increase in UPF intake, out of in the total diet (g/day), was associated with a 6% higher risk of cancer-related mortality [5]. Studies of overall cancer incidence have found no association while cancer-specific analyses have been inconsistent [6]. Assessing UPF intake in observational studies is challenging, and it is likely that non-differential exposure misclassification biases risk estimates toward the null. This is due, in part, to the fact that food frequency questionnaires (FFQ) as well as short-form questionnaires and recalls, often lack sufficient detail such as ingredient level information to reliably classify all food items according to the extent and purpose of industrial food processing, as per Nova. Recently, we developed a method for disaggregating FFQ food items into food codes and then linking them to a Nova database developed using 24-hour dietary recall data in the National Health and Nutrition Examination Survey. We used this method to estimate UPF intake using the 124-item NIH-AARP FFQ, an early version of the DHQ, and then validated it using two non-consecutive 24-hour dietary recalls as the reference [7]. We will apply our method to the PLCO DHQ data to derive estimated intake based on Nova.
1. Steele, E.M., et al., Identifying and Estimating Ultra-processed Food Intake in the US NHANES According to the Nova Classification System of Food Processing. J Nutr, 2023. 153(1): p. 225-241.
2. Sung, H., et al., Emerging cancer trends among young adults in the USA: analysis of a population-based cancer registry. Lancet Public Health, 2019. 4(3): p. e137-e147.
3. Stuckler, D., et al., Manufacturing epidemics: the role of global producers in increased consumption of unhealthy commodities including processed foods, alcohol, and tobacco. PLoS Med, 2012. 9(6): p. e1001235.
4. Schnabel, L., et al., Association Between Ultraprocessed Food Consumption and Risk of Mortality Among Middle-aged Adults in France. JAMA Intern Med, 2019. 179(4): p. 490-498.
5. Chang, K., et al., Ultra-processed food consumption, cancer risk and cancer mortality: a large-scale prospective analysis within the UK Biobank. EClinicalMedicine, 2023. 56: p. 101840.
6. Lane, M.M., et al., Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses. BMJ, 2024. 384: p. e077310.
7. Loftfield, E., et al., Performance of a Food Frequency Questionnaire for Estimating Ultraprocessed Food Intake According to the Nova Classification System in the United States NIH-American Association of Retired Persons (AARP) Diet and Health Study. J Nutr, 2025.
1) We aim to estimate dietary intake in the PLCO cohort, using DHQ data, according to degree of food processing as defined by the Nova classification system. We will estimate UPF exposure based on gram weight and proportion of energy intake for Nova groups 1-4 and for Nova subgroups. We will also estimate the contribution of Nova groups to intake of nutrients (e.g., mg/day of calcium) and foods groups (e.g., servings/day of dairy) to explore potential heterogeneity of associations by degree of food processing with cancer and mortality risk.
2) We will use Cox Proportional Hazards regression models to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for UPF intake with cancer and mortality risk. Models will be adjusted for potential confounders including age, BMI, race/ethnicity, alcohol consumption, physical activity, self-rated health, detailed smoking status, and dietary factors, which will also be explored as potential mediators. We will consider substitution analyses in which total intake is held constant to estimate the effect of replacing UPF with minimally or less precessed foods and derived nutrients. We will conduct stratified analyses to explore potential differences by sex, age, BMI, smoking status, self-rated health, and race/ethnicity. We will conduct sensitivity analyses (e.g., excluding those with <2 years of follow-up).
Lisa Kahle, IMS