The American Cancer Society Cancer Prevention Study II Nutrition Cohort

The CPS-II Nutrition Cohort is a prospective cohort established in 1992–1993 as a subcohort of the Cancer Prevention Study II to investigate the relation between diet and other lifestyle factors and the risk of cancer, mortality, and survival. Baseline: 162,408 cancer-free participants (77,048 men; 85,360 women) aged 50–74 years completed a mailed questionnaire in 1992–1993. Follow-up: biennial questionnaires update exposure information and identify new cancer cases; response rates were about 90% for follow-up questionnaires in 1997 and 1999. Cancer outcomes are verified through medical records, registry linkage, or death certificates. Biospecimens: blood samples collected from a subcohort of 40,000; DNA from buccal cells among approximately 50,000; LifeLink linkage and centralized storage; A second wave of DNA collection and long-term storage planned. The baseline dietary assessment used a semi-quantitative food frequency questionnaire; repeated diet-related questionnaires were administered in 1997 and 1999 to update exposure data. The study design emphasizes collecting detailed dietary and lifestyle information to evaluate associations with cancer occurrence and mortality.

The Nutrition Cohort provides a large, ongoing resource for prospective analyses of how diet and lifestyle relate to cancer occurrence, mortality, and survival. Baseline characteristics of cancer-free participants indicate an older, predominantly White population with varying education and BMI, and notable use of vitamin supplements. Follow-up procedures aim to identify incident cancers and deaths, enabling future analyses of risk factors. The authors conclude the CPS-II Nutrition Cohort will be particularly valuable for studying associations between obesity/weight change, physical activity across life, vitamin supplement use, exogenous hormone use, other medications (such as aspirin and NSAIDs), and cancer screening modalities in relation to cancer outcomes. As an observational cohort, findings will reflect associations rather than causal effects.

Not a random sample; generalizability to the broader population may be limited. Dietary assessment relied on a baseline brief FFQ that underestimated total energy and many nutrients; validity for several micronutrients was modest (e.g., some micronutrients showed low correlations with reference measures), though reproducibility for many measures was high (>0.7). Data are largely self-reported, introducing potential measurement error and misclassification; follow-up and updates are ongoing to capture incident cancers and deaths.