Meat Consumption in Relation to Colorectal Cancer Incidence in Anatomical Subsites in the National Institutes of Health-AARP Diet and Health Study.

Background: In the literature, the consumption of red and processed meat is associated with colorectal cancer (CRC) incidence. It still remains unclear whether the associations of different types of meat intake with CRC vary by subsite of the colorectum. Objectives: We investigated the association of red, white, and processed meat intakes with CRC incidence across anatomical subsites of the colon and rectum. Methods: We analyzed data from 454,505 cancer-free adults in the National Institutes of Health-AARP Diet and Health Study cohort. During 6,896,619 person-years of follow-up, 9,461 CRC cases were diagnosed. Type and amount of meat consumed were assessed via food frequency questionnaire. Cox proportional hazards models were used to estimate associations of meat intake, by type, with CRC for an estimated fixed-unit measure of 50 g/1000 kcal, applying both a substitution and addition approach. The substitution model estimated the effect of replacing a given meat type with another type while keeping total meat constant. The addition model evaluated the independent effects of each meat type, keeping all other individual meat types, but not total meat, constant. Results: Replacing 50 g/1000 kcal of white meat with red meat was associated with 21% higher CRC incidence. Higher incidence was also observed for proximal colon (16%), distal colon (25%), and rectal (30%) cancers, driven by positive associations in the cecum (19%), transverse colon (32%), sigmoid colon (28%), and rectosigmoid junction (65%). Every 50 g/1000 kcal of red meat intake was independently associated with higher CRC incidence in the proximal colon (11%), distal colon (22%), and rectum (19%). However, white meat was only associated with a lower incidence when replacing red meat. Replacing unprocessed with processed meat intake was associated with higher CRC incidence overall (15%), namely in the proximal colon (18%) and the cecum (33%). Processed meat intake was also independently associated with higher CRC incidence. Conclusions: We found that replacing white meat with red meat was associated with higher CRC incidence. The lack of independent inverse associations for white meat intake indicates that lower CRC incidence is derived from lower red meat intake.