Objective: Results associating dairy and Ca intakes with colorectal cancer (CRC) risk have been mixed. Most previous analyses have suffered from confounding between dairy and Ca intakes. We examined independent associations between these variables, also dairy foods, and CRC incidence in a population with a large range of dairy intakes.

Design: Adventist Health Study-2 is a cohort study where subjects were enrolled 2002-2007. Proportional hazard regression analyses were performed to estimate hazard ratios (HR). Regression calibration was used to correct for dietary measurement error.

Setting: The population lived in all states of the USA.

Subjects: There were 77712 analytic subjects, all of whom were Seventh-day Adventists. Much of their dietary Ca came from non-dairy sources.

Subjects: During a mean follow-up of 7·8 years, 380 incident colon cancer and 111 incident rectal cancer cases were observed.

Results: Comparing extreme quintiles of intake in measurement error-corrected analyses, dairy intake (HR=0·31; 95 % CI 0·09, 0·88), independent of total Ca, was inversely related with risk of rectal cancer but gave little indication of association with colon cancer. However, total Ca intake (independent of dairy) was associated with risk of colon cancer (HR=0·55; 95 % CI 0·28, 0·98) and there was little indication of association with rectal cancer. Traditional regression analyses and associations with macronutrients from dairy generally supported these results. Milk intake was also negatively associated with CRC (HR=0·63; 95 % CI 0·43, 0·89).

Conclusions: Dairy intake may decrease the risk of rectal cancer, and Ca may reduce risk of colon cancer and CRC.