In a prospective study of 10 011 men with 815 prostate cancer cases, despite plausible biological mechanisms, neither increasing intake levels of dairy products nor calcium from dairy products (P trend; 0.23 and 0.64, respectively), or calcium supplements was associated with prostate cancer risk (relative risk, 1.05; 95% confidence interval, 0.84 -1.31). (Gronberg, 2003). Although its incidence is increasing steadily in almost all countries, few modifiable predictors for the disease have been identified (Gronberg, 2003). Well-established risk factors, such as age, family history, race, and country of residence, are not amenable to modification and have limited utility as primary prevention strategies.Calcium intake, a modifiable dietary factor, has recently been proposed as a risk factor for prostate cancer. Evidence from in vitro and animal studies suggests that risk may be increased by higher levels of calcium, since these suppress 1,25-dihydroxyvitamin D 3 , the most active form of vitamin D 3 , that inhibits the proliferation of prostate cancer cells and promotes their differentiation (Schwartz et al, 1995(Schwartz et al, , 1997Giovannucci, 1998). However, relevant studies have yielded inconsistent results (Gao et al, 2005), although few studies have considered intake from calcium supplementation, that can provide higher doses than dietary intake. Given calcium's importance in the prevention of diseases such as osteoporosis, we conducted a study of the question in a large cohort of men.
MATERIALS AND METHODS
SubjectsThe Harvard Alumni Health Study, initiated in the 1960s, is an ongoing cohort study of men who entered Harvard University, United States, as undergraduates between 1916 and 1950. Health information is obtained from subjects via questionnaires mailed at periodic intervals. The present study utilises data from a questionnaire mailed in 1988 that included detailed information on diet, and eligible subjects were 12 805 men who responded. We excluded men reporting cancer at baseline (n ¼ 1731) and those not providing information on dairy product intake (n ¼ 60). Of the remaining 11 014 men, we successfully followed 10 011 (90.9%) for the development of prostate cancer; these men represent the subjects for the present analysis.
Assessment of calcium intake and other factorsThe 1988 questionnaire assessed diet using an abbreviated version (23 food items) of a validated semiquantitative food frequency questionnaire (Willett et al, 1985(Willett et al, , 1987. Men indicated their usual daily intake of seven different dairy products (whole milk, low fat milk, cream, ice cream, yogurt, cheese and butter) using pre-specified responses ranging from 'almost never' to 'six or more times per day.' On the original, longer version of the questionnaire, 13 dairy items were assessed. However, we believe our assessment of calcium intake using the seven items s reasonably valid. While we did not directly validate our 7-item questionnaire against the longer questionnaire, a previous study that assessed fewer items -mil...