To test the hypothesis that increased risk of coronary heart disease, stroke, peripheral vascular disease, and cancer is related to diets high in saturated fat, animal protein, and hydrogenated vegetable oil, and low in polyunsaturated fat, fiber, vitamins A, C, and E, calcium, selenium, and chromium.

BACKGROUND:

The diet-heart hypothesis, that high dietary saturated fat and cholesterol intake increase the risk and high polyunsaturated fat reduces the risk of coronary heart disease in man is supported by ecologic studies, by experiments in rodents and non-human primates, by voluminous literature relating dietary factors to serum lipids, by several secondary prevention trials, and by the Lipid Research Clinics Trial demonstrating a reduction in coronary heart disease among participants assigned to cholestyramine.

Despite the substantial scientific interest and the obvious public health implications of the diet and heart disease issue, relatively few observational cohort or case-control investigations had been published prior to 1985. Although these observational studies were not entirely consistent, taken collectively, they tended to provide important general support for the diet-heart hypothesis. However, due to study design, limited numbers of endpoints, or methods of analysis, many central questions remained unanswered. The most important issue was the quantitative relationship between specific dietary factors and risk of coronary heart disease.

DESIGN NARRATIVE:

In this prospective cohort study, participants completed a mailed general medical and health questionnaire at baseline and an intensively validated semiquantitative food frequency questionnaire (SFFQ). At one year, tissue specimens were collected and catalogued for future nested case-control analyses of coronary heart disease risk in relation to levels of calcium, selenium, and chromium. Follow-up questionnaires to update exposure information and ascertain non-fatal endpoints were mailed at two-year intervals. All reported cases of non-fatal myocardial infarction, stroke, and cancer were documented with hospital records and/or pathology reports. Fatal events were ascertained with the National Death Index and documented. To standardize SFFQ nutrient scores against measurements of absolute intake, two one-week diet records were obtained from a random sample of 150 Boston-area participants.

The study was renewed in 1991, 1997, and in 2003 to continue the follow-up of 51,529 male health professionals. The cohort is followed by questionnaires mailed at two-year intervals to update exposure information and ascertain nonfatal events. Complete dietary assessments are included every four years.

No eligibility criteria