POOLED ANALYSES OF PROSPECTIVE DIET AND CANCER STUDIES

点击翻译按钮获取中文摘要

We propose to build on an established collaboration which has successfully pooled data on dietary fat, fiber, and alcohol, and breast cancer risk. We will pool data from nine major prospective studies that have used a comprehensive assessment of diet, to relate dietary factors to the incidence of breast cancer among women, and colorectal and renal cell cancer among men and women. We will examine further the unanticipated increase in risk of breast cancer among women reporting <15% Of energy from fat observed in our first analyses, as well as a decreased risk among women consuming more monounsaturated, rather than saturated or polyunsaturated, fat. Data from other studies suggest that preformed vitamin A, carotenoids, vitamin C and high vegetable intake may be protective against breast cancer. First, we will assess the consistency among the studies of the relations between these dietary factors and breast cancer incidence. We will then determine the best overall estimates of association between these dietary variables and incidence of breast cancer in a combined analysis of all the prospective data including approximately 9,400 cases. For colorectal cancer we will examine associations with red and other meat intake, fiber, folate, calcium, and antioxidant vitamins, in particular, vitamin E. Analyses will be conducted for colon cancer (4,600 cases) and rectal cancer (1,850 cases) separately. We will also pool data to analyze associations between diet and renal cancer. Other studies have suggested that meat intake, animal protein, and coffee and tea consumption may be associated with renal cell cancer. Approximately 1,070 cases of renal cell cancer are expected, which will allow us to examine these associations several years before any individual study would have adequate power. To conduct these analyses, data will be sent from each study center to the Channing Laboratory. Investigators from study centers will meet annually to determine approaches for analyzing of these data, and to discuss the interpretation of findings. Pooling the primary data in this way permits more precise estimates of the strength of these associations than previously possible, allows us to examine more extreme contrasts in intake, and enables us to check for the consistency of interactions with other cancer risk factors across studies to avoid overinterpretation of chance interactions within a single study. Initial work has involved developing a combined data set from the complex nutrient information collected at each center. We believe we are now in an excellent position to capitalize on this work to more rapidly test these dietary hypotheses.