Key points are not available for this paper at this time.
The European Prospective Investigation of Cancer (EPIC) was initiated in 1989 in order to define more clearly the relationship between diet and the risk of developing cancer at a range of sites (Riboli, 1992). The hypothesis that diet might be responsible for a substantial proportion of total cancer incidence was widely publicized by a report in the early 1980s (Doll, 1981) which suggested that 30% of cancer in the USA might be preventable by dietary modification. However, the uncertainty bounds around the estimate of 30% were acknowledged to be wide, and no association between a specific dietary constituent and risk for a specific cancer type was considered conclusively established. Since then, evidence has continued to accrue supporting the general association between diet and cancer, but two recently published comprehensive reviews only found sufficient evidence for a few specific relationships (World Cancer Research Fund, 1997; Department of Health. 1998). The most convincing evidence is for a protective role for fruit and vegetables for a number of cancers, including oesophagus, stomach, large bowel, lung and breast. It is not clear, however, that the same food constituents are involved for each cancer, nor even that the same type of fruit and vegetable are implicated. The rationale for EPIC was that in order to generate data which improved the definition of diet-cancer associations, studies had to be prospective, large-scale and based on populations with wide variation both in diet and in cancer incidence. In addition, studies had to use improved methods for dietary assessment including biological markers. EPIC comprises cohorts established in France, Germany, Greece, Italy, The Netherlands, Spain and the UK, with associated cohorts in Denmark, Norway and Sweden. Recruitment of these cohorts with baseline data and biological samples is now almost complete (as shown in Table 1), achieving the target set earlier (Riboli, 1992). As an example of the range in cancer rates seen across the populations participating in EPIC, Figure 1 shows mortality rates for breast cancer. The study in Norfolk was initially planned as a diet and cancer cohort and later joined as a collaborative component cohort of EPIC. It has now, however, broadened its scope to include end points other than cancer, and exposures other than diet. The intention now is to investigate the majority of causes of disability and death in middle and later life, and to include additional lifestyle exposures, notably exercise, physical activity and psychosocial variables.
Day et al. (Thu,) studied this question.