Epidemiological research on cancer and diet seeks to associate exposure to certain dietary factors with the occurrence of cancer in selected population groups. Studies can be descriptive, reporting the occurrence of cancers in populations, in subgroups of a given population, or in a certain population over time. Observed patterns may be related to particular variables, such as diet, but although the results of such investigations are suggestive, they are certainly not definitive. Their use is especially valuable, however, in identifying populations at risk.
Correlation studies investigate the possible relationships between more-or-less crude exposure data and cancer incidence data in different populations or nations in order to generate new hypotheses. But they are generally of limited value, because national per-capita food intake and cancer incidence data are only approximations and differ from place to place in terminological definition and accuracy. Results should be used mainly as an indication of trends or relationships.
Another type of investigation is the case control study, in which the investigator is able to collect data from individuals instead of groups, and confounding variables can be controlled to a certain extent. In studies of a specific type of cancer, food intake data are collected and are compared to data similarly obtained in matched controls. When known biases and chance can be excluded, associations can be made between exposure and disease. Case control studies are relatively cheap and are of short duration but never prove causal relationships.
Much more expensive, time consuming, and elaborate are prospective cohort studies that focus on individuals in a given (large) population and establish exposure or nutrition data before the occurrence of the disease. Thus, at the time of disease occurrence, exposure data of the patients are compared to data from people not having the disease, thereby providing evidence for a possible causal relationship. Drawbacks in such studies are that the exposure assessment at any given time may not be representative of the individual’s whole life, and this is especially likely in the case of diet.
Finally, intervention studies provide the investigator with the possibility of a random assignment of subjects of a given (often high-risk) population into groups that are treated or fed differently under controlled conditions. In this type of trial, causal relationships can be established.
In investigations focused on the relationship between nutritional factors and cancer, the methods used for determining dietary intake are crucial and difficult. All dietary intake measures share certain limitations, because people vary in their abilities to estimate the amount of something they have eaten. Indeed, sometimes they may even fail to notice or report consumption of certain foods, and they usually possess insufficient knowledge about the ingredients in the foods they consume. A further handicap is that accuracy in dietary recall deteriorates over time. Finally, in the case of cancer, it is very difficult to relate varying diets to the disease because the latter has a long course of development before becoming clinically manifest (10 to 40 years).
The major strength of epidemiological studies, however, is their focus on human populations, thus avoiding the need to extrapolate from other species. They tend to afford an opportunity to examine different effects at different exposure levels, and they are always realistic, in contrast to the high exposure levels usually employed with animals (National Research Council 1982; International Agency for Research on Cancer 1990;Weisburger and Kroes 1994). In this connection, however, it should be noted that in specific investigations into the role of nutrients such as fat, the experimental design in laboratory animals usually faithfully mimics the situation of human populations at high or low risk, thus providing relatively reliable comparative data. But at the same time, as noted, it should also be recognized that the normal food patterns of laboratory animals differ from those of humans.
World History
