Skip to main content
         

Vol. 10 Issue 4, Fall 2005

Research Commentary: Preschool Diet and Adult Risk of Breast Cancer: Considerations of Fundamental Issues in Carcinogenesis and Acrylamide Exposure
The Ribbon 

Barbour S. Warren, Ph.D.
Research Associate, BCERF Sprecher Institute for Comparative Cancer Research

Carol M. Devine, Ph.D.
Associate Professor, Division of Nutritional Sciences

A great deal of attention has recently been given to a study that examined the association of adult breast cancer risk with various foods in the diets of women during their preschool years (ages three to five years) (Michels et al., 2005). This study found a moderate increase in breast cancer risk (27%) for each weekly serving of french fries in the preschool diets of the women being studied. This is a preliminary report that will require further verification before recommendations can be made. Nevertheless, it has a good scientific foundation and its review provides fertile ground for the discussion of a number of fundamental issues important to breast cancer. I will build the review of this study around the discussion of these issues.

1. There are likely to be critical periods of susceptibility to the initiation of breast cancer. This study examines the association of 30 different foods eaten during women's preschool years with their adult breast cancer risk. Previous studies have examined women's diets during puberty but this study is the first to examine diet at this early life period. This time may well be an important period for breast cancer susceptibility. Studies of breast cancer risk in women exposed to radiation from the Hiroshima and Nagasaki bombings reported the greatest risk for women exposed at ages up to five years (Tokunaga et al., 1987). Examinations of potential breast cancer risk factors during this time period of women's life are important.

2. Cancer takes decades to develop following the initial exposures to cancer causing agents. Study of these initial exposures, which occurred decades earlier, is difficult to carry out but nonetheless critical. This study is part a group of studies known as the Nurses' Mothers' Health Study. These studies gathered information from the mothers of nurses being followed in two large cohort studies. This case-control study evaluated the breast cancer risk associated with nurses' preschool diets as remembered by their mothers. The design of the study, built around gathering information about exposures far in the past, has validity as it fits well with the biological time course of the cancer formation. However, this design also has several important weaknesses. The results are dependent on how well the mothers of these nurses remember their daughter's diet at ages three to five, a time 40 to 80 years earlier. In addition, there is concern that the memory of mothers whose daughters developed breast cancer might be selectively affected by their daughter's disease, leading to inaccurate reporting of certain foods. For example, a mother might take on some responsibility for her daughter's breast cancer and report lower values for healthy food and higher values for unhealthy foods. Because of these weaknesses, it will be important to see the results of other studies that examine diet and breast cancer risk during this life period.

3. An association found in an epidemiological study does not necessarily indicate a cause and effect relationship. The association of weekly servings of french fries at ages three to five and adult breast cancer risk may indicate that the french fries, themselves, affect breast cancer risk. On the other hand, it may also be that the association merely reflects other behaviors associated with breast cancer risk. Girls who ate more french fries may have had poorer overall diets and it is this, rather then the french fry consumption, which contributed to the moderate increase in breast cancer risk.

These limitations are well understood in epidemiological studies and the evidence necessary for a cause and effect relationship has been defined. This topic was discussed in an earlier issue of The Ribbon (see "Evaluation of the Evidence: What Does It Take to Show a Cause and Effect Relationship Between Carcinogen Exposure and Cancer Formation?" The Ribbon, Volume 7, Number 4). A key piece of such evidence is the presence of a biologically plausible explanation, as is described below.

4. A cause and effect relationship between an exposure and cancer formation must be biologically plausible. The authors of this study postulate that since they did not find an association with potatoes, the preparation of the french fries may be important. The typical use of frying oils high in saturated fats and trans-fatty acids was noted, but recent cohort studies have downplayed such an effect. What has gotten more attention has been the presence of high levels of acrylamide in french fries. The International Agency for Research on Cancer (IARC) has reviewed the scientific literature on acrylamide. This agency has classified acrylamide as probably carcinogenic to humans based on its carcinogenicity in animals and its genetic toxicity.

Exposure to acrylamide was, until recently, thought to occur only in industrial environments but this idea dramatically changed following a Swedish contamination incident in 1997 (Reynolds, 2002). Acrylamide is present in grouts used to make watertight seals and workers were exposed to high levels of acrylamide in a Swedish tunnel construction project. Studies to document the levels of acrylamide exposure examined and compared markers of acrylamide exposure in these workers to those of "unexposed" members of the general population. Surprisingly, substantial levels of the markers of acrylamide exposure were seen in both groups. The source of the levels in the "unexposed" individuals was ultimately traced to a dietary exposure. Acrylamide can be generated when carbohydrate foods are cooked at high temperatures, and substantial levels have been found in numerous foods (Dybing and Sanner, 2003; Jagerstad and Skog, 2005; Ruden, 2004). The highest levels of acrylamide exposure are thought to arise from french fries, potato chips, coffee and some cookies and breakfast cereals. But levels can vary considerably depending on the source of the food, and increase with the extent of browning during cooking.

Epidemiological studies examining acrylamide exposure and various types of cancer have not consistently detected an associated cancer risk (Rice, 2005). However, many of these studies have re-evaluated existing data in light of reported acrylamide levels in various foods. This approach is limited as it does not take into account the differences in acrylamide levels that can occur in foods depending on their source and preparation. These, as well as other design shortcomings, have raised questions about the validity of these studies. Studies of men exposed to acrylamide during its industrial production of acrylamide have also been inconsistent (Erdreich and Friedman, 2004; Rice, 2005).

Currently, there is little contention that acrylamide presents a potential human carcinogenic hazard. But the size of the associated human cancer risk is not agreed upon. Some regulatory agencies such as the World Health Organization do feel that the methods to predict human cancer risk are of sufficient reliability to be applied. The US Environmental Protection Agency does not agree with this assessment of these methods and has estimated a lifetime cancer risk of 4.5 per 1,000 at high dietary levels of acrylamide exposure (i.e., if 10,000 people were exposed it would be expected that 45 of these exposed individuals would get cancer).

The California Attorney General has also weighed into this issue. California's proposition 65 requires that the state's Office of Environmental Health Hazard Assessment (OEHHA) maintain a list of substances that cause cancer, birth defects or other reproductive harm. More importantly, it also requires that if a business exposes its customers to any of these substances, it must provide them with a clear and reasonable warning. Acrylamide is on this OEHHA list and the current California Attorney General has filed a suit against nine manufacturers of potato chips and French fries seeking a warning on these products indicating that they contain a cancer-causing chemical.

5. If everyone is exposed to a cancer-causing agent, case-control and cohort studies will fail to detect the risk and will only identify markers of susceptibility. Case-control and cohort studies are designed to determine the relative risk of cancer linked to some agent or factor. Relative risk is determined by comparing the levels of absolute cancer risk of two groups, one group exposed to the agent being studied and another group not exposed to the agent. If this comparison is between groups that are both exposed to the agent being studied, the risk of each group will be the same and no risk will be found. However, what can be detected in this situation is risk resulting from susceptibility to the disease. For example, an epidemiological study of smoking and lung cancer risk could not be carried out if all of its subjects smoked; the groups being compared would both be exposed and their lung cancer risks would be close to the same. However, an increase in risk would be detected if an examination were carried out in which one of the groups contained a large number of people susceptible to lung cancer. Asbestos increases the susceptibility to smoking induced lung cancer (Williams and Sandler, 2001). So if one of the examined groups contained many asbestos exposed auto brake repair technicians, this group would have a higher lung cancer risk. The increased risk would arise due to the higher susceptibility of the asbestosis exposed individuals.

Could a similar case be operating with french fries in these girls' diets? It is likely that most of the population is exposed to acrylamide, as it is present in many commonly eaten foods. As in the above example, epidemiological studies would have difficulty detecting cancer risk unless they examined a group with many susceptible individuals, like the brake repair technicians. It is possible that age, particularly ages three to five years, is such a marker of susceptibility for breast cancer risk and dietary acrylamide exposure. At this point this is no more than a theoretical suggestion and much more study is needed before it can begin to be verified.

Where does this leave women, and especially mothers, now? Regardless of their potential cancer association, french fries and potato chips are high in fat and sodium and for these reasons alone should be eaten infrequently. This is no less true for children, as atherosclerosis has been described as a nutritional disease of childhood. Further, fried potatoes have been found to be the source for three quarters or more of the acrylamide in the diets of people in Sweden, Holland and Switzerland (Pelucchi et al., 2005).

There are currently no recommendations regarding other foods containing acrylamide. This is an area of high interest and the result of more examinations should become available in the near future.

Dybing, E., and Sanner, T. (2003). Risk assessment of acrylamide in foods. Toxicological Science 75, 7-15.

Erdreich, L. S., and Friedman, M. A. (2004). Epidemiologic evidence for assessing the carcinogenicity of acrylamide. Regulatory Toxicology and Pharmacology 39, 150-157.

Jagerstad, M., and Skog, K. (2005). Genotoxicity of heat-processed foods. Mutation Research 574, 156-172.

Michels, K. B., Rosner, B. A., Chumlea, W. C., Colditz, G. A., and Willett, W. C. (2005). Preschool diet and adult risk of breast cancer. International Journal of Cancer Aug 10; [Epub ahead of print].

Pelucchi, C., Galeone, C., Levi, F., Negri, E., Franceschi, S., Talamini, R., Bosetti, C., Giacosa, A., and La Vecchia, C. (2005). Dietary acrylamide and human cancer. International Journal of Cancer 7, 7.

Reynolds, T. (2002). Acrylamide and cancer: tunnel leak in Sweden prompted studies. Journal of the National Cancer Institute 94, 876-878.

Rice, J. M. (2005). The carcinogenicity of acrylamide. Mutation Research 580, 3-20.

Ruden, C. (2004). Acrylamide and cancer risk-expert risk assessments and the public debate. Food Chemistry and Toxicology 42, 335-349.

Tokunaga, M., Land, C. E., and Yamamoto, T. (1987). Incidence of female breast cancer among atomic bomb survivors, Hiroshima and Nagasaki, 1950-1980. Radiation Research 112, 243-272.

Williams, M. D., and Sandler, A. B. (2001). The epidemiology of lung cancer. Cancer Treatment and Research 105, 31-52.

Back to the top

Last Update 12.02.05