Dietary Fat and Breast Cancer

The statistics are alarming. One out of eight women will develop breast cancer. As the incidence of breast cancer grows every year, so does the interest on whether a high fat diet is linked somehow to breast cancer.

A high fat intake has been associated with diseases such as cardiovascular disease, diabetes, and colon cancer. But the debate continues about whether this holds true for breast cancer. The few studies conducted have been conflicting. Several reports show no correlation between eating a high fat diet in the development and progression of breast cancer, while other studies say there is a correlation, although very moderate.

PERSONAL VIEWPOINTS

I chose this topic because as a woman eating a high fat diet, I wanted to know what the risks were of contracting this disease. From the journals and magazine articles that I have researched, I found no definite correlation between a high fat diet and breast cancer due to research design, sampling procedures, validity of measures, types of analysis, and the adequacy of the controls.

EPIDEMIOLOGIC STUDIES

There are three types of epidemiologic studies that have provided evidence on an association between fat intake and breast cancer. For the purpose of this paper and the research that I have found, only two out of the three epidemiologic studies will be discussed.

RESEARCH DESIGN: COHORT STUDIES

ÒThe first type of epidemiologic evidence comes from cohort studies in which diets measured at baseline are compared with subsequent risk of breast cancer in a large group of women.Ó (Howe, 1992, 2080) The NursesÕ Health Study is a model of this type of research design.

PROSPECTIVE COHORT STUDY: NURSESÕ HEALTH STUDY

To test the hypothesis that dietary fat increases the risk of breast cancer, a prospective cohort study using a self-administered food frequency questionnaire was conducted. The participants included 89,494 nurses who were 34 through 59 years of age in 1980 and were followed up for 8 years. (Williett, Hunter, Stampfer, Colditz, Manson, Spiegelman, Rosner, Hennekens, and Speizer, 1992, 2037) This study demonstrates the complete lack of any positive correlation between fat intake and breast cancer.

ANALYZING THE RESEARCH DESIGN

Before accepting the results of this study found in the Journal of the American Medical Association , I first have to question how this research was designed. I found it possible contained both selection and recall bias. I think they chose the wrong subjects to conduct this study because nurses are knowledgeable health professionals that have the necessary tools to modify their diet to prevent breast cancer or to modify their diet just long enough for this study to be completed. If the researchers wanted to test their theory, the subjects chosen should have been women with high fat diets only and avoid those who are in the health professions.

I also feel that a questionnaire is not a reliable method to base a study on. Subjects can easily withhold information or modify what they write. But I also realize that the researchers cannot be with the participants 24 hours a day and so this method is the most cost effective while sacrificing some credibility in its findings.

RESEARCH DESIGN: CASE-CONTROL STUDIES

ÒThe second group of epidemiological studies consists of case-control studies in which pre-diagnosis diets of women with breast cancer, ascertained following diagnosis, are compared with diets of women without breast cancer.Ó (Howe, 1992, 2080) To compare prospective and retrospective data, a case-control study of breast cancer among NursesÕ Health Study participants was conducted and its findings written in the American Journal of Epidemiology .

CASE-CONTROL STUDY: BREAST CANCER PATIENTS AMONG NURSESÕ HEALTH STUDY PARTICIPANTS AND THEIR CONTROLS

Among those who completed the food frequency questionnaire and were free of cancer in 1986, 398 were diagnosed with breast cancer during two years of follow-up. These cases were matched with 798 age-matched controls. (Giovannucci, Stampfer, Colditz, Manson, Rosner, Lognecker, Speizer, and Willett, 1993, 502) When another food frequency questionnaire was sent to the subjects in this study in 1989 inquiring about their diet in 1985, 300 cases and 602 controls responded.

CONCLUSIONS ON THE CASE-CONTROL STUDY

ÒThe age-adjusted analysis using the prospective (1986) questionnaire demonstrated no appreciable associations between breast cancer incidence and intakes of total fat. The age-adjusted analysis using the retrospective (1989) questionnaire suggested positive associations between breast cancer and intakes of fat.Ó (Giovannucci, 1993, 502: et. al.) These results suggest that there are biased associations in case-control studies between total fat intake and risk of breast cancer. The biases in this research design are similar to that in the NursesÕ Health Study that I mentioned before. The only exception is that this research design contains a greater percentage of recall bias because participants in 1989 had to remember what their diets were in 1985.

CASE-CONTROL STUDY AMONG JAPANESE WOMEN

In a second case-control study, 212 Japanese women with breast cancer were interviewed one to three months after surgical operation for breast cancer. Information was obtained about their reproductive histories, dietary histories, and socioeconomic status. The interview consisted of item by item recall in a typical week before the onset of the disease. Clinical information was obtained through medical records for each patient as well. The patients were followed up 5 and 10 years after the initial start of the study. The investigation showed no support for the hypothesis that a high fat diet is a survival determinant for breast cancer patients.

CONCLUSIONS ON THE JAPANESE WOMEN'S STUDY

As with the previous reports, the investigation had selection bias in choosing Japanese women who have low fat diets as compared with American women with higher fat intakes. To test the theory that a high fat diet is related to the patientsÕ prognosis of breast cancer, a society with a high fat intake should have been selected.

CASE-CONTROL STUDY: WOMEN WITH A HISTORY OF BREAST TUMORS

In a third study written in Science News magazine, 220 women who had surgery to remove a malignant tumor were studied. They were interviewed by a nutritionist about their diet one year prior to diagnosis. All were followed for at least four years.

CONCLUSIONS BASED ON THE STUDY IN SCIENCE NEWS MAGAZINE

'A statistical analysis revealed that fat intake did influence the outcome of breast cancer treatment, but only in women who had tumors with lots of estrogen receptors.' (Fackelmann, 1993, 23) The research design for this study, like other studies, contained recall bias, but I think the proper group was selected.

RESEARCH DESIGN: MAGAZINE ARITCLES

In three out of the five magazine articles that I have found, they all relate to the Nurses' Health Study. Little or no information was given on how the study was designed and what sampling procedures were used.

STRUCTURE OF LAY PUBLICATIONS

Magazine articles are organized in such a manner that they are supposed to be easy to read for the average individual. In the process of doing this, valuable information is lost and the consumer is left only with the opinion of the author.

MS MAGAZINE: NO STATISTICAL DATA PROVIDED

In a fourth magazine article in Ms. magazine, the author tells us there is a link between fat intake and breast cancer without studies or statistical data. (National WomensÕ health Network, 1993, 47) The article proceeds by telling the consumer what dink of diet is conducive in preventing breast cancer.

PERSONAL VIEWPOINTS: STRUCTURE OF MAGAZINE ARTICLES

Most people believe what they read. That is why it is so important to think critically as we read and not to accept everything as the truth. I feel if the magazines (for simplistic reasons) fail to mention important studies or statistical data, they should at least provide references where we can research the information for ourselves if we are interested in the subject.

The magazine articles that I have just mentioned with the exception of Science News magazine, will no longer be discussed throughout the rest of this paper because they either refer to the journal articles such as the NursesÕ Health Study or contain no studies at all.

DEFINITIONS

As defined by Russell T. Hurlburt, author of Comprehending Behavioral Statistics , a random sample is a sample chosen so that each member of a population has an equal chance of being included in the sample. A representative sample is a sample of a population that reflects that characteristics of the parent population.

REPRESENTATIVE SAMPLING: NURSESÕ HEALTH STUDY

In the NursesÕ Health Study and the case-control study conducted on several members from the nursesÕ study, there was no mentioning on the sampling procedure. But from reading the journal articles, I would assume that the nurses are a representative sample because they were specifically chosen to represent the population of women. As in a random sample, chance did not play a role. All the subjects chosen were nurses from eleven states.

REPRESENTATIVE SAMPLE: JAPANESE WOMENSÕ STUDY

The study of Japanese women, written in the International Journal of Nutrition and Cancer , was also a representative sample. They represent the population of women who underwent surgical operation for breast cancer. The 212 women were specifically chosen with these characteristics so that the case-control study could test its hypothesis.

REPRESENTATIVE SAMPLE: STUDY OF WOMEN WITH MALIGNANT TUMORS

The investigative report in Science News magazine was a representative sample as well. The 220 women studied all had surgery to remove a malignant breast tumor. They were chosen for the same characteristics as the Japanese womenÕs study.

PURPOSE OF SAMPLING

The overall purpose of sampling is to estimate the population value. But sometimes, as I feel in the case of the NursesÕ Health Study, the sample population does not truly represent the total population of women because of their health profession. That is why when any research is being conducted, extreme care must be taken so that bias does not creep into the selection procedure.

VALIDITY OF MEASURES: INCIDENCE AND MORTALITY RATES

The validity of the measures is extremely important when accepting a positive or negative correlation. Due to international differences of incidence and mortality rates in breast cancer, it is very difficult to generalize the study, for example on Japanese women, to the population of women in the world, particularly in the United States. Japanese women have a greater survival rate possible due to the extent or spread of the disease or the kinds of therapy used. Thus, factors influencing the occurrence of breast cancer might also affect the course of the disease. (Kyogoku, Hirohata, Nomura, Shigematsu, Takeshita, and Hirohata, 1992, 272)

VALIDITY OF MEASURES: FOOD CONSUMPTION

There are also international studies on food consumption. Japanese people consume about 23% of calories as fat. (Kyogoku 1992, 275; et al.) On the other hand, Americans consume about 36% of calories as fat. (Napier 1993, 91) Nutritionists recommend restricting fat to 30%. This wide range between Japanese and American societies proves that a study may be valid for one country but may not hold valid for another.

VALIDITY OF MEASURES FOR THE TOTAL POPULATION

Besides a study being valid on an international level, the study needs to be valid for the population in questions. Research design, sampling procedure, statistical analysis, and the adequacy of the controls are all important measures proving the validity of the research.

STATISTICAL ANALYSIS

I have already discussed research design and sampling procedure. Next, I will discuss types of analysis. The types of analysis conducted on studies may vary greatly. Statistical analysis are figures used to describe the findings of a sample population. For example, the mean intake of fat may be used to describe the average amount of saturated, monosaturated, or polysaturated fat consumed. Statistical analysis can be extremely confusing to understand especially for the average individual.

PROSPECTIVE AND RETROSPECTIVE ANALYSIS

Another type of analysis is prospective and retrospective analysis. Statistics are also incorporated but a prospective analysis deals mainly with the situation at hand. An example would be a study conducted on the NursesÕ Health Study of women who were diagnosed with breast cancer. They were given a prospective analysis in 1986 of their diet in that year. A retrospective analysis would be another food frequency questionnaire given to this sample and having them recall their dietary habits four years ago. the retrospective analysis is subject to greater recall bias than the prospective analysis.

ADEQUACY OF THE CONTROLS

Finally, the adequacy of the controls, can also determine any correlation between a high fat intake and breast cancer. Sometimes it is difficult to link the two because of an external factor. In the study of 220 women who had surgery to remove a malignant breast tumor, a boost in the concentrations of estrogen in the blood may be the true cause of breast cancer versus a high fat intake only. However, high fat diets may boost these concentrations. So it is not adequate to say that dietary fat leads to breast cancer.

Other measures must be controlled as well. The control group should match the sample population as much as possible (age, height, weight, sex...) in order to receive more accurate measurements. As mentioned before, the construction of the research must be carefully examined before conducting the experiment. This includes the areas I have already discussed: research design, sampling procedure, validity of measures, types of analysis, and adequacy of the controls.

OVERALL CONCLUSIONS: STUDIES PROVE INCONCLUSIVE

The debate between those who believe a high dietary fat intake is associated with breast cancer and those who believe there is no correlation will likely continue. The studies conducted are inconclusive and more research needs to be done. However, the general consensus in the journals and magazine articles suggest limiting the amount of fat because of the high risk of getting other diseases such as cardiovascular disease and colon cancer.

LIBRARY RESEARCH

My search for a topic began in Hamilton Library. I researched for possible topics in the Reader's Guide to Periodical Literature . After much elimination, I decided on the topic of dietary fat and breast cancer.

I found five current magazine articles on high fat intake and breast cancer. I then decided that I would find these articles and check if it was pertinent to my topic. It took me a few hours from the time I started to find all the articles. But once I found all five magazine articles, it was well worth the time spent because they were very useful to my topic.

Since I was already at the library, I decided to do more research. All the CD ROM terminals were in use, so I went to one of the many computer terminals and researched in UnCover and ERIC. There I found two journal articles that would assist with the controversy between a high fat diet and breast cancer. Both journals were found on the third floor. After making xerox copies, I returned to my home and began reading all the articles.

Once I had everything, I realized that I needed to find more information because three out of the five articles referred to this one study that I did not have.

About two weeks after my initial visit to Hamilton Library, I returned to hopefully find more journal articles and to find the study that three out of the five magazines referred to. Once again I researched in UnCover and ERIC. This time, though, I knew exactly what I wanted and needed. I not only found the study that I wanted, but I also found another journal article that dealt with the controversy. As my research concluded, I found a total of five magazine and four journal articles on the controversy between a high fat diet and the possible link to breast cancer.

REFERENCE LIST

Fackelmann, K. A. 1993. Dietary fat predicts breast cancer rouse. Science News, 9 January, 23.

Giovannucci, Edward, Meir J. Stampfer, Graham A. Colditz, JoAnn E. Manson, Bernard A. Rosner, Matt Lognecker, Frank E. Speizer and Walter C. Willett. 1993. A comparison of prospective and retrospective assessments of diet in the study of breast cancer. American Journal of epidemiology 137 (March):512- 510.

Howe, Geoffrey R. 1992. High fat diets and breast cancer risk: The epidemiologic evidence. Journal of the American Medical Association 268 (October):2080-2081.

Kyogoku, Shinji, Tomio Hirohata, Yasuo Nomura, Takao Shigematsu, Setsuko Takeshita and Itsuyo Hirohata. 1992. Diet and prognosis of breast cancer. International Journal of Nutrition and Cancer 17 (January):271-276.

Napier, Kristine. 1993. Fat and breast cancer. American Health March, 91.

National Women's Health Network. 1993. The diet that may save your life. Ms., May/June, 47-51.

Willett, Walter C., David J. Hunter, Meir J. Stampfer, Graham Colditz, JoAnn E. Manson, Donna Spielgelmann, Bernard Rosner, Charles H. Hennekens and Frank E. Speizer. 1992. Dietary fat and fiber in relation to risk of breast cancer: An 8-year follow-up. Journal of the American Medical Association 268 (October):2037-2043.

APPENDIX A: References I found useful

Fackelmann, K. A. "Dietary Fat: No Link to Breast Cancer." Science News, 24 October 1992, 276.

"Relax, Mrs. Sprat: High-Fat, Low-Fiber Diets May Not Cause Breast Cancer After All." Time, 2 November 1992, 23.

To Home Page

Back One Page

Continue with Tour

Dietary Fat and Breast Cancer

Dr. Leon James Psychology 210: Statistics June 22, 1994

TABLE OF CONTENTS

Dr. Leon James
Psychology 210: Statistics
June 22, 1994