Main Menu                Health Library     

 DIETARY FATS and CANCER

Abstract:  Saturated fat in diets increases risk of cancer at most or all sites based on analysis of results from 25 specific studies plus three meta analyses.  Average cancer risk increases by about 6% for each 1% increase in dietary saturated fat.  But this rate of increase varies from only 1% for breast cancer to 10% for colorectal cancer.  Life Ahead assumes that cancer develops and recedes slowly over a 15 year period from changes in dietary fat similar to its change in risk that is associated with cigarette smoking.

Background:  Saturated fats have been widely implicated as a key cause of heart disease. Yet it now seems likely that their negative affect on health may be as large or larger for cancer than for heart disease.  Life Ahead now identifies from more than adequate research that the effect of various dietary fats on cancer is a very significant factor that can affect our Well-Days of life.

How Dietary Fat Produces Cancer:   Research relating dietary fat to cancer shows that dietary fat is either a direct carcinogen or a carcinogen promoter, and is an important contributor to cancer both in the US and worldwide. As described in the section on cancer, the contact of carcinogenic chemicals with body flesh maintained over long periods of time gradually produces the cellular damage that results in cancer.  A carcinogenic component in food contacts the mouth, the esophagus, the stomach, and the intestines in its passage through the body, and also may pass through the bladder and genital organs.  Lung and breast tissue is not directly contacted via this passage, but these sites can be contacted secondarily by passage of carcinogens into the circulating system. 

Fats include a wide range of differing related chemicals.  It seems likely that certain of these chemical constituents are fairly potent carcinogens, and that others are benign or may even inhibit cancer.  Because we do not have detailed analyses and information on each of these specific fat-chemicals in all of our foods, it is necessary to make best possible estimates of the carcinogenic potency of broad classes of fats on specific cancer sites.   

Cancer Usually Develops only Slowly over Long Time Periods:  Cancer is produced very gradually, probably as a result of a continuous process of conflict between its production by carcinogens and the defeat of this process by the immune system.  A minimum time needed for cancer development in the average human body appears to be about 15 years for carcinogens of the potency of those in cigarettes, with about 40 years of such contact needed for near maximum effect. The overall carcinogenic potency of dietary fats is probably lower than that of the tar in cigarettes.  Thus it should take a minimum of 15 years of exposure for dietary fats to develop cancer.  An  important observation here is that cancer from dietary fat and other food components is produced not by a most recent diet, but by some average cumulative diet of the past two or three decades. This makes research on dietary fat very difficult because records of long past diets rarely are available.

Dieting and diet changes of every kind and description have been rampant during the past 3 decades, and especially among women. Further, the vast amount of advice on better dietary habits in recent years has encouraged many but by no means all in our population to observe more healthful habits. Thus if we select those having the most healthful diets now, it is likely that this group included many of those that changed diets from earlier less healthful habits.  Thus their disease suffered recently was in part due to diets in much earlier years.

A net result of this is that today’s research on dietary fat probably underestimates and may significantly underestimate its true importance to future health. One attempt to compare diets taken at modestly differing times vs. a measurement of disease (Giovannucci,  Am J Epidemiol 137:502)  showed that quite different results were obtained using dietary assessments at times only a few years apart.  It may not be possible to obtain summed diets of useful accuracy for long periods of past years.  In the meantime there is need for a better recognition of this problem and of its probably impact on results of research.  

Men and women in different countries of the world however can have different eating habits over their full lifetimes.  Differences in disease vs. difference in dietary habits in different countries thus may provide a better measure of the long term effect of diets than that potential from practical conventional research in a given country.

The Research Data on Dietary Fat and Cancer:  The Table RDC following provides the key research on dietary fats vs. cancer selected for analysis.  This includes 38 risk ratios from 25 different studies on 8 differing sites of disease, plus 3 meta-analyses including at least 23 studies of breast cancer.  This listing probably captures most of the major research on this subject published to date. 

Nearly every available study identifies total or saturated fat to be a significant contributor to cancer, with an average risk ratio of above 2.0 for the highest vs. the lower 5th of the populations on most sites.  This risk is associated with an estimated average difference of about 7% of total calories as saturated fat.  The risk for meat appears similar to that for saturated fats. This overall result from at least 48 research comparisons provides a dramatic illustration that dietary saturated fat is a major contributor to the cancer of our population. 

The results for unsaturated fats are confusing.  Three comparisons showed unsaturated fats to convey the same cancer risk as saturated fats.  Yet results on olive oil that is very high in mono-unsaturates show it to be beneficial with risk ratios average 0.8.   Unsaturated fat rich peanut, soya, and safflower oils also were found beneficial in one study.  Animal data confirm that olive oil is protective, that unsaturated N3 acids protect, but N6 unsaturates do not protect.  Because of these conflicts, unsaturated fats are now included in Life Ahead as having a neutral or no effect on cancer.  It is hoped that an improved basis for these valuing these fats can be developed later.

Colon & Rectal Cancer:  The three observation studies listed indicate an average risk ratio of 2.3 for saturated fats on colorectal cancer. This works out to about 1.10 ratio per percent difference in saturated fat over the 15 year time frame used in Life Ahead.  But a fourth and much the largest study B1 found little effect of fat on this cancer. Figure 6 of the B1 paper did show the expected effects of saturated fat, fiber, and other factors directionally in dose related response. A possible effect of a 1.21 risk ratio for about 6% lower saturated fat was indicated there.  This would be about 1.032 times per percent of calories change in saturated fat.  But this randomized study was for an 8 year period, or only about one half of the usual time frame assumed for results of the observation studies that is used in Life Ahead.  Thus for the 15 year time frame used in Life Ahead for cancer development a comparable risk ratio from this study would be a maximum of about 1.06 per percent difference in % saturated fat.  In view of the results of this largest study this lower value is used in Life Ahead for risk of colorectal cancer.

Prostate Cancer:  A mean risk ratio selected from the 5 studies included for the effect of saturated fat on prostate cancer is 1.7.  This provides a risk ratio of 1.079 per percent difference in % saturated fat as a percent of calories.  Carcinogens in body water remain in contact with body tissue in the prostate area for substantial periods of time each day. Results of the studies are all consistently positive, although margins of error are rather wide on the individual studies. As for colorectal cancer, this estimate is probably conservative.

Breast Cancer: The effect of fat on breast cancer has been a subject not only of much study but has been of controversy because of the conflicting results from some studies.  But the effect of fat on breast cancer is much smaller than that it is for most other cancers.  This is consistent with the fact the dietary carcinogens would not be in direct contact with breast tissue but would arrive there via secondary paths.  The risk from a careful analysis of 7 quality prospective studies for saturated and total fat showed risk ratios of only 1.07 and 1.05 respectively, with upper 95% limits of 1.20 and 1.16.  The risk from a large meta-analysis of 23 studies of total fat and breast cancer was only 1.12, again with an upper limit of 1.21.  Although a small average risk ratio of 1.10 would be quite consistent with most of this research, most case control studies show higher risks than this.  Howe found an average risk factor of 1.42 for 12 case control studies, and the Boyd meta-analysis found case control studies to show a risk factor of 1.21.  A more recent large randomized study found a risk ratio of 1.10 for what was about a 3.2% difference in saturated fat as % of calories over 8 years but with marginal significance. 

These differences in result for all studies reported should be well beyond those likely from statistical error.  It is possible that these differences are due to the different timing of the diet information vs. the incidence of cancer in the two study methods.  Although a true risk of breast cancer probably is higher than the above 1.10 ratio this risk probably still is small. The risk ratio value used for Life Ahead is a 1.014 ratio per percent difference in dietary saturated fat is probably quite conservative..

Endometrial and Ovarian Cancer:  The research on endometrial cancer considered the effect of total fat, and the research on ovarian cancer considered saturated fat.  Results are all positive for risk but quite variable and involve substantial error margins. An estimated average risk ratio of 1.8 was selected from reviewing eight of these studies for what is a probable typical difference of 6% in saturated fats as percent of calories.  The risk value now used in Life Ahead is 1.10 per difference of 1% in saturated fat as % of calories. 

Lung Cancer:  The limited research on fat vs. lung cancer appears confounded by the large effect of smoking.  Study #2 was particularly interesting because it included a large sample of 429 non-smoking women who contracted lung cancer. Those having highest saturated dietary fat had a 6-fold increase in lung cancer vs. those having lowest fat.  Although other studies that included smokers failed to find an effect of fat on cancer, it seems likely in view of study #2 that lung cancer is not immune to the effect of fat.  Tentatively and because of inadequate data the effect of saturated fat on lung cancer has been assigned a risk equal to that of all other cancers that follow.

Other Cancer:  Three studies in the table following show results on the effect of dietary fat on testicular, bladder, and pancreatic cancer.  Again all results are positive, indicating risks.  The risks found from these studies plus those found from the research on the other sites suggests that saturated fat probably will increase risk of all or at least nearly types of cancer.  For appropriate dietary concern it is estimated that cancer sites other than those cited above will incur a risk ratio of 1.5, or 1.06 per % change in dietary saturated fat.

Dietary Fat and its Duration of Exposure:  Life Ahead forecasts the effect on Well-Days for both current habits and for changes in dietary habits.  It is unlikely that overall risk of cancer will either increase or decline immediately when a change in diet is made.  As example, near full benefits from the stopping of smoking are achieved only after 10-15 years. And it took 15 or more years for cancer from smoking to develop.  As a conservative and logical approach, it is assumed that cancer risk in response to a change in dietary fat will gradually take place similarly over a 15 year period.  Although not shown in Table RDC, most other research studies included in Life Ahead probably considered average exposures to factors of at least 10-15 years. 

Life Ahead now predicts the risk of cancer at various sites at present age using the above described risk ratios.  But in response to a dietary change as assumed in the good habits valuation or in response to an individual's change in diet, the program assumes that this change in risk will take place stepwise over 15 years of time, similar to that found for risks of cancer from cigarettes.  Life Ahead analyses initiate with risks for an average US population assumed to have 13% of dietary saturated fat in total calories.  Thus Life Ahead adopts a risk factor of 1.0 for this average population value of 13% saturated fat.  The risk of colorectal cancer increases by 1.104 times for each % increase in dietary saturated fat.  Thus s forecast risk of colorectal cancer after 15 years for say 15% fat would be 1.104 ^ (15-13) or 1.22.  If dietary fat were reduced to 9% the risk at present age, the eventual risk of this cancer would become 1.104 ^ (9-13) or 0.67.  Life Ahead now forecasts for this second example that risk of this cancer would progress down each year by an amount of 0.9734 per year that in 15 years would reach the new risk level of 0.67.  After 15 years will the risk will stay at 0.67.  Lower and upper limits accepted for saturated fat for computation are those represented by research of 7.5% and 16% respectively.  It is possible that higher or lower values of this fat or longer times of exposure will produce higher or lower risks of cancer than that now forecast.  This method now used thus produces a conservative but probably more correct valuation of a change in habits than will the more usual statistical type valuation that assumes risk changes fully and instantly in response to each factor change.

Research on Cancer vs. Saturated Fat in Different Countries Could Help:  As noted above, historical populations in different countries have used quite different usual diets than those usual recently in the US. A problem in using such information is that other habits in different countries also varied substantially.  Thus a direct comparison of cancer rates and saturated fat although possibly useful may not identify correctly the effect of dietary fat and cancer per se. 

Life Ahead may provide the most complete and rigorous method now available for correcting risks of cancer and other included diseases for differences in diet and other habits.  Thus inputting best estimates of average country life style habits and other dietary factors into the program at one or more ages should forecast values of disease risks that could be compared with actual country population risks in these countries.  Although it is not expected that the present program will be capable of predicting risks of each disease precisely  for different countries, a clear trend in actual vs. forecast values vs. factors could indicate areas of factor confirmation or deficiency.  For example, if there is no trend of actual vs. predicted values with substantial actual differences in country average saturated fat, this will confirm further that the values of risk developed above from research are correct.  If the actual vs. predicted values show say a clear trend of larger than estimated risk with saturated fat, this will suggest that actual risks are measurably higher than those estimated and need further upward adjustment.   

                                                                                             TABLE RDC, CONTINUED