Letters

Diet and the prevention of cancer

Author's recommendations are not justified

Whether meat is a risk factor for cancer remains uncertain

Consumption of oily fish should be encouraged

No evidence has linked ovarian cancer with high intakes of fat and meat

Authors' reply

Author's recommendations are not justified

EDITORThe epidemiological literature justifies only two of the conclusions that Cummings and Bingham draw in their review about diet and the prevention of cancer: the recommendations to avoid (high doses of) vitamin supplements and mouldy foods.¹ Even the cited report of the World Cancer Research Fund shows that the overall evidence for dietary recommendations is weak if one takes into account the more reliable data from prospective cohort and intervention studies.²

Cummings and Bingham give an excellent example of publication bias in their section on colorectal cancer and red meat: they cite two prospective studies that support a role for red meat in colorectal carcinogenesis. What they do not mention is thatbeside at least three other studiesfive prospective studies cited in the World Cancer Research Fund's report did not find a significant association with red meat.

It is not yet proved that heterocyclic amines or N-nitroso compounds definitely increase rates of colon cancer.³ Bingham herself has shown that the endogenous production of N-nitroso compounds varies widely between individuals and also depends on other components of the diet (for example, resistant starch).⁴ Recent research has shown that chicken, which is often recommended as a healthy substitute for red meat, can contribute heavily to the uptake of heterocyclic amines.⁵

Cummings and Bingham's statement that "non-starch polysaccharides (fibre) and vegetables are established factors that reduce risk" is also misleading. As is shown in the World Cancer Research Fund's report, none of four prospective cohort studies on non-starch polysaccharides showed a significant reducing effect on colon or rectal cancer.

The protective effect of vegetables is also far from proved. Of four prospective cohort studies cited in the World Cancer Research Fund's report, one found no effect with green salad; one found a significant reduction in risk with rising vegetable consumption only in women; one found an increasing risk with increasing amounts of dark green vegetables in men; and one found no significant effect with any of 15 kinds of vegetables and fruits.

These few examples show that there is no evidence in the prospective literature for an upper limit of 140 g of red meat a day, nor for a general protective effect of fibre or vegetables. Public interest in cancer prevention is high, and scientists should be careful with statements or recommendations.

Ulrike Gonder, Nutritionist. 
European Institute of Food and Nutrition Sciences, D-65239 Hochheim, Germany ugonder{at}aol.com

   Competing interests: None declared.

Whether meat is a risk factor for cancer remains uncertain

EDITORMeat looks decreasingly likely to be a direct cause of cancer. Concerned by the discrepancy between evidence on meat and cancer and authoritative statements such as those of Cummings and Bingham,¹ last December I chaired a workshop of cancer epidemiologists, nutrition experts, and researchers on the colon from Australia, New Zealand, and Britain.

We found that since the report by the Committee on Medical Aspects of Food and Nutrition Policy² two more prospective studies have failed to show an association of meat intake with colorectal cancer.³ There are now 12 prospective studies reporting meat intake and subsequent large bowel cancer, but in only two was a significant association found.⁴ Even in these the association was weak (relative risk <2.0) and seen only in people with the highest fifth of meat intake. These two studies come from groups in the United States.

As well as this accumulation of mostly negative prospective studies, a multinational combined report of five follow up studies of vegetarians and socially matched omnivore controls (total 76 000 subjects) found the relative risk of colorectal cancer in the vegetarians to be 0.99 (indistinguishable from 1.00).⁵ This is as near as we are likely to get to randomised controlled trials of meat eating.

Our review concentrated on meat and large bowel cancer because, as the report of the Committee on Medical Aspects of Food concluded, the evidence is weak that lower intakes of meat would lower the incidences of breast, lung, prostate, or pancreatic cancer.²

If meat made no positive contribution to the diet it would be straightforward to warn that in a minority of subjects in a minority of epidemiological studies there seemed to be a small risk of colorectal cancer. But nutritional advice to the general public can't be as simple as this when it concerns one of the central foods groups of most people. Meat is the major source of available iron, vitamin B-12, zinc, and protein in Britain and most other affluent countries. There is of course no need to char or heavily brown meat, poultry, or fish when frying or grilling, and our workshop recommended not to.

The paper's conclusions from our workshop are published in the European Journal of Cancer Prevention.⁶ Our conclusionthat it remains uncertain whether meat is a risk factor for cancercoincides with British opinions reported last December in the Sunday Times Magazine.⁷

A Stewart Truswell, Professor of human nutrition. 
Human Nutrition Unit, University of Sydney, NSW 2006, Australia S.Truswell{at}biochem.usyd.edu.au

   Competing interests: Professor Truswell received a fee from Meat and Livestock Australia for his time acting as chairman of the workshop mentioned.

Consumption of oily fish should be encouraged

EDITORIn their review of dietary factors in the development of cancer Cummings and Bingham emphasised the importance of lifestyle factors in most cancers of the large bowel, breast, and prostate.¹ They failed to mention the value of fish in the diet; as a source of n-3 polyunsaturated fatty acids, fish have been suggested to have a protective effect in breast and colorectal cancer in particular.

Ecological studies have suggested a reduced incidence of breast and colorectal cancer in those populations consuming higher proportions of fish. ^{2 3} n-3 Fatty acids have also been shown to increase the resistance of cultured cells to transformation by irradiation and transfection in contrast to the effects of n-6 fatty acids.⁴ A large rise in mortality from colorectal and breast cancer has accompanied the increase in the ratio of n-6 to n-3 fatty acids in the Japanese diet in the past 50 years.⁵

As with most of the dietary advice given in the review, there is no evidence to confirm that increasing the proportion of n-3 fatty acids in the diet will reduce the risk of cancer. Advice to encourage the consumption of oily fish may, however, be of benefit.

Matthew D Barber, Senior house officer. 
University Department of Surgery, Royal Infirmary of Edinburgh, Edinburgh EH3 9YW MDB{at}srv1.med.ed.ac.uk

1.	Cummings JH, Bingham SA. Diet and the prevention of cancer. BMJ 1998; 317: 1636-1640. (12 December.)
2.	Kaizer L, Boyd NF, Kriukov V, Tritchler D. Fish consumption and breast cancer risk: an ecological study. Nutr Cancer 1989; 12: 61-68[Medline].
3.	Caygill CPJ, Charlett A, Hill MJ. Fat, fish, fish oil and cancer. Br J Cancer 1996; 74: 159-164[Medline].
4.	Takahashi M, Przetakiewicz M, Ong A, Borek C, Lowenstein JM. Effect of w3 and w6 fatty acids on transformation of cultured cells by irradiation and transfection. Cancer Res 1992; 52: 154-162[Abstract/Free Full Text].
5.	Okuyama H, Kobayashi T, Watanabe S. Dietary fatty acidsthe n-6/n-3 balance and chronic elderly diseases. Prog Lipid Res 1997; 35: 409-457.

   Competing interest: Dr Barber has received financial support for research and conference attendance from Scotia Pharmaceuticals and Ross Products Division, Abbott Laboratories; he has undertaken research on the effects of n-3 fatty acids in cancer.

No evidence has linked ovarian cancer with high intakes of fat and meat

EDITORCummings and Bingham state that their review on diet and the prevention of cancer¹ was prompted by reports from the Committee on Medical Aspects of Food² and the World Cancer Research Fund.³ In their otherwise careful review the authors state that "high intakes of fat, milk products, and meat increase risk" of ovarian cancer.

The Committee on Medical Aspects of Food concluded that "there are too few studies which have examined the relationship between meat, fat and dairy products and ovarian cancer to draw conclusions."² Likewise, the World Cancer Research Fund concluded that "the evidence relating milk consumption and the risk of ovarian cancer is limited and inconsistent: no judgement is possible."³ Both reports cited Engle et al's paper as indicating no significant association of this cancer with dairy products.⁴

Thus there is no scientific support for Cummings and Bingham's statement, which was in any case unreferenced. If their review becomes widely quoted the erroneous notion that milk products (which many regard as highly nutritious and acceptable foods) cause ovarian cancer may become accepted. Surely this is not something that responsible nutritionists would want.

M I Gurr, Visiting professor in human nutrition, University of Reading, Reading. 
Vale View Cottage, St Mary's, Isles of Scilly TR21 0NU

   Competing interests: Professor Gurr has received consulting fees for advice given to the dairy industry.

Authors' reply

EDITORDefining the precise relation between diet and the risk of cancer is difficult. There are many cancers, multiple foods and nutrients, and other confounding factors such as exercise, smoking, and alcohol. It is therefore essential to have consistency in the evidence from population studies, animal and human studies, and the molecular and cellular pathology of cancer.

The two recent major reviews of diet and cancer independently showed a large measure of agreement about food items such as fruit, vegetables, and meat and the risk of cancer. ^{1 2} Some of the results used in these reviews, however, were from epidemiological studies, particularly cohort studies, set up several years ago. Accurate methods for dietary assessment have only recently been developed, and earlier crude assessments of dietary intake, based on short lists of food and food frequency questionnaires, were used in most of these studies. These crude assessments give a substantial degree of measurement error that is not amenable to correction.³ Thus, for several reasons, these epidemiological studies cannot alone form the basis of recommendations on diet and cancer.

The Committee on Medical Aspects of Food concluded that "there is moderately consistent evidence that diets with less red meat are associated with reduced risk of colorectal cancer" (page 4),¹ and the World Cancer Research Fund concluded that "red meat probably increases risk" (page 246).² In the recent meta-analysis of five cohorts referred to by Truswell, meat eaters were not at greater risk of colon cancer than those who were not meat eaters.⁴ The amount of meat consumed by meat eaters was not, however, established; it may well have been low since the cohorts were recruited from people who shopped at health food shops or read vegetarian magazines, friends and relatives of vegetarians, and Seventh Day Adventists.

Cancer rates in vegetarians are 41% of those in the general British population.⁵ Nevertheless, neither the Committee on Medical Aspects of Food nor the World Cancer Research Fund recommended that no meat should be eaten. The Committee on Medical Aspects of Food recommended that intakes should not rise and that consumers eating more than 140 g a day should consider a reduction. The World Cancer Research Fund recommended that individuals consume less than 80 g daily.

Barber asks for special mention to be given to fish and n-3 fatty acids. However, both the Committee on Medical Aspects of Food and the World Cancer Research Fund were unconvinced that either fish or polyunsaturated acids were protective. Moreover, selecting individual nutrients, as opposed to foods, to protect against cancer can be dangerous, as trials of -carotene and vitamin E have shown.

Gurr is correct. It was not our intention to identify dairy products as significant risk factors for ovarian cancer. In fact, the opening sentence of this part of our review stated that "the major known risk factors for other sites are non-dietary."

John H Cummings, Honorary consultant physician. 
Department of Molecular and Cellular Pathology, Ninewells Hospital and Medical School, Dundee DD1 9SY

Sheila A Bingham, Head, diet and cancer group. 
MRC Dunn Human Nutrition Centre, Cambridge CB2 2DH