I am currently Consultant in Radiotherapy and Oncology at St Bartholomew's Hospital London. Prior to this I was Senior Lecturer in Clinical Oncology at the University of Edinburgh. There over a period of 10 years I conducted clinical research into various forms of cancer including breast cancer and colo-rectal cancer. I have published numerous articles on colo-rectal cancer and have contributed to chapters on large bowel cancer to major United Kingdom text books of oncology.
I have been commissioned by Barlow Lyde and Gilbert to produce a report on the leaflet "What's wrong with "McDonalds". In particular the allegation contained in the leaflet linking McDonalds food with cancer's of the breast and bowel.
Multiple factors are involved in the causation of both breast cancer and colo-rectal cancer in the same way as multiple steps are implicated in the development of these tumours. It is probable that both environmental and constitutional factors are interwoven within the causation but that the precise composition of events varies from one individual to another. As a result neither one factor nor one set of factors will underlie the causation of all cases of breast and bowel cancer. Cancer is ultimately a disorder at the level of the gene and although the behaviour of a malignant cell will depend upon the way in which genetic abnormalities are expressed this may well be influenced by environmental factors. Undoubtedly in a number of cases some constitutional factor leads to the development of these tumours. This may be the inheritance of some distinct genetic abnormality. This for example is seen in familial polyposis which is a precursor to the development of bowel cancer. In addition breast cancer is seen to run to families and although in most instances no specific genetic abnormality has been discovered undoubtedly inheritance plays an important role.
While diet is believed to be important in the development of a number of diseases the exact mechanism by which these diseases are produced remains unclear. The multi-factorial nature of the causation of both breast and large bowel cancers therefore makes it extremely difficult to identify what if any influence diet has on their causation.
The possibility of a relationship between dietary fat intake and the development of breast cancer has been extensively investigated in recent years. This hypothesis arose because in animal research into the effects of diet it appeared that increased fat intake caused breast tumours to develop. In international studies examining the mortality from breast cancer in women it was shown that the mortality in different countries strongly correlated with the fat consumption per head of population as crudely measured in food balance tables. These correlations applied only to animal fat intake, not vegetable, and were stronger for post-menopausal than pre menopausal women. Other findings that seemed to be relevant were that in migrant studies it was noted that there was frequently an increase in the incidence of breast cancer amongst women moving from areas of low incidence to countries where the incidence was high. For example Japanese women who moved to the United States of America and who at the same time increased their intake of fat had increased levels of breast cancer. Such evidence is however crude and indirect and for a direct relationship to be shown between fat intake and the causation of cancer individual women must have been shown to have had a higher intake than average. In addition since this may simply reflect higher energy requirements as determined by body size, physical activity or individual metabolism, affected women should ideally be shown to obtain a higher proportion of calories from total fat than do control subjects.
Case Control Studies
Case Control Studies are those where information relating to diet is obtained in patients with a specific type of cancer. The same information is then collected from a similar group of women who do not have the disease and therefore act as control subjects. The evidence from Disease Control Studies has been conflicting. Many have had serious limitations in the methodology employed and in others the proportion of patients who have been excluded from the study makes any conclusions drawn, suspect. On the whole, studies of total fat intake have not generally found that women with Beast cancer report a significantly higher consumption than controls. The serious limitations of these studies is highlighted by that conducted by Van't Veer et al (1990) where only 55% of the controls actually participated in the study. Other studies have examined only particular fatty foods such as dairy products, fried potatoes, cheese and milk. These studies have been less rigorously carried out and have not allowed adjustment for total calorie intake. In other investigations limitations such as a poor response rate amongst controls and differences between the circumstances of their interviews and those of the cases have cast doubt on their relevance. One recent review of a combined analysis of 12 case contra studies did suggest a relationship between fat intake and the development of breast cancer (Howe et al 1990). However the methodology used in this investigation has been strongly criticised and a more appropriate analysis which has subsequently been conducted indicated that there was no significant relationship (Willet and Stampfer 1990).
This lack of a convincing relationship from Case Control Studies has relevance to the evidence cited by the defendants. They quote epidemiological studies all carried out a number of years ago. The main criticism of such investigations is that there are real difficulties in accurately assessing dietary intake. One of the studies quoted (Bristol et al 1985) specifically mentions the imprecise nature of the methodology used In the study. It is stated quite clearly that sizes of portions were estimated rather than measured. In addition 60 to 70% of the population of patients had to be excluded in order to meet the authors criteria and this itself introduces further bias.
These criticisms apply to the other studies cited (Drasar and Irvin 1973; Kolonel et al 1981). One of the authors quoted, subsequently produced a further report on a study designed to confirm their original findings (Kolonel et al 1983). The conclusions of this subsequent review were that there was a much weaker association between fat and the possibility of subsequent breast cancer and they drew the additional conclusion that this was consistent with most of the literature.
Diet is difficult to describe at the best of times and especially to recall from the past. In Cohort Studies dietary derails are collected at the very start of the study before the development of any disease and the individuals are then followed over a number of years. In this way these studies avoid not only the possible effects of the disease, diagnosis and treatment on recall but also the effects of recent diet which may further indirectly influence recall of past diet. As a general rule these studies have produced less support for the hypothesis than Case Control studies. The largest of these has examined 89,494 women in the nurses health study in the United States (Willet et al t992). The study was set up in 1980 and investigated a range of women between the ages of 34 and 59. The most recent report (Willet et al 1992) describes an 8 year follow-up period. 1,439 cases of breast cancer were diagnosed in these women which included 774 amongst post menopausal women. Adjustments were made for age, established risk factors, and total energy intake. After allowing for these adjustments no evidence was discovered for any positive association between total fat intake and the incidence of breast cancer.
The strongest association between diet and the development of breast cancer described in Case Control Studies has been in post menopausal women. The post menopausal women in the Nurses Health Study were examined in detail and no such association could be found. It has also been suggested that if fat intake could be reduced below 25% of total energy intake then this would be protective for the development of breast cancer. This specific point was examined in detail in the Nurses Health Study but even when those women having the lowest fat intake were compared with those having the highest fat intake there was a complete absence of any positive association. Other possible associations with diet were examined in this study in particular the possibility of a protective effect of fibre. However when this was examined no protective effect of dietary fibre could be discovered.
The only positive association between fat intake and breast cancer in Cohort Studies has come from Japan (Hirayama 1978). However the number of women studied in these Japanese investigations was extremely small and therefore the confidence limits are very wide.
A number of other Cohort Studies have looked at the question of a relationship between fat intake and the development of breast cancer. None has shown any clear causal relationship between fat intake and breast cancer (Kinlen 1992).
The evidence therefore from Cohort studies is quite clear h that there is no association whatsoever between dietary fat in relation to the risk of the development of breast cancer.
Breast cancer is the most frequent cancer amongst sporran in North America and Western Europe in terms of both incidence and mortality. However the incidence of breast cancer h countries such as Japan is extremely low. Studies of Japanese women who moved to the United States of America showed an increase in breast cancer and it was noted that in this group their fat intake had also increased. Breast cancer is a hormonally mediated disease and several factors that influence hormonal status, for example age at first birth, or are an indication of a change in hormonal status, eg, age at menarche or age of menopause have been shown to be associated with breast cancer risk. These factors cannot be readily modified and as a result attention has focused on factors that may be more amenable to change and whist' may possibly account for international variations in the incidence of the disease. National characteristics of diet have been related to breast cancer incidence in the individual countries concerned and the findings of such correlation together with animals studies have led to the hypothesis relating dietary factors, particularly fat intake and the development of breast cancer. The international correlations with fat are so striking that they apparently explain about 80% of the international variation in breast cancer mortality. However the dietary data that are involved are particularly crude. They consist of food balance or food disappearance data. That is home production plus imports minus exports and expressed per capita by dividing by the total population of the relevant country. The relationship between fat and breast cancer compared to other nutrients may partly relate to the fact that fat provides most calories in food. This could therefore influence 3 of the principal risk factors for the development of breast cancer. These are body size, age at monarchs during the years of growth and obesity in adult life.
A strong international correlation has been found between adult height in women and the incidence of breast cancer. This may reflect the effects of energy restriction during the period of growth (Kinlen 1991). Individual studies comparing height and breast cancer incidence in Western Countries have not produced consistent results. This is almost certainly because within country variations are likely to be mainly genetic in origin. However a Norwegian study (Vatten and Kvinnsland 1990) found evidence of a relationship but only amongst women who had been at the age of puberty during the 1939 - 1945 war years when calorie intake was restricted. Such energy restriction is also more likely to be found in developing countries in which Here is a high calorie expenditure on physical work but a low proportion of energy derived from fat.
The age of menarche is a well established risk factor for breast cancer and later menarche which has a lower risk attached, like height, can reflect calorie restriction during the years of growth.
Late age at first birth, low patty and late menopause are commoner in high fat consuming countries. They may therefore indirectly contribute to the international correlations.
The lack of adjustment for the amount of calories expended in physical work may be relevant to the weaker international correlation which has been found between breast cancer and calories than with fat (Kinlen 1987). In largely western societies with a high fat consumption and high breast cancer rate obesity is frequent. In virtually all the studies which have looked at the incidence of breast cancer in post menopausal women there has been shown to be a positive relationship with body weight. This is probably an oestrogen effect since oestrogen precursors are produced in adipose tissue.
The crude nature of food balance tables is highlighted if comparisons of food disappearance data are made with nutritional studies of individuals in the same countries. There is no consistent pattern across different countries and even within a given country. In rich countries which have the highest incidence of breast cancer the greatest food wastages are found. It is impossible to make f proper adjustment for this in Mod balance tables. Wastage of fat may contribute therefore to Me apparent association with breast cancer. The international correlation between fat and breast cancer is stronger for mortality data than for the incidence of the disease. It is possible that this reflects, to a certain extent. the adverse effect of obesity on survival. This also may explain the longer survival rates recorded amongst Japanese than western women who develop breast cancer.
An increased risk of breast cancer in animals with high fat consumption was first demonstrated by Tannenbaum (1942). Father animal studies have shown that this effect occurs with both saturated and unsaturated fats. These animal studies have been influential in the fat hypothesis of human breast cancer. However more recent experiments have looked at the question of a high calorie diet as well as the relationship of fat intake. These have shown that animals fed a high calorie diet but with a low fat intake developed more mammary tumours than those fed a high fat low calorie diet (Kritchevsky et al 1984). h a similar way to the data on human breast cancer the interpretation of the evidence overall depends upon whether fat has an effect independent to that of the calories it contributes to total intake. The evidence from animal experiments on this point is controversial (Albanes 1987).
The contradictory nature of the evidence concerning fat and breast cancer casts real doubt over there being any causal relationship. T. he largest prospective studies which have been carefully conducted do not show any evidence of a positive association between the intake of fat and risk of breast cancer amongst either premenopausal or post-menopausal women (millet et al 1992). Although international data do suggest that a relationship does exist several factors could explain the apparent discrepancies between these data and those of prospective studies. The data, on which the international studies are based are themselves unreliable and do not take into account food wastage. For example there is good evidence that fat intake in the United States has actually decreased steadily since the early 1 950's. During this period of time the breast cancer incidence in the United States has increased by about 40%. (Stephen and Wald 1990) Although fat intake has been low historically in many countries ninth a low incidence of breast cancer, such as Japan, numerous other differences in diet and life style could account for the variations seen in breast cancer incidence rates. Earlier dietary energy restriction may be important in this respect. This hypothesis is also supported by more recent animal studies which demonstrate a stronger and more consistent effect of energy restriction rather than dietary fat intake on the development of mammary tumours.
Taking in to account all me evidence currently available it is therefore difficult to draw the conclusion that there is a causal relationship between fat intake and the development of breast cancer.
LARGE BOWEL CANCER
Cancer of the colon and rectum is one of the most common forms of malignant disease. The mortality rates in the United Kingdom are second only to lung cancer for males and breast cancer for females. There are nearly 20,000 deaths from colo-rectal cancer in the United Kingdom each year.
There is an anatomical distinction between the rectum and the colon with the rectum being the terminal part of the bowel which begins opposite the first part of the sacrum. There is some growing evidence that cancers of the right and left colon may be epidemiologically distinct although much of what applies to colon cancer also applies to rectal cancer in terms of causation.
There is a striking variation in the frequency of colo-rectal cancer throughout the world. The disease is especially common in North America, Northern Europe and Australasia.
It is rare in Asia and particularly uncommon in Sub-Saharan Africa. The disease is commoner in men than in women. The greatest variation between low and high risk areas is seen for coon cancer in men. Colon cancer is especially frequent in the North of Scotland.
In countries where the rates were initially low, such as Japan, substantial increases have been observed win the passage of time. Where rates were moderately high at the outset, small increases have occurred. In those countries where rates were initially very high gradual falls in mortality have been noted with time. (Morson et al 1990) A number of factors have been implicated in the development of colorectal cancers. There may be differences in the relative importance of these factors with regard to the development of right sided compared with left sided cancers. A number of circumstances have been highlighted in which the risk of right sided colon cancer appears to be increased selectively. These are-female sex, previous cholecystectomy - especially in women, a low blood cholesterol prior to surgery, nulliparity and following hormone therapy for prostatic cancer. A common denominator with respect to these factors might be altered patterns of the metabolism of bile acids which are secreted into the bowel to help with the digestion of food. The amounts or types of bile acids entering the right colon may be relevant to carcinogenesis at this site whereas the faecal concentration of bile acids in the left side of the colon will be modified by other dietary factors such as fibre content. Other factors which have been implicated in the development of right sided cancers include selenium deficiency and genetic factors.
Cancer of the colo-rectum usually presents in the seventh decade. However in low risks areas the average age at diagnosis is about 50.
Multiple factors are involved in the aetiology and development of colo-rectal cancer. Genetic factors are important in a small group of people with familial adenomatous polyposis. However genetic factors obviously play a part in other families where there is a strong family history of large bowel cancer.
The considerable geographical variation in the incidence of colorectal cancer indicates the potential importance of environmental Actors in the development of the disease. The variation does not appear to be rented to racial differences as migrants who move from low to high risk areas acquire the pattern of incidence of bowel cancer of the country to which they move. Diet has formed the main focus of attention of epidemiologists and experimentalists and excess consumption of meat and animal fat, refined carbohydrate and beer have been implicated as harmful. Fibre, vegetables and trace elements have all been suggested as being protective.
It has been postulated that carcinogens might be produced by Me action of bacteria on certain substances contained within the intestine. Diet could naturally influence both the concentration and the composition of these substances within the bowel. It has generally been assumed that these substances are the bile acids produced by the liver to help with the digestion of food. There is some evidence that bile acids act as tumour promoters (Morson et al 1990).
In animal experiments surgical and dietary manipulations that increase the bile acid concentration in the faeces also increase the development of colo-rectal cancers.
The level of bile acids in faeces has been Ensured in high and low risk populations. These epidemiological studies strongly support the role for bile acids in the development of large bowel cancer. It has been suggested that increased consumption of meat and hence animal fat would account for the high levels of faecal bile acids detected in high risk populations.
As with breast cancer however the picture is far from clear. One of the first suggestions that there was a relationship between diet and the development of colo-rectal cancer was that of Burkitt in 1971. He had noted the low incidence of large bowel cancer in Africa and suggested that the. high fibre diet of the Bantu's might protect them from the development of colorectal cancer by increasing the speed of intestinal transit. This he implied would reduce the exposure of the gut to the chemicals present within the contained faeces which might be responsible for the causation of bowel cancer. At the same time it was implied that the f carcinogens contained within the faeces would be diluted by the greater bulk of the stool resulting from fibre intake. However the link between fibre, transit time and colo-rectal cancer has never been substantiated.
Other studies have investigated the differences in colo-rectal cancer incidence and mortality in migrants. For example the incidence and mortality rate from large bowel cancer in Puerto Ricans migrating to the United States are 2 fold higher than for those living in Puerto Rico. Life style factors such as diet are believed to play an important role in the development of bowel cancer in these circumstances. A number of studies have examined different populations looking at the question of dietary fibre and cancer incidence. Although not entirely conclusive there is a suggestion that dietary fibre acts in a protective fashion for the development of cancer of the large bowel. (Doll 19881
CLINICAL DIETARY STUDIES
A number of studies in human nutrition have been carried out over the years. The results of these have been somewhat controversial (Greenwald 1992). One of the main reasons for this is that drastic changes in long established eating patterns are difficult achieve. Studies which have attempted to investigate alterations in dietary patterns have suffered from a high attrition rate. Also it is known that there is a low long term success rate of most weight management programmes (Vargas and Alberts 1992)
One of the largest studies which has attempted to identify risk factors for fatal colon cancer was in a prospective study investigating 764,343 adults (Thun et al 1992). This study started in 1982. The mortality of those people included in the investigation was studied over a six year period until August 1988. One thousand one hundred and fifty deaths from colon cancer occurred during this time. Six hundred and eleven in men and five hundred and thirty nine in women. Multivariate analyses were then used to compare these patients with five thousand seven hundred and forty six matched control subjects drawn from the same study. The results of this investigation support the hypothesis that a diet high in vegetables and grains may protect against fatal colon-cancer.
Although the exact vegetables and high fibre grains which were most closely associated with the reduced risk could not be identified the data supported growing evidence that a variety of plant products may be beneficial. The relative risk for those men Skirt the highest vegetable intake compared with those having the lowest intake was 0.76. In women it was 0.62. However weaker associations were found for physical inactivity, obesity and total dietary fat. There was also a weaker association with family history. Importantly in this very large prospective study no association could be found with a consumption of red meat or saturated fat and the development of bowel cancer.
A further study which looked at the intake of vegetables, fruits and vitamin supplements and the subsequent development of cancer was carried out on 11,580 resident of a retirement community, near Los Angeles in the USA. (Shibata al at 1992). A total of 1,335 cases of cancer were diagnosed in these residents between 1981 and 1989. When those patients who developed large bowel cancer were examined it was found that in contrast to other large American studies no benefit could be shown for even high vegetable and fruit intake amongst male residents. Interestingly in males an increased risk of coon cancer was seen in those who consumed large quantities of darn green vegetables. In contrast in women a significant protective effect was noted for all vegetables and fruit as far as colo-rectal cancer was concerned.
The relationship of diet and coon cancer was also investigated in the study of 89,494 women in the nurses health study which primarily examined breast cancer (Willet et al 1992). Although no association was found between fat intake and breast cancer they did describe an association between animal fat intake and the risk of colon cancer in the same group of women.
These 3 studies highlight the difficulties of drawing any firm conclusions about any possible association between diet and the development of large bowel cancer. The clearest evidence seems to be that a good intake of fruits and vegetables may be protective. However the relationship between fat intake and coon cancer remains controversial. Undoubtedly no one factor can account for the development of all cases of large bowel cancer. An inter-play of Actors including genetic factors is involved.
SCIENTIFIC EVIDENCE CITED BY THE DEFENDANTS
As with breast cancer the evidence supporting a relationship between fat intake and the development of large bowel cancer is controversial. Although there appears to be a slightly grease' association between fat intake and colo-rectal cancer than there is in breast cancer the evidence is by no means clear. This is of particular relevance to the claims made by the defendants.
From all this it can concluded that it is impossible to state that there is a specific causal relationship between eating McDonalds food and the development of either breast or bowel cancer. In addition McDonalds are criticised for their nutritional guide and for the fact that their food is unappetising and does not require chewing. It is alleged that chewing aids subsequent digestion. The secretion of saliva is under complex control of visual as well as sensory stimuli. Were is no evidence whatsoever to state that McDonalds foods will not stimulate these sensations and there is no evidence whatsoever that McDonalds food is difficult to digest.
The dietary information given by McDonalds food leaflets clearly states the content of the food and it also emphasises that variety and balance in ones diet is important to good health. No suggestion is made that people should eat only McDonalds foods.
In conclusion it can be stated that the evidence relating fat intake and the development of breast and bowel cancer remains controversial. Whilst diet may play a part in the development of these tumours the exact items in the diet which are involved and the mechanisms by which they stimulate the development of tumours is unknown. There is therefore no substantiation whatsoever for the allegations that have been made regarding McDonalds food and cancer development.
exhibits: Not applicable/ available