By far the largest reliably known statistic is the percentage of current US cancer deaths that are due to tobacco, which cannot be less than 25 percent or more than 40 percent, although it is possible that some nutritional factor(s) may eventually be found to be of comparable importance.
Tobacco and Smoking
The percentage of US cancer deaths due to tobacco is still increasing, and must be expected to continue to increase for some years, due to the delayed effects of the adoption of cigarettes in earlier decades. Of the approximately 155,000 US deaths attributed to cancer of the lung, only about 12,000 are not due to tobacco.
Other than for smoking (for which the cause-and-effect is proven beyond reasonable doubt), the researchers prefer to talk of percentages of cancers that “might be avoidable” in various ways, rather than to the percentages that are due to various “extrinsic” or “environmental” factors.
Overall, among people under 65 years of age, most of the national trends in recorded death from cancer are downward (except those due to smoking). Another exception important is the increase in melanoma. Some decreases are due to improved treatment (for example, in the case of Hodgkin’s disease) while some are for unknown reasons (as in the ease of cancer of the stomach) but cannot be accounted for by improvements in the outcome of treatment.
However, although some thousands of Americans every year are now dying of asbestos cancer, this public health disaster cannot be clearly seen in the national trends (except for the still rare asbestos-induced Mesothelioma), so an analysis of trends is clearly a crude tool.
While national trends are useful in setting priorities, they are clearly unsuitable for dealing with specific problems. Detailed analyses of individual agents yield some fascinating insights.
Because the risk of lung cancer in later life depends on the age at which smoking was begun, a delay of a couple of years in this can reduce the later risk by as much as 20 percent. The larger lung cancer rates in smokers in urban compared to country areas may be due to different social norms earlier this century rather than an interaction between smoking and urban pollution. Because the mortality rate for non-smokers is so very low, irrespective of residence, the effect of urban pollution is small.
Alcohol cannot be shown by laboratory tests to cause cancer. However, it “interacts” with smoking, each agent enhancing the other’s effects. Alcohol possibly exerts its effect by dissolving cancer-causing chemicals and bringing them in contact with body cells. An important distinction must be made between mortality (death) and incidence (occurrence) of cancer.
The US trends in age-standardized female breast mortality are almost constant, though very slightly upwards [decreasing for women under 65, increasing among older women overall 0.2 percent a year upward]. It is likely that these trends are a reasonably accurate reflection of the underlying onset rates of serious breast cancer because the primary treatment of breast cancer (by surgery and/or radiotherapy) did not alter much, while chemotherapy of early breast cancer was not in common use, and fatal breast cancer is not a difficult disease to diagnose.
In contrast, the available data on breast cancer incidence exhibit bizarre fluctuations far too large to be due to chance. These fluctuations can have little or no biological reality and must, in part at least, be determined by fluctuations in public and professional interest, causing many lumps that would not otherwise have been diagnosed as breast cancer to be removed, classified as “locally invasive” and counted among the incidence data.
It has been noted that although 15 – 20 percent of women with cancer in one breast would be found to have cancer in the opposite breast if the opposite breast were immediately biopsied, the rate of clinical appearance of cancer in the opposite breast, if no biopsy is taken, is only 0.5 percent a year over many subsequent years.
It was once thought that one of the causes of breast cancer was the contraceptive hormone pill. However, the only well established link with breast cancer incidence is a low but definite correlation with the age at first pregnancy.
The earlier the pregnancy, the lower the risk of cancer in middle and old age, so any effective form of contraception will appear to increase risk (including abstinence!) a nice example of the care that needs to be taken in postulating as to causes.
Food Additives and Chemical Carcinogens
The importance of gross aspects of nutrition were suggested many years ago in experiments on mice which showed that restricting the intake of food without modifying the proportions could reduce the incidence of spontaneous tumors of the breast and lung and of a variety of cancer produced experimentally by known carcinogens. This work has since been repeated on many occasions.
More interest could have been aroused by this work if the freely fed mice had been described as obese instead of the mice on the restricted diet being described as small! Although work has been done on human studies, the results are confounded because weight is associated with a variety of social and behavioral characteristics which are difficult to standardize for.
Food additives is an area in which considerable effort has been expended in laboratory studies but effects on humans are unknown. However, their effects could be more serious than experiments might suggest. Nitrite used as a food additive for example, reaches the stomach in concentrated pockets and reacts chemically in proportion to the square of its concentration.
Because people tend to be conservative in their diets, there is merit in investigating the possible protective action of certain food components such as B-carotene (carrots, dark green leafy vegetables) vitamin E and vitamin C.
If any minor components of diet do indeed materially reduce cancer risk, then their discovery might be of more immediate practical value than the discovery of harmful factors, for prescription may be more acceptable than proscription.
There is no doubt that most human carcinogens have been identified through epidemiological study of occupational exposure. This is not surprising, since industrial chemicals have been used freely in the past unless a few simple tests on animals or a brief experience of their use had shown them to be acutely poisonous.
The hazards that have been recognized thus far tend to be those which increase the relative risk of some particular type(s) of cancer very substantially, and important occupational hazards may quite possibly exist that have not yet been detected because the added risk is small compared with that due to other causes, or because only a few people have been exposed, or simply because a hazard has not been suspected and so not looked for.
It must also be borne in mind that cancer in humans seldom develops until one or more decades after the beginning of exposure to a carcinogen, and it is too soon to be sure whether agents are human carcinogens if they were introduced into industry only during the last 20 years.
Of interest to the current herbicide debate is occupational exposure to phenoxy acid and chlorophenol herbicides (or their impurities) is a reasonably well-established cause of soft tissue sarcomas or perhaps lymphomas.
What significance massive industrial exposure has for the lesser exposure of the general population (or Vietnam veterans) is still being argued. In any event, the problem with any such studies is that the questions they answer are only those already asked.
It would be nice to answer these, but one wonders how far current cancer research may be from even knowing the right questions to ask. This is particularly true in regard to dietary factors which would appear to loom as large as smoking in their total effects.
Because many smokers do not get lung cancer and a few non-smokers do, it is often erroneously assumed that this invalidates the link. Outside of pure mathematics, all proof is stochastic, that is based on probability. Sometimes the probabilities are very high but they are probabilities none the less. There is no absolute proof that the sun will rise tomorrow morning.
Because children are more likely to suffer the effects of smoking in later life the earlier they start, the time may be ripe to take legal action for damages against government authorities who condone the advertising and unmonitored sale of cigarettes to minors, and against sporting bodies whose public image of health and vitality contrasts with their private addiction to the income of tobacco companies. After all, the evidence for smoking causing cancer is much stronger than the evidence reported as being deemed sufficient in a recent highly publicized murder trial.
The other major area of concern is melanoma. This cancer is linked with exposure to sunlight but in no straightforward manner. It is therefore pleasing to note that after four years of argument, a consensus has finally been reached between industry, government and consumers on a standard for sunscreens which incorporates the latest research results (some done locally).
The Causes of Cancer, by Richard Doll and Richard Peto, is published by Oxford University Press.