CAUSES OF CANCER
People who migrate from low to high organ-specific cancer incidence countries tend to acquire the cancer incidence profile of their new country within one to three generations; for example, breast cancer risk among Asian American women with grandparents born in the United States is actually higher than for white U.S. women. These data suggest that the large international differences in cancer incidence rates are caused by environment and lifestyle differences between countries, rather than ethnic-specific differences in genetic susceptibility. Substantial changes in cancer incidence within countries over time, such as the rise in U.S. lung cancer incidence in the last 50 years (while stomach cancer incidence has fallen steadily) also suggest changes in the environmental determinants of these diseases. Thus, a high proportion of cancer cases are due to environment and lifestyle, and cancer epidemiologists have tried to establish these causative factors over the last 60 years.
Smoking is the major modifiable cause of cancer in many countries, estimated to cause one third of cancer deaths in the United States and 21% of cancer deaths worldwide. While smoking prevalence has fallen in the United States over the past several decades, it has increased in many other countries, including the world’s largest, China; the number of cancer deaths from tobacco is likely to be much higher in the 21st century than in the 20th century. Although lung cancer dominates the spectrum of smoking-related cancers, cancers at many other anatomic sites have been convincingly related to smoking, including those of the oropharynx, larynx, oesophagus, stomach, liver, pancreas, kidney and ureter, cervix, bladder, and colon/rectum, as well as acute myeloid leukaemia. Second-hand (or passive) smoking has been associated with lung cancer. Smoking cessation results in rapidly reduced rates of lung cancer; cessation before age 30 reduces lifetime risk by more than 90% compared with continuing tobacco smoking. However, the latency between smoking initiation and cancer occurrence is several decades, so the emergence of large numbers of smoking-related cancers occurs decades after smoking prevalence increases.
A substantial fraction of cancers, estimated to be about 16% globally, are caused by infectious agents, particularly in less developed countries (estimated to be ≈ 23%). On the other hand, only 3% of cancers were estimated to be due to infections in the United Kingdom. The major infectious agents are Helicobacter pylori (stomach cancer), human papillomavirus (HPV [cervical cancer]), and hepatitis B and C viruses (liver cancer).
Perhaps no field of epidemiology has been as complex and controversial as the relationship between diet and cancer. Authoritative sources have estimated that a large fraction of cancer incidence is associated with dietary factors; however, a definitive understanding of the mechanisms involved has been difficult, mainly because of difficulties in obtaining valid estimates of intake of specific foods and nutrients and dietary patterns. Foods or nutrients associated with one type of cancer may not be associated with other types of cancer; it is also difficult to distinguish between true etiologic differences between cancer at different sites and false-positive results for a specific cancer type. The World Cancer Research Fund has issued two large-scale summaries of the evidence, most recently in 2007, and the critical recommendations on diet were to: (1) limit consumption of energy-dense foods (avoid sugary drinks); (2) eat mostly foods of plant origin; (3) limit intake of red meat, and avoid processed meat; and (4) limit consumption of salt, and avoid mouldy cereals (grains) and legumes.
Part of the difficulty in making generalizations about the percent of cancer attributable to diet is the relationship between diet and body size, and whether to “count” this influence as a “dietary” or “anthropometric” factor. Well-nourished, more affluent societies tend to have taller and heavier populations, both factors positively related to risk of cancer at several sites.
Body Weight and Height
The role of overweight and obesity as risk factors for a wide variety of cancers has become more apparent in the last decade. Even this relationship can be complex. For example, obesity is inversely related to the incidence of premenopausal breast cancer but positively related to postmenopausal breast cancer. Greater height is associated with modest increases in the risk of cancer at many sites. The secular trends of increasing height and weight in more affluent societies explain a substantial fraction of the increases in cancer rates over time.
An independent role for physical activity in cancer causation is difficult to dissect from the fact that sedentary lifestyles are associated with overweight and obesity. Physical activity has occupational, recreational, and activities of daily living components and can be difficult to measure in epidemiologic studies. A majority of studies suggest that higher levels of physical activity are associated with lower colorectal cancer risk, independent of body weight, and many studies have also reported an inverse relationship with breast cancer. Whether or not this is due to uncontrolled confounding by body weight, a more active lifestyle is a key component of prevention of weight gain, and thus is recommended for prevention of at least these two types of cancer.
Consumption of alcohol is associated with cancer at several sites, notably cancers of the mouth, pharynx and larynx, oesophagus, liver, breast, and rectum. Globally, about 5% of cancers were estimated to be due to alcohol, and for the United Kingdom the estimate is 4%. For breast cancer, the risk appears to increase linearly with increasing alcohol consumption, whereas for the aerodigestive cancers, such as oesophageal cancer, the risk is most apparent for heavy drinking. Most authorities recommend no more than one alcoholic drink per day for women, and two for men.
Ionizing radiation is a clear cause of leukaemia and thyroid cancer; however, minimization of exposure means that it is a relatively infrequent cause of cancer, estimated to account for about 2% of cancers in Western societies.
Ultraviolet radiation is the major cause of non-melanoma skin cancers and melanoma, and is estimated to account for 3 to 4% of cancers in the Western world. Avoidance of sunburns in early life may be particularly important for reducing the risk of melanoma.
Starting with the work of Sir Perceval Potts on scrotal cancer in chimney sweeps, it has been appreciated that cancer at certain sites is more frequent in specific occupations. A wide variety of industrial chemicals, more common in certain occupations, are associated with risk of specific cancers. In advanced economies, efforts to minimize exposure usually ensure that risks are minimal, although it was estimated that almost 4% of cancers in the United Kingdom are still due to occupational exposures. In less developed economies, risks may be higher, owing to migration of “dirty” industries to less regulated environments, and a lower level of appreciation of the risks and need to protect workers from carcinogenic exposures.
Use of oral contraceptives increases the relative risk of breast cancer by about 30% while women are taking them; however, since most women taking oral contraceptives are in the 15- to 40-year age group, when breast cancer incidence rates are low, the absolute number of additional cases is small. Ten or more years of use of oral contraceptives reduces ovarian cancer risk by more than 50%, as well as reducing risk of endometrial cancer.
Use of postmenopausal hormones, particularly those combining oestrogen and progestins, increases risk of breast cancer, although risk declines quickly after cessation and decreases risk of colon cancer. After the publication of the Women’s Health Initiative findings of a positive association, use of postmenopausal hormones declined dramatically, and the incidence rate of breast cancer (mostly oestrogen receptor positive) declined over the next few years, one of the few examples in which a rapid change in prevalence of a risk factor can be linked to a short-term change in cancer incidence.
A variety of reproductive factors have been associated with risk of cancer in women. These include early age at menarche, late age at first birth, nulliparity or low parity, and late age at menopause, short duration of lactation (all increase risk of breast cancer), and nulliparity or low parity (increased risk of ovarian and endometrial cancer).