Health benefits and risk analysis with reference to cancer studies
H P Lee
In epidemiological studies, a major objective is the elucidation of risk factors (i.e., risk-raising factors) and protective factors (i.e., risk-reducing factors) as part of the total understanding of disease etiology. Once the factors are identified, strategies can be devised to either remove the deleterious ones or promote the beneficial ones. Risk analysis is done with the aid of case-control and cohort studies, often referred to as analytical studies.
Risk factors are identified by the odds ratio or the relative risk, which indicates the strength of association between a defined risk factor and the disease. Conversely, if the odds ratio or the relative risk shows that a factor is more likely to occur in control subjects than in case subjects, then that factor may be protective. Case-Control Studies
Case-control studies involve the comparison of case subjects (those with confirmed disease) with matched control subjects (those without disease). Researchers examine the exposure history of both groups for factors that are likely to explain the way the disease is distributed.
Case-control studies are popular because they are very close to the clinical situation of interviewing patients and control subjects. Results can be obtained rather rapidly, and these studies are inexpensive to conduct. Their main drawbacks are their heavy dependence on subject recall and the proper selection of comparable control subjects.
Examples of two case-control studies completed in Singapore are a colorectal cancer study with a total of 628 subjects and a breast cancer study with a total of 606 subjects. In the colorectal cancer study, 203 subjects with histologically confirmed incident cases of colorectal cancer were compared with 425 age-matched hospitalized control subjects. High intake of cruciferous vegetables had a significant protective effect. A high ratio of meat-to-vegetable consumption had a significant predisposing effect. In the breast cancer study, 200 female subjects with histologically confirmed cases of breast cancer were compared with 406 age-matched hospitalized control subjects. Dietary findings were mainly confined to younger premenopausal women, who had greater variations in their diet. Increased risks were seen with high intakes of red meat. Decreased risks were seen with high intakes of beta-carotene and a high ratio of soya to total protein. Cohort Studies
Cohort studies are longitudinal studies that start with apparently disease-free subjects who either have or do not have exposure to certain known risk factors (e.g., people with high versus low caloric intake, people who eat meat versus those who are vegetarians). With time, some of them will develop the disease in question, or even unexpected diseases, because such studies are really open ended.
By observing the onset or incidence of disease in the exposed versus the nonexposed groups (usually, high intake versus low intake), we can estimate the relative risk. Again, the relative risk measures the strength of the association between a factor and the presence of disease in either a positive or a negative direction.
Cohort studies do not depend as much on patients’ memories as do case-control studies. Documentation of exposure histories can be quite objective. However, they take a long time, require large numbers of subjects, and are very expensive. For those reasons, they are not often done, although the results would give greater confidence to the researchers’ conclusions.
Examples of two ongoing cohort studies are the Nurses’ Health Study Cohort and the Singapore Cohort Study. The Nurses’ Health Study Cohort is being conducted by Harvard University, Boston, Massachusetts, and comprises about 120,000 female nurses aged 30-55 years recruited since 1976. These researchers have reported on a number of outcomes, including colorectal and breast cancer. The Singapore Cohort Study was launched in 1993, and researchers hope to recruit 60,000 male and female subjects aged 45-74 years. The general objective is to look at various links between diet and disease, after a reasonable follow-up period of at least 8-10 years. Problems with Dietary Studies
Dietary studies have some inherent problems, among them imprecision in subjects’ dietary histories, small dietary variations, and weak associations. These problems are described below.
Dietary studies are blunt instruments because many epidemiological studies depend on subjects’ recollection of their dietary histories. Direct weighing and recording of food portions, although more accurate, are not useful in these studies and often pose serious logistical problems. However, it is important to remember that group-based studies are more concerned with usual intakes than particular intakes. In epidemiology, our analysis of results depends more on the ranking of subjects (e.g., high versus low consumers) than on the actual intake quantities. Dietary variations tend to be small within homogeneous populations because most people tend to be influenced by others in the same community. Weak associations are indicated by relative risks of 1.2-2.0. Their weakness may be compounded by the lack of accuracy about dietary intake and also by multicollinearity between nutrients, which refers to foods (e.g., vegetables) that contain many nutrients. Because of this problem, rigorous statistical methods are vital in the analysis of the results. Impact of Dietary Factors
Although the strength of association for a particular dietary factor is likely to be weak, the overall impact of dietary factors is much greater. This impact is measured by the population-attributable risk or fraction, which is also influenced by the prevalence of exposure. In other words, although the risk for an individual may be low, the fact that many people are exposed shows that the overall impact in the community can be quite great. Since the 1960s, studies in cancer epidemiology have identified important risk factors such as cigarette smoke and occupational carcinogens (e.g., asbestos, benzene, polycyclic aromatic hydrocarbons, beta-naphthylamine). Dietary factors are now considered to constitute yet another major group that influences the incidence of cancer. These factors may be nutrients or nonnutrients. Some are not present in food in its raw state, but are the projects of subsequent treatment (e.g., cooking, which gives rise to aromatic amines in meats).
The important risk factors implicated in cancer include saturated fats, total calories, heterocyclic amines, and some preserved foods (e.g., Cantonese-style salted fish, preserved vegetables). All of these findings are still inconclusive. The relative risks are in the region of 1.2 to 1.8. We will need further confirmation of these results, perhaps in the form of meta-analysis of all the studies done so far.
One of the most consistent findings so far is that vegetable intake has a protective effect against a number of epithelial cancers (e.g., lung, larynx, oral cavity, esophagus, stomach, large bowel, urinary bladder, female breast). The nutritional factors include vitamins A, C, and E, betacarotene and other carotenoids, fiber, trace elements (e.g., zinc, selenium) and substances such as indoles, phenols, isoflavones, and flavonoids. Carotenoids, vitamins C and E, and selenium are known antioxidants. It is increasingly clear that the protective effect of vegetable consumption cannot be attributed to a particular active agent. Instead, the “cocktail” effect is at work; in other words, the sum total of all agents present in vegetables exerts the beneficial effect. This being the case, it would be more prudent to advise people to eat vegetables regularly rather than to take pills that contain each of the many diverse compounds present in the food.
A recent statement by The World Health Organization (WHO) Working Panel on the Prevention of Colorectal Cancer published in the WHO Bulletin (1995) contained the following recommendations for primary prevention:’ Fat consumption should not exceed 20% of total calories.
A balanced diet should be consumed, including five to eight servings daily of fruits, vegetables, legumes, and whole-grain cereals.
Dietary fiber from all sources should be at least 25 g/day. Consumption of excess calories should be avoided. Tobacco use should be avoided.
People should engage in physical activities on a daily basis.
1. World Health Organization Working Panel on the Prevention of Colorectal Cancer. Prevention of colorectal cancer-guidelines. WHO Bull 1995;73:7-10
Dr. Lee is Professor and Head, Department of Community, Occupational, and Family Medicine, National University of Singapore, Singapore.
Copyright International Life Sciences Institute and Nutrition Foundation Nov 1996
Provided by ProQuest Information and Learning Company. All rights Reserved