Finding healthier diets

Professor Stewart Truswell is Boden Professor of Human Nutrition at the University of Sydney. He discussed some modern scientific methods that help explain the role of food components in the aetiology of chronic degenerative diseases.

Researchers in animal nutrition have an easier task than ours. They have clear-cut goals, for example, achieving the maximum weight in farm animals by 12 months at the least cost (or even the finest possible wool, regardless of cost). They work with genetically homogenous subjects, can make 100 per cent of subjects participate, closely control experimental conditions and measure body composition whenever needed by the most direct method.

In human feeding our goals are multiple, complex and not fully agreed upon. Maximum growth for least cost is not a goal with high priority. Common nutritional goals include feeling well, functioning well, looking good and avoiding disease in the future. As well, people expect to enjoy eating the foods that accomplish these things; production and preparation of these foods provides many people with employment.

These different goals are one reason for confusion and controversy about human nutrition. Paediatricians are still concerned primarily to avoid effects of nutritional deficiency on growth and development, though the breast-feeding policy has made this much simpler. In countries like Australia, nutrition research for adults has become focussed on preventing, that is, delaying, the chronic degenerative diseases. The methodology of research has to be different between these different goals - and so does public health management.

Epidemiology is an important tool in researching what components in the diet, and hence what foods, should reduce the risk of chronic diseases. It may start with a traveller's tale about Eskimos, the Masai, the Bushmen, the Japanese or the French, who don't seem to suffer from a disease. This impression needs then to be observed in a more systematic way. The problem often gets broken up into soluble sections. Epidemiological studies may look at easily measured associates or correlates of a disease rather than the disease itself. Serum cholesterol is associated with coronary heart disease so a study of the epidemiology of serum cholesterol is an indirect contribution to the epidemiology of coronary heart disease - as is the epidemiology of smoking and high blood pressure.

In using epidemiology it is essential to first differentiate between a hypothesis and data of association. The data must be analysed to determine if chance, bias or confounding are more likely explanations of the findings. If an association passes these tests it should then be examined against criteria of strength, reproducibility and consistency with other findings.

No one should be eating more or less of a food because of a hypothesis without data, for example, margarine and asthma, or because of an association which has not been confirmed and critically examined, for example, some margarines and coronary heart disease.

When an association been repeatedly found, prevention trials provide the most convincing confirmation that the factor is causal. There are good designs and inadequate designs of dietary trials. One positive trial is encouraging but not enough. Again people can be misled by reports in the media of one positive trial followed by a negative one. The new technique of meta-analysis, appropriately used, is very helpful but is not yet understood by journalists. It is instructive to compare the results of trials for dietary prevention of coronary heart disease, of oat fibre and plasma cholesterol and of antioxidant vitamins and coronary heart disease.

Whether in controlled trials or in experiments of nature, the independent variable in nutrition research is the amount of foods and their components that individual subjects consume. in recent years nutrition researchers have been searching for biomarkers, quantitative biochemical tests, to provide objective confirmation of food component intake which is otherwise subjective, depending on what people say they eat. One of the best is the 24-hour urinary sodium test (best collected over several days) to reflect salt intake.

Doubly labelled water

Invention of the ingenious doubly labelled water method now provides an objective measure of total energy expenditure over a period of about a week, a measurement which up to now has been very tedious and inaccurate. If a person's energy intake is appreciably less than energy expenditure, either they will lose weight or they have understated their food intake. the doubly labelled water method is very expensive but from it a much simpler check on fook intake has been devised. It is physiologically impossible for the food intake of a sedentary worker to be less than about 140 per cent of basal metabolic rate. The basal metabolic rate can be measured directly from oxygen consumption at rest or it can be extimated knowing age, gender and body size. All modern reports of food and energy intake now have to be checked against the individual's basal metabolic rate. This phenomenon, that a percentage of any random selection of people will understate their usual energy intake, affects the intake of all nutrients and other food components. Researchers have to try and estimate whether they are affected proportionately and whether they can be corrected by correcting for understatement of the energy intake.

We are a long way from having all the answers about the role of food components in the aetiology of chronic degenerative diseases but the picture available using these modern scientific methods is much more complete than the sad material fed to the general public by the daily papers, women's magazines and electronic media.