Over fifty years ago nutrition was defined by C. Glen King as the science of food and its relationship to health. This quite accurate definition implies that nutrition is not a single science but a cluster of sciences relating to the production and utilization of food. Furthermore, nutrition, like medicine, is a field for both scientists and practitioners. Nutrition scientists are as diverse as molecular biologists who study nutrient-related gene expression and epidemiologists who track the movement of nutrient-related diseases in populations. The thread that unites them is the study of various aspects of food. The nutritional sciences, in fact, include essentially all biological sciences that can be applied to the study of nutritional problems.

This observation is well illustrated by Volume 12 of the Annual Review of Nutrition. Approximately 45% of the reviews are concerned with the basic sciences, 40% with clinical science, and 15% with epidemiology and public health. Of the ten basic science reviews, three deal with various aspects of the retinoids, including molecular biology of signal transduction, physiological and biochemical processing of vitamin A, and cancer prevention. The other reviews are devoted to the regulation of iron homeostasis, the regulation of enzymes of the urea cycle, dietary effects on biliary lipids, cellular and molecular aspects of adipose tissue development, multisite regulation of cellular energy metabolism, nutritional aspects of collagen metabolism, and the dietary impact of food processing. Seven of the ten use the techniques of molecular biology to elucidate nutrient effects on gene expression.

Of the nine clinical reviews, one deals with the physiological effects of dietary fiber, two are directed toward human milk and lactation, one discusses the role of homocysteine in the pathogenesis of atherosclerosis, and another unravels the pathogenesis of the eosinophilic-myalgia syndrome in patients taking tryptophan as a health food. Two reviews are devoted to vitamins: one focuses on the effects of cobalamin on the nervous system and the other on carotenoids, vitamin E, and vitamin C as anti-cancer agents. Completing the clinical group are a review of the physiology of placental nutrition and a discussion of the nutritional requirements of humans in space.

Three reviews concern epidemiology and public health. The preface by Dr. Gopalan, president of the Nutrition Foundation of India and former president of the International Union of Nutritional Sciences, describes the contribution of nutrition research to the control of undernutrition in India. He recounts the important research discoveries that have aided in the reduction of protein-energy malnutrition, xerophthalmia, pellagra, goiter, lathyrism, and iron deficiency in India during the past three decades. The second review is an epidemiologic study of hypocholesterolemia and cancer, and the third paper discusses the safety and efficacy of fat substitutes and the role of the Food and Drug Administration in monitoring new foods.

The breadth and the variety of technologies used in the nutritional sciences, as revealed by any thorough study of nutritional research (including the range of subjects considered in Volume 12), tend to create varying standards for reaching conclusions in different scientific areas. The precision of measurement is high and the average error of conclusions is fairly low in the basic science investigation of nutritional events. As one examines clinical investigations, generally, and specifically those dealing with nutrition, one finds somewhat less precision in measurements and a larger mean error in the derived conclusions. This finding is partly due to genetic heterogeneity in the human studies and less control over the experimental variables in the investigation.

Finally, as one surveys the results of epidemiologic investigations of nutritional disease, the precision of measurement is lowest, often because of the need to measure life style events that are difficult to document like diet and drug (including alcohol) intake, smoking, exercise, and work environment. Finally, the potential errors in conclusions reached are the sum of the errors in input factors, like diet, and output factors, like disease incidence, that require long-term follow-up for accuracy.

Although epidemiologists freely admit in seminars that associations of input and output variables do not constitute evidence of cause and effect, when an epidemiologic study is viewed in the light of public health practice the tendency is for epidemiologists to use their studies as a basis for public health action. For example, there is currently no acceptable scientific evidence that a diet containing 30% of calories from fat, 10% of calories from saturated fat, and 300 mg of cholesterol per day will extend life by preventing coronary heart disease (CRD) and cancer. Furthermore, the hypothesis that low-fat diets will prevent CHD has been tested mainly in middle-aged males. Yet most cardiologic and health societies the world over, including the US Committee on Diet and Health of the National Academy of Sciences and various expert panels of the National Heart, Lung, and Blood Institute, as well as a WHO study group, have recommended that this low-fat diet become a health standard for all populations over two years of age worldwide. This recommendation is made in the belief that this dietary regimen will improve health and reduce the incidence of noncommunicable diseases even though scientific data are lacking.

This recommendation is particularly erroneous when applied to children who, as a group, are not at risk for the chronic degenerative diseases, who need more fat calories to supply their energy requirement, and who, when placed on low-fat diets, have been shown to be undernourished with respect to other essential nutrients. Within medical practice, however, one can identify some adults at high risk for CHD who have benefited from such a dietary regimen, usually coupled with other measures to reduce risk.

The bottom line is that when data from epidemiologic studies and clinical trials are in conflict, and they often are, a consensus of experts in epidemiology usually recommends a course of public health practice that is scientifically radical rather than conservative. Although the nutritional sciences vary in terms of their study materials and research objectives, there should be a uniform standard for the evaluation of data, the development of conclusions, and the extrapolation of findings to generate nutrition policy.

I thank my associate editors, Dennis M. Bier and Donald B. McCormick, for reviewing manuscripts, all members of the Editorial Committee for their help in assembling the list of topics and authors, and the authors who contributed the excellent reviews that appear in Volume 12. Production editor Joan Cohen in Palo Alto, California, deserves our thanks for her diligence in producing this volume.

Robert E. Olson, Editor