Perspective Does social class predict diet quality? 1–3 Nicole Darmon and Adam Drewnowski ABSTRACT A large body of epidemiologic data show that diet quality follows a socioeconomic gradient. Whereas higher-quality diets are associ- ated with greater affluence, energy-dense diets that are nutrient-poor are preferentially consumed by persons of lower socioeconomic status (SES) and of more limited economic means. As this review demonstrates, whole grains, lean meats, fish, low-fat dairy products, and fresh vegetables and fruit are more likely to be consumed by groups of higher SES. In contrast, the consumption of refined grains and added fats has been associated with lower SES. Although mi- cronutrient intake and, hence, diet quality are affected by SES, little evidence indicates that SES affects either total energy intakes or the macronutrient composition of the diet. The observed associations between SES variables and diet-quality measures can be explained by a variety of potentially causal mechanisms. The disparity in energy costs ($/MJ) between energy-dense and nutrient-dense foods is one such mechanism; easy physical access to low-cost energy- dense foods is another. If higher SES is a causal determinant of diet quality, then the reported associations between diet quality and bet- ter health, found in so many epidemiologic studies, may have been confounded by unobserved indexes of social class. Conversely, if limited economic resources are causally linked to low-quality diets, some current strategies for health promotion, based on recommend- ing high-cost foods to low-income people, may prove to be wholly ineffective. Exploring the possible causal relations between SES and diet quality is the purpose of this review. Am J Clin Nutr 2008; 87:1107–17. INTRODUCTION Morbidity and mortality rates in industrialized societies fol- low a socioeconomic gradient (1–3). The more disadvantaged groups suffer from higher rates of obesity (4 – 6), diabetes (7, 8), cardiovascular disease (9), osteoporosis (10, 11), dental caries (12), and some forms of cancer (13). All of these diseases have a direct link to nutrition and diet (14). It has been suggested, more than once, that dietary factors may help explain some of the observed social inequities in health (15, 16). The more affluent population subgroups are not only healthier and thinner, but they also consume higher-quality diets than do the poor (17). Diet quality is affected not only by age and sex, but also by occupation, education, and income levels (18 –20)—the conven- tional indexes of socioeconomic status (SES) or social class (21). The different socioeconomic indicators appear to have similar, although independent, effects on nutrition and diet (18 –20, 22). However, a convincing causal relation between SES indicators and diet quality still remains to be established. Given that SES variables are likely to affect all aspects of energy balance, from access to healthy foods to opportunities for physical activity, there is a pressing need to address them directly in the context of epidemiologic research. It may well turn out that the reported associations between diet quality and better health, found in so many epidemiologic studies, may have been confounded by un- observed indexes of socioeconomic status. EVIDENCE OF A SOCIAL GRADIENT IN DIET QUALITY Dietary energy density is one index of the overall quality of the diet (23). Diets high in whole grains, lean meats, fish, and fresh vegetables and fruit have a low energy density (defined as the available dietary energy per unit weight) and a high content of vitamins and minerals (23, 24). In many epidemiologic studies, their consumption has been associated with better health (14). Conversely, diets high in refined grains, added sugars, and added fats tend to be energy-dense but nutrient-poor (24). Such diets have been associated with higher energy intakes and with lower intakes of several micronutrients (23, 24). In epidemiologic stud- ies, their consumption has been associated with higher disease risk and higher mortality rates (14). In some studies, dietary energy density was an independent predictor of obesity and the metabolic syndrome (25). Studies suggest that energy-dense foods and energy-dense diets may predispose the consumer to overeating (26). Palatabil- ity is one explanation. Energy-dense foods, especially mixtures of sugars and fat, tend to be more palatable than foods of low energy density and high water content (27). A reduced volume of energy-dense foods is said to suppress satiation and satiety (26). 1 From INRA, UMR1260, Nutriments Lipidiques et Prévention des Mal- adies Métaboliques, Marseille, France (ND); INSERM, U476, Marseille, France (ND); Univ Aix-Marseille 1, Faculté de Médecine, IPHM-IFR 125, Marseille, France (ND); and the University of Washington Center for Obesity Research and the Nutritional Sciences Program, School of Public Health and Community Medicine, University of Washington, Seattle, WA (AD). 2 Supported by the Agence Nationale de la Recherche, The French Na- tional Research Agency under the Programme National de Recherche en Alimentation et Nutrition Humaine (ANR-05-PNRA, 012), and the National Research Initiative of the USDA Cooperative State Research Education and Extension Service grant 2004-35215-14441. 3 Reprints not available. Address correspondence to N Darmon, Nutrition Research Unit 476 INSERM/1260 INRA, 27 Bd Jean Moulin, 13385 Mar- seille Cedex 05, France. E-mail: nicole.darmon@univmed.fr. Received August 1, 2007. Accepted for publication November 7, 2007. 1107 Am J Clin Nutr 2008;87:1107–17. Printed in USA. © 2008 American Society for Nutrition at Univ of Washington Health Sciences Library on May 13, 2008 www.ajcn.org Downloaded from