Is consumption of breakfast associated with body mass index in US adults?

OBJECTIVE: To test the hypothesis that breakfast consumption is associated with weight status measured by body mass index in US adults. DESIGN: Analyses of data from the National Health and Nutrition Examination Survey, 1999-2000. PARTICIPANTS/SETTING: Men and women aged > or = 19 years (N=4,218), excluding pregnant and/or lactating women. STATISTICAL ANALYSES PERFORMED: SAS (release 8.1, 2000, SAS Institute Inc, Cary, NC) and SUDAAN (release 8.0.2, 2003, Research Triangle Institute, Research Triangle Park, NC) software were used to calculate sample-weighted means, standard errors, and population percentages of breakfast consumers. Multiple logistic and linear regression models, with controls for covariates, were used to determine the predictability of body mass index from breakfast consumption and from inclusion of ready-to-eat cereal (RTEC) in the breakfast meal. RESULTS: Breakfast consumers were more likely than breakfast nonconsumers to be older, female, white, nonsmokers, regular exercisers, and trying to control their weight. For women, daily energy intake was higher among breakfast consumers than among breakfast nonconsumers; for both men and women, energy intake from fat among RTEC breakfast consumers was significantly lower than among non-RTEC breakfast consumers, whereas energy from carbohydrate among RTEC breakfast consumers was significantly higher than among non-RTEC breakfast consumers. For women, the odds ratios for BMI > or = 25 were lower for breakfast consumers (odds ratio = 0.76) and RTEC breakfast consumers (odds ratio = 0.70) compared with breakfast nonconsumers and non-RTEC breakfast consumers, respectively, after adjusting the models for covariates. When RTEC consumption was added as a covariate, breakfast consumers no longer exhibited significantly lower odds ratios compared with breakfast nonconsumers. Furthermore, regression analyses supported an inverse association between RTEC breakfast consumption and body mass index in women (regression coefficient = -0.37, P<.01) after adjusting for covariates. CONCLUSIONS: When we document the association of breakfast consumption with lower prevalence of overweight and obesity, types of meal should be considered as an important determinant. RTEC breakfast consumption, associated with a desirable macronutrient profile for preventing obesity, predicted weight status in women, but not in men. In addition to sex difference in the association of breakfast consumption and RTEC breakfast consumption with lower prevalence of overweight, the effects of physiological variables and health-related behaviors on the relationship between total and RTEC intake at breakfast and weight status, remain to be established.