Dietary patterns are associated with body mass index in multiethnic women

Abstract

This cross-sectional study investigated the relationship between dietary patterns and body mass index among 514 women with different ethnic backgrounds who completed a validated food-frequency questionnaire. An exploratory factor analysis with orthogonal rotation started with 23 food items and resulted in four factors that accounted for 93% of the total variance. Confirmatory factor analysis with the 16 items that had factor loadings of at least 0.60 validated the four dietary patterns. The most significant dietary pattern, 'meat,' was characterized by high intake of processed and red meats, fish, poultry, eggs, fats and oils, and condiments. The 'vegetable' pattern loaded high on different vegetables, whereas the third pattern named 'bean' was high in legumes, tofu and soy protein. The major components of the 'cold foods' pattern were fruit, fruit juice and cold breakfast cereals. Although the 'meat' pattern was predominant among Hawaiians and the 'bean' pattern very common among Chinese and Japanese women, factors two and four were not related to ethnicity. After adjustment for daily energy intake, the 'meat' pattern was positively associated with body mass index (r = 0.17, P = 0.0001), whereas the other three patterns showed negative relationships to body mass index (r = -0.076, P = 0.084, r = -0.13, P = 0.003, and r = -0.13, P = 0.003) for vegetables, beans and cold foods, respectively. The associations were similar in direction and magnitude for all ethnic groups. The study results support the ideas that choosing the right foods may be important in weight control and that food-based dietary patterns may be useful in dietary counseling.