The association between food patterns and the metabolic syndrome using principal components analysis: The ATTICA Study.J Am Diet Assoc 2007; 107(6):979-87; quiz 997JA
Dietary habits have been associated with the prevalence of the metabolic syndrome.
The associations between foods or food patterns and the characteristics of the metabolic syndrome were evaluated.
During 2001 to 2002, 1,514 men (aged 18 to 87 years) and 1,528 women (aged 18 to 89 years) without any clinical evidence of cardiovascular disease were randomly enrolled, from the Attica region in Greece.
MAIN OUTCOME MEASURES
Dietary habits were evaluated using a semiquantitative, food frequency questionnaire. Characteristics of the metabolic syndrome (ie, blood pressure, waist circumference, glucose, triglycerides, and high-density lipoprotein cholesterol) were also measured.
Principal components analysis was applied to extract dietary patterns from 22 foods or food groups. Multivariate regression analysis evaluated the associations between the extracted dietary patterns and characteristics of the metabolic syndrome.
Six components were derived explaining 56% of the total variation in intake. Component 1 was characterized by the consumption of cereals, fish, legumes, vegetables, and fruits (explained variation 19.7%); component 2 was characterized by the intake of potatoes and meat (explained variation 11.7%), component 6 was characterized by alcohol intake (explained variation 4.8%), whereas the other components were mainly characterized by consumption of dairy and sweets. After adjusting for various confounders, component 1 was inversely associated with waist circumference, systolic blood pressure, triglycerides, positively associated with high-density lipoprotein cholesterol levels, and inversely with the likelihood of the metabolic syndrome (odds ratio [OR] 0.87, 95% confidence interval [CI] 0.79 to 0.97), whereas components 2 and 6 were positively correlated with the previous indexes, and the likelihood of having the metabolic syndrome (OR 1.13, 95% CI 1.05 to 1.21 and OR 1.26, 95% CI 1.21 to 1.33).
A dietary pattern that includes cereals, fish, legumes, vegetables, and fruits was independently associated with reduced levels of clinical and biological markers linked to the metabolic syndrome, whereas meat and alcohol intake showed the opposite results.