Factor analysis in the identification of dietary patterns and their predictive role in morbid and fatal events

Abstract

Objective The purpose was to examine the role of dietary patterns derived from factor analysis and their association with health and disease. Design Longitudinal population study, with measurement of diet (dietary history method), cardiovascular risk factors and a follow-up of 20 years for CHD incidence and 40 years for mortality. Setting Two population samples in rural villages in northern and central Italy. Subjects Men (n 1221) aged 45-64 years were examined and followed up. Results One of the factors identified with factor analysis, run on seventeen food groups, was converted into a factor score (Factor 2 score) and used as a possible predictor of morbid and fatal events. High values of Factor 2 score were characterized by higher consumption of bread, cereals (pasta), potatoes, vegetables, fish and oil and by lower consumption of milk, sugar, fruit and alcoholic beverages. In multivariate analysis, Factor 2 score (mean 0·0061; sd 1·3750) was inversely and significantly associated (hazard ratio for a 1 sd increase; 95% CI) with 20-year CHD incidence (0·88; 0·73, 0·96) and 40-year mortality from CHD (0·79; 0·66, 0·95), CVD (0·87; 0·78, 0·96), cancer (0·84; 0·74, 0·96) and all causes (0·89; 0·83, 0·96), after adjustment for five other risk factors. Men in quintile 5 of Factor 2 score had a 4·1 years longer life expectancy compared with men in quintile 1. Conclusions A dietary pattern derived from factor analysis, and resembling the characteristics of the Mediterranean diet, was protective for the occurrence of various morbid and fatal events during 40 years of follow-up. © 2012 The Authors.