Fig 2 Scree plot for identification of dietary patterns (components) by principal component analysis. Food intakes (g/day) were aggregated into 58 food groups and used as input variables. Factors considered appropriate for patterns shown in table 1 are the three factors with eigenvalues >2