New challenges in dietary pattern analysis: combined dietary patterns and calorie adjusted factor analysis in type 2 diabetic patients

Background: Some variability for dietary pattern analysis due to subjective procedures (e.g. arbitrary food categorization and number of factors extraction) was reported. The aim of this study was to present or design a new approach to challenge the conventional dietary pattern analysis through new classification of dietary patterns according to the possibility of the high adherence to more than one dietary pattern and calorie adjusted factor extracting. Methods: This cross-sectional study conducted on 734 type2 diabetic patients. Factor analysis defined three major dietary patterns (Western like, Asian like and Traditional like) and the associations of each pattern were assessed with glycemic control and lipid profiles among tertiles of each pattern. In order to compare variables in highest tertile of three defined dietary patterns, eight new different groups were classified according to the high adherence to one or more patterns and ANOVA and ANCOVA were used to compare them. Also, calorie adjusted factor extracting were done to find out if the same factor loadings would be extract. Results: Among three major dietary patterns, only Western like showed a significant association with fasting blood sugar (p=0.03, 12.49±5.99), serum total cholesterol (p=0.02, 8.71± 3.81) and LDL cholesterol (p=0.04, 5.04± 2.40). While comparison of new classified patterns, showed no significant differences, except a high blood glucose in Western like- Asian like versus traditional like dietary pattern (p=0.04). Also, calorie adjusted factor extracting showed different factor loadings. Conclusions: Results showed that the conventional dietary pattern analysis method may have substantial limitations in interpreting the results and may lead to inappropriate conclusions.