PLoS One 2014 5;9(6):e98647. Epub 2014 Jun 5.
Dietary Exposure Assessment Group, International Agency for Research on Cancer, Lyon, France.
Background: Compared to food patterns, nutrient patterns have been rarely used particularly at international level. We studied, in the context of a multi-center study with heterogeneous data, the methodological challenges regarding pattern analyses.
Methodology/principal Findings: We identified nutrient patterns from food frequency questionnaires (FFQ) in the European Prospective Investigation into Cancer and Nutrition (EPIC) Study and used 24-hour dietary recall (24-HDR) data to validate and describe the nutrient patterns and their related food sources. Associations between lifestyle factors and the nutrient patterns were also examined. Principal component analysis (PCA) was applied on 23 nutrients derived from country-specific FFQ combining data from all EPIC centers (N = 477,312). Harmonized 24-HDRs available for a representative sample of the EPIC populations (N = 34,436) provided accurate mean group estimates of nutrients and foods by quintiles of pattern scores, presented graphically. An overall PCA combining all data captured a good proportion of the variance explained in each EPIC center. Four nutrient patterns were identified explaining 67% of the total variance: Principle component (PC) 1 was characterized by a high contribution of nutrients from plant food sources and a low contribution of nutrients from animal food sources; PC2 by a high contribution of micro-nutrients and proteins; PC3 was characterized by polyunsaturated fatty acids and vitamin D; PC4 was characterized by calcium, proteins, riboflavin, and phosphorus. The nutrients with high loadings on a particular pattern as derived from country-specific FFQ also showed high deviations in their mean EPIC intakes by quintiles of pattern scores when estimated from 24-HDR. Center and energy intake explained most of the variability in pattern scores.
Conclusion/significance: The use of 24-HDR enabled internal validation and facilitated the interpretation of the nutrient patterns derived from FFQs in term of food sources. These outcomes open research opportunities and perspectives of using nutrient patterns in future studies particularly at international level.