Frontiers in Nutrition
School of Medical and Health Sciences / Institute for Nutrition Research
MINECO (Spain, PCIN-2017-076) Instituto de Salud Carlos III European Regional Development Fund's Away, Europe and the Generalitat de Catalunya's Agency AGAUR (2017SGR1546) Danish Cancer Society, Knæk Cancer 2012 Møllerske Støttefond (grant no 10619) FORMAS (DNR 2016-00314) National Agency for Research and Development (ANID)/Food and Nutrition Doctoral Program/DOCTORADO BECAS CHILE/2019 – 72200061 Miguel Servet program (CPII20/00009), Institute of Health Carlos III [co-funded by the European Social Fund (ESF) – ESF investing in your future]
Flavonoids are bioactive plant compounds that are widely present in the human diet. Estimating flavonoid intake with a high degree of certainty is challenging due to the inherent limitations of dietary questionnaires and food composition databases. This study aimed to evaluate the degree of reliability among flavonoid intakes estimated using four different approaches based on the two most comprehensive flavonoid databases, namely, United States Department of Agriculture (USDA) and Phenol Explorer (PE). In 678 individuals from the MAX study, a subcohort of the Diet, Cancer and Health-Next Generations cohort, dietary data were collected using three 24-h diet recalls over 1 year. Estimates of flavonoid intake were compared using flavonoid food content from PE as (1) aglycones (chromatography with hydrolysis), (2) aglycones transformed (converted from glycosides by chromatography without hydrolysis), (3) as they are in nature (glycosides, aglycones, and esters), and 4) using flavonoid content from USDA as aglycones (converted). Spearman's intra-class correlation (ICC) coefficient and weighted kappa (K) coefficient were calculated for the reliability analysis. When comparing PE total aglycones to USDA total aglycones, there was a moderate reliability when a continuous variable was used [ICC: 0.73, 95% confidence interval (CI): 0.70–0.76] and an excellent reliability when flavonoid intake was modeled as a categorical variable (K: 0.89, 95% CI: 0.88–0.90). The degree of reliability among all methods of estimated flavonoid intakes was very similar, especially between database pairs, for the flavanol subclass, while larger differences were observed for flavone, flavonol, and isoflavone subclasses. Our findings indicate that caution should be taken when comparing the results of the associations between flavonoid intakes and health outcomes from studies, when flavonoid intakes were estimated using different methods, particularly for some subclasses.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.