Prevalence of fermented foods in the Dutch adult diet and validation of a food frequency questionnaire for estimating their intake in the NQplus cohort

BACKGROUND: Humans have a long history of consuming fermented foods. However, their prevalence in human diets remains largely undetermined, and there is a lack of validated dietary assessment tools assessing the intake of different fermented products. This study aimed to identify fermented foods consumed in The Netherlands and determine the relative validity of a food frequency questionnaire (FFQ) compared to multiple 24-h recalls for estimating their intake. METHODS: The validation population consisted of 809 participants (53.1 ± 11.9 years) from a Dutch observational cohort (NQplus) who completed a FFQ and multiple 24-h recalls. Fermented foods from the FFQ and recalls were identified and aggregated into conventional food groups. Percent difference in mean intakes, quintile cross-classification, Spearman’s correlations, and Bland-Altman analyses were used to evaluate the agreement between the two dietary assessment methods. RESULTS: Approximately 16–18% of foods consumed by this population were fermented, and a further 9–14% were dishes containing a fermented ingredient. Fermented foods with the highest consumption included coffee (~ 453 g/day;~ 0.5% of daily energy intake), yoghurts (~ 88 g/day;~ 2.2%), beer (~ 84 g/day;~ 1.7%), wholegrain bread (~ 81 g/day;~ 9.4%), wine (~ 65 g/day;~ 2.7%), and cheese (~ 32 g/day;~ 5.0%). Mean percent difference between the FFQ and recalls was small for fermented beverages (coffee), breads (brown, white, wholegrain, rye), and fermented dairy (cheeses) (0.3–2.8%), but large for buttermilk and quark (≥53%). All fermented food groups had > 50% of participants classified into the same or adjacent quintile of intake (58%-buttermilk to 89%-fermented beverages). Strong Spearman’s correlations (crude/energy-adjusted r(s) ≥ 0.50) were obtained for fermented beverages (coffee, beer, wine), cereals/grains (wholegrain bread), and dairy (yoghurts). For ‘other bread’, quark, and buttermilk, correlations were low (r(s) < 0.20). Bland-Altman analyses revealed good agreement for fermented beverages (coffee, beer), breads (brown, wholegrain, rye, other), pastries, chocolate, and fermented dairy (cheeses) (mean difference: 0.1–9.3). CONCLUSIONS: Fermented food groups with acceptable or good validity across all measures included commonly consumed foods in The Netherlands: fermented beverages (coffee), wholegrain and rye bread, and fermented dairy (cheeses). However, for less frequently consumed foods, such as quark and buttermilk, the levels of agreement were poor and estimates of intake should be interpreted with caution. This report provides the basis for developing a FFQ specific for fermented foods. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s40795-020-00394-z.