Validation of a food frequency questionnaire to determine vitamin D intakes using the method of triads

• Published in: Journal of human nutrition and dietetics 2016-04, Vol.29 (2), p.255-261
• Main Authors: Weir, R. R., Carson, E. L., Mulhern, M. S., Laird, E., Healy, M., Pourshahidi, L. K.
• Format: Article
• Language: English
• Subjects: Adolescent; Adult; Biomarkers - blood; Chromatography, Liquid; Diet; Diet Records; Dietary supplements; Energy Intake; Female; food frequency questionnaire; Humans; Male; method of triads; Middle Aged; Nutrition Assessment; Nutritional Status; Questionnaires; Reproducibility of Results; Sample Size; Seasons; Surveys and Questionnaires; Tandem Mass Spectrometry; triangulation; validation; Vitamin D; Vitamin D - administration & dosage; Vitamin D - blood; Young Adult
• ISSN: 0952-3871; 1365-277X
• DOI: 10.1111/jhn.12328

Background Dietary sources of vitamin D (both natural and fortified) are increasingly contributing to consumers’ vitamin D intake and status. Therefore, the present study aimed to validate a vitamin D food frequency questionnaire (FFQ) for the assessment of habitual vitamin D intake. Methods A total of 49 apparently healthy consenting adults (aged 18–64 years) from the local community were sampled at the end of winter. Dietary intakes were recorded using a 4‐day weighed food record (4d‐WFR) and a 17‐item FFQ based on foods known to contribute to dietary vitamin D intake. Fasting vitamin D status was quantified by serum 25‐hydroxyvitamin D [25(OH)D] using liquid chromatography tandem mass spectrometry. The method of triads was applied using these three measurements to determine the overall validity of the FFQ. Results Vitamin D intakes from 4d‐WFR ranged between 0.42 and 31.65 μg day−1, whereas intakes determined from the FFQ ranged from 1.03 to 36.08 μg day−1. Serum 25(OH)D concentrations ranged between 12.89 and 279.00 nmol L−1. The mean (SD) difference between the FFQ and 4d‐WFR was +1.62 ( 3.86). There were strong correlations between the vitamin D intake estimated by the FFQ and that from the 4d‐WFR (r = 0.562) and also with serum 25(OH)D concentrations (r = 0.567). Vitamin D intake estimated from the 4d‐WFR was also strongly correlated with serum 25(OH)D concentrations (r = 0.411). The overall validity coefficient calculated using the method of triads was high (0.881). Conclusions The vitamin D FFQ has been validated for use in future studies aiming to assess habitual vitamin D intake.