Relative Validity and Reproducibility of a Food Frequency Questionnaire for Assessing Dietary Intakes in a Multi-Ethnic Asian Population Using 24-h Dietary Recalls and Biomarkers.

The assessment of diets in multi-ethnic cosmopolitan settings is challenging. A semi-quantitative 163-item food frequency questionnaire (FFQ) was developed for the adult Singapore population, and this study aimed to assess its reproducibility and relative validity against 24-h dietary recalls (24 h DR) and biomarkers. The FFQ was administered twice within a six-month interval in 161 adults (59 Chinese, 46 Malay, and 56 Indian). Fasting plasma, overnight urine, and 24 h DR were collected after one month and five months. Intra-class correlation coefficients between the two FFQ were above 0.70 for most foods and nutrients. The median correlation coefficient between energy-adjusted deattenuated FFQ and 24 h DR nutrient intakes was 0.40 for FFQ1 and 0.39 for FFQ2, highest for calcium and iron, and lowest for energy and carbohydrates. Significant associations were observed between urinary isoflavones and soy protein intake (r = 0.46), serum carotenoids and fruit and vegetable intake (r = 0.34), plasma eicosapentaenoic acid and docosahexaenoic acid (EPA + DHA) and fish/seafood intake (r = 0.36), and plasma odd chain saturated fatty acids (SFA) and dairy fat intake (r = 0.25). Associations between plasma EPA + DHA and fish/seafood intake were consistent across ethnic groups (r = 0.28-0.49), while differences were observed for other associations. FFQ assessment of dietary intakes in modern cosmopolitan populations remains feasible for the purpose of ranking individuals' dietary exposures in epidemiological studies.