Validity and Reproducibility of a Self-Administered Food Frequency Questionnaire for the Assessment of Sugar Intake in Middle-Aged Japanese Adults

We evaluated the validity and reproducibility of estimated sugar intakes using a food frequency questionnaire (FFQ) among middle-aged Japanese adults in the Japan Public Health Centre-Based Prospective (JPHC) study. In subsamples of the JPHC study (Cohorts I and II in multiple areas), we computed Spearman's correlations of FFQ results with urine sugar concentrations and dietary records (DR) for validity; we evaluated correlations between two FFQs for reproducibility. During 1994⁻1998, participants (Cohort I: = 27 [men], = 45 [women]) provided two (spring and fall) 24-h urine samples and completed 7-consecutive-day DR per season (I: = 99, = 113; II: = 168, = 171) and two FFQs (147 food items) at yearly intervals (I: = 101, = 108; II: = 143, = 146). Sugar intakes from FFQ were correlated with urinary sugar (de-attenuated correlations: 0.40; 95%CI: 0.19, 0.58). After adjustment for sociodemographic and lifestyle variables, correlations between FFQ and DR for men and women were 0.57 (0.42, 0.69) and 0.41 (0.24, 0.55) (I) and 0.56 (0.44, 0.65) and 0.34 (0.20, 0.47) (II), respectively. Correlations between FFQs for men and women were 0.63 (0.49, 0.73) and 0.55 (0.41, 0.67) (I) and 0.66 (0.55, 0.74) and 0.63 (0.52, 0.72) (II). In conclusion, our study showed moderate FFQ validity and reproducibility for sugar intake evaluation.