RESUMEN
PURPOSE: A biomarker for dietary flavonoid intake from fruit and vegetables (FlavFV) is needed to elucidate the relevance of flavonoids from these sources for the prevention of chronic diseases. Urinary hippuric acid (HA)-a major metabolite of flavonoids-is promising in this respect as it was shown to satisfyingly indicate fruit and vegetable consumption in different age groups. Therefore, we validated urinary HA as a biomarker for intake of FlavFV. METHODS: Analyses included data from 287 healthy adolescents of the DONALD Study (aged 9-16 years) for whom a minimum of two pairs of HA measurements from 24-h urine samples (test method) and FlavFV intake estimated from 3-day weighed dietary records (reference method) existed. Agreement between both methods was assessed by Spearman correlation and cross-classification analyses. Possible confounders of the association were identified by linear regression models. Analyses were performed using a split-sample approach allowing for consecutive exploration (n = 192) and confirmation (n = 95) of results. RESULTS: Agreement between urinary HA excretion and FlavFV intake was moderate according to correlation analysis in the exploratory sample (r unadjusted = 0.47, P < 0.0001). Yet, 79 % of the subjects were classified into same/adjacent quartiles, and only 5 % were misclassified into opposite quartiles. These findings were corroborated by analyses in the confirmatory sample (r unadjusted = 0.64; 88 % in same/adjacent vs. 4 % in opposite quartiles). Body surface area (BSA) was the only relevant covariate in the exploratory sample, and its adjustment improved cross-classification estimates in both subsamples. CONCLUSIONS: BSA-adjusted 24-h urinary HA excretion represents a suitable biomarker of habitual FlavFV intake in healthy adolescents.