Biomarkers of Dietary Blueberry Intake: Insights from an Untargeted Metabolomics Study of Human Urine Using...

Biomarkers of food intake (BFIs) can improve dietary assessment by providing objective measures of intake, yet no validated specific biomarkers are currently available for blueberries. In a randomized, controlled, crossover feeding study, 20 healthy young adults completed a 2-day run-in and then consumed fresh blueberries with standard meals (390 g on day 3; 780 g on day 4). Urine samples were collected at fasting, postprandial, and follow-up time points. Using untargeted dual-column ultrahigh-performance liquid chromatography-high-resolution mass spectrometry (UHPLC-HRMS) metabolomics, together with a multistage feature selection pipeline and analysis of fruit and in vitro digested samples, we identified 28 candidate BFIs, including novel metabolites such as 4-hydroxy-3-prenylbenzaldehyde and kirenol. XGBoost models showed that a three-compound panel comprising 4-hydroxy-3-prenylbenzaldehyde, kirenol, and p-coumaric acid-4-O-glucoside best predicted recent blueberry intake (AUC 0.81) beyond 24 h after intake and showed good specificity relative to other fruits. These findings support further development of BFIs for blueberry intake.