Urinary Hypaphorine as a Candidate Marker of FFQ-Derived Legume Intake Within the EAT-Lancet Framework in Free-Living Adults

Abstract

Dietary assessment is affected by recall, reporting, portion-size error, and food-composition inconsistencies. Hypaphorine, a tryptophan-derived betaine alkaloid found in pulses and other foods, has been proposed as a marker of legume exposure, but its performance in free-living populations remains uncertain. In this cross-sectional study of 138 adults, dietary intake was assessed by food-frequency questionnaire, and legume intake was represented by an FFQ-derived variable within the EAT-Lancet food-group framework. Urinary hypaphorine was quantified by targeted hydrophilic-interaction liquid chromatography-mass spectrometry. Associations were examined using Spearman correlations and multivariable linear models with HC3 robust standard errors. Predictive performance was assessed using repeated 10-fold cross-validation. Hypaphorine correlated with legume intake (ρ = 0.240; p = 0.0047). After adjustment for age, sex, BMI, energy intake, and log10-transformed urinary creatinine signal, each 10 g/day increment in legume intake was associated with 19.1% higher hypaphorine (95% CI, 7.2–32.2%; p = 0.0014). Estimation of the FFQ-derived legume-intake variable was limited (cross-validated R2 = 0.061), whereas discrimination between intake above and at or below the sample median was modest (AUC = 0.653). In the exploratory clinical screen, the inverse correlation with leukocyte count did not remain below the 0.05 FDR threshold (q = 0.090). A focused adjusted model yielded an inverse association (β = −1.38 × 103/µL; p = 0.030), but exclusion of the most influential observation attenuated statistical support (p = 0.061). Urinary hypaphorine may provide modest information for ranking FFQ-derived legume exposure but should not be considered a quantitative measure of individual intake without further validation.