Background: Chronic non-communicable diseases (NCDs) are a major global
health challenge, with unhealthy diets contributing significantly to
their burden. Metabolomics data offer new possibilities for identifying
nutritional biomarkers, as demonstrated in short-term intervention
studies. This study investigated associations between habitual dietary
intake and urinary metabolites, a not well-studied area. Methods: Data
were available from 496 participants of the population-based MEIA study.
Linear and median regression models examined associations between
habitual dietary intake and metabolites, adjusted for possible
confounders. K-means clustering identified urinary metabolite clusters,
and multinomial regression models were applied to analyze associations
between food intake and metabolite clusters. Results: Using linear
regression models, previously reported associations could be replicated,
including citrus intake with proline betaine, protein intake with urea,
and fiber intake with hippurate. Novel findings include positive
associations of poultry intake with taurine, indoxyl sulfate,
1-methylnicotinamide, and trimethylamine-N-oxide. Milk substitutes were
positively associated with urinary uracil, pseudouridine,
4-hydroxyhippurate, and 3-hydroxyhippurate, and inversely associated
with quinic acid. Dietary fiber intake showed a positive association
with 3-(3-hydroxyphenyl)-3-hydroxypropionic acid and a negative
association with indoxyl sulfate. We identified sucrose and taurine as
key metabolites differentiating metabolite clusters. Multinomial
regression analysis confirmed significantly different dietary
associations across clusters, particularly for fruits, processed meat,
poultry, and alcoholic beverages. Conclusions: This study highlights
established and novel food–metabolite associations, demonstrating the
potential of urinary metabolomics for use as nutritional biomarkers in
individuals from the general population.