Plasma metabolites, dietary intakes, and breast cancer incidence: a prospective case-cohort study in NutriNet-Santé.

BACKGROUND

Breast cancer (BC) is the most common cancer in females. Combining metabolomic and dietary data provides an opportunity to explore metabolic pathways underlying BC risk, but few prospective studies have examined these relationships.

OBJECTIVES

This study aimed to identify plasma metabolites associated with incident BC, explore diet-metabolite associations, and assess whether metabolites mediate diet-BC associations.

METHODS

Fifty-three plasma metabolites were measured by liquid chromatography-mass spectrometry in females from a prospective case-cohort nested within the French NutriNet-Santé study (2762 subcohort members, 306 total BC cases; median age 53.6 y). Dietary intake was assessed using repeated 24-h dietary records. Principal component (PC) analysis derived metabolite and dietary patterns. Cox models estimated hazard ratios (HRs) for metabolite-BC associations, adjusted for established risk factors. ElasticNet and linear regression identified dietary predictors of metabolites. Mediation analysis evaluated whether metabolites mediated diet-BC associations.

RESULTS

A metabolite pattern enriched in conjugated bile acids (PC04) was positively associated with BC [HR: 1.14; 95% confidence interval (CI): 1.00, 1.30], including glycohyocholic acid (HR: 1.10; 95% CI: 1.03, 1.17), glycoursodeoxycholic acid (HR: 1.03; 95% CI: 1.00, 1.06), and taurocholic acid (HR: 1.04; 95% CI: 1.00, 1.09). A pattern enriched in unconjugated bile acids was inversely associated (HR: 0.86; 95% CI: 0.75, 0.99), as was linoleyl-carnitine (HR: 0.81; 95% CI: 0.68, 0.96) and, among premenopausal females, lysophosphatidylcholine (18:1) (HR: 0.60; 95% CI: 0.38, 0.96). Several dietary factors were associated with these metabolites; for example, vitamin K intake was inversely associated with glycohyocholic acid (β: ‒0.12 per standard deviation, 125.1 μg/d) and PC04 (β: ‒0.07). Mediation analysis indicated vitamin K intake was inversely associated with BC (HR: 0.82; 95% CI: 0.68, 0.93), with reductions in glycohyocholic acid and PC04 each explaining ∼4% of this association.

CONCLUSIONS

BC incidence was positively associated with conjugated bile acids, including the novel glycohyocholic acid, and inversely associated with unconjugated bile acids, suggesting a potential role for bile acid conjugation in carcinogenesis. Multiple dietary factors were linked to these metabolites, highlighting potential metabolic pathways connecting diet and BC. This trial was registered at clinicaltrials.gov as NCT03335644.