Abstract
BACKGROUND: Dairy intake has been implicated in later ovarian aging but mechanism underlying the association is unknown. This study aimed to investigate (1) associations between dairy intake and metabolites previously shown related to anti-müllerian hormone (AMH) decline rate; (2) mediating roles of these metabolites in the prospective association of total dairy consumption with odds of AMH fast decline rate. METHODS: The participants comprised 186 reproductive-aged women randomly selected from the Tehran Lipid and Glucose Study. AMH was measured at baseline (1999-2001) and the 5th follow-up (2014-2017), and dietary data was collected at the second follow-up (2005-2008) using a food frequency questionnaire. Untargeted metabolomics was performed by gas chromatography-mass spectrometry using fasting-serum samples of the second follow-up. We analyzed dairy intake in association with the eight metabolites linked to the higher odds of AMH fast decline rate using linear regression with the Benjamini-Hochberg false discovery correction. Mediatory roles of the metabolites were assessed by bootstrapping. RESULTS: Mean age and BMI of the participants at metabolomics assessment were 44.7 ± 5.87 years and 28.8 ± 4.88 kg/m(2), respectively. Phosphate, branched-chain amino acids (BCAAs), and proline decreased significantly from the first to the third tertile of total dairy intake. Total dairy as a continuous variable inversely associated with phosphate (beta = -0.166; p value = 0.018), valine (beta = -0.176; p value = 0.016), leucine (beta = -0.226; p value = 0.002), proline (beta = -0.219; p value = 0.003), and urea (beta = -0.156; p = 0.035) after accounting for all potential covariates and correction for multiplicity (q-value < 0.1). Fermented dairy showed similar results, but milk did not associate with any of the metabolites. Simple mediation showed significant indirect effects for phosphate, proline, and BCAAs but not urea. Entering the sum of phosphate, proline, and BCAAs as a mediator, the metabolites' total indirect effects were significant [β = -0.12 (95% CIs - 0.26, - 0.04)]. In contrast, the direct association of total dairy intake with the fast decline in AMH was non-significant [β = -0.28 (95% CIs - 0.67, 0.10)]. CONCLUSIONS: Total dairy was inversely associated with AMH decline rate-related metabolites. Inverse association of dairy intakes with the odds of AMH fast decline rate was indirectly mediated by lower phosphate, proline, and BCAAs.