Metals and Maternal Glucose Intolerance: Individual and Joint Associations

[摘要] Objective: Studies suggest that essential metal(loid)s within normal levels may help sustain glucose homeostasis, while some non-essential metal(loid)s are linked to disrupted glucose metabolism. However, few studies have examined these associations among pregnant women, and the joint associations of these essential and non-essential metal(loid)s are unclear. In this thesis, we explored both the individual and joint associations between metal(loid)s and gestational glucose levels and compared the findings from different statistical methods.Methods:For Study I and II, we used data from Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Fetal Growth Studies - Singleton cohort, a prospective pregnancy/birth cohort study conducted between July 2009 and January 2013. Overall 2,334 non-obese healthy women were enrolled in the cohort. Our analyses included 1,857 of these women in Study I, and 1,720 in Study II. The exposures of interest for the purpose of this thesis were zinc, selenium, copper, and molybdenum concentrations measured by inductively coupled plasma mass spectrometry using plasma collected during the 1st trimester. The main outcome was glucose levels measured from non-fasting, 50-gram glucose tests from later pregnancy gestational diabetes screening test. In Study I, linear regression models and quantile regression models were fitted for each metal(loid)s in association with gestational glucose levels, adjusting for maternal socio-demographic characteristics, life style factors, and reproductive history. In Study II, three statistical approaches – Bayesian Kernel Machine Regression (BKMR), adaptive Least Absolute Shrinkage and Selection Operator (LASSO), and generalized additive model (GAM) – were used to evaluate the joint associations between the metal(loid)s mixture and glucose levels, adjusting for the same covariates as Study I.For Study III, we used data from Project Viva, a prospective pregnancy/birth cohort in eastern Massachusetts. Concentrations of 11 essential and non-essential metal(loid)s – arsenic, barium, cadmium, cesium, copper, magnesium, manganese, lead, selenium, zinc and mercury – were measured using red blood cells collected in early pregnancy. Glucose levels were measured from non-fasting, 50-gram glucose tests in later pregnancy. BKMR models were applied to model the joint associations between metal(loid)s mixtures and glucose levels, GAM and multivariable linear regression were subsequently fitted to examine the reproducibility of BKMR results.Results: In Study I, we found that higher plasma copper concentrations in early pregnancy were associated with significantly higher glucose levels (i.e., per 50% increase in copper was related to 4.9 mg/dL higher mean glucose level, 95% confidence interval (CI): 2.2, 7.5), and higher plasma molybdenum concentrations were associated with lower glucose levels (i.e. 1.2 mg/dL lower per 50% increase in molybdenum, 95% CI: -2.3, -0.1). In Study II, in addition to these associations, a super-additive interaction between copper and zinc was observed, with consistent results from three methods.In Study III, our BKMR findings suggested a modest nonlinear association between early pregnancy erythrocyte barium and glucose levels (2.1 mg/dL higher mean glucose level comparing women in the 75th to the 25th percentile of barium concentrations, 95% credible interval: -0.2, 4.4; 1.54 mg/dL lower mean glucose level comparing women in the 90th to the 75th percentile of barium concentration, 95% credible interval: -3.3, 0.2). We also observed a modest inverse association between mercury (Hg) and glucose levels (1.9 mg/dL lower mean glucose level comparing women in the 75th to the 25th percentile of mercury concentration, 95% credible interval: -4.2, 0.4), and suggestive interactions for barium with lead, magnesium, and arsenic.Conclusions: We evaluated early pregnancy metal(loid)s mixture and found that some metal(loid)s (i.e. copper, molybdenum, barium, mercury) may alter normal glucose levels in later pregnancy, with different directions of associations. Some metal(loid)s may also interact with each other (i.e. copper and zinc, barium and lead/magnesium/arsenic). If replicated and better quantified in future work among different populations and using different biomarkers, these findings will have important implications for the management of glycemia during pregnancy.

[发布机构] University:Havard University;Department:Population Health Sciences

[效力级别] metal

[关键词] environmental mixtures;metal;gestational glucose levels;Bayesian Kernel Machine Regression;Generalized Additive Model