Preterm Birth and Metal Mixture Exposure among Pregnant Women from the Navajo Birth Cohort Study

Abstract

Background: Preterm birth (PTB), defined as birth before 37 wk gestation, is associated with hypertension, diabetes, inadequate prenatal care, unemployment or poverty, and metal exposure. Indigenous individuals are more likely to have maternal risk factors associated with PTB compared with other populations in the United States; however, the role of environmental metals on PTB among pregnant Indigenous women remains uncertain. Previous research identified associations between PTB and individual metals, but there is limited investigation on metal mixtures and this birth outcome.

Objectives: We used a mixtures analysis framework to investigate the association between metal mixtures and PTB among pregnant Indigenous women from the Navajo Birth Cohort Study (NBCS).

Methods: Maternal urine and blood samples were collected at the time of study enrollment and analyzed for metals by inductively coupled plasma dynamic reaction cell mass spectrometry. Bayesian Profile Regression was used to identify subgroups (clusters) of individuals with similar patterns of coexposure and to model association with PTB.

Results: Results indicated six subgroups of maternal participants with distinct exposure profiles, including one group with low exposure to all metals and one group with total arsenic, cadmium, lead, and uranium concentrations exceeding representative concentrations calculated from the National Health and Nutrition Examination Survey (NHANES). Compared with the reference group (i.e., the lowest exposure subgroup), the subgroup with the highest overall exposure had a relative risk of PTB of 2.9 times (95% credible interval: 1.1, 6.1). Exposures in this subgroup were also higher overall than NHANES median values for women 14-45 years of age.

Discussion: Given the wide range of exposures and elevated PTB risk for the most exposed subgroups in a relatively small study, follow-up investigation is recommended to evaluate associations between metal mixture profiles and other birth outcomes and to test hypothesized mechanisms of action for PTB and oxidative stress caused by environmental metals. https://doi.org/10.1289/EHP10361.

Figure 1.

Map of the Navajo Nation (darker gray shaded area) located in the Southwestern United States. The lighter, yellow shaded areas on and adjacent to the Navajo Nation represent previous uranium mining areas.

Figure 2.

Classification of maternal Navajo Birth Cohort Study (NBCS) participants into exposure clusters (ECs) by (A) quantiles of measured metal and metalloid biomarkers and (B) relative to median metal biomarker concentrations measured in NHANES. (A) Each column represents an individual element and each row represents a distinct subgroup (or EC). The numeric values in each box represents a quantile score (1 is the lowest, and 4 is the highest). iHg, THg, and Mn were classified as not detected or detected. Asterisks indicate that that particular element was classified into tertiles. (B) Each column represents an individual element and each row represents a distinct subgroup (or EC). The colors and symbols correspond to the NBCS median concentration relative to the NHANES median for the same analyze; green ($<$) indicates that the NBCS median is less than the NHANES median; yellow ($\sim$) indicates that the NBCS median is similar to the NHANES median; and red ($>$) indicates that the NBCS median is greater than the NHANES median. Seven analytes were not visualized here because the variable selection method suggested very low contribution for the observe clustering outcome; this included blood Cd, Mn, Pb, and Se and serum Cu, Se, and Zn. See Table 3 for summary statistics. Note: Ba, barium; Cd, cadmium; Ce, cesium; Co, cobalt; Cu, copper; D, detected; iHg, inorganic mercury; NHANES, National Health and Nutrition Examination Survey; Mg, mercury; Mn, manganese; Mo, molybdenum; ND, not detected; Pb, lead; Sb, antimony; Se, selenium; Sn, tin; Sr, strontium; Tl, thallium; THg, total mercury, U, uranium; W, tungsten; Zn, zinc.

Figure 3.

Pseudo profiles generated from the Bayesian Profile Regression analysis for maternal Navajo Birth Cohort Study participants (2013–2018). The figure illustrates the posterior distribution of PTB risk for individual elements (urinary total As and urine U), and combinations of urine total $\mathrm{As}+\mathrm{U}$, and urine total $\mathrm{As}+\mathrm{U}+\mathrm{Hg}$ (urine Hg, total blood Hg, and inorganic blood Hg). The posterior distribution of predicted PTB is plotted by the quantile score of each metal or metal mixture. The black line overlapping each bean plot represents the mean predicted probability of PTB. Numeric data represented in this figure are also provided in Table S15. Note: As, arsenic; Hg, mercury; PTB, preterm birth; U, uranium.