Table 2.

Summary of the included studies using mixture analysis methods

Author, year of publication; title	Location	Year at baseline; study name if available	Study population	Metals or metalloids included	Exposure	Outcomes	Mixture analysis method(s)	Covariates	Key findings
Kupsco et al. 2019; “Prenatal metal concentrations and childhood cardio-metabolic risk using Bayesian Kernel Machine Regression to assess mixture and interaction effects”	Mexico City, Mexico	2007; PROGRESS cohort	548 mother–child pairs	As, Cd, Co, Cr, Cs, Cu, Mn, Pb, Sb, Se, and Zn	Maternal blood samples (2nd trimester)	SBP, DBP, cardio-metabolic risk score, TG, HbA1c, non-HDL-C, leptin, and adiponectin at 4–6 years of age	BKMR	Maternal age, education, socioeconomic status, parity, environmental tobacco smoke, and date of follow-up visit (for HbA1c, global risk score, non-HDL cholesterol, SBP and DBP outcomes only). Birth weight, gestational age, sex, and pre-pregnancy BMI included as covariates in sensitivity analyses	•Inverse joint effect of metal mixture on SBP was observed • Higher Se was associated with lower TG • Sb and As were associated with lower leptin • No interaction among the metals or nonlinear responses detected
Warembourg et al. 2019; “Early-Life Environmental Exposures and Blood Pressure in Children”	Europe (United Kingdom, France, Spain, Lithuania, Norway, Greece)	NA; HELIX European consortium	1,277 mother–child pairs	As, Cd, Co, Cs, Cu, Hg, Mn, Mo, Pb, and Tl (out of 89 prenatal exposures)	Maternal blood samples	SBP and DBP at 6 to 11 years	DSA	Cohort, maternal age, maternal education level, maternal pre-pregnancy BMI, parity, parental country of birth, child age, child sex, and child height in the multi-exposure linear regression models	•The DSA method selected 5 and 2 prenatal exposures for SBP and DBP in childhood, respectively • None of these exposures included prenatal exposure to metals
Zanobetti et al. 2020; Ambient particle component and newborn blood pressure in Project Viva	Massachusetts, USA	1999; Project Viva	1,131 mother-infant pairs	Al, Si, K, Ca, Ti, Fe, Mg, As, Cu, Zn, Br, Pb, V, Ni, and Na (in addition to PM2.5, BC, S, and Cl)	The PM2.5 components measured at the Harvard Supersite linked to each mother’s residential address	SBP and DBP at a mean age of 30 h	LASSO and BKMR	Maternal age, third-trimester maternal BP, race/ethnicity, and educational level; infan’s postnatal age and birth weight, infant state at BP, median neighborhood income, same day temperature, and sine and cosine terms of day of year	•For SBP, the adaptive LASSO selected S, Ni, Zn, and Cl • Using BKMR, these components have a linear dose–response relationship with the outcome, with Ni and Zn showing the highest PIPs • Ni and S were associated with higher SBP, whereas Zn and Cl were associated with lower SBP. Only Ni and Zn were significantly associated with BP • For DBP, the adaptive LASSO selected Ni, Zn, S, Si, As, Cu, and Pb • The BKMR analysis indicated Ni and Zn as the most influential components •Ni was positively associated with DBP, whereas Zn was inversely associated with the outcome. Other associations were suggestive
Howe et al., 2021; “Prenatal metal mixtures and child blood pressure in the Rhea mother–child cohort in Greece”	Heraklion, Greece	2007; Rhea Cohort	176 mother–child pairs	Mg, Co, Se, Mo, As, Cd, Sb, and Pb	Maternal urine samples	SBP and DBP at 4 and 11 years of age, BP change from ages 4 to 11, and elevated BP at 11 years of age	BKMR	Maternal age, maternal education, maternal pre-pregnancy BMI, maternal smoking during pregnancy, child’s sex, child’s age, and child’s height	•Mo and Co were associated with increased SBP and DBP at baseline (age 4). J-shaped associations were identified • Cd was inversely associated with DBP at baseline (age 4) • Co was associated with lower per-year increases in both SBP and DBP from ages 4 to 11 • Mo was associated with lower per-year increases in DBP from ages 4 to 11 • Mg was associated with higher per-year increases in both SBP and DBP from ages 4 to 11, but not with BP at baseline (age 4) • Mo and Pb were associated with BP at age 11 (J-shaped) • A possible synergistic interaction between Mo and Pb was shown for BP at ages 4 and 11
Zhang et al., 2021; “In Utero exposure to heavy metals and trace elements and childhood blood pressure in a US urban, low-income, minority birth cohort”	Boston, USA	2002; Boston Birth Cohort	1,194 mother–child pairs	Pb, Hg, Cd, Se, and Mn	Maternal blood samples (2nd trimester)	Child SBP and DBP between 3 and 15 years of age	BKMR	Maternal age, at delivery, race/ethnicity, educational level, pre-pregnancy body mass index, and cigarette smoking history	•No joint association observed • The hierarchical variable selection indicated that essential elements (highest conditional PIP for Se) were more strongly associated with SBP than heavy metals (largest conditional PIP for Pb) • In the BKMR individual analysis, Se and Mn were inversely associated with child SBP percentiles; no association found for DBP • In utero exposure to Mn and Cd interacted with each other in relation to child blood pressure

Al, aluminum; As, arsenic; BC, black carbon; BKMR, Bayesian kernel machine regression; BMI, body mass index; BP, blood pressure; Br, bromine; Ca, calcium; Cd, cadmium; Cl, chlorine; Co, cobalt; Cs, cesium; Cu, copper; DBP, diastolic blood pressure; DSA, deletion-substitution-addition algorithm; Fe, iron; Hg, mercury; K, potassium; LASSO, Least Absolute Shrinkage and Selection Operator; Mg, magnesium; Mn, manganese; Mo, molybdenum; NA, not applicable; Na, sodium; Ni, nickel; Pb, lead; PIP, posterior inclusion probability; PM2.5, particle mass with diameter < 2.5 μm; S, sulfur; Sb, antimony; SBP, systolic blood pressure; Se, selenium; Si, silicon; Ti, titanium; Tl, thallium; V, vanadium; Zn, zinc