Bayesian Kernel Machine Regression Analysis Research Articles

The effects of chemical mixtures on lipid profiles in the Korean adult population: threshold and molecular mechanisms for dyslipidemia involved.

A scarcity of research assesses the effects of exposure to a combination of chemicals on lipid profiles as well as molecular mechanisms related to dyslipidemia. A cross-sectional study of 3692 adults aims to identify the association between chemical mixtures, including blood and urine 26 chemicals, and lipid profiles among Korean adults (aged ≥ 18) using linear regression models, weighted quantile sum (WQS) regression, quantile g-computation (qgcomp), and Bayesian kernel machine regression (BKMR). In silico toxicogenomic data-mining, we assessed molecular mechanisms linked with dyslipidemia, including genes, miRNAs, pathways, biological processes, and diseases. In the linear regression models, heavy metals, volatile organic compound metabolites, and phthalate metabolites were found to be related to HDL-C, triglycerides, LDL-C, total lipids, and total cholesterol, and significant trends were observed for these chemical quartiles (p < 0.01). The WQS index was significantly linked with HDL-C, triglycerides, LDL-C, total cholesterol, and total lipids. The qgcomp index also found a significant association between chemicals and HDL-C, triglycerides, and total lipids. In BKMR analysis, the overall effect of the chemical mixture was significantly associated with HDL-C, triglycerides, total cholesterol, and total lipids. We found that mixed chemicals interacted with the PPARA gene and were linked with dyslipidemia. Several pathways ("SREBF and miR33 in cholesterol," "estrogen receptor pathway and lipid homeostasis," and "regulation of PGC-1α"), "negative regulation of hepatocyte apoptotic process," "negative regulation of sequestering of triglycerides," "regulation of hepatocyte apoptotic process," and "negative regulation of cholesterol storage," and "abdominal obesity metabolic syndrome" were identified as key molecular mechanisms that may be affected by mixed chemicals and implicated in the development of dyslipidemia. The highest interaction and expression of miRNAs involved in the process of dyslipidemia were also described. Especially, the cutoff levels for chemical exposure levels related to lipid profiles were also provided.

Mixed effects of perfluoroalkyl and polyfluoroalkyl substances exposure on cognitive function among people over 60years old from NHANES.

The relationship between perfluoroalkyl and polyfluoroalkyl substances (PFAS) and cognitive function are inconsistent, and the mixed effects of PFAS on cognitive function are still unclear. We aimed to evaluate the joint effects of PFAS on cognitive function assessed using four tests as follows: the Consortium to Establish a Registry for Alzheimer's Disease Immediate Recall Test (IRT), Delayed Recall Test (DRT), the Animal Fluency Test (AFT), and the Digit Symbol Substitution Test (DSST) in the US elderly. A total of 777 individuals aged ≥ 60 from the National Health and Nutrition Examination Survey (NHANES) 2011-2014 were included in this study. Multivariable logistic regression and Bayesian kernel machine regression (BKMR) were constructed to estimate the overall and the individual effects of PFAS exposure on cognitive function. There were 21.36%, 22.65%, 21.62%, and 21.24% participants with cognitive decline in IRT, DRT, AFT, and DSST, respectively. After multivariable adjustment, perfluorooctanoic acid (PFOA) was inversely associated with cognitive decline in IRT, DRT, and AFT, while no significant association was observed between any other PFAS and cognitive decline. Compared with the lowest quartile, the adjusted odds ratio of cognitive decline with a 95% confidence interval (CI) for the highest quartile of PFOA was 0.33 (95% CI: 0.15-0.69) in IRT, 0.50 (0.26-0.96) in DRT, and 0.45 (0.21-0.95) in AFT. In BKMR analysis, the overall effect of mixtures was significantly protective on cognitive decline in IRT, of which PFOA made the greatest contribution. The consistent protective effect in DRT and DSST was observed when all the chemicals were at their 50th percentile or below it. No significant interaction was observed among PFAS for cognitive function. These findings suggested that PFAS mixture at a low level of current exposure of the US population may have a protective effect on cognitive function.

The association between urine elements and fasting glucose levels in a community-based elderly people in Beijing.

Epidemiological studies have demonstrated that various kinds of urinary element concentrations were different between healthy, prediabetes, and diabetes patients. Meanwhile, many studies have explored the relationship between element concentration and fasting blood glucose (FBG), but the association between joint exposure to co-existing elements and FBG level has not been well understood. The study explored the associations of joint exposure to co-existing urinary elements with FBG level in a cross-sectional design. 275 retired elderly people were recruited from Beijing, China. The questionnaire survey was conducted, and biological samples were collected. The generalized linear model (GLM) and two-phase Bayesian kernel machine regression (BKMR) model were used to perform in-depth association analysis between urinary elements and FBG. The GLM analysis showed that Zn, Sr, and Cd were significantly correlated with the FBG level, under control potential confounding factors. The BKMR analysis demonstrated 8 elements (Zn, Se, Fe, Cr, Ni, Cd, Mn, and Al) had a higher influence on FBG (posterior inclusion probabilities > 0.1). Further intensive analyses result of the BKMR model indicated that the overall estimated exposure of 8 elements was positively correlated with the FBG level and was statistically significant when all creatinine-adjusted element concentrations were at their 65th percentile. Meanwhile, the BKMR analysis showed that Cd and Zn had a statistically significant association with FBG levels when other co-existing elements were controlled at different levels (25th, 50th, or 75th percentile), respectively. The results of the GLM and BKMR model were inconsistent. The BKMR model could flexibly calculate the joint exposure to co-existing elements, evaluate the possible interaction effects and nonlinear correlations. The meaningful conclusions were found that it was difficult to get by traditional methods. This study will provide methodological reference and experimental evidence for the association between joint exposure to co-existing elements and FBG in elderly people.

Influence of combined exposure levels of total arsenic and inorganic arsenic on arsenic methylation capacity among university students: findings from Bayesian kernel machine regression analysis.

The arsenic (As) methylation capacity is an important determinant of susceptibility to As-related diseases. Total As (TAs) or inorganic As (iAs) was reported to associated with As methylation capacity. We measured urinary concentrations of iAs, monomethylarsonic acid (MMA), and dimethylarsinic acid (DMA) by using HPLC-HG-AFS and calculated the primary methylation capacity index (PMI) and secondary methylation capacity index (SMI) in 209 university students in Hefei, China, a non-As endemic area. Volunteers were given a standardized questionnaire asking about their sociodemographic characteristics. Bayesian kernel machine regression (BKMR) analysis was used to estimate the association of lnTAs and lniAs levels with methylation indices (ln%MMA, ln%DMA, lnPMI, lnSMI). The median concentrations of iAs, MMA, and DMA were 1.22, 0.92, and 12.17μg/L, respectively; the proportions of iAs, MMA, and DMA were 8.76%, 6.13%, and 84.84%, respectively. Females had higher %DMA and lower %MMA than males. The combined levels of lnTAs and lniAs showed a decrease in the changes in ln%DMA and lnSMI. With regard to the single exposure level, the lnTAs showed positive correlations with ln%DMA, lnPMI, and lnSMI when lniAs was set at a specific level, while lniAs showed negative correlations with ln%DMA, lnPMI, and lnSMI when lnTAs was set at a specific level; all the dose-response relationships were nonlinear. Our results suggested that combined levels of TAs and iAs play an important role in reducing As methylation capacity, especially iAs, and the reduction only occurs when TAs and iAs are present up to a certain combined level.

Mixtures modeling identifies vitamin B1 and B3 intakes associated with depression

BackgroundWe aim to identify the association between a mixture of vitamin B1, B2, and B3 intakes and depression. MethodsDaily intake of vitamins was measured by a one-day 24 h recall. Multivariate logistic regression, weighted quantile sum (WQS), quantile g-computation (qgcomp), and Bayesian kernel machine regression (BKMR) were used. ResultsOf 9,848 adults included in the final analysis, 4.38% had depression. In the logistic regression model, daily vitamin B1 and B3 intakes were associated with depression, and significant trends were observed for these vitamin intake tertiles (p < 0.001). The WQS index was significantly associated with depression (OR = 0.24, 95% CI: 0.23–0.24). The gqcomp index also found a significant association between a mixture of vitamin B1 and B3 intake and depression (OR = 0.67, 95% CI: 0.44–0.98). Vitamin B1 intake was the most heavily weighed vitamin intake in this model. In BKMR analysis, the overall effects of vitamin B1 and B3 intake mixture were negatively associated with depression. Vitamin B1 and B3 intake showed negative trends and was observed as the most important factor associated with depression. The cutoff levels for B vitamin intake levels related to depression were reported. LimitationsA 24-hour recall and cross-sectional design were used. ConclusionsGiven the rising prevalence of depressive symptoms in Korea, an increase in daily intake of vitamin B1 and/or B3 through regular diets may help to reduce the risk of depression. Therefore, there is an ongoing need to investigate these associations between B vitamin supplementation and depression, either separately or jointly, in well-characterized cohorts of depression population.

Association between mixed dietary B vitamin intake and insulin resistance in US middle-aged and older adults without diabetes: The Bayesian kernel machine regression approach.

In daily life, the intake of dietary nutrients is mixed. However, evidence for the association between mixed dietary B vitamin intake and insulin resistance is limited. In this study, we estimated the joint effect of intake of various dietary B vitamins on insulin resistance. This cross-sectional study used data from the National Health and Nutrition Examination Survey 2011-2018. We included 1,628 middle-aged and 1,058 older adults without diabetes. Multivariable logistic regression and Bayesian kernel machine regression models were constructed. In the multivariable logistic regression, when all B vitamins were included in the model, the ORs (95% CIs) of insulin resistance were 3.06 (1.00-9.37) and 0.42 (0.19- 0.93) for the highest quartile of vitamin B-1 and B-12 intake in the middle-aged group when the lowest quartile was the reference. In the older group, no significant association was observed. In the Bayesian kernel machine regression analysis, a negative trend was noted between mixed B vitamin intake and insulin resistance in both examined groups. The univariate exposure-response function indicated that vitamin B-12 intake was negatively associated with insulin resistance in the middle-aged group, and that vitamin B-6 and dietary folate equivalent intakes were negatively associated with insulin resistance in older group. The bivariate exposure-response function indicated a potential interaction effect between dietary intake of vitamin B-12 and those of vitamin B-1, B-2, niacin, and dietary folate equivalent on insulin resistance in older people. Our results suggest that mixed dietary B vitamin intake tends to decrease the OR of insulin resistance both in middle-aged and older people.

The Relationship Between Mild Cognitive Impairment and Anti-Inflammatory/Pro-Inflammatory Nutrients in the Elderly in Northern China: A Bayesian Kernel Machine Regression Approach.

ObjectiveThis study aims to explore the role of nutrients with pro-inflammatory or anti-inflammatory potential in the risk of mild cognitive impairment in the elderly.MethodsWe evaluated the intake of 22 nutrients in the baseline survey data of 612 elderly people in northern China. Meanwhile, the least absolute contraction and selection operator (LASSO) regression analysis was used to screen the nutrients with strong correlation with MCI. Bayesian kernel machine regression (BKMR) was devoted to explore the possible associations between various nutrients with different inflammatory potentials and the risk of mild cognitive impairment.ResultsA total of 253 people (41.3%) were diagnosed with mild cognition. Ten nutrients are significantly related to the risk of MCI and were screened by a lasso regression model, including 5 pro-inflammatory nutrients (inflammation effect score >0) and 5 anti-inflammatory nutrients (inflammation effect score <0). We incorporated the inflammatory effect scores of 10 nutrients into the BKMR model, and the results showed that the inflammatory effect of 10 nutrients continued to rise with the increase in inflammation scores, proposing that the overall effect is pro-inflammatory. The BKMR analysis results of the pro-inflammatory group and the anti-inflammatory group showed that multiple nutrients in the two groups had a significant combined effect on mild cognitive impairment. We found that by comparing the overall effect of inflammation and the effect of a single group, we found that the inflammation effect of the pro-inflammatory diet and the anti-inflammatory diet had a certain offsetting effect (P < 0.005).ConclusionIn the elderly population in northern China, pro-inflammatory diets are associated with an increased risk of mild cognitive impairment. However, these results need to be further evaluated and verified in more prospective studies.

Association between prenatal exposure to metal mixtures and early childhood allergic diseases

The association between prenatal exposure to the metal mixture and allergic diseases is poorly understood. We aimed to explore the individual effect and the combined effect of prenatal exposure to vanadium (V), chromium (Cr), nickel (Ni), arsenic (As), cadmium (Cd), thallium (Tl), and lead (Pb) on early childhood allergic diseases based on a birth cohort study that included 628 mother-infant pairs. Metals were measured in maternal urine samples collected in the first, second, and third trimesters. Children were prospectively followed up at age 4 years to collect information on allergic rhinitis, wheeze, and eczema status. By applying logistic regression models, weighted quantile sum (WQS) regression, and Bayesian kernel machine regression (BKMR), the different statistical analyses revealed urinary metals were only associated with early childhood allergic rhinitis. The averaged prenatal As exposure was significantly associated with an increased OR for allergic rhinitis in both single-metal (OR = 2.04, 95% CI: 1.35, 3.07) and multiple-metal logistic regression models (OR = 1.78, 95% CI: 1.15, 2.78). The WQS index of mixed metal exposure was positively associated with allergic rhinitis (OR = 1.66, 95% CI: 1.26, 2.19), and As and Tl had the largest weights in the WQS index (weighted 0.51 and 0.29, respectively). The BKMR analysis also showed the overall effect of the metal mixture was significantly associated with allergic rhinitis when all the metals were at their 55th percentile or above, compared to their 50th percentile. The effect of As and Tl on the risk of allergic rhinitis was significant when all of the other metals were fixed at the specific percentiles. Our findings suggest that prenatal co-exposure to higher levels of the seven metals increases the risk of allergic rhinitis in children, and As and Tl may contribute most to the combined risk.

Association between exposure to chemical mixtures in relation to serum total IgE among adults 19–86 years old

There is a scarcity of studies on the effects of mixed chemicals on total IgE. We aim to assess whether there is a link between chemical mixtures (blood and urine of 26 chemicals including lead, mercury, cadmium, t,t-muconic acid, benzylmercapturic acid , 1-hydroxypyrene, 2-naphthol, 2-hydroxyfluorene, 1-hydroxyphenanthrene, mono-(2-ethyl-5-hydroxyhexyl) phthalate, mono-(2-ethyl-5-oxohexyl) phthalate, mono-n-butyl phthalate, mono-benzyl phthalate, mono-(2-ethyl-5-carboxypentyl) phthalate, mono-carboxyoctyl phthalate, mono-carboxy-isononly phthalate, mono (3-carboxypropyl) phthalate, bisphenol A, bisphenol F, bisphenol S, triclosan, methylparaben, ethylparaben, propylparaben, 3-phenoxybenzoic acid, and cotinine), and total IgE in 3,642 Korean adults aged ≥ 19. The effects of mixed chemical exposure on total IgE were identified using linear regression models, weighted quantile sum (WQS) regression, quantile g-computation (qgcomp), and Bayesian kernel machine regression (BKMR). The most relevant factors linked with IgE, according to the linear regression models, were blood or urine mercury and urine bisphenol A levels, with significant trends detected for these chemical tertiles (p < 0.01). The WQS index was significantly linked with ln2-transformed levels of serum total IgE (β = 0.30, 95 %CI 0.25–0.32). The qgcomp index also found a significant link between chemicals and ln2-transformed levels of serum total IgE (β = 0.52, 95 %CI 0.21–0.82), and elevated serum total IgE levels (OR = 2.55, 95 %CI 1.14–5.71). In BKMR analysis, the overall effect of the mixture was significantly associated with ln2-transformed levels of serum total IgE. The cutoff levels for exposure levels related to serum total IgE levels/elevated serum total IgE levels were reported. We discovered that whole-body exposure to 26 chemicals was associated with serum total IgE levels after assessing the findings of these four models. More research is needed in the future to gain a better understanding of the impact of mixed chemical exposure on allergic disorders and how to minimize chemical exposure, especially for people under the age of 18.

Toxic Metals and Subclinical Atherosclerosis in Carotid, Femoral, and Coronary Vascular Territories: The Aragon Workers Health Study.

Studies evaluating the association of metals with subclinical atherosclerosis are mostly limited to carotid arteries. We assessed individual and joint associations of nonessential metals exposure with subclinical atherosclerosis in 3 vascular territories. Approach and Results: One thousand eight hundred seventy-three Aragon Workers Health Study participants had urinary determinations of inorganic arsenic species, barium, cadmium, chromium, antimony, titanium, uranium, vanadium, and tungsten. Plaque presence in carotid and femoral arteries was determined by ultrasound. Coronary Agatston calcium score ≥1 was determined by computed tomography scan. Median arsenic, barium, cadmium, chromium, antimony, titanium, uranium, vanadium, and tungsten levels were 1.83, 1.98, 0.27, 1.18, 0.05, 9.8, 0.03, 0.66, and 0.23 μg/g creatinine, respectively. The adjusted odds ratio (95% CI) for subclinical atherosclerosis presence in at least one territory was 1.25 (1.03-1.51) for arsenic, 1.67 (1.22-2.29) for cadmium, and 1.26 (1.04-1.52) for titanium. These associations were driven by arsenic and cadmium in carotid, cadmium and titanium in femoral, and titanium in coronary territories and mostly remained after additional adjustment for the other relevant metals. Titanium, cadmium, and antimony also showed positive associations with alternative definitions of increased coronary calcium. Bayesian Kernel Machine Regression analysis simultaneously evaluating metal associations suggested an interaction between arsenic and the joint cadmium-titanium exposure. Our results support arsenic and cadmium and identify titanium and potentially antimony as atherosclerosis risk factors. Exposure reduction and mitigation interventions of these metals may decrease cardiovascular risk in individuals without clinical disease.

Association of both prenatal and early childhood multiple metals exposure with neurodevelopment in infant: A prospective cohort study

BackgroundImpaired neurodevelopment of children has become a growing public concern; however, the associations between metals exposure and neurocognitive function have remained largely unknown. ObjectivesWe systematically evaluated the associations of multiple metals exposure during pregnancy and childhood on the neurodevelopment of children aged 2–3 years. MethodsWe measured 22 metals in the serum and urine among703 mother-child pairs from the Guangxi Birth Cohort Study. The neurocognitive development of children was assessed by the Gesell Development Diagnosis Scale (GDDS; Chinese version). Multiple linear regression models were used to evaluate the relationship between the metals (selected by elastic net regression) and the outcomes. The Bayesian kernel machine regression (BKMR) was used to evaluate the possible joint effect between the multiple metal mixture and the outcomes. ResultsPrenatal aluminum (Al) exposure was negatively associated with the fine motor developmental quotient (DQ) (β = −1.545, 95%CI: 2.231, −0.859), adaption DQ (β = −1.182, 95%CI: 1.632, −0.732), language DQ (β = −1.284, 95% CI: 1.758, −0.809), and social DQ (β = −1.729, 95% CI: 2.406, −1.052) in the multi-metal model. Prenatal cadmium (Cd) exposure was negatively associated with gross motor DQ (β = −2.524, 95% CI: 4.060, −0.988), while postpartum Cd exposure was negatively associated with language DQ (β = −1.678, 95% CI: 3.227, −0.129). In stratified analyses, infants of different sexes had different sensitivities to metal exposure, and neurobehavioral development was more significantly affected by metal exposure in the first and second trimester. BKMR analysis revealed a negative joint effect of the Al, Cd, and selenium (Se) on the language DQ score; postpartum Cd exposure played a major role in this relationship. ConclusionPrenatal exposure to Al, Ba, Cd, molybdenum (Mo), lead (Pb), antimony (Sb), and strontium (Sr), and postpartum exposure to cobalt (Co), Cd, stannum (Sn), iron (Fe), nickel (Ni), and Se are associated with neurological development of infants. The first and second trimester might be the most sensitive period when metal exposure affects neurodevelopment.

Associations between prenatal multiple metal exposure and preterm birth: Comparison of four statistical models

BackgroundExposure to some heavy metals has been demonstrated to be related to the risk of preterm birth (PTB). However, the effects of multi-metal mixture are seldom assessed. Thus, we aimed to investigate the associations of maternal exposure to metal mixture with PTB, and to identify the main contributors to PTB from the mixture. MethodsThe population in the nested case-control study was from a prospective cohort enrolled in Wuhan, China between 2012 and 2014. Eighteen metals were measured in maternal urine collected before delivery. Logistic regression, elastic net regularization (ENET), weighted quantile sum regression (WQSR), and Bayesian kernel machine regression (BKMR) were used to estimate the overall effect and identify important mixture components that drive the associations with PTB. ResultsLogistic regression found naturally log-transformed concentrations of 13 metals were positively associated with PTB after adjusting for the covariates, and only V, Zn, and Cr remained the significantly positive associations when additionally adjusting for the 13 metals together. ENET identified 11 important metals for PTB, and V (β = 0.23) had the strongest association. WQSR determined the positive combined effect of metal mixture on PTB (OR: 1.44, 95%CI: 1.32, 1.57), and selected Cr and V (weighted 0.41 and 0.32, respectively) as the most weighted metals. BKMR analysis confirmed the overall mixture was positively associated with PTB, and the independent effect of V was the most significant. Besides, BKMR showed the non-linear relationships of V and Cu with PTB, and the potential interaction between Zn and Cu. ConclusionApplying different statistical models, the study found that exposure to the metal mixture was associated with a higher risk of PTB, and V was identified as the most important risk factor among co-exposed metals for PTB.

The Effect of Mixture of Heavy Metals on Obesity in Individuals ≥50 Years of Age.

Little is known about the association between a mixture of heavy metals and obesity among individuals ≥50 years of age with comorbidities. Thus, we identified the associations of serum cadmium (Cd), lead (Pb), and mercury (Hg) with obesity using linear regression models; weighted quantile sum (WQS) regression, quantile g-computation (qgcomp), and Bayesian kernel machine regression (BKMR) were conducted as secondary analyses. Of the 6434 subjects included in the analysis, 13.8% had obesity and 44.6% had abdominal obesity. In the logistic regression model, serum Hg was associated with obesity and abdominal obesity, and significant trends were observed for these heavy metal tertiles (p < 0.001). Serum Hg levels were also associated with body mass index (BMI) and waist circumference (WC). The WQS index was significantly associated with both obesity (OR = 1.43, 95% CI: 1.40-1.46) and abdominal obesity (β = 1.51, 95% CI: 1.48-1.54). The qgcomp index also found a significant association between heavy metals and both obesity (OR = 1.35, 95% CI: 1.12-1.63) and abdominal obesity (OR = 1.34, 95% CI: 1.12-1.60). Serum Hg was the most heavily weighed heavy metal in these models. In BKMR analysis, the overall effect of the mixture was significantly associated with obesity, BMI, and WC. Serum Hg showed positive trends and was observed as the most important factor associated with obesity, BMI, and WC. Our findings were largely robust to secondary analyses that used three novel mixture modeling approaches: WQS, qpcomp, and BKMR. Given increasing exposure to heavy metals, well-characterized cohorts of individuals aged ≥50 years are required to determine the mixed effects of heavy metals on obesity and related diseases.

Exposure to perfluoroalkyl substances in early pregnancy and the risk of hypertensive disorders of pregnancy: A nested case-control study in Guangxi, China

BackgroundPrevious epidemiological studies have examined the associations between exposure to perfluoroalkyl substances (PFASs) and the risk of hypertensive disorders of pregnancy (HDP). However, these studies have drawn discrepant conclusions and have some limitations. MethodsA nested case-control study was conducted with the Guangxi Zhuang Birth Cohort (GZBC), a prospective, ongoing birth cohort that was implemented in Guangxi, China, in June 2015. Maternal serum concentrations of nine PFASs were measured using ultra-high-performance liquid chromatography-tandem mass spectrometry (UPLC-MS). The associations between PFAS exposure and the risk of HDP were assessed using logistic regression (single-exposure), weighted quantile sum (WQS) regression, and Bayesian kernel machine regression (BKMR) models. ResultsA total of 136 HDP cases and 408 controls were enrolled in this study. In logistic regression models, perfluoroundecanoic acid (PFUnA), perfluorooctane sulfonate (PFOS), perfluorononanoic acid (PFNA) and perfluorobutanesulfonic acid (PFBS) were positively associated with HDP, while perfluorohexane sulfonate (PFHxS) was inversely associated with HDP. In the BKMR analysis, the joint effect of PFASs was positively associated with HDP. PFOS and PFBS showed positive trends, while PFHxS and PFHpA showed inverse trends. In WQS regression analysis, we calculated two WQS indices that were estimated using constraints in both the positive and negative directions of effects. Both WQS indices were significantly associated with HDP (OR: 2.663, 95% CI: 1.795–3.951; OR: 0.338, 95% CI: 0.229–0.499, respectively). PFBS, PFOS and PFUnA had significant weights in the positive effect direction; PFHxS, perfluoroheptanoic acid (PFHpA) and perfluorododecanoic acid (PFDoA) had significant weights in the negative effect direction. ConclusionConsidering all model results, we found that combined exposure to nine PFASs had a positive effect on the development of HDP. Moreover, PFOS and PFBS were positively associated with the HDP risk, while PFHxS and PFHpA were negatively associated with the HDP risk in women in Guangxi, China.

The associations between urinary metals and metal mixtures and kidney function in Chinese community-dwelling older adults with diabetes mellitus

BackgroundPrevious studies have found associations between single toxic metals, such as arsenic and cadmium, and kidney function in adults with diabetes. However, studies with regards to other metals and metal mixtures are still limited. ObjectiveOur study aimed to investigate the associations between urinary concentrations of 5 selected metals and metal mixtures and kidney function using a sample of older adults with diabetes mellitus in Chinese communities. MethodsIn a sample of older adults (n = 5186), 592 eligible subjects were included in this study. Urinary concentrations of 5 metals, i.e., arsenic (As), cadmium (Cd), vanadium (V), cobalt (Co), and thallium (Tl), were measured by inductively coupled plasma mass spectrometer (ICP-MS). Estimated glomerular filtration rate (eGFR) was calculated and dichotomized into indicator of chronic kidney disease (CKD). Logistic analysis and Bayesian kernel machine regression (BKMR) were used to explore the associations between single metals and metal mixtures and CKD, respectively. ResultsUrinary levels of As and V were positively correlated with CKD (OR=2.37, 95% CI: 1.31–4.30 for As; OR=2.24, 95% CI: 1.25–4.03 for V), when compared the 4th quartile with the 1st quartile. After adjustment for potential confounders, the significant association between As and CKD still existed (OR=2.73, 95% CI: 1.23–6.07). BKMR analyses showed strong linear positive associations between As and V and CKD. Higher urinary levels of the mixture were significantly associated with higher odds of CKD in a dose-response pattern. As and V showed the highest posterior inclusion probabilities. ConclusionUrine As and V were positively associated with CKD in older adults with diabetes mellitus, separately and in a mixture. The metals mixture showed a linear dose-response association with the odds of CKD. The analyses of mixtures, rather than of single metals, may provide a real-world perspective on the relationship between metals and kidney function.

Associations between multiple heavy metals exposure and glycated hemoglobin in a Chinese population

BackgroundHeavy metals may play an important role as environmental risk factors in diabetes mellitus. This study aimed to explore the association of HbA1c with As, Cd, Cu, Ni, Pb, and Zn in single-metal exposure and multi-metal co-exposure models. MethodsA cross-sectional study involving 3472 participants was conducted. Plasma concentrations of heavy metals were determined by inductively coupled plasma mass spectrometry. We estimated the association of each metal with HbA1c by linear regression. Potential heterogeneities by sex, age, and smoking were investigated, and metal mixtures and interactions were assessed by the Bayesian kernel machine regression (BKMR). ResultsIn linear regression, Cu (β = 0.324, p < 0.05) and Ni (β = −0.19, p < 0.05) showed significant association with HbA1c in all participants. In BKMR analyses, all exposure–response relationships were approximately linear. Cu was significantly and positively associated with HbA1c levels in overall participants, women, participants aged 60 years old and above, and nonsmokers. Ni was significantly and negatively associated with HbA1c levels in overall participants. We did not observe the overall effect of plasma metal mixtures on HbA1c or the interaction effect of the metals on HbA1c. ConclusionCu was positively correlated with HbA1c, whereas Ni was negatively correlated with HbA1c, when evaluated individually or in a metal mixture. Additional studies are necessary to confirm these correlations and to control for exposure to different metals in the general population.

Health Effects of Air Pollutant Mixtures on Overall Mortality Among the Elderly Population Using Bayesian Kernel Machine Regression (BKMR)

BACKGROUND AND AIM: It is well documented that air pollutants are associated with a range of adverse health outcomes. In reality, humans are exposed to multiple pollutants that are likely highly correlated with each other. However, the majority of the existing research has primarily focused on the health effect of a single pollutant, partially because we lack the methodology to evaluate the complicated relationship of mixture exposures. METHODS: We collected the health data of 1,406,185 Medicare enrollees residing in North Carolina, South Carolina, and Georgia through the Centers for Medicare and Medicaid Services. Ambient PM2.5, O3, and NO2 concentrations, derived from a well-validated ensemble machine learning model, were assigned to individuals based on their ZIP codes. We assessed the individual and joint effect of air pollutant mixtures (PM2.5, O3, and NO2) on all-cause mortality by applying a novel mixture modeling approach, Bayesian Kernel machine regression (BKMR). RESULTS:We observed a statistically significant adverse effect of the multi-pollutant mixture (PM2.5, O3, and NO2) on overall mortality. We found significant evidence for the association between PM2.5 and increased mortality on the population level in this study; the positive association with mortality appears stronger at lower percentiles of other pollutants. An interquartile range (IQR) change in PM2.5 concentration was associated with a significant increase in mortality of 1.7 (95% CI: 0.5, 2.9), 1.6 (95% CI: 0.4, 2.7), and 1.4 (95% CI: 0.1, 2.6) SDs when O3 and NO2 were set at the 25th, 50th, and 75th percentiles, respectively. CONCLUSIONS:This finding suggests a strong association between pollutant mixture and all-cause mortality, mainly driven by PM2.5. BKMR analysis did not identify statistically significant interactions among PM2.5, O3, and NO2. However, since the small sub-population might weaken the study power, additional studies (in larger sample size and other regions in the US) are in need to reinforce the current finding. KEYWORDS: Air pollution, Mixtures, Long-term exposure, Mortality, Bayesian kernel machine regression (BKMR)

Prenatal exposure to mixtures of endocrine-disrupting chemicals and autism spectrum disorder

BACKGROUND AND AIM: Mixtures methods have the potential to address the unclear relationship between prenatal exposure to endocrine-disrupting chemicals (EDCs) and risk of autism spectrum disorder (ASD). However, mixtures studies in this area are currently sparse. This study examines the relationship of prenatal exposure to a mixture of 25 EDCs, representing 4 classes of persistent pollutants, and ASD. METHODS: We analyzed data from the Early Markers for Autism (EMA) Study, a population-based, case-control study of children born during 2000-2003 in Southern California. Children diagnosed with ASD (n=491) were identified from clinical records of the Department of Developmental Services and were confirmed by expert review. General population (GP) controls (n=373) were randomly sampled from birth certificate files and frequency matched to children with ASD by sex, birth month, and birth year. Eleven polychlorinated biphenyls (PCBs), 6 brominated flame retardants (BFRs), 2 organochlorine pesticides (OCPs), and 6 polyfluoroalkyl substances (PFAS) were detectable in maternal serum samples collected during the second trimester. To examine the relationship between ASD and EDC mixtures, both within each chemical subclass and across all EDCs, we applied two METHODS: weighted quantile sums regression (WQSR) with repeated-holdout validation and Bayesian kernel machine regression (BKMR). RESULTS:In WQSR adjusted for covariates, PCB, PFAS, and BFR mixtures were not associated with ASD when constrained to the positive direction, but BFRs were associated with ASD in the negative direction (aOR: 0.82; 95%CI: 0.71, 0.95 per quartile increase). Results from BKMR were generally consistent with WQSR, with PBDE 153 contributing the most to both BFR mixture models. WQSR and BKMR analyses showed null associations between the aggregate EDC mixture and ASD. BKMR analyses did not indicate interaction between any of the EDC biomarkers. CONCLUSIONS:Results suggest that prenatal exposure to the mixture of these EDCs is not adversely associated with overall ASD risk. KEYWORDS: autism, mixtures, endocrine-disrupting compounds, prenatal

Associations of Perfluoroalkyl and Polyfluoroalkyl Substances (PFAS) and PFAS Mixtures with Adipokines in Midlife Women

BACKGROUND AND AIM: Perfluoroalkyl and polyfluoroalkyl substances (PFAS) exposure have been associated with obesity and related comorbidities. However, underlying mechanisms are not well understood. We determined if serum PFAS concentrations were associated with adipokine profiles in midlife women. METHODS: We examined 1,245 women aged 45-56 years from the Study of Women’s Health Across the Nation. Concentrations of 11 PFAS were quantified in baseline serum samples collected in 1999-2000. Linear and branched perfluorooctane sulfonic acid isomers (n-PFOS and Sm-PFOS) and their sum (PFOS), linear perfluorooctanoic acid (n-PFOA), perfluorononanoic acid (PFNA), perfluorohexane sulfonic acid (PFHxS), 2-(N-methyl-perfluorooctane sulfonamido) acetic acid (MeFOSAA), and 2-(N-ethyl-perfluorooctane sulfonamido) acetic acid (EtFOSAA) with detection frequencies 60% were included in the analysis. Adipokines including leptin, soluble leptin receptor (sOB-R), free leptin index (FLI, the ratio of leptin to sOB-R), total and high molecular weight (HMW) adiponectin were assessed in 2002-2003. We utilized multivariable linear regressions and Bayesian kernel machine regression (BKMR) to assess individual and overall joint effects of PFAS on adipokines with adjustment for age, race/ethnicity, study site, education, smoking status, physical activity, menopausal status, and waist circumference. RESULTS:A doubling of PFAS concentrations was associated with 7.8% (95% CI: 2.5%, 13.4%) higher FLI for PFOS, 9.4% (95% CI: 3.7%, 15.3%) for n-PFOA, 5.5% (95% CI: 2.2%, 9.0%) for EtFOSAA and 7.4% (95% CI: 2.8%, 12.2%) for MeFOSAA. Similar associations were found for leptin. Only EtFOSAA was associated with lower sOB-R concentrations (1.4% lower, 95% CI: -2.7%, -0.1%). In BKMR analysis, women with PFAS concentrations at the median and the 90th percentile had 30.9% (95% CI: 15.6%, 48.3%) and 52.1% (95% CI: 27.9%, 81.0%) higher FLI, respectively, compared with those with concentrations fixed at the 10th percentile. CONCLUSIONS:Some PFAS may alter circulating levels of leptin. Understanding associations between PFAS and adipokines may help elucidate whether PFAS can influence obesity and metabolic disease. KEYWORDS: PFAS, Mixtures, Mixture analysis, Obesity and metabolic disorders

Environmental metals and metal mixtures and renal disease in an adult population from Spain: the Aragon Workers Health Study

BACKGROUND AND AIM: Disproportionally high exposures to arsenic (As), cadmium (Cd) and lead are nephrotoxic. However, epidemiological studies evaluating low-chronic exposure, especially for other metals, are limited. We assessed the cross-sectional and prospective associations of a panel of metals, individually and as mixtures, with albuminuria and estimated glomerular filtration rate (eGFR), two well-established markers of renal function. METHODS: 1519 participants from the Aragon Worker’s Health Study had available cobalt (Co), copper (Cu), molybdenum (Mo), zinc (Zn), As, barium (Ba), Cd, chromium (Cr), antimony (Sb), titanium (Ti), uranium (U), vanadium (V) and tungsten (W) urinary measurements. In a subset of 707 participants with available follow-up, we estimated annual-average increase of urine albumin and eGFR. The joint association of metals were evaluated implementing Bayesian Kernel Machine Regression (BKMR) methods. RESULTS:Median metal levels were 0.24, 7.0, 18.6, 295, 3.1, 1.9, 0.25, 0.99, 0.20, 9.7, 0.03, 0.50, 0.22 μg/g creatinine for Co, Cu, Mo, Zn, As, Ba, Cd, Cr, Sb, Ti, U, V and W, respectively. Increased urinary Co, Cu, Zn, As, Cr, Ti, U and W were associated with higher odds of albuminuria ≥30 mg/g. In BKMR analysis, these associations seemed to be driven by W, As and Zn. In prospective analyses, increased Cu, As and Cr were also related with higher odds of annual-increase albumin excretion ≥20%. For eGFR, the prospective association with annual-eGFR change [mean difference in change (95% confidence interval)] was not statistically significant for most metals except for As [-0.35 (-0.70, -0.001)], Zn [-0.31 (-0.61, -0.01)] and Ba [0.38 (0.09, 0.67)], directionally consistent with and without urine creatinine adjustment. CONCLUSIONS:We identified W, As, Zn and Cr exposures as potential risk factors of renal dysfunction. Exposure reduction and mitigation interventions of metals may decrease renal disease risk. KEYWORDS: Heavy metals, Epidemiology, Outcomes