Latent classes for chemical mixtures analyses in epidemiology: an example using phthalate and phenol exposure biomarkers in pregnant women

Estimating effects of longitudinal and cumulative exposure to PFAS mixtures on early adolescent body composition

Few methods have been used to characterize repeatedly measured biomarkers of chemical mixtures. We applied latent profile analysis (LPA) to serum concentrations of four perfluoroalkyl and polyfluoroalkyl substances (PFAS) at four timepoints from gestation to age 12 years. We evaluated the relations between profiles and z-scores of height, body mass index, fat mass index, and lean body mass index at age 12 years (n = 218). We compared LPA findings with an alternative approach for cumulative PFAS mixtures using g-computation to estimate the effect of simultaneously increasing the area under the curve (AUC) for all PFAS. We identified two profiles: a higher (35% of sample) and lower PFAS profile (relative to each other), based on their average PFAS concentrations at all timepoints. The higher PFAS profile had generally lower z-scores for all outcomes, with somewhat larger effects for males, though all CIs crossed the null. For example, the higher PFAS profile was associated with a -0.50 (95% CI: -1.07, 0.08) lower BMI z-score among males but not females (0.04; 95% CI: -0.45, 0.54). We observed similar patterns with AUCs. We found that higher childhood PFAS profile and higher cumulative PFAS mixtures may be associated with altered growth in early adolescence.

Personal Care Product Use during Puberty and Incident Breast Cancer among Black, Hispanic/Latina, and White Women in a Prospective US-Wide Cohort

BACKGROUND

Some personal care products (PCPs) contain endocrine-disrupting chemicals that may affect breast cancer (BC) risk. Patterns of use vary by race and ethnicity. Use often starts in adolescence, when rapidly developing breast tissue may be more susceptible to environmental carcinogens. Few studies have examined associations of BC with PCP use during this susceptible window.

OBJECTIVES

We characterized race and ethnicity-specific patterns of PCP use at 10-13 years of age and estimated associations of use with incident BC.

METHODS

At enrollment (2003-2009), Sister Study participants (n = 4,049 Black, 2,104 Latina, and 39,312 White women) 35-74 years of age reported use of 37 "everyday" PCPs during the ages of 10-13 y (did not use, sometimes, or frequently used). We conducted race and ethnicity-specific latent class analyses to separately identify groups of women with similar patterns of beauty, hair, and skincare/hygiene product use. We estimated hazard ratios (HRs) and 95% confidence intervals (CIs) for associations of identified PCP classes and single products with incident BC using Cox proportional hazards regression.

RESULTS

During a mean follow-up time of 10.8 y, 280 Black, 128 Latina, and 3,137 White women were diagnosed with BC. Classes of adolescent PCP use were not clearly associated with BC diagnosis among Black, Latina, or White women. HRs were elevated but imprecise for frequent nail product and perfume use in Black women (HR = 1.34; 95% CI: 0.85, 2.12) and greater hair product use in Black (HR = 1.28; 95% CI: 0.91, 1.80) and Latina (HR = 1.42; 95% CI: 0.81, 2.48) women compared with lighter use. In single-product models, we observed higher BC incidence associated with frequent use of lipstick, nail products, pomade, perfume, makeup remover, and acne/blemish products in at least one group.

DISCUSSION

This work provides some support for the hypothesis that PCP use during puberty is associated with BC risk. More research is needed to confirm these novel findings. https://doi.org/10.1289/EHP13882.

Exposure to different residential indoor characteristics during childhood and asthma in adolescence: a latent class analysis of the Danish National Birth Cohort

BACKGROUND

Many residential indoor environments may have an impact on children's respiratory health.

OBJECTIVES

The aims of this study were to identify latent classes of children from the Danish National Birth Cohort (DNBC) who share similar patterns of exposure to indoor home characteristics, and to examine the association between membership in the latent classes and asthma in adolescence.

METHODS

We included data on residential indoor characteristics of offspring from the DNBC whose mothers had responded to the child's 11-year follow-up and who had data on asthma from the 18-year follow-up. Number of classes and associations were estimated using latent class analysis. To account for sample selection, we applied inverse probability weighting.

RESULTS

Our final model included five latent classes. The probability of current asthma at 18 years was highest among individuals in class one with higher clustering on household dampness (9, 95%CI 0.06-0.13). Individuals in class four (with higher clustering on pets ownership and living in a farm) had a lower risk of current asthma at age 18 compared to individuals in class one (with higher clustering on household dampness) (OR 0.53 (95%CI 0.32-0.88), p = .01).

CONCLUSION

Our findings suggest that, in a high-income country such as Denmark, groups of adolescents growing up in homes with mold and moisture during mid-childhood might be at increased risk of current asthma at age 18. Adolescents who grew-up in a farmhouse and who were exposed to pets seem less likely to suffer from asthma by age 18.

Toward Advancing Precision Environmental Health: Developing a Customized Exposure Burden Score to PFAS Mixtures to Enable Equitable Comparisons Across Population Subgroups, Using Mixture Item Response Theory

Quantifying a person's cumulative exposure burden to per- and polyfluoroalkyl substances (PFAS) mixtures is important for risk assessment, biomonitoring, and reporting of results to participants. However, different people may be exposed to different sets of PFASs due to heterogeneity in the exposure sources and patterns. Applying a single measurement model for the entire population (e.g., by summing concentrations of all PFAS analytes) assumes that each PFAS analyte is equally informative to PFAS exposure burden for all individuals. This assumption may not hold if PFAS exposure sources systematically differ within the population. However, the sociodemographic, dietary, and behavioral characteristics that underlie systematic exposure differences may not be known, or may be due to a combination of these factors. Therefore, we used mixture item response theory, an unsupervised psychometrics and data science method, to develop a customized PFAS exposure burden scoring algorithm. This scoring algorithm ensures that PFAS burden scores can be equitably compared across population subgroups. We applied our methods to PFAS biomonitoring data from the United States National Health and Nutrition Examination Survey (2013-2018). Using mixture item response theory, we found that participants with higher household incomes had higher PFAS burden scores. Asian Americans had significantly higher PFAS burden compared with non-Hispanic Whites and other race/ethnicity groups. However, some disparities were masked when using summed PFAS concentrations as the exposure metric. This work demonstrates that our summary PFAS burden metric, accounting for sources of exposure variation, may be a more fair and informative estimate of PFAS exposure.

Temporal trends and predictors of phthalate, phthalate replacement, and phenol biomarkers in the LIFECODES Fetal Growth Study

BACKGROUND

Exposure to many phthalates and phenols is declining as replacements are introduced. There is little information on temporal trends or predictors of exposure to these newer compounds, such as phthalate replacements, especially among pregnant populations.

OBJECTIVE

Examine temporal trends and predictors of exposure to phthalates, phthalate replacements, and phenols using single- and multi-pollutant approaches.

METHODS

We analyzed data from 900 singleton pregnancies in the LIFECODES Fetal Growth Study, a nested case-cohort with recruitment from 2007 to 2018. We measured and averaged concentrations of 12 phthalate metabolites, four phthalate replacement metabolites, and 12 phenols in urine at three timepoints during pregnancy. We visualized and analyzed temporal trends and predictors of biomarker concentrations. To examine chemical mixtures, we derived clusters of individuals with shared exposure profiles using a finite mixture model and examined temporal trends and predictors of cluster assignment.

RESULTS

Exposure to phthalates and most phenols declined across the study period, while exposure to phthalate replacements (i.e., di(isononyl) cyclohexane-1,2-dicarboxylic acid, diisononyl ester [DINCH] and di-2-ethylhexyl terephthalate [DEHTP]) and bisphenol S (BPS) increased. For example, the sum of DEHTP biomarkers increased multiple orders of magnitude, with an average concentration of 0.92 ng/mL from 2007 to 2008 and 61.9 ng/mL in 2017-2018. Biomarkers of most chemical exposures varied across sociodemographic characteristics, with the highest concentrations observed in non-Hispanic Black or Hispanic participants relative to non-Hispanic White participants. We identified five clusters with shared exposure profiles and observed temporal trends in cluster membership. For example, at the end of the study period, a cluster characterized by high exposure to phthalate replacements was the most prevalent.

SIGNIFICANCE

In a large and well-characterized pregnancy cohort, we observed exposure to phthalate replacements and BPS increased over time while exposure to phthalates and other phenols decreased. Our results highlight the changing nature of exposure to consumer product chemical mixtures.

Using Latent Profile Analysis to Identify Associations Between Gestational Chemical Mixtures and Child Neurodevelopment

BACKGROUND

Unsupervised machine learning techniques have become increasingly popular for studying associations between gestational exposure mixtures and human health. Latent profile analysis is one method that has not been fully explored.

METHODS

We estimated associations between gestational chemical mixtures and child neurodevelopment using latent profile analysis. Using data from the Maternal-Infant Research on Environmental Chemicals (MIREC) research platform, a longitudinal cohort of pregnant Canadian women and their children, we generated latent profiles from 27 gestational exposure biomarkers. We then examined the associations between these profiles and child Verbal IQ, Performance IQ, and Full-Scale IQ, measured with the Wechsler Preschool and Primary Scale of Intelligence, Third Edition (WPPSI-III). We validated our findings using k-means clustering.

RESULTS

Latent profile analysis detected five latent profiles of exposure: a reference profile containing 61% of the study participants, a high monoethyl phthalate (MEP) profile with moderately low persistent organic pollutants (POPs) containing 26%, a high POP profile containing 6%, a low POP profile containing 4%, and a smoking chemicals profile containing 3%. We observed negative associations between both the smoking chemicals and high MEP profiles and all IQ scores and between the high POP profile and Full-Scale and Verbal IQ scores. We also found a positive association between the low POP profile and Full-Scale and Performance IQ scores. All associations had wide 95% confidence intervals.

CONCLUSIONS

Latent profile analysis is a promising technique for identifying patterns of chemical exposure and is worthy of further study for its use in examining complicated exposure mixtures.

Exposure to phthalates, phenols, and parabens mixture and alterations in sex steroid hormones among adolescents

Exposure to phthalates (PAEs), phenols, and parabens has been linked with sex hormone imbalance; however, previous studies were predominantly limited to adults and failed to examine the combined effects of these chemicals mixture among adolescents. Thus, we used the data from the National Health and Nutrition Examination Survey (2013-2016) to explore the associations of urinary PAEs, phenols, and parabens biomarkers with sex hormones among participants aged 12-19 years old (n = 613). Latent class analysis (LCA) and quantile-based g-computation (QGC) were applied to assess the associations of the latent exposure profiles and chemicals mixture with sex hormone indicators, including steroid hormones and sex hormone binding globulin (SHBG), in adolescents and by sex. Using LCA, four latent classes were identified among all participants. Compared with the class characterized by "Low exposure", the class represented by "High PAEs" [mono (2-ethyl-5-carboxypentyl) phthalate (MECPP), mono (2-ethyl-5-hydroxyhexyl) phthalate (MEHHP), and monobenzyl phthalate (MBZP)] had lower level of estradiol (E₂) [β = -0.249, 95% confidence interval (CI): -0.419, -0.08], free androgen index (FAI) (β = -0.258, 95%CI: -0.512, -0.005) and free testosterone (FT) (β = -0.248, 95%CI: -0.496, -0.001) among male adolescents. These results were echoed by the results in QGC analyses, where PAEs mixture was negatively associated with E₂ (β = -0.137, 95% CI: -0.263, -0.011), FAI (β = -0.198, 95%CI: -0.387, -0.008) and FT (β = -0.189, 95%CI: -0.375, -0.002) among male adolescents. By contrast, the associations of the identified latent classes or chemicals mixture with sex hormone indicators were generally nonsignificant among female counterparts, except for a positive association between PAEs mixture and SHBG (β = 0.121, 95%CI: 0.012, 0.23). Our study demonstrated that exposure to PAEs, particularly MECPP, MEHHP, and MBZP, would be a threat to the sex hormone homeostasis of male adolescents.

Association Between Gestational Exposure to Toxicants and Autistic Behaviors Using Bayesian Quantile Regression

Autism spectrum disorder, which is characterized by impaired social communication and stereotypic behaviors, affects 1%-2% of children. Although prenatal exposure to toxicants has been associated with autistic behaviors, most studies have been focused on shifts in mean behavior scores. We used Bayesian quantile regression to assess the associations between log2-transformed toxicant concentrations and autistic behaviors across the distribution of behaviors. We used data from the Maternal-Infant Research on Environmental Chemicals study, a pan-Canadian cohort (2008-2011). We measured metal, pesticide, polychlorinated biphenyl, phthalate, bisphenol-A, and triclosan concentrations in blood or urine samples collected during the first trimester of pregnancy. Using the Social Responsiveness Scale (SRS), in which higher scores denote more autistic-like behaviors, autistic behaviors were assessed in 478 children aged 3-4 years old. Lead, cadmium, and most phthalate metabolites were associated with mild increases in SRS scores at the 90th percentile of the SRS distribution. Manganese and some pesticides were associated with mild decreases in SRS scores at the 90th percentile of the SRS distribution. We identified several monotonic trends in which associations increased in magnitude from the bottom to the top of the SRS distribution. These results suggest that quantile regression can reveal nuanced relationships and, thus, should be more widely used by epidemiologists.

Assessment of Multipollutant Exposures During Pregnancy Using Silicone Wristbands

Silicone wristbands can assess multipollutant exposures in a non-invasive and minimally burdensome manner, which may be suitable for use among pregnant women. We investigated silicone wristbands as passive environmental samplers in the New Hampshire Birth Cohort Study, a prospective pregnancy cohort. We used wristbands to assess exposure to a broad range of organic chemicals, identified multipollutant exposure profiles using self-organizing maps (SOMs), and assessed temporal consistency and determinants of exposures during pregnancy. Participants (n = 255) wore wristbands for 1 week at 12 gestational weeks. Of 1,530 chemicals assayed, 199 were detected in at least one wristband and 16 were detected in >60% of wristbands. A median of 23 (range: 12,37) chemicals were detected in each wristband, and chemicals in commerce and personal care products were most frequently detected. A subset of participants (n=20) wore a second wristband at 24 gestational weeks, and concentrations of frequently detected chemicals were moderately correlated between time points (median intraclass correlation: 0.22; range: 0.00,0.69). Women with higher educational attainment had fewer chemicals detected in their wristbands and the total number of chemicals detected varied seasonally. Triphenyl phosphate concentrations were positively associated with nail polish use, and benzophenone concentrations were highest in summer. No clear associations were observed with other a priori relations, including certain behaviors, season, and socioeconomic factors. SOM analyses revealed 12 profiles, ranging from 2 to 149 participants, captured multipollutant exposure profiles observed in this cohort. The most common profile (n = 149) indicated that 58% of participants experienced relatively low exposures to frequently detected chemicals. Less common (n ≥ 10) and rare (n < 10) profiles were characterized by low to moderate exposures to most chemicals and very high and/or very low exposure to a subset of chemicals. Certain covariates varied across SOM profile membership; for example, relative to women in the most common profile who had low exposures to most chemicals, women in the profile with elevated exposure to galaxolide and benzyl benzoate were younger, more likely to be single, and more likely to report nail polish use. Our study illustrates the utility of silicone wristbands for measurement of multipollutant exposures in sensitive populations, including pregnant women.