Estimation of exposure to particulate matter in pregnant individuals living in an area of unconventional oil and gas operations: Findings from the EXPERIVA study.

Northeastern British Columbia (Canada) is an area of oil and gas exploitation, which may result in release of fine (PM2.5) and inhalable (PM10) particulate matter. The aims of this study were to: 1) apply extrapolation methods to estimate exposure to PM2.5 and PM10 concentrations among EXPERIVA (Exposures in the Peace River Valley study) participants using air quality data archives; and 2) conduct exploratory analyses to investigate correlation between PM exposure and metrics of oil and gas wells density, proximity, and activity. Gestational exposure to PM2.5 and PM10 of the EXPERIVA participants (n = 85) was estimated by averaging the concentrations measured at the closest or three closest air monitoring stations during the pregnancy period. Drilling metrics were calculated based upon the density and proximity of conventional and unconventional oil and gas wells to each participant's residence. Phase-specific metrics were determined for unconventional wells. The correlations (ρ) between exposure to PM2.5 and PM10 and metrics of well density/proximity were determined using Spearman's rank correlation test. Estimated PM ambient air concentrations ranged between 4.73 to 12.13 µg/m3 for PM2.5 and 7.14 to 26.61 µg/m3 for PM10. Conventional wells metrics were more strongly correlated with PM10 estimations (ρ between 0.28 and 0.79). Unconventional wells metrics for all phases were positively correlated with PM2.5 estimations (ρ between 0.23 and 0.55). These results provide evidence of a correlation between density and proximity of oil and gas wells and estimated PM exposure in the EXPERIVA participants.

Maternal occupational exposure to chemicals and child cognitive function.

BACKGROUND: Limited data exist regarding child neurodevelopment in relation to maternal occupational exposure to endocrine-disrupting chemicals (EDCs). METHODS: We included 1058 mother-child pairs from the INfancia y Medio Ambiente (INMA) project (2003-2008). Using a job-exposure matrix, exposure probability scores for ten EDC groups were assigned to each mother based on her longest held job during pregnancy. At the child's 5-year visit, the McCarthy Scales of Children's Abilities was administered, yielding the general cognitive index and scales for specific cognitive domains. We analyzed region-specific associations between EDC exposures and each outcome separately using adjusted linear regression and combined region-specific effect estimates using random-effects meta-analyses. RESULTS: Approximately 24% of women were exposed to at least one EDC group, but exposure to most individual EDC groups was low (<5%). Maternal organic solvent exposure was associated with lower quantitative scores among children (-5.8 points, 95% confidence interval: -11.0, -0.5). Though statistically non-significant, exposures to polycyclic aromatic hydrocarbons, phthalates, alkylphenolic compounds, and miscellaneous chemicals were associated with poorer offspring performance for most or all cognitive domains. CONCLUSIONS: This study found limited evidence for a role of maternal occupational EDC exposures on child cognition. Further research is needed to better characterize exposures among pregnant workers. IMPACT: Using data from a prospective birth cohort, we help fill an important research gap regarding the potential consequences of work-related exposure to endocrine-disrupting chemicals (EDCs) among pregnant women on child neurodevelopment. We expand on existing literature-largely limited to pesticide and organic solvent exposures-by using a job-exposure matrix to estimate exposure to several EDC groups. We found limited evidence of an association between maternal occupational EDC exposure and children's overall cognition. We did observe specific associations between exposure to organic solvents and lower quantitative reasoning scores.

Proximity and density of unconventional natural gas wells and mental illness and substance use among pregnant individuals: An exploratory study in Canada.

BACKGROUND: Hydraulic fracturing (fracking) is a method used to extract unconventional natural gas (UNG). Living near UNG operations has been associated with various health outcomes, but few have explored the association between UNG and mental health and substance use. Our objective was to evaluate the association between metrics of residential UNG well density/proximity and mental illness and substance use among pregnant individuals in Northeastern British Columbia, Canada. METHODS: Individuals who gave birth at the Fort St John hospital between December 30, 2006 and December 29, 2016 (n = 6278) were included in the study. Exposure was determined using inverse distance weighting (IDW) to calculate the density and proximity of UNG wells to the postal code centroid ofindividual's residential address at delivery. Four exposure metrics, categorized by quartiles, were calculated based on 50, 10, 5 and 2.5 km buffer zones around each postal code centroid. Logistic regression was used to separately evaluate associations between IDW quartiles of each metric and diagnosis of depression and anxiety prior to or during pregnancy, and self-reported substance use during pregnancy, controlling for relevant and available confounders. RESULTS: The second and third quartile (Q) of the 10 km IDW were associated with greater odds of depression (Q2: adjusted (aOR) 1.30, 95% (confidence interval) CI 1.03-1.64; Q3: aOR 1.35, 95% CI 1.07-1.70) compared to the first quartile, but not the fourth. Using the 5 km IDW, we observed a suggestive positive association with depression in the second and third quartile (aOR Q2: 1.21, 95% CI 0.96-1.53; aOR Q3: 1.24, 95% CI 0.98-1.57) compared to the first quartile. No statistically significant association was observed using the 2.5 km IDW exposure metric. CONCLUSION: We observed some evidence of greater odds of mental illness prior to or during pregnancy, and substance use during pregnancy in pregnant individuals living in postal codes with increased UNG well density/proximity, although associations were not observed in smaller buffer zones. This study adds to the growing literature on the adverse health outcomes surrounding living in proximity to UNG operations.

Heavy metal pollution of soils and risk assessment in Houston, Texas following Hurricane Harvey.

In August 2017, after Hurricane Harvey made landfall, almost 52 inches of rain fell during a three-day period along the Gulf Coast Region of Texas, including Harris County, where Houston is located. Harris County was heavily impacted with over 177,000 homes and buildings (approximately 12 percent of all buildings in the county) experiencing flooding. The objective of this study was to measure 13 heavy metals in soil in residential areas and to assess cancer and non-cancer risk for children and adults after floodwaters receded. Between September and November 2017, we collected 174 surface soil samples in 10 communities, which were classified as "High Environmental Impact" or "Low Environmental Impact" communities, based on a composite metric of six environmental parameters. A second campaign was conducted between May 2019 and July 2019 when additional 204 soil samples were collected. Concentrations of metals at both sampling campaigns were higher in High Environmental Impact communities than in Low Environmental Impact communities and there was little change in metal levels between the two sampling periods. The Pollution Indices of lead (Pb), zinc, copper, nickel, and manganese in High Environmental Impact communities were significantly higher than those in Low Environmental Impact communities. Further, cancer risk estimates in three communities for arsenic through soil ingestion were greater than 1 in 1,000,000. Although average soil Pb was lower than the benchmark of the United States Environmental Protection Agency, the hazard indices for non-cancer outcomes in three communities, mostly attributed to Pb, were greater than 1. Health risk estimates for children living in these communities were greater than those for adults.

Exposure to DDT from indoor residual spraying and biomarkers of inflammation among reproductive-aged women from South Africa.

BACKGROUND: Evidence from animal studies suggests that DDT and DDE can adversely affect immuno-competence while human data are less conclusive. We aimed to assess the association of plasma concentrations of DDT and DDE with biomarkers of inflammation among reproductive-aged women residing in homes sprayed with DDT through Indoor Residual Spraying (IRS). METHODS: This study included 416 women from the Study of Women and Babies, South Africa (2010-2011). DDT, DDE, and biomarkers of inflammation (immunoglobulins A, G and M, interleukins 1ß, 6, and 8, tumor necrosis factor-α, C-reactive protein, serum amyloid-A, intercellular adhesion molecule-1, vascular cell adhesion molecule-1) were quantified in plasma. Linear regression was used to assess associations of DDT and DDE with each natural log-transformed biomarker. Models were adjusted for age, body mass index, parity, income, and season; beta estimates were expressed as percent differences. RESULTS: Compared to women with the lowest plasma concentrations of DDT and DDE, those with the highest concentrations of both compounds had higher levels IL-1ß, IL6, and TNF- α. While associations were statistically significant for both DDT and DDE, the magnitude of the associations was slightly stronger for DDT. Compared to women in the lowest quintile of DDT, women in the highest quintile were estimated to have 53.0% (95%CI: 21.7%, 84.4%), 28.1% (95%CI: 6.4%, 49.8%), and 26.6% (95%CI: 12.0%, 41.1%) higher levels of IL-1ß, IL6, and TNF- α, respectively. CONCLUSIONS: Our results suggest that increased plasma concentrations of DDT and DDE resulting from exposure to IRS may increase concentrations of pro-inflammatory biomarkers among reproductive-aged women in South Africa.

Association of organic solvents and occupational noise on hearing loss and tinnitus among adults in the U.S., 1999-2004.

PURPOSE: Exposure to organic solvents and noise may be causal agents in the development of hearing loss and tinnitus. The objectives of the present study were to examine the association of organic solvents with hearing loss and tinnitus and to assess the interaction of organic solvent and occupational noise exposure on hearing loss and tinnitus. METHODS: A secondary data analysis of data from the National Health and Nutrition Examination Survey and Occupational Information Network (O*NET) among a study population ranging from 1085 to 2471 study participants from 1999 to 2004. Multiple multivariate logistic regression models were used to assess the associations of individual organic solvent exposures as measured by blood biomarkers (1,4-dichlorobenzene, benzene, ethylbenzene, styrene, toluene, o-xylene, and m-/p-xylene) with self-reported hearing loss, audiometrically assessed hearing loss, and self-reported tinnitus. Models were adjusted for age, gender, race/ethnicity, diabetes, non-occupational noise exposure, smoking, and income. Organic solvents found to be statistically significantly associated with the outcome after adjusting for covariates were tested for interaction with occupational noise exposure. RESULTS: Solvent exposure was not statistically significantly associated with self-reported tinnitus. Benzene (OR 1.43, 95% CI 1.15-1.78), ethylbenzene (OR 1.24, 95% CI 1.02-1.50), and toluene (OR 1.27, 95% CI 1.06-1.52) concentrations were statistically significantly associated with increased adjusted odds of high-frequency hearing loss. No statistically significant interaction was observed between these solvents and occupational noise on high-frequency hearing loss. CONCLUSIONS: We found no evidence of an association between organic solvents and tinnitus; however, there was evidence of an association between organic solvent exposure and prevalence of high-frequency hearing loss.

Exploring Disparities in Maternal Residential Proximity to Unconventional Gas Development in the Barnett Shale in North Texas.

Background: This study explores sociodemographic disparities in residential proximity to unconventional gas development (UGD) among pregnant women. Methods: We conducted a secondary analysis using data from a retrospective birth cohort of 164,658 women with a live birth or fetal death from November 2010 to 2012 in the 24-county area comprising the Barnett Shale play, in North Texas. We considered both individual- and census tract-level indicators of sociodemographic status and computed Indexes of Concentration at the Extremes (ICE) to quantify relative neighborhood-level privilege/disadvantage. We used negative binomial regression to investigate the relation between these variables and the count of active UGD wells within 0.8 km of the home during gestation. We calculated count ratios (CR) and 95% confidence intervals (CI) to describe associations. Results: There were fewer wells located near homes of women of color living in low-income areas compared to non-Hispanic white women living in more privileged neighborhoods (ICE race/ethnicity + income: CR = 0.51, 95% CI = 0.48⁻0.55). Conclusions: While these results highlight a potential disparity in residential proximity to UGD in the Barnett Shale, they do not provide evidence of an environmental justice (EJ) issue nor negate findings of environmental injustice in other regions.

Household fuel use and biomarkers of inflammation and respiratory illness among rural South African Women.

Though literature suggests a positive association between use of biomass fuel for cooking and inflammation, few studies among women in rural South Africa exist. We included 415 women from the South African Study of Women and Babies (SOWB), recruited from 2010 to 2011. We obtained demographics, general medical history and usual source of cooking fuel (wood, electricity) via baseline questionnaire. A nurse obtained height, weight, blood pressure, and blood samples. We measured plasma concentrations of a suite of inflammatory markers (e.g., interleukins, tumor necrosis factor-α, C-reactive protein). We assessed associations between cooking fuel and biomarkers of inflammation and respiratory symptoms/illness using crude and adjusted linear and logistic regression models. We found little evidence of an association between fuel-use and biomarkers of inflammation, pre-hypertension/hypertension, or respiratory illnesses. Though imprecise, we found 41% (95% confidence interval (CI) = 0.72-2.77) higher odds of self-reported wheezing/chest tightness among wood-users compared with electricity-users. Though studies among other populations report positive findings between biomass fuel use and inflammation, it is possible that women in the present study experience lower exposures to household air pollution given the cleaner burning nature of wood compared with other biomass fuels (e.g., coal, dung).

Menstrual cycle characteristics as determinants of plasma concentrations of perfluoroalkyl substances (PFASs) in the Norwegian Mother and Child Cohort (MoBa study).

INTRODUCTION: Perfluoroalkyl substances (PFASs) are fluorinated organic compounds that have been used in a variety of industrial and consumer applications. Menstruation is implicated as a possible route of elimination for PFASs in women. The overall purpose of this study was to examine menstrual cycle characteristics as determinants of plasma PFAS concentrations in women. METHODS: Our study sample consisted of 1977 pregnant women from the Norwegian Mother and Child Cohort (MoBa) study. The women were asked about menstrual cycle regularity in the year before the pregnancy and typical menstrual cycle length as well as other demographic and reproductive characteristics in a questionnaire completed during the pregnancy. Blood samples were collected around 17-18 weeks gestation and PFAS concentrations were measured in plasma. We examined the association between menstrual cycle characteristics and seven PFASs (perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA), perfluorodecanoic acid (PFDA), perfluoroundecanoic acid (PFUnDA), perfluorohexane sulfonate (PFHxS), perfluoroheptane sulfonate (PFHpS), and perfluorooctane sulfonate (PFOS)) using multiple linear regression, adjusted for age, pre-pregnancy body mass index, smoking, education, income, parity, oral contraceptive use, inter-pregnancy interval, and breastfeeding duration. RESULTS: Irregular cycles were not associated with PFAS concentrations. Overall, we found no evidence of associations between menstrual cycle length and PFAS concentrations. In subgroup analyses we found some evidence, among parous women, of decreased PFHpS and PFOS with short menstrual cycles; we also found, among recent OC users (in the 12 months before the questionnaire) increased PFNA and PFUnDA with long cycle length. Limitations of our study include misclassification of menstrual cycle characteristics, small sample sizes in the sub-group analyses, and a lack of information on duration and volume of menses. CONCLUSIONS: In the entire study sample, we found little evidence of menstrual cycle characteristics as determinants of PFAS concentrations. However, we observed some associations between cycle length and PFAS concentrations with some select PFAS compounds in subgroup analyses.

The Association Between PM2.5 and Ozone and the Prevalence of Diabetes Mellitus in the United States, 2002 to 2008.

OBJECTIVE: To examine the association between air pollution and diabetes prevalence in the United States, 2002 to 2008. METHODS: Annual average particulate matter (PM2.5) and ozone concentrations were calculated using daily county-level data from the CDC's Tracking Network. Individual-level outcome and covariate data were obtained from the Centers for Disease Control and Prevention (CDC) Behavioral Risk Factor Surveillance System for 862,519 individuals. We used Poisson regression analyses to examine associations between each air pollutant (per 10-unit increase) with diabetes, including regional sub-analyses. Analyses were adjusted for year, age, sex, race, ethnicity, education, income, smoking status, body mass index, exercise, and asthma. RESULTS: Positive associations between each pollutant and diabetes were found (PM2.5: prevalence ratio [PR] = 1.10; 95% confidence interval [CI] = 1.03, 1.17; ozone: PR = 1.06; 95% CI = 1.03, 1.09). There was limited evidence of effect modification by region. CONCLUSIONS: Interventions to reduce ambient air pollution may help alleviate the diabetes burden in the US.

Exploring the links between groundwater quality and bacterial communities near oil and gas extraction activities.

Bacterial communities in groundwater are very important as they maintain a balanced biogeochemical environment. When subjected to stressful environments, for example, due to anthropogenic contamination, bacterial communities and their dynamics change. Studying the responses of the groundwater microbiome in the face of environmental changes can add to our growing knowledge of microbial ecology, which can be utilized for the development of novel bioremediation strategies. High-throughput and simpler techniques that allow the real-time study of different microbiomes and their dynamics are necessary, especially when examining larger data sets. Matrix-assisted laser desorption-ionization (MALDI) time-of-flight mass spectrometry (TOF-MS) is a workhorse for the high-throughput identification of bacteria. In this work, groundwater samples were collected from a rural area in southern Texas, where agricultural activities and unconventional oil and gas development are the most prevalent anthropogenic activities. Bacterial communities were assessed using MALDI-TOF MS, with bacterial diversity and abundance being analyzed with the contexts of numerous organic and inorganic groundwater constituents. Mainly denitrifying and heterotrophic bacteria from the Phylum Proteobacteria were isolated. These microorganisms are able to either transform nitrate into gaseous forms of nitrogen or degrade organic compounds such as hydrocarbons. Overall, the bacterial communities varied significantly with respect to the compositional differences that were observed from the collected groundwater samples. Collectively, these data provide a baseline measurement of bacterial diversity in groundwater located near anthropogenic surface and subsurface activities.

Maternal residential proximity to unconventional gas development and perinatal outcomes among a diverse urban population in Texas.

OBJECTIVE: To assess associations between unconventional natural gas development (UGD) and perinatal outcomes. METHODS: We conducted a retrospective birth cohort study among 158,894 women with a birth or fetal death from November 30, 2010-November 29, 2012 in the Barnett Shale, in North Texas. We constructed three UGD-activity metrics by calculating the inverse distance-weighted sum of active wells within three separate geographic buffers surrounding the maternal residence: ≤½, 2, or 10-miles. We excluded women if the nearest well to her residence was >20 miles. Metrics were categorized by tertiles among women with ≥1 well within the respective buffer; women with zero wells ≤10 miles (the largest buffer) served as a common referent group. We used logistic or linear regression with generalized estimating equations to assess associations between UGD-activity and preterm birth, small-for-gestational age (SGA), fetal death, or birthweight. Adjusted models of fetal death and birthweight included: maternal age, race/ethnicity, education, pre-pregnancy body mass index, parity, smoking, adequacy of prenatal care, previous poor pregnancy outcome, and infant sex. Preterm birth models included all of the above except parity; SGA models included all of the above except previous poor pregnancy outcome. RESULTS: We found increased adjusted odds of preterm birth associated with UGD-activity in the highest tertiles of the ½- (odds ratio (OR) = 1.14; 95% confidence interval 1.03, 1.25), 2- (1.14; 1.07, 1.22), and 10-mile (1.15; 1.08, 1.22) metrics. Increased adjusted odds of fetal death were found in the second tertile of the 2-mile metric (1.56; 1.16, 2.11) and the highest tertile of the 10-mile metric (1.34; 1.04-1.72). We found little indication of an association with SGA or term birthweight. CONCLUSIONS: Our results are suggestive of an association between maternal residential proximity to UGD-activity and preterm birth and fetal death. Quantifying chemical and non-chemical stressors among residents near UGD should be prioritized.

Antimüllerian hormone in relation to tobacco and marijuana use and sources of indoor heating/cooking.

OBJECTIVE: To evaluate exposure to tobacco, marijuana, and indoor heating/cooking sources in relation to antimüllerian hormone (AMH) levels. DESIGN: Cross-sectional analysis in a sample of premenopausal women (n = 913) enrolled in the Sister Study cohort (n = 50,884). SETTING: Not applicable. PATIENT(S): Women, ages 35-54 years at time of enrollment, with an archived serum sample and at least one intact ovary and classified as premenopausal. INTERVENTION(S): Not applicable. MAIN OUTCOME MEASURE(S): Serum AMH (ng/mL) levels ascertained by ultrasensitive ELISA assay. RESULT(S): Lower AMH levels were associated with sources of indoor heating, including burning wood (-36.0%; 95% confidence interval [CI], -55.7%, -7.8%) or artificial fire logs (-45.8%; 95% CI, -67.2%, -10.4%) at least 10 times/year in a residential indoor stove/fireplace. Lower AMH levels were also observed in women who were current smokers of ≥20 cigarettes/day relative to nonsmokers (-56.2%; 95% CI, -80.3%, -2.8%) and in women with 10+ years of adult environmental tobacco smoke (ETS) exposure (-31.3%; 95% CI, -51.3%, -3.1%), but no associations were observed for marijuana use. CONCLUSION(S): We confirmed previously reported findings of lower AMH levels in current heavy smokers and also found associations for long-term ETS exposure and indoor burning of wood or artificial fire logs. These findings suggest that combustion by-products from common exposures can have toxic effects on the human ovary.

Anti-Müllerian hormone concentrations in premenopausal women and breast cancer risk.

Laboratory models support an inverse association between anti-Müllerian hormone (AMH) and breast tumor development. Human studies are lacking; one study (N = 105 cases, 204 controls) with prospectively collected serum reported the opposite-an approximate 10-fold increase in breast cancer risk comparing fourth with first quartile AMH levels. We investigated the relation between serum AMH levels and breast cancer risk in a case-control (N = 452 cases, 902 controls) study nested within the prospective Sister Study cohort of 50,884 women. At enrollment, participants were ages 35 to 54, premenopausal, and completed questionnaires on medical and family history, lifestyle factors, and demographics. AMH (ng/mL) was measured by ultrasensitive ELISA in serum collected at enrollment and log-transformed for analysis. Multivariate conditional logistic regression was used to calculate ORs and 95% confidence intervals (CI) to account for matching on age and enrollment year. Mean age at enrollment was 46.8 years with an average 2.9 years from blood draw to breast cancer diagnosis (SD = 1.9). AMH concentrations were below the limit of detection (0.003 ng/mL) for approximately 25% of samples. Compared with samples below the LOD, women with AMH >2.84 ng/mL (90th percentile among controls) had a 2-fold increase in breast cancer odds (OR, 2.25; 95% CI, 1.26-4.02). For each 1-unit increase in lnAMH, overall breast cancer odds increased by 8% (OR, 1.08; 95% CI, 1.02-1.15) and odds of estrogen receptor-positive, invasive disease increased by 15% (OR, 1.15; 95% CI, 1.05-1.25). Our findings demonstrate an overall positive relation between AMH and breast cancer.

Exposure to tobacco smoke in utero and subsequent plasma lipids, ApoB, and CRP among adult women in the MoBa cohort.

BACKGROUND: Recent findings suggest that maternal smoking during pregnancy may play a role in the development of metabolic alterations in offspring during childhood. However, whether such exposure increases the risk of developing similar metabolic alterations during adulthood is uncertain. OBJECTIVE: We evaluated the association of in utero exposure to maternal tobacco smoke with plasma lipids, apolipoprotein B (apoB), and C-reactive protein (CRP) in adulthood. METHODS: The study was based on a subsample of the Norwegian Mother and Child Cohort Study (MoBa) and included 479 pregnant women with plasma lipids, apoB, and CRP measurements. Information on in utero exposure to tobacco smoke, personal smoking, and other factors were obtained from the women by a self-completed questionnaire at enrollment, at approximately 17 weeks of gestation. RESULTS: Women exposed to tobacco smoke in utero had higher triglycerides [10.7% higher; 95% confidence interval (CI): 3.9, 17.9] and lower high-density lipoprotein cholesterol (HDL) (-1.9 mg/dL; 95% CI: -4.3, 0.5) compared with unexposed women, after adjusting for age, physical activity, education, personal smoking, and current body mass index (BMI). Exposed women were also more likely to have triglycerides ≥ 200 mg/dL [adjusted odds ratio (aOR) = 2.5; 95% CI: 1.3, 5.1] and HDL < 50 mg/dL (aOR = 2.3; 95% CI: 1.1, 5.0). Low-density lipoprotein cholesterol, total cholesterol, and apoB were not associated with the exposure. CRP was increased among exposed women; however, after adjustment for BMI, the association was completely attenuated. CONCLUSIONS: In this population, in utero exposure to tobacco smoke was associated with high triglycerides and low HDL in adulthood, 18-44 years after exposure.

Kriged and modeled ambient air levels of benzene in an urban environment: an exposure assessment study.

BACKGROUND: There is increasing concern regarding the potential adverse health effects of air pollution, particularly hazardous air pollutants (HAPs). However, quantifying exposure to these pollutants is problematic. OBJECTIVE: Our goal was to explore the utility of kriging, a spatial interpolation method, for exposure assessment in epidemiologic studies of HAPs. We used benzene as an example and compared census tract-level kriged predictions to estimates obtained from the 1999 U.S. EPA National Air Toxics Assessment (NATA), Assessment System for Population Exposure Nationwide (ASPEN) model. METHODS: Kriged predictions were generated for 649 census tracts in Harris County, Texas using estimates of annual benzene air concentrations from 17 monitoring sites operating in Harris and surrounding counties from 1998 to 2000. Year 1999 ASPEN modeled estimates were also obtained for each census tract. Spearman rank correlation analyses were performed on the modeled and kriged benzene levels. Weighted kappa statistics were computed to assess agreement between discretized kriged and modeled estimates of ambient air levels of benzene. RESULTS: There was modest correlation between the predicted and modeled values across census tracts. Overall, 56.2%, 40.7%, 31.5% and 28.2% of census tracts were classified as having 'low', 'medium-low', 'medium-high' and 'high' ambient air levels of benzene, respectively, comparing predicted and modeled benzene levels. The weighted kappa statistic was 0.26 (95% confidence interval (CI) = 0.20, 0.31), indicating poor agreement between the two methods. CONCLUSIONS: There was a lack of concordance between predicted and modeled ambient air levels of benzene. Applying methods of spatial interpolation for assessing exposure to ambient air pollutants in health effect studies is hindered by the placement and number of existing stationary monitors collecting HAP data. Routine monitoring needs to be expanded if we are to use these data to better assess environmental health risks in the future.

Childhood lymphohematopoietic cancer incidence and hazardous air pollutants in southeast Texas, 1995-2004.

BACKGROUND: Cancer is the second leading cause of death among U.S. children with few known risk factors. There is increasing interest in the role of air pollutants, including benzene and 1,3-butadiene, in the etiology of childhood cancers. OBJECTIVE: Our goal was to assess whether census tracts with the highest benzene or 1,3-butadiene ambient air levels have increased childhood lymphohematopoietic cancer incidence. METHODS: Our ecologic analysis included 977 cases of childhood lymphohematopoietic cancer diagnosed from 1995-2004. We obtained the U.S. Environmental Protection Agency's 1999 modeled estimates of benzene and 1,3-butadiene for 886 census tracts surrounding Houston, Texas. We ran Poisson regression models by pollutant to explore the associations between pollutant levels and census-tract cancer rates. We adjusted models for age, sex, race/ethnicity, and community-level socioeconomic status (cSES). RESULTS: Census tracts with the highest benzene levels had elevated rates of all leukemia [rate ratio (RR) = 1.37; 95% confidence interval (CI), 1.05, 1.78]. This association was higher for acute myeloid leukemia (AML) (RR = 2.02; 95% CI, 1.03-3.96) than for acute lymphocytic leukemia (ALL) (RR = 1.24; 95% CI, 0.92-1.66). Among census tracts with the highest 1,3-butadiene levels, we observed RRs of 1.40 (95% CI, 1.07-1.81), 1.68 (95% CI, 0.84-3.35), and 1.32 (95% CI, 0.98-1.77) for all leukemia, AML, and ALL, respectively. We detected no associations between benzene or 1,3-butadiene levels and lymphoma incidence. Results that examined joint exposure to benzene and 1,3-butadiene were similar to those that examined each pollutant separately. CONCLUSIONS: Our ecologic analysis suggests an association between childhood leukemia and hazardous air pollution; further research using more sophisticated methodology is warranted.