Performance of multi-city land use regression models for nitrogen dioxide and fine particles

Development of land use regression, dispersion, and hybrid models for prediction of outdoor air pollution exposure in Barcelona

BACKGROUND

Air pollution is the leading environmental risk factor for health. Assessing outdoor air pollution exposure with detailed spatial and temporal variability in urban areas is crucial for evaluating its health effects.

AIM

We developed and compared Land Use Regression (LUR), dispersion (DM), and hybrid (HM) models to estimate outdoor concentrations for NO2, PM2.5, black carbon (BC), and PM2.5-constituents (Fe, Cu, Zn) in Barcelona.

METHODS

Two monitoring campaigns were conducted. In the first, NO2 concentrations were measured twice at 984 home addresses and in the second, NO2, PM2.5, and BC were measured four times at 34 points across Barcelona. LUR and DM were constructed using conventional techniques, while HM was developed using Random Forest (RF). Model performance was evaluated using leave-one-out cross-validation (LOOCV) and 10-fold cross-validation (10-CV) for LUR and HM, and by comparing DM and LUR estimates with routine monitoring stations. NO2 levels estimated by all models were externally validated using the home monitoring campaign. Agreement between models was assessed using Spearman correlation (rs) and Bland-Altman (BA) plots.

RESULTS

Models showed moderate to good performance. LUR exhibited R2LOOCV of 0.62 (NO2), 0.45 (PM2.5), 0.83 (BC), and 0.85 to 0.89 (PM2.5-constituents). DM model comparison showed R2 values of 0.39 (NO2), 0.26 (PM2.5), and 0.65 (BC). HM models had higher R210-CV 0.64 (NO2), 0.66 (PM2.5), 0.86 (BC), and 0.44 to 0.70 (PM2.5-constituents). Validation for NO2 showed R2 values of 0.56 (LUR), 0.44 (DM), and 0.64 (HM). Correlations between models varied from -0.38 to 0.92 for long-term exposure, and - 0.23 to 0.94 for short-term exposure. BA plots showed good agreement between models, especially for NO2 and BC.

CONCLUSIONS

Our models varied substantially, with some models performing better in validation samples (NO2 and BC). Future health studies should use the most accurate methods to minimize bias from exposure measurement error.

Air pollution, lung function and mortality: survival and mediation analyses in UK Biobank

Background

Air pollution is associated with lower lung function, and both are associated with premature mortality and cardiovascular disease (CVD). Evidence remains scarce on the potential mediating effect of impaired lung function on the association between air pollution and mortality or CVD.

Methods

We used data from UK Biobank (n∼200 000 individuals) with 8-year follow-up to mortality and incident CVD. Exposures to particulate matter <10 µm (PM10), particulate matter <2.5 µm (PM2.5) and nitrogen dioxide (NO2) were assessed by land-use regression modelling. Lung function (forced expiratory volume in 1 s (FEV1), forced vital capacity (FVC) and the FEV1/FVC ratio) was measured between 2006 and 2010 and transformed to Global Lung Function Initiative (GLI) z-scores. Adjusted Cox proportional hazards and causal proportional hazards mediation analysis models were fitted, stratified by smoking status.

Results

Lower FEV1 and FVC were associated with all-cause and CVD mortality, and incident CVD, with larger estimates in ever- than never-smokers (all-cause mortality hazard ratio per FEV1 GLI z-score decrease 1.29 (95% CI 1.24-1.34) for ever-smokers and 1.16 (95% CI 1.12-1.21) for never-smokers). Long-term exposure to PM2.5 or NO2 was associated with incident CVD, with similar effect sizes for ever- and never-smokers. Mediated proportions of the air pollution-all-cause mortality estimates driven by FEV1 were 18% (95% CI 2-33%) for PM2.5 and 27% (95% CI 3-51%) for NO2. Corresponding mediated proportions for incident CVD were 9% (95% CI 4-13%) for PM2.5 and 16% (95% CI 6-25%) for NO2.

Conclusions

Lung function may mediate a modest proportion of associations between air pollution and mortality and CVD outcomes. Results likely reflect the extent of either shared mechanisms or direct effects relating to lower lung function caused by air pollution.

Impact assessment of spatial-temporal distribution of riverine dust on air quality using remote sensing data and numerical modeling

Soil erosion is a severe problem in Taiwan due to the steep terrain, fragile geology, and extreme climatic events resulting from global warming. Due to the rapidly changing hydrological conditions affecting the locations and the amount of transported sand and fine particles, timely impact evaluation and riverine dust control are difficult, particularly when resources are limited. To comprehend the impact of desertification in estuarine areas on the variation of air pollutant concentrations, this study utilized remote sensing technology coupled with an air pollutant dispersion model to determine the unit contribution of potential pollution sources and quantify the effect of riverine dust on air quality. The images of the downstream area of the Beinan River basin captured by Formosat-2 in May 2006 were used to analyze land use and land cover (LULC) composition. Subsequently, the diffusion model ISCST-3 based on Gaussian distribution was utilized to simulate the transport of PM across the study area. Finally, a mixed-integer programming model was developed to optimize resource allocation for dust control. Results reveal that sand deposition in specific river sections significantly influences regional air quality, owing to the unique local topography and wind field conditions. The present optimal plan model for regional air quality control further showed that after implementing engineering measures including water cover, revegetation, armouring cover, and revegetation, total PM concentrations would be reduced by 51%. The contribution equivalent calculation, using the air pollution diffusion model, was effectively integrated into the optimization model to formulate a plan for reducing riverine dust with limited resources based on air quality requirements.

A comprehensive review of the development of land use regression approaches for modeling spatiotemporal variations of ambient air pollution: A perspective from 2011 to 2023

Land use regression (LUR) models are widely used in epidemiological and environmental studies to estimate humans' exposure to air pollution within urban areas. However, the early models, developed using linear regressions and data from fixed monitoring stations and passive sampling, were primarily designed to model traditional and criteria air pollutants and had limitations in capturing high-resolution spatiotemporal variations of air pollution. Over the past decade, there has been a notable development of multi-source observations from low-cost monitors, mobile monitoring, and satellites, in conjunction with the integration of advanced statistical methods and spatially and temporally dynamic predictors, which have facilitated significant expansion and advancement of LUR approaches. This paper reviews and synthesizes the recent advances in LUR approaches from the perspectives of the changes in air quality data acquisition, novel predictor variables, advances in model-developing approaches, improvements in validation methods, model transferability, and modeling software as reported in 155 LUR studies published between 2011 and 2023. We demonstrate that these developments have enabled LUR models to be developed for larger study areas and encompass a wider range of criteria and unregulated air pollutants. LUR models in the conventional spatial structure have been complemented by more complex spatiotemporal structures. Compared with linear models, advanced statistical methods yield better predictions when handling data with complex relationships and interactions. Finally, this study explores new developments, identifies potential pathways for further breakthroughs in LUR methodologies, and proposes future research directions. In this context, LUR approaches have the potential to make a significant contribution to future efforts to model the patterns of long- and short-term exposure of urban populations to air pollution.

Intra-urban variability of long-term exposure to PM2.5 and NO2 in five cities in Colombia

Rapidly urbanizing cities in Latin America experience high levels of air pollution which are known risk factors for population health. However, the estimates of long-term exposure to air pollution are scarce in the region. We developed intraurban land use regression (LUR) models to map long-term exposure to fine particulate matter (PM2.5) and nitrogen dioxide (NO2) in the five largest cities in Colombia. We conducted air pollution measurement campaigns using gravimetric PM2.5 and passive NO2 sensors for 2 weeks during both the dry and rainy seasons in 2021 in the cities of Barranquilla, Bucaramanga, Bogotá, Cali, and Medellín, and combined these data with geospatial and meteorological variables. Annual models were developed using multivariable spatial regression models. The city annual PM2.5 mean concentrations measured ranged between 12.32 and 15.99 µg/m3 while NO2 concentrations ranged between 24.92 and 49.15 µg/m3. The PM2.5 annual models explained 82% of the variance (R2) in Medellín, 77% in Bucaramanga, 73% in Barranquilla, 70% in Cali, and 44% in Bogotá. The NO2 models explained 65% of the variance in Bucaramanga, 57% in Medellín, 44% in Cali, 40% in Bogotá, and 30% in Barranquilla. Most of the predictor variables included in the models were a combination of specific land use characteristics and roadway variables. Cross-validation suggests that PM2.5 outperformed NO2 models. The developed models can be used as exposure estimate in epidemiological studies, as input in hybrid models to improve personal exposure assessment, and for policy evaluation.

Exposure to Multiple Air Pollutants and the Risk of Fractures: A Large Prospective Population-Based Study

Atmospheric chemistry studies suggest air pollution impedes ultraviolet B photons and thus reduces cutaneous vitamin D3 synthesis. Biological evidence shows that inhaled pollutants disrupt circulating 25-hydroxyvitamin D (25[OH]D) metabolism and ultimately impact bone health. The hypothesis is that higher air pollution concentrations are associated with a higher risk of fractures, mediated by lower circulating 25(OH)D. The study included participants of the UK Biobank who were free of fracture history at enrollment (2006 to 2010) and analyzed their environmental exposure data (2007 to 2010). Air pollution measurements included the annual averages of air particulate matter (PM2.5 , PM2.5-10 , and PM10 ), nitrogen oxides (NO2 and NOx ), and a composite air pollution score. Multivariable Cox proportional hazard models were used to assess the associations of the individual pollutants and the score with fracture risks. Mediation analyses were conducted to assess the underlying role of serum 25(OH)D in such associations. Among 446,395 participants with a median of 8-year follow-up, 12,288 incident fractures were documented. Participants living in places with the highest quintile of air pollution score had a 15.3% increased risk of fractures (hazard ratio [95%CI]: 1.15[1.09,1.22]) compared to those in the lowest, and 5.49% of this association was mediated through serum 25(OH)D (pmediation  < 0.05). Pollutant-specific hazard of top-to-bottom quintiles was 16% for PM2.5 , 4% for PM2.5-10 , 5% for PM10 , 20% for NO2 , and 17% for NOx , with a 4% to 6% mediation effect of serum 25(OH)D concentrations. The associations of the air pollution score with fracture risks were weaker among female participants, those who drank less alcohol, and consumed more fresh fruit than their counterparts (pinteraction  < 0.05). © 2023 American Society for Bone and Mineral Research (ASBMR).

High-resolution patterns and inequalities in ambient fine particle mass (PM2.5) and black carbon (BC) in the Greater Accra Metropolis, Ghana

Growing cities in sub-Saharan Africa (SSA) experience high levels of ambient air pollution. However, sparse long-term city-wide air pollution exposure data limits policy mitigation efforts and assessment of the health and climate effects. In the first study of its kind in West Africa, we developed high resolution spatiotemporal land use regression (LUR) models to map fine particulate matter (PM2.5) and black carbon (BC) concentrations in the Greater Accra Metropolitan Area (GAMA), one of the fastest sprawling metropolises in SSA. We conducted a one-year measurement campaign covering 146 sites and combined these data with geospatial and meteorological predictors to develop separate Harmattan and non-Harmattan season PM2.5 and BC models at 100 m resolution. The final models were selected with a forward stepwise procedure and performance was evaluated with 10-fold cross-validation. Model predictions were overlayed with the most recent census data to estimate the population distribution of exposure and socioeconomic inequalities in exposure at the census enumeration area level. The fixed effects components of the models explained 48-69 % and 63-71 % of the variance in PM2.5 and BC concentrations, respectively. Spatial variables related to road traffic and vegetation explained the most variability in the non-Harmattan models, while temporal variables were dominant in the Harmattan models. The entire GAMA population is exposed to PM2.5 levels above the World Health Organization guideline, including even the Interim Target 3 (15 μg/m3), with the highest exposures in poorer neighborhoods. The models can be used to support air pollution mitigation policies, health, and climate impact assessments. The measurement and modelling approach used in this study can be adapted to other African cities to bridge the air pollution data gap in the region.

Analysis of spatial distribution characteristics and main influencing factors of heavy metals in road dust of Tianjin based on land use regression models

Land use regression (LUR) models are mainly used for the simulation and prediction of conventional atmospheric pollutants. Whether the LUR models can be expanded to study more toxic and hazardous pollutants (such as heavy metals) remains to be verified. Combined with the factors of road, land use type, population, pollution enterprise, meteorology, and terrain, the LUR models were used to simulate the spatial distribution characteristics of heavy metals in road dust and determine the main influencing factors. Samples of road surface dust were collected from 144 evenly distributed points in Tianjin, China, with 108 modelling points and 36 verification points. The R² values of the LUR models of Cd, Cr, Cu, Ni, and Pb contents were 0.301, 0.412, 0.399, 0.496, and 0.377, and their error rates were 2.72%, 4.96%, 4.64%, 8.91%, and 4.94%, respectively. The error rates of the kriging interpolation models were 3.33%, 6.50%, 5.14%, 18.30%, and 22.87%, which were all greater than those of the LUR models. The estimation effect of the LUR models was more refined than that of the kriging interpolation models. The contents of most heavy metals (except Ni) in road dust of the central area in Tianjin were generally higher than those of the surrounding areas. The heavy metal contents in road dust of Tianjin were mainly affected by road variables and meteorological variables. The LUR models were suitable for small-scale spatial prediction of heavy metals in urban road dust within urban areas.

Development of transferable neighborhood land use regression models for predicting intra-urban ambient nitrogen dioxide (NO2) spatial variations

Land use regression (LUR) models have been extensively used to predict air pollution exposure in epidemiological and environmental studies. The lack of dense routine monitoring networks in big cities places increased emphasis on the need for LUR models to be developed using purpose-designed neighborhood-scale monitoring data. However, the unsatisfactory model transferability limits these neighborhood LUR models to be then applied to other intra-urban areas in predicting air pollution exposure. In this study, we tackled this issue by proposing a method to develop transferable neighborhood NO₂ LUR models with comparable predictive power based on only micro-scale predictor variables for modeling intra-urban ambient air pollution exposure. Taking Auckland metropolis, New Zealand, as a case study, the proposed method was applied to three neighborhoods (urban, central business district, and dominion road) and compared with the corresponding counterpart models developed using pools of (a) only macro-scale predictor variables and (b) a mixture of both micro- and macro-scale predictor variables (traditional method). The results showed that the models using only macro-scale variables achieved the lowest accuracy (R²: 0.388-0.484) and had the worst direct (R²: 0.0001-0.349) and indirect transferability (R²: 0.07-0.352). Those models using the traditional method had the highest model fitting R² (0.629-0.966) with lower cross-validation R² (0.495-0.941) and slightly better direct transferability (R²: 0.0003-0.386) but suffered poor model interpretability when indirectly transferred to new locations. Our proposed models had comparable model fitting R² (0.601-0.966) and the best cross-validation R² (0.514-0.941). They also had the strongest direct transferability (R²: 0.006-0.590) and moderate-to-good indirect transferability (R²: 0.072-0.850) with much better model interpretability. This study advances our knowledge of developing transferable LUR models for the very first time from the perspective of the scale of the predictor variables used in the model development and will significantly benefit the wider application of LUR approaches in epidemiological and environmental studies.

Urban environment and cognitive and motor function in children from four European birth cohorts

BACKGROUND

The urban environment may influence neurodevelopment from conception onwards, but there is no evaluation of the impact of multiple groups of exposures simultaneously. We investigated the association between early-life urban environment and cognitive and motor function in children.

METHODS

We used data from 5403 mother-child pairs from four population-based birth-cohorts (UK, France, Spain, and Greece). We estimated thirteen urban home exposures during pregnancy and childhood, including: built environment, natural spaces, and air pollution. Verbal, non-verbal, gross motor, and fine motor functions were assessed using validated tests at five years old. We ran adjusted multi-exposure models using the Deletion-Substitution-Addition algorithm.

RESULTS

Higher greenness exposure within 300 m during pregnancy was associated with higher verbal abilities (1.5 points (95% confidence interval 0.4, 2.7) per 0.20 unit increase in greenness). Higher connectivity density within 100 m and land use diversity during pregnancy were related to lower verbal abilities. Childhood exposure to PM2.5 mediated 74% of the association between greenness during childhood and verbal abilities. Higher exposure to PM2.5 during pregnancy was related to lower fine motor function (-1.2 points (-2.1, -0.4) per 3.2 μg/m3 increase in PM2.5). No associations were found with non-verbal abilities and gross motor function.

DISCUSSION

This study suggests that built environment, greenness, and air pollution may impact child cognitive and motor function at five years old. This study adds evidence that well-designed urban planning may benefit children's cognitive and motor development.

Prenatal and childhood exposure to air pollution and traffic and the risk of liver injury in European children

Nonalcoholic fatty liver disease is the most prevalent pediatric chronic liver disease. Experimental studies suggest effects of air pollution and traffic exposure on liver injury. We present the first large-scale human study to evaluate associations of prenatal and childhood air pollution and traffic exposure with liver injury.

METHODS

Study population included 1,102 children from the Human Early Life Exposome project. Established liver injury biomarkers, including alanine aminotransferase, aspartate aminotransferase, gamma-glutamyl transferase, and cytokeratin-18, were measured in serum between ages 6-10 years. Air pollutant exposures included nitrogen dioxide, particulate matter <10 μm (PM10), and <2.5 μm. Traffic measures included traffic density on nearest road, traffic load in 100-m buffer, and inverse distance to nearest road. Exposure assignments were made to residential address during pregnancy (prenatal) and residential and school addresses in year preceding follow-up (childhood). Childhood indoor air pollutant exposures were also examined. Generalized additive models were fitted adjusting for confounders. Interactions by sex and overweight/obese status were examined.

RESULTS

Prenatal and childhood exposures to air pollution and traffic were not associated with child liver injury biomarkers. There was a significant interaction between prenatal ambient PM10 and overweight/obese status for alanine aminotransferase, with stronger associations among children who were overweight/obese. There was no evidence of interaction with sex.

CONCLUSION

This study found no evidence for associations between prenatal or childhood air pollution or traffic exposure with liver injury biomarkers in children. Findings suggest PM10 associations maybe higher in children who are overweight/obese, consistent with the multiple-hits hypothesis for nonalcoholic fatty liver disease pathogenesis.

Fusion of land use regression modeling output and wireless distributed sensor network measurements into a high spatiotemporally-resolved NO2 product

Land use regression modeling is a common method for assessing exposure to ambient pollutants, yet it suffers from very coarse temporal resolution. Wireless distributed sensor networks (WDSN) is a promising technology that can provide extremely high spatiotemporal pollutant patterns but is known to suffer from several limitations that put into question its data reliability. This study examines the advantages of fusing data from these two methods and obtaining high spatiotemporally-resolved product that can be used for exposure assessment. We demonstrate this approach by estimating nitrogen dioxide (NO₂) concentrations at a sub-urban scale, with the study area limited by the deployment of the WDSN nodes. Specifically, hourly-resolved fused-data estimates were obtained by combining a stationary traffic-based land use regression (LUR) model with observations (15 min sampling frequency) made by an array of low-cost sensor nodes, with the sensors' readings mapped over the whole study area. Data fusion was performed by merging the two independent information products using a fuzzy logic approach. The performance of the fused product was examined against reference hourly observations at four air quality monitoring (AQM) stations situated within the study area, with the AQM data not used for the development of any of the underlying information layers. The mean hourly RMSE between the fused data product and the AQM records was 9.3 ppb, smaller than the RMSE of the two base products independently (LUR: 14.87 ppb, WDSN: 10.45 ppb). The normalized Moran's I of the fused product indicates that the data-fusion product reveals more realistic spatial patterns than those of the base products. The fused NO₂ concentration product shows considerable spatial variability relative to that evident by interpolation of both the WDSN records and the AQM stations data, with significant non-random patterns in 74% of the study period.

Development of Europe-Wide Models for Particle Elemental Composition Using Supervised Linear Regression and Random Forest

We developed Europe-wide models of long-term exposure to eight elements (copper, iron, potassium, nickel, sulfur, silicon, vanadium, and zinc) in particulate matter with diameter <2.5 μm (PM_2.5) using standardized measurements for one-year periods between October 2008 and April 2011 in 19 study areas across Europe, with supervised linear regression (SLR) and random forest (RF) algorithms. Potential predictor variables were obtained from satellites, chemical transport models, land-use, traffic, and industrial point source databases to represent different sources. Overall model performance across Europe was moderate to good for all elements with hold-out-validation R-squared ranging from 0.41 to 0.90. RF consistently outperformed SLR. Models explained within-area variation much less than the overall variation, with similar performance for RF and SLR. Maps proved a useful additional model evaluation tool. Models differed substantially between elements regarding major predictor variables, broadly reflecting known sources. Agreement between the two algorithm predictions was generally high at the overall European level and varied substantially at the national level. Applying the two models in epidemiological studies could lead to different associations with health. If both between- and within-area exposure variability are exploited, RF may be preferred. If only within-area variability is used, both methods should be interpreted equally.

Drivers of seasonal and annual air pollution exposure in a complex urban environment with multiple source contributions

Outdoor air pollution is a global health concern, but detailed exposure information is still limited for many parts of the world. In this study, high-resolution exposure surfaces were generated for annual and seasonal fine particulate matter (PM_2.5), coarse particulate matter (PM₁₀), and carbon monoxide (CO) for the Greater Beirut Area (GBA), Lebanon, an urban zone with a complex topography and multiple source contributions. Land use regression models (LUR) were calibrated and validated with monthly data collected from 58 locations between March 2017 and March 2018. The annual mean (±1 SD) concentrations of PM_2.5, PM₁₀, and CO across the monitoring locations were 68.1 (±15.7) μg/m³, 83.5 (±19.5) μg/m³, and 2.48 (±1.12) ppm, respectively. The coefficients of determination for LUR models ranged from 56 to 67% for PM_2.5, 44 to 63% for the PM₁₀ models, and 50 to 60% for the CO. LUR model structures varied significantly by season for both PM_2.5 and PM₁₀ but not for CO. Traffic emissions were consistently the main source of CO emissions throughout the year. The relative importance of industrial emissions and power generation sources towards predicted PM levels increased during the hot season while the contribution of the international airport diminished. Moreover, the complex topography of the study area along with the seasonal changes in the predominant wind directions affected the spatial predicted concentrations of all three pollutants. Overall, the predicted exposure surfaces were able to conserve the inter-pollution correlations determined from the field monitoring campaign, with the exception of the cold season. Our pollution surfaces suggest that the entire population of Beirut is regularly exposed to concentrations exceeding the World Health Organization (WHO) air quality standards for both PM_2.5 and PM₁₀.

Early-Life Environmental Exposures and Childhood Obesity: An Exposome-Wide Approach

BACKGROUND

Chemical and nonchemical environmental exposures are increasingly suspected to influence the development of obesity, especially during early life, but studies mostly consider single exposure groups.

OBJECTIVES

Our study aimed to systematically assess the association between a wide array of early-life environmental exposures and childhood obesity, using an exposome-wide approach.

METHODS

The HELIX (Human Early Life Exposome) study measured child body mass index (BMI), waist circumference, skinfold thickness, and body fat mass in 1,301 children from six European birth cohorts age 6-11 y. We estimated 77 prenatal exposures and 96 childhood exposures (cross-sectionally), including indoor and outdoor air pollutants, built environment, green spaces, tobacco smoking, and biomarkers of chemical pollutants (persistent organic pollutants, metals, phthalates, phenols, and pesticides). We used an exposure-wide association study (ExWAS) to screen all exposure-outcome associations independently and used the deletion-substitution-addition (DSA) variable selection algorithm to build a final multiexposure model.

RESULTS

The prevalence of overweight and obesity combined was 28.8%. Maternal smoking was the only prenatal exposure variable associated with higher child BMI (z-score increase of 0.28, 95% confidence interval: 0.09, 0.48, for active vs. no smoking). For childhood exposures, the multiexposure model identified particulate and nitrogen dioxide air pollution inside the home, urine cotinine levels indicative of secondhand smoke exposure, and residence in more densely populated areas and in areas with fewer facilities to be associated with increased child BMI. Child blood levels of copper and cesium were associated with higher BMI, and levels of organochlorine pollutants, cobalt, and molybdenum were associated with lower BMI. Similar results were found for the other adiposity outcomes.

DISCUSSION

This first comprehensive and systematic analysis of many suspected environmental obesogens strengthens evidence for an association of smoking, air pollution exposure, and characteristics of the built environment with childhood obesity risk. Cross-sectional biomarker results may suffer from reverse causality bias, whereby obesity status influenced the biomarker concentration. https://doi.org/10.1289/EHP5975.

Prenatal air pollution exposure and growth and cardio-metabolic risk in preschoolers

OBJECTIVES

We investigated the association between outdoor air pollutants exposure in the first trimester of pregnancy, and growth and cardio-metabolic risk at four years of age, and evaluated the mediating role of birth weight.

METHODS

We included mother-child pairs (N = 1,724) from the Spanish INMA birth cohort established in 2003-2008. First trimester of pregnancy nitrogen dioxide (NO₂) and fine particles (PM_2.5) exposure levels were estimated. Height, weight, waist circumference, blood pressure, and lipids were measured at four years of age. Body mass index (BMI) trajectories from birth to four years were identified.

RESULTS

Increased PM_2.5 exposure in the first trimester of pregnancy was associated with decreased z-scores of weight (zWeight) and BMI (zBMI) (zWeight change per interquartile range increase in PM_2.5 exposure = -0.12; 95% CI: -0.23, -0.01; zBMI change = -0.12; 95% CI: -0.23, -0.01). Higher NO₂ and PM_2.5 exposure was associated to a reduced risk of being in a trajectory with accelerated BMI gain, compared to children with the average trajectory. Birth weight partially mediated the association between PM_2.5 and zWeight and zBMI. PM_2.5 and NO₂ were not associated with the other cardio-metabolic risk factors.

CONCLUSIONS

This comprehensive study of many growth and cardio-metabolic risk related outcomes suggests that air pollution exposure during pregnancy may be associated with delays in physical growth in the early years after birth. These findings imply that pregnancy exposure to air pollutants has a lasting effect on growth after birth and require follow-up at later child ages.

Associations between air pollution and pediatric eczema, rhinoconjunctivitis and asthma: A meta-analysis of European birth cohorts

BACKGROUND

Uncertainly continues to exist regarding the role of air pollution on pediatric asthma and allergic conditions, especially as air pollution levels have started to decrease in recent decades.

OBJECTIVE

We examined associations of long-term air pollution levels at the home address with pediatric eczema, rhinoconjunctivitis and asthma prevalences in five birth cohorts (BIB, EDEN, GASPII, RHEA and INMA) from seven areas in five European countries.

METHODS

Current eczema, rhinoconjunctivitis and asthma were assessed in children aged four (N = 6527) and eight years (N = 2489). A multi-morbidity outcome (≥2 conditions versus none) was also defined. Individual outdoor levels of nitrogen dioxide (NO₂), nitrogen oxides, mass of particulate matter with an aerodynamic diameter <10 μm (PM₁₀), 10-2.5 μm (PM_coarse) and <2.5 μm (PM_2.5), and PM_2.5 absorbance were assigned to the birth, four- and eight-year home addresses using highly defined spatial air pollution exposure models. Cohort-specific cross-sectional associations were assessed using logistic regression models adjusted for demographic and environmental covariates and combined in a random effects meta-analysis.

RESULTS

The overall prevalence of pediatric eczema, rhinoconjunctivitis and asthma at four years was 15.4%, 5.9% and 12.4%. We found no increase in the prevalence of these outcomes at four or eight years with increasing air pollution exposure. For example, the meta-analysis adjusted odds ratios (95% confidence intervals) for eczema, rhinoconjunctivitis and asthma at four years were 0.94 (0.81, 1.09), 0.90 (0.75, 1.09), and 0.91 (0.74, 1.11), respectively, per 10 μg/m³ increase in NO₂ at the birth address, and 1.00 (0.81, 1.23), 0.70 (0.49, 1.00) and 0.88 (0.54, 1.45), respectively, per 5 μg/m³ increase in PM_2.5 at the birth address.

DISCUSSION

In this large meta-analysis of five birth cohorts, we found no indication of adverse effects of long-term air pollution exposure on the prevalence of current pediatric eczema, rhinoconjunctivitis or asthma.

Hourly Exposure to Ultrafine Particle Metrics and the Onset of Myocardial Infarction in Augsburg, Germany

BACKGROUND

Epidemiological evidence on the health effects of ultrafine particles (UFP) remains insufficient to infer a causal relationship that is largely due to different size ranges and exposure metrics examined across studies. Moreover, evidence regarding the association between UFP and cardiovascular disease at a sub-daily timescale is lacking.

OBJECTIVE

We investigated the relationship between different particle metrics, including particle number (PNC), length (PLC), and surface area (PSC) concentrations, and myocardial infarction (MI) at an hourly timescale.

METHODS

We collected hourly air pollution and meteorological data from fixed urban background monitoring sites and hourly nonfatal MI cases from a MI registry in Augsburg, Germany, during 2005-2015. We conducted a time-stratified case-crossover analysis with conditional logistic regression to estimate the association between hourly particle metrics and MI cases, adjusted for air temperature and relative humidity. We also examined the independent effects of a certain particle metric in two-pollutant models by adjusting for copollutants, including particulate matter (PM) with an aerodynamic diameter of ≤10μm or 2.5μm (PM10 and PM2.5, respectively), nitrogen dioxide, ozone, and black carbon.

RESULTS

Overall, a total of 5,898 cases of nonfatal MI cases were recorded. Exploratory analyses showed similar associations across particle metrics in the first 6-12 h. For example, interquartile range increases in PNC within the size range of 10-100 nm, PLC, and PSC were associated with an increase of MI 6 h later by 3.27% [95% confidence interval (CI): 0.27, 6.37], 5.71% (95% CI: 1.79, 9.77), and 5.84% (95% CI: 1.04, 10.87), respectively. Positive, albeit imprecise, associations were observed for PNC within the size range of 10-30 nm and 100-500 nm. Effect estimates for PLC and PSC remained similar after adjustment for PM and gaseous pollutants.

CONCLUSIONS

Transient exposure to particle number, length, and surface area concentrations or other potentially related exposures may trigger the onset of nonfatal myocardial infraction. https://doi.org/10.1289/EHP5478.

A comparison of linear regression, regularization, and machine learning algorithms to develop Europe-wide spatial models of fine particles and nitrogen dioxide

Empirical spatial air pollution models have been applied extensively to assess exposure in epidemiological studies with increasingly sophisticated and complex statistical algorithms beyond ordinary linear regression. However, different algorithms have rarely been compared in terms of their predictive ability. This study compared 16 algorithms to predict annual average fine particle (PM_2.5) and nitrogen dioxide (NO₂) concentrations across Europe. The evaluated algorithms included linear stepwise regression, regularization techniques and machine learning methods. Air pollution models were developed based on the 2010 routine monitoring data from the AIRBASE dataset maintained by the European Environmental Agency (543 sites for PM_2.5 and 2399 sites for NO₂), using satellite observations, dispersion model estimates and land use variables as predictors. We compared the models by performing five-fold cross-validation (CV) and by external validation (EV) using annual average concentrations measured at 416 (PM_2.5) and 1396 sites (NO₂) from the ESCAPE study. We further assessed the correlations between predictions by each pair of algorithms at the ESCAPE sites. For PM_2.5, the models performed similarly across algorithms with a mean CV R² of 0.59 and a mean EV R² of 0.53. Generalized boosted machine, random forest and bagging performed best (CV R²~0.63; EV R² 0.58-0.61), while backward stepwise linear regression, support vector regression and artificial neural network performed less well (CV R² 0.48-0.57; EV R² 0.39-0.46). Most of the PM_2.5 model predictions at ESCAPE sites were highly correlated (R² > 0.85, with the exception of predictions from the artificial neural network). For NO₂, the models performed even more similarly across different algorithms, with CV R²s ranging from 0.57 to 0.62, and EV R²s ranging from 0.49 to 0.51. The predicted concentrations from all algorithms at ESCAPE sites were highly correlated (R² > 0.9). For both pollutants, biases were low for all models except the artificial neural network. Dispersion model estimates and satellite observations were two of the most important predictors for PM_2.5 models whilst dispersion model estimates and traffic variables were most important for NO₂ models in all algorithms that allow assessment of the importance of variables. Different statistical algorithms performed similarly when modelling spatial variation in annual average air pollution concentrations using a large number of training sites.

Associations of Combined Exposures to Surrounding Green, Air Pollution, and Road Traffic Noise with Cardiometabolic Diseases

BACKGROUND

Surrounding green, air pollution, and noise have been associated with cardiometabolic diseases, but most studies have assessed only one of these correlated exposures.

OBJECTIVES

We aimed to evaluate associations of combined exposures to green, air pollution, and road traffic noise with cardiometabolic diseases.

METHODS

In this cross-sectional study, we studied associations between self-reported physician-diagnosed diabetes, hypertension, heart attack, and stroke from a Dutch national health survey of 387,195 adults and residential surrounding green, annual average air pollutant concentrations [including particulate matter with aerodynamic diameter [Formula: see text] ([Formula: see text]), PM with aerodynamic diameter [Formula: see text] ([Formula: see text]), nitrogen dioxide ([Formula: see text]), and oxidative potential (OP) with the dithiothreitol (DTT) assay ([Formula: see text])] and road traffic noise. Logistic regression models were used to analyze confounding and interaction of surrounding green, air pollution, and noise exposure.

RESULTS

In single-exposure models, surrounding green was inversely associated with diabetes, while air pollutants ([Formula: see text], [Formula: see text]) and road traffic noise were positively associated with diabetes. In two-exposure analyses, associations with green and air pollution were attenuated but remained. The association between road traffic noise and diabetes was reduced to unity when adjusted for surrounding green or air pollution. Air pollution and surrounding green, but not road traffic noise, were associated with hypertension in single-exposure models. The weak inverse association of surrounding green with hypertension attenuated and lost significance when adjusted for air pollution. Only [Formula: see text] was associated with stroke and heart attack.

CONCLUSIONS

Studies including only one of the correlated exposures surrounding green, air pollution, and road traffic noise may overestimate the association of diabetes and hypertension attributed to the studied exposure. https://doi.org/10.1289/EHP3857.

Prenatal and Childhood Traffic-Related Air Pollution Exposure and Telomere Length in European Children: The HELIX Project

BACKGROUND

Telomere length is a molecular marker of biological aging.

OBJECTIVE

Here we investigated whether early-life exposure to residential air pollution was associated with leukocyte telomere length (LTL) at 8 y of age.

METHODS

In a multicenter European birth cohort study, HELIX (Human Early Life Exposome) ([Formula: see text]), we estimated prenatal and 1-y childhood exposure to nitrogen dioxide ([Formula: see text]), particulate matter with aerodynamic diameter [Formula: see text] ([Formula: see text]), and proximity to major roads. Average relative LTL was measured using quantitative real-time polymerase chain reaction (qPCR). Effect estimates of the association between LTL and prenatal, 1-y childhood air pollution, and proximity to major roads were calculated using multiple linear mixed models with a random cohort effect and adjusted for relevant covariates.

RESULTS

LTL was inversely associated with prenatal and 1-y childhood [Formula: see text] and [Formula: see text] exposures levels. Each standard deviation (SD) increase in prenatal [Formula: see text] was associated with a [Formula: see text] (95% CI: [Formula: see text], [Formula: see text]) change in LTL. Prenatal [Formula: see text] was nonsignificantly associated with LTL ([Formula: see text] per SD increase; 95% CI: [Formula: see text], 0.6). For each SD increment in 1-y childhood [Formula: see text] and [Formula: see text] exposure, LTL shortened by [Formula: see text] (95% CI: [Formula: see text], [Formula: see text]) and [Formula: see text] (95% CI: [Formula: see text], 0.1), respectively. Each doubling in residential distance to nearest major road during childhood was associated with a 1.6% (95% CI: 0.02, 3.1) lengthening in LTL.

CONCLUSION

Lower exposures to air pollution during pregnancy and childhood were associated with longer telomeres in European children at 8 y of age. These results suggest that reductions in traffic-related air pollution may promote molecular longevity, as exemplified by telomere length, from early life onward. https://doi.org/10.1289/EHP4148.

Distribution of PM2.5 Air Pollution in Mexico City: Spatial Analysis with Land-Use Regression Model

In this study, the spatial distribution of PM2.5 air pollution in Mexico City from 37 personal exposures was modeled. Meteorological, demographic, geographic, and social data were also included. Geographic information systems (GIS), spatial analysis, and Land-Use Regression (LUR) were used to generate the final predictive model and the spatial distribution map which revealed two areas with very high concentrations (up to 109.3 µg/m3) and two more with lower concentrations (between 72 to 86.5 µg/m3) (p < 0.05). These results illustrate an overview trend of PM2.5 in relation to human activity during the studied periods in Mexico City and show a general approach to understanding the spatial variability of PM2.5.

Estimates of the 2016 global burden of kidney disease attributable to ambient fine particulate matter air pollution

OBJECTIVE

To quantitate the 2016 global and national burden of chronic kidney disease (CKD) attributable to ambient fine particulate matter air pollution ≤ 2.5 μm in aerodynamic diameter (PM_2.5).

DESIGN

We used the Global Burden of Disease (GBD) study data and methodologies to estimate the 2016 burden of CKD attributable to PM_2.5 in 194 countries and territories. Population-weighted PM_2.5 levels and incident rates of CKD for each country were curated from the GBD study publicly available data sources.

SETTING

GBD global and national data on PM_2.5 and CKD.

PARTICIPANTS

194 countries and territories.

MAIN OUTCOME MEASURES

We estimated the attributable burden of disease (ABD), years living with disability (YLD), years of life lost (YLL) and disability-adjusted life-years (DALYs).

RESULTS

The 2016 global burden of incident CKD attributable to PM_2.5 was 6 950 514 (95% uncertainty interval: 5 061 533-8 914 745). Global YLD, YLL and DALYs of CKD attributable to PM_2.5 were 2 849 311 (1 875 219-3 983 941), 8 587 735 (6 355 784-10 772 239) and 11 445 397 (8 380 246-14 554 091), respectively. Age-standardised ABD, YLL, YLD and DALY rates varied substantially among geographies. Populations in Mesoamerica, Northern Africa, several countries in the Eastern Mediterranean region, Afghanistan, Pakistan, India and several countries in Southeast Asia were among those with highest age-standardised DALY rates. For example, age-standardised DALYs per 100 000 were 543.35 (391.16-707.96) in El Salvador, 455.29 (332.51-577.97) in Mexico, 408.41 (283.82-551.84) in Guatemala, 238.25 (173.90-303.98) in India and 178.26 (125.31-238.47) in Sri Lanka, compared with 5.52 (0.82-11.48) in Sweden, 6.46 (0.00-14.49) in Australia and 12.13 (4.95-21.82) in Canada. Frontier analyses showed that Mesoamerican countries had significantly higher CKD DALY rates relative to other countries with comparable sociodemographic development.

CONCLUSIONS

Our results demonstrate that the global toll of CKD attributable to ambient air pollution is significant and identify several endemic geographies where air pollution may be a significant driver of CKD burden. Air pollution may need to be considered in the discussion of the global epidemiology of CKD.

Assessing the transferability of landuse regression models for ultrafine particles across two Canadian cities

Land use regression (LUR) models have been increasingly used to predict intra-city variations in the concentrations of different air pollutants. However, limited research assessing the transferability of these models between cities has been published to date. In this study, LUR models were generated for Ultra-Fine Particles (UFP) (<0.1 um) using data collected from mobile monitoring campaigns in two Canadian cities, Montreal and Toronto. City-specific models were first generated for each city before the models were transferred to the second city with and without recalibration. The calibrated transferred models showed only a slight decrease in performance, with the coefficient of determination (R²), dropping from 0.49 to 0.36 for Toronto and from 0.41 to 0.38 for Montreal. Transferring models between cities with no calibration resulted in low R²; 0.11 in Toronto and 0.18 in Montreal. Moreover, two additional models were generated by combining data from the two cities. The first combined model (CM1) assumed a spatially invariant effect of the predictors, while the second (CM2) relaxed the assumption of spatial invariance for some of the model coefficients. The performance of both combined models (R² ranged between 0.41 for CM1 and 0.43 for CM2; root mean squared error (RMSE) ranged between 0.34 for CM1 and 0.33 for CM2) was found to be on par with the Toronto city-specific model and outperformed the Montreal model. The results of this study highlight that the UFP LUR models appear to support transferability of model structures between cities with similar geographical characteristics, with a minor drop in model fit and predictive skill.

Influence of the Urban Exposome on Birth Weight

BACKGROUND

The exposome is defined as the totality of environmental exposures from conception onwards. It calls for providing a holistic view of environmental exposures and their effects on human health by evaluating multiple environmental exposures simultaneously during critical periods of life.

OBJECTIVE

We evaluated the association of the urban exposome with birth weight.

METHODS

We estimated exposure to the urban exposome, including the built environment, air pollution, road traffic noise, meteorology, natural space, and road traffic (corresponding to 24 environmental indicators and 60 exposures) for nearly 32,000 pregnant women from six European birth cohorts. To evaluate associations with either continuous birth weight or term low birth weight (TLBW) risk, we primarily relied on the Deletion-Substitution-Addition (DSA) algorithm, which is an extension of the stepwise variable selection method. Second, we used an exposure-by-exposure exposome-wide association studies (ExWAS) method accounting for multiple hypotheses testing to report associations not adjusted for coexposures.

RESULTS

The most consistent statistically significant associations were observed between increasing green space exposure estimated as Normalized Difference Vegetation Index (NDVI) and increased birth weight and decreased TLBW risk. Furthermore, we observed statistically significant associations among presence of public bus line, land use Shannon's Evenness Index, and traffic density and birth weight in our DSA analysis.

CONCLUSION

This investigation is the first large urban exposome study of birth weight that tests many environmental urban exposures. It confirmed previously reported associations for NDVI and generated new hypotheses for a number of built-environment exposures. https://doi.org/10.1289/EHP3971.

Development of land-use regression models for fine particles and black carbon in peri-urban South India

Land-use regression (LUR) has been used to model local spatial variability of particulate matter in cities of high-income countries. Performance of LUR models is unknown in less urbanized areas of low-/middle-income countries (LMICs) experiencing complex sources of ambient air pollution and which typically have limited land use data. To address these concerns, we developed LUR models using satellite imagery (e.g., vegetation, urbanicity) and manually-collected data from a comprehensive built-environment survey (e.g., roads, industries, non-residential places) for a peri-urban area outside Hyderabad, India. As part of the CHAI (Cardiovascular Health effects of Air pollution in Telangana, India) project, concentrations of fine particulate matter (PM_2.5) and black carbon were measured over two seasons at 23 sites. Annual mean (sd) was 34.1 (3.2) μg/m³ for PM_2.5 and 2.7 (0.5) μg/m³ for black carbon. The LUR model for annual black carbon explained 78% of total variance and included both local-scale (energy supply places) and regional-scale (roads) predictors. Explained variance was 58% for annual PM_2.5 and the included predictors were only regional (urbanicity, vegetation). During leave-one-out cross-validation and cross-holdout validation, only the black carbon model showed consistent performance. The LUR model for black carbon explained a substantial proportion of the spatial variability that could not be captured by simpler interpolation technique (ordinary kriging). This is the first study to develop a LUR model for ambient concentrations of PM_2.5 and black carbon in a non-urban area of LMICs, supporting the applicability of the LUR approach in such settings. Our results provide insights on the added value of manually-collected built-environment data to improve the performance of LUR models in settings with limited data availability. For both pollutants, LUR models predicted substantial within-village variability, an important feature for future epidemiological studies.

Comparison of land use regression models for NO2 based on routine and campaign monitoring data from an urban area of Japan

BACKGROUND

Typically, land use regression (LUR) models have been developed using campaign monitoring data rather than routine monitoring data. However, the latter have advantages such as low cost and long-term coverage. Based on the idea that LUR models representing regional differences in air pollution and regional road structures are optimal, the objective of this study was to evaluate the validity of LUR models for nitrogen dioxide (NO₂) based on routine and campaign monitoring data obtained from an urban area.

METHOD

We selected the city of Suita in Osaka (Japan). We built a model based on routine monitoring data obtained from all sites (routine-LUR-All), and a model based on campaign monitoring data (campaign-LUR) within the city. Models based on routine monitoring data obtained from background sites (routine-LUR-BS) and based on data obtained from roadside sites (routine-LUR-RS) were also built. The routine LUR models were based on monitoring networks across two prefectures (i.e., Osaka and Hyogo prefectures). We calculated the predictability of the each model. We then compared the predicted NO₂ concentrations from each model with measured annual average NO₂ concentrations from evaluation sites.

RESULTS

The routine-LUR-All and routine-LUR-BS models both predicted NO₂ concentrations well: adjusted R²=0.68 and 0.76, respectively, and root mean square error=3.4 and 2.1ppb, respectively. The predictions from the routine-LUR-All model were highly correlated with the measured NO₂ concentrations at evaluation sites. Although the predicted NO₂ concentrations from each model were correlated, the LUR models based on routine networks, and particularly those based on all monitoring sites, provided better visual representations of the local road conditions in the city.

CONCLUSION

The present study demonstrated that LUR models based on routine data could estimate local traffic-related air pollution in an urban area. The importance and usefulness of data from routine monitoring networks should be acknowledged.

The Urban Exposome during Pregnancy and Its Socioeconomic Determinants

BACKGROUND

The urban exposome is the set of environmental factors that are experienced in the outdoor urban environment and that may influence child development.

OBJECTIVE

The authors' goal was to describe the urban exposome among European pregnant women and understand its socioeconomic determinants.

METHODS

Using geographic information systems, remote sensing and spatio-temporal modeling we estimated exposure during pregnancy to 28 environmental indicators in almost 30,000 women from six population-based birth cohorts, in nine urban areas from across Europe. Exposures included meteorological factors, air pollutants, traffic noise, traffic indicators, natural space, the built environment, public transport, facilities, and walkability. Socioeconomic position (SEP), assessed at both the area and individual level, was related to the exposome through an exposome-wide association study and principal component (PC) analysis.

RESULTS

Mean±standard deviation (SD) NO2 levels ranged from 13.6±5.1 μg/m3 (in Heraklion, Crete) to 43.2±11 μg/m3 (in Sabadell, Spain), mean±SD walkability score ranged from 0.22±0.04 (Kaunas, Lithuania) to 0.32±0.07 (Valencia, Spain) and mean±SD Normalized Difference Vegetation Index ranged from 0.21±0.05 in Heraklion to 0.51±0.1 in Oslo, Norway. Four PCs explained more than half of variation in the urban exposome. There was considerable heterogeneity in social patterning of the urban exposome across cities. For example, high-SEP (based on family education) women lived in greener, less noisy, and less polluted areas in Bradford, UK (0.39 higher PC1 score, 95% confidence interval (CI): 0.31, 0.47), but the reverse was observed in Oslo (-0.57 PC1 score, 95% CI: -0.73, -0.41). For most cities, effects were stronger when SEP was assessed at the area level: In Bradford, women living in high SEP areas had a 1.34 higher average PC1 score (95% CI: 1.21, 1.48).

CONCLUSIONS

The urban exposome showed considerable variability across Europe. Pregnant women of low SEP were exposed to higher levels of environmental hazards in some cities, but not others, which may contribute to inequities in child health and development. https://doi.org/10.1289/EHP2862.

Outdoor air pollution, exhaled 8-isoprostane and current asthma in adults: the EGEA study

Associations between outdoor air pollution and asthma in adults are still scarce, and the underlying biological mechanisms are poorly understood. Our aim was to study the associations between 1) long-term exposure to outdoor air pollution and current asthma, 2) exhaled 8-isoprostane (8-iso; a biomarker related to oxidative stress) and current asthma, and 3) outdoor air pollution and exhaled 8-iso.Cross-sectional analyses were conducted in 608 adults (39% with current asthma) from the first follow-up of the French case-control and family study on asthma (EGEA; the Epidemiological study of the Genetic and Environmental factors of Asthma). Data on nitrogen dioxide, nitrogen oxides, particulate matter with a diameter ≤10 and ≤2.5 µm (PM₁₀ and PM_2.5), road traffic, and ozone (O₃) were from ESCAPE (European Study of Cohorts for Air Pollution Effects) and IFEN (French Institute for the Environment) assessments. Models took account of city and familial dependence.The risk of current asthma increased with traffic intensity (adjusted (a)OR 1.09 (95% CI 1.00-1.18) per 5000 vehicles per day), with O₃ exposure (aOR 2.04 (95% CI 1.27-3.29) per 10 µg·m^-3) and with exhaled 8-iso concentration (aOR 1.50 (95% CI 1.06-2.12) per 1 pg·mL^-1). Among participants without asthma, exhaled 8-iso concentration increased with PM_2.5 exposure (adjusted (a)β 0.23 (95% CI 0.005-0.46) per 5 µg·m^-3), and decreased with O₃ and O_3-summer exposures (aβ -0.20 (95% CI -0.39- -0.01) and aβ -0.52 (95% CI -0.77- -0.26) per 10 µg·m^-3, respectively).Our results add new insights into a potential role of oxidative stress in the associations between outdoor air pollution and asthma in adults.

Design of an Air Pollution Monitoring Campaign in Beijing for Application to Cohort Health Studies

No cohort studies in China on the health effects of long-term air pollution exposure have employed exposure estimates at the fine spatial scales desirable for cohort studies with individual-level health outcome data. Here we assess an array of modern air pollution exposure estimation approaches for assigning within-city exposure estimates in Beijing for individual pollutants and pollutant sources to individual members of a cohort. Issues considered in selecting specific monitoring data or new monitoring campaigns include: needed spatial resolution, exposure measurement error and its impact on health effect estimates, spatial alignment and compatibility with the cohort, and feasibility and expense. Sources of existing data largely include administrative monitoring data, predictions from air dispersion or chemical transport models and remote sensing (specifically satellite) data. New air monitoring campaigns include additional fixed site monitoring, snapshot monitoring, passive badge or micro-sensor saturation monitoring and mobile monitoring, as well as combinations of these. Each of these has relative advantages and disadvantages. It is concluded that a campaign in Beijing that at least includes a mobile monitoring component, when coupled with currently available spatio-temporal modeling methods, should be strongly considered. Such a campaign is economical and capable of providing the desired fine-scale spatial resolution for pollutants and sources.

Land-use regression with long-term satellite-based greenness index and culture-specific sources to model PM2.5 spatial-temporal variability

This study utilized a long-term satellite-based vegetation index, and considered culture-specific emission sources (temples and Chinese restaurants) with Land-use Regression (LUR) modelling to estimate the spatial-temporal variability of PM_2.5 using data from Taipei metropolis, which exhibits typical Asian city characteristics. Annual average PM_2.5 concentrations from 2006 to 2012 of 17 air quality monitoring stations established by Environmental Protection Administration of Taiwan were used for model development. PM_2.5 measurements from 2013 were used for external data verification. Monthly Normalized Difference Vegetation Index (NDVI) images coupled with buffer analysis were used to assess the spatial-temporal variations of greenness surrounding the monitoring sites. The distribution of temples and Chinese restaurants were included to represent the emission contributions from incense and joss money burning, and gas cooking, respectively. Spearman correlation coefficient and stepwise regression were used for LUR model development, and 10-fold cross-validation and external data verification were applied to verify the model reliability. The results showed a strongly negative correlation (r: -0.71 to -0.77) between NDVI and PM_2.5 while temples (r: 0.52 to 0.66) and Chinese restaurants (r: 0.31 to 0.44) were positively correlated to PM_2.5 concentrations. With the adjusted model R² of 0.89, a cross-validated adj-R² of 0.90, and external validated R² of 0.83, the high explanatory power of the resultant model was confirmed. Moreover, the averaged NDVI within a 1750 m circular buffer (p < 0.01), the number of Chinese restaurants within a 1750 m buffer (p < 0.01), and the number of temples within a 750 m buffer (p = 0.06) were selected as important predictors during the stepwise selection procedures. According to the partial R², NDVI explained 66% of PM_2.5 variation and was the dominant variable in the developed model. We suggest future studies consider these three factors when establishing LUR models for estimating PM_2.5 in other Asian cities.

Land Use Regression Models for Ultrafine Particles in Six European Areas

Long-term ultrafine particle (UFP) exposure estimates at a fine spatial scale are needed for epidemiological studies. Land use regression (LUR) models were developed and evaluated for six European areas based on repeated 30 min monitoring following standardized protocols. In each area; Basel (Switzerland), Heraklion (Greece), Amsterdam, Maastricht, and Utrecht ("The Netherlands"), Norwich (United Kingdom), Sabadell (Spain), and Turin (Italy), 160-240 sites were monitored to develop LUR models by supervised stepwise selection of GIS predictors. For each area and all areas combined, 10 models were developed in stratified random selections of 90% of sites. UFP prediction robustness was evaluated with the intraclass correlation coefficient (ICC) at 31-50 external sites per area. Models from Basel and The Netherlands were validated against repeated 24 h outdoor measurements. Structure and model R² of local models were similar within, but varied between areas (e.g., 38-43% Turin; 25-31% Sabadell). Robustness of predictions within areas was high (ICC 0.73-0.98). External validation R² was 53% in Basel and 50% in The Netherlands. Combined area models were robust (ICC 0.93-1.00) and explained UFP variation almost equally well as local models. In conclusion, robust UFP LUR models could be developed on short-term monitoring, explaining around 50% of spatial variance in longer-term measurements.

Land Use Regression Models for Ultrafine Particles in Six European Areas

Long-term ultrafine particle (UFP) exposure estimates at a fine spatial scale are needed for epidemiological studies. Land use regression (LUR) models were developed and evaluated for six European areas based on repeated 30 min monitoring following standardized protocols. In each area; Basel (Switzerland), Heraklion (Greece), Amsterdam, Maastricht, and Utrecht ("The Netherlands"), Norwich (United Kingdom), Sabadell (Spain), and Turin (Italy), 160-240 sites were monitored to develop LUR models by supervised stepwise selection of GIS predictors. For each area and all areas combined, 10 models were developed in stratified random selections of 90% of sites. UFP prediction robustness was evaluated with the intraclass correlation coefficient (ICC) at 31-50 external sites per area. Models from Basel and The Netherlands were validated against repeated 24 h outdoor measurements. Structure and model R² of local models were similar within, but varied between areas (e.g., 38-43% Turin; 25-31% Sabadell). Robustness of predictions within areas was high (ICC 0.73-0.98). External validation R² was 53% in Basel and 50% in The Netherlands. Combined area models were robust (ICC 0.93-1.00) and explained UFP variation almost equally well as local models. In conclusion, robust UFP LUR models could be developed on short-term monitoring, explaining around 50% of spatial variance in longer-term measurements.

A critical review of the ESCAPE project for estimating long-term health effects of air pollution

The European Study of Cohorts for Air Pollution Effects (ESCAPE) is a13-nation study of long-term health effects of air pollution based on subjects pooled from up to 22 cohorts that were intended for other purposes. Twenty-five papers have been published on associations of various health endpoints with long-term exposures to NOx, NO2, traffic indicators, PM10, PM2.5 and PM constituents including absorbance (elemental carbon). Seven additional ESCAPE papers found moderate correlations (R2=0.3-0.8) between measured air quality and estimates based on land-use regression that were used; personal exposures were not considered. I found no project summaries or comparisons across papers; here I conflate the 25 ESCAPE findings in the context of other recent European epidemiology studies. Because one ESCAPE cohort contributed about half of the subjects, I consider it and the other 18 cohorts separately to compare their contributions to the combined risk estimates. I emphasize PM2.5 and confirm the published hazard ratio of 1.14 (1.04-1.26) per 10μg/m3 for all-cause mortality. The ESCAPE papers found 16 statistically significant (p<0.05) risks among the125 pollutant-endpoint combinations; 4 each for PM2.5 and PM10, 1 for PM absorbance, 5 for NO2, and 2 for traffic. No PM constituent was consistently significant. No significant associations were reported for cardiovascular mortality; low birthrate was significant for all pollutants except PM absorbance. Based on associations with PM2.5, I find large differences between all-cause death estimates and the sum of specific-cause death estimates. Scatterplots of PM2.5 mortality risks by cause show no consistency across the 18 cohorts, ostensibly because of the relatively few subjects. Overall, I find the ESCAPE project inconclusive and I question whether the efforts required to estimate exposures for small cohorts were worthwhile. I suggest that detailed studies of the large cohort using historical exposures and additional cardiovascular risk factors might be productive.

Land use regression modeling of ultrafine particles, ozone, nitrogen oxides and markers of particulate matter pollution in Augsburg, Germany

Important health relevance has been suggested for ultrafine particles (UFP) and ozone, but studies on long-term effects are scarce, mainly due to the lack of appropriate spatial exposure models. We designed a measurement campaign to develop land use regression (LUR) models to predict the spatial variability focusing on particle number concentration (PNC) as indicator for UFP, ozone and several other air pollutants in the Augsburg region, Southern Germany. Three bi-weekly measurements of PNC, ozone, particulate matter (PM₁₀, PM_2.5), soot (PM_2.5abs) and nitrogen oxides (NO_x, NO₂) were performed at 20 sites in 2014/15. Annual average concentration were calculated and temporally adjusted by measurements from a continuous background station. As geographic predictors we offered several traffic and land use variables, altitude, population and building density. Models were validated using leave-one-out cross-validation. Adjusted model explained variance (R²) was high for PNC and ozone (0.89 and 0.88). Cross-validation adjusted R² was slightly lower (0.82 and 0.81) but still indicated a very good fit. LUR models for other pollutants performed well with adjusted R² between 0.68 (PM_coarse) and 0.94 (NO₂). Contrary to previous studies, ozone showed a moderate correlation with NO₂ (Pearson's r=-0.26). PNC was moderately correlated with ozone and PM_2.5, but highly correlated with NO_x (r=0.91). For PNC and NO_x, LUR models comprised similar predictors and future epidemiological analyses evaluating health effects need to consider these similarities.

Independent Validation of National Satellite-Based Land-Use Regression Models for Nitrogen Dioxide Using Passive Samplers

Including satellite observations of nitrogen dioxide (NO₂) in land-use regression (LUR) models can improve their predictive ability, but requires rigorous evaluation. We used 123 passive NO₂ samplers sited to capture within-city and near-road variability in two Australian cities (Sydney and Perth) to assess the validity of annual mean NO₂ estimates from existing national satellite-based LUR models (developed with 68 regulatory monitors). The samplers spanned roadside, urban near traffic (≤100 m to a major road), and urban background (>100 m to a major road) locations. We evaluated model performance using R² (predicted NO₂ regressed on independent measurements of NO₂), mean-square-error R² (MSE-R²), RMSE, and bias. Our models captured up to 69% of spatial variability in NO₂ at urban near-traffic and urban background locations, and up to 58% of variability at all validation sites, including roadside locations. The absolute agreement of measurements and predictions (measured by MSE-R²) was similar to their correlation (measured by R²). Few previous studies have performed independent evaluations of national satellite-based LUR models, and there is little information on the performance of models developed with a small number of NO₂ monitors. We have demonstrated that such models are a valid approach for estimating NO₂ exposures in Australian cities.

Development of West-European PM2.5 and NO2 land use regression models incorporating satellite-derived and chemical transport modelling data

Satellite-derived (SAT) and chemical transport model (CTM) estimates of PM_2.5 and NO₂ are increasingly used in combination with Land Use Regression (LUR) models. We aimed to compare the contribution of SAT and CTM data to the performance of LUR PM_2.5 and NO₂ models for Europe. Four sets of models, all including local traffic and land use variables, were compared (LUR without SAT or CTM, with SAT only, with CTM only, and with both SAT and CTM). LUR models were developed using two monitoring data sets: PM_2.5 and NO₂ ground level measurements from the European Study of Cohorts for Air Pollution Effects (ESCAPE) and from the European AIRBASE network. LUR PM_2.5 models including SAT and SAT+CTM explained ~60% of spatial variation in measured PM_2.5 concentrations, substantially more than the LUR model without SAT and CTM (adjR²: 0.33-0.38). For NO₂ CTM improved prediction modestly (adjR²: 0.58) compared to models without SAT and CTM (adjR²: 0.47-0.51). Both monitoring networks are capable of producing models explaining the spatial variance over a large study area. SAT and CTM estimates of PM_2.5 and NO₂ significantly improved the performance of high spatial resolution LUR models at the European scale for use in large epidemiological studies.

A New Technique for Evaluating Land-use Regression Models and Their Impact on Health Effect Estimates

BACKGROUND

Leave-one-out cross-validation that fails to account for variable selection does not properly reflect prediction accuracy when the number of training sites is small. The impact on health effect estimates has rarely been studied. The objective of this study was to develop an improved validation procedure for land-use regression models with variable selection and investigate health effect estimates in relation to land-use regression model performance.

METHODS

We randomly generated 10 training and test sets for nitrogen dioxide and particulate matter. For each training set, we developed models and evaluated them using a cross-holdout validation approach. Cross-holdout validation develops new models for each evaluation compared with refitting the model without variable selection, as in standard leave-one-out cross-validation. We also implemented holdout validation, which evaluates model predictions using independent test sets. We evaluated the relationship between cross-holdout validation and holdout validation R and estimates of the association between air pollution and forced vital capacity in the Dutch birth cohort.

RESULTS

Cross-holdout validation Rs were generally identical to holdout validation Rs, but were notably smaller than leave-one-out cross-validation Rs. Decreases in forced vital capacity in relation to air pollution exposure were larger for land-use regression models that had larger holdout validation and cross-holdout validation Rs rather than leave-one-out cross-validation R.

CONCLUSION

Cross-holdout validation accurately reflects predictive ability of land-use regression models and is a useful validation approach for small datasets. Land-use regression predictive ability in terms of holdout validation and cross-holdout validation rather than leave-one-out cross-validation was associated with the magnitude of health effect estimates in a case study.

Back-extrapolated and year-specific NO2 land use regression models for Great Britain - Do they yield different exposure assessment?

Robust methods to estimate historic population air pollution exposures are important tools for epidemiological studies evaluating long-term health effects. We developed land use regression (LUR) models for NO2 exposure in Great Britain for 1991 and explored whether the choice of year-specific or back-extrapolated LUR yields 1) similar LUR variables and model performance, and 2) similar national and regional address-level and small-area concentrations. We constructed two LUR models for 1991using NO2 concentrations from the diffusion tube monitoring network, one using 75% of all available measurement sites (that over-represent industrial areas), and the other using 75% of a subset of sites proportionate to population by region to study the effects of monitoring site selection bias. We compared, using the remaining (hold-out) 25% of monitoring sites, the performance of the two 1991 models with back-extrapolation of a previously published 2009 model, developed using NO2 concentrations from automatic chemiluminescence monitoring sites and predictor variables from 2006/2007. The 2009 model was back-extrapolated to 1991 using the same predictors (1990 & 1995) used to develop 1991 models. The 1991 models included industrial land use variables, not present for 2009. The hold-out performance of 1991 models (mean-squared-error-based-R(2): 0.62-0.64) was up to 8% higher and ~1μg/m(3) lower in root mean squared error than the back-extrapolated 2009 model, with best performance from the subset of sites representing population exposures. Year-specific and back-extrapolated exposures for residential addresses (n=1.338,399) and small areas (n=10.518) were very highly linearly correlated for Great Britain (r>0.83). This study suggests that year-specific model for 1991 and back-extrapolation of the 2009 LUR yield similar exposure assessment.

Development of land use regression models for nitrogen dioxide, ultrafine particles, lung deposited surface area, and four other markers of particulate matter pollution in the Swiss SAPALDIA regions

BACKGROUND

Land Use Regression (LUR) is a popular method to explain and predict spatial contrasts in air pollution concentrations, but LUR models for ultrafine particles, such as particle number concentration (PNC) are especially scarce. Moreover, no models have been previously presented for the lung deposited surface area (LDSA) of ultrafine particles. The additional value of ultrafine particle metrics has not been well investigated due to lack of exposure measurements and models.

METHODS

Air pollution measurements were performed in 2011 and 2012 in the eight areas of the Swiss SAPALDIA study at up to 40 sites per area for NO2 and at 20 sites in four areas for markers of particulate air pollution. We developed multi-area LUR models for biannual average concentrations of PM2.5, PM2.5 absorbance, PM10, PMcoarse, PNC and LDSA, as well as alpine, non-alpine and study area specific models for NO2, using predictor variables which were available at a national level. Models were validated using leave-one-out cross-validation, as well as independent external validation with routine monitoring data.

RESULTS

Model explained variance (R(2)) was moderate for the various PM mass fractions PM2.5 (0.57), PM10 (0.63) and PMcoarse (0.45), and was high for PM2.5 absorbance (0.81), PNC (0.87) and LDSA (0.91). Study-area specific LUR models for NO2 (R(2) range 0.52-0.89) outperformed combined-area alpine (R (2)  = 0.53) and non-alpine (R (2)  = 0.65) models in terms of both cross-validation and independent external validation, and were better able to account for between-area variability. Predictor variables related to traffic and national dispersion model estimates were important predictors.

CONCLUSIONS

LUR models for all pollutants captured spatial variability of long-term average concentrations, performed adequately in validation, and could be successfully applied to the SAPALDIA cohort. Dispersion model predictions or area indicators served well to capture the between area variance. For NO2, applying study-area specific models was preferable over applying combined-area alpine/non-alpine models. Correlations between pollutants were higher in the model predictions than in the measurements, so it will remain challenging to disentangle their health effects.

Long-Term Trends Worldwide in Ambient NO2 Concentrations Inferred from Satellite Observations

BACKGROUND

Air pollution is associated with morbidity and premature mortality. Satellite remote sensing provides globally consistent decadal-scale observations of ambient nitrogen dioxide (NO2) pollution.

OBJECTIVE

We determined global population-weighted annual mean NO2 concentrations from 1996 through 2012.

METHODS

We used observations of NO2 tropospheric column densities from three satellite instruments in combination with chemical transport modeling to produce a global 17-year record of ground-level NO2 at 0.1° × 0.1° resolution. We calculated linear trends in population-weighted annual mean NO2 (PWMNO2) concentrations in different regions around the world.

RESULTS

We found that PWMNO2 in high-income North America (Canada and the United States) decreased more steeply than in any other region, having declined at a rate of -4.7%/year [95% confidence interval (CI): -5.3, -4.1]. PWMNO2 decreased in western Europe at a rate of -2.5%/year (95% CI: -3.0, -2.1). The highest PWMNO2 occurred in high-income Asia Pacific (predominantly Japan and South Korea) in 1996, with a subsequent decrease of -2.1%/year (95% CI: -2.7, -1.5). In contrast, PWMNO2 almost tripled in East Asia (China, North Korea, and Taiwan) at a rate of 6.7%/year (95% CI: 6.0, 7.3). The satellite-derived estimates of trends in ground-level NO2 were consistent with regional trends inferred from data obtained from ground-station monitoring networks in North America (within 0.7%/year) and Europe (within 0.3%/year). Our rankings of regional average NO2 and long-term trends differed from the satellite-derived estimates of fine particulate matter reported elsewhere, demonstrating the utility of both indicators to describe changing pollutant mixtures.

CONCLUSIONS

Long-term trends in satellite-derived ambient NO2 provide new information about changing global exposure to ambient air pollution. Our estimates are publicly available at http://fizz.phys.dal.ca/~atmos/martin/?page_id=232.

Effect of time-activity adjustment on exposure assessment for traffic-related ultrafine particles

Exposures to ultrafine particles (<100 nm, estimated as particle number concentration, PNC) differ from ambient concentrations because of the spatial and temporal variability of both PNC and people. Our goal was to evaluate the influence of time-activity adjustment on exposure assignment and associations with blood biomarkers for a near-highway population. A regression model based on mobile monitoring and spatial and temporal variables was used to generate hourly ambient residential PNC for a full year for a subset of participants (n=140) in the Community Assessment of Freeway Exposure and Health study. We modified the ambient estimates for each hour using personal estimates of hourly time spent in five micro-environments (inside home, outside home, at work, commuting, other) as well as particle infiltration. Time-activity adjusted (TAA)-PNC values differed from residential ambient annual average (RAA)-PNC, with lower exposures predicted for participants who spent more time away from home. Employment status and distance to highway had a differential effect on TAA-PNC. We found associations of RAA-PNC with high sensitivity C-reactive protein and Interleukin-6, although exposure-response functions were non-monotonic. TAA-PNC associations had larger effect estimates and linear exposure-response functions. Our findings suggest that time-activity adjustment improves exposure assessment for air pollutants that vary greatly in space and time.

A retrospective analysis of environmental risk factors for the diagnosis of deep stromal abscess in 390 horses in North Central Florida from 1991 to 2013

PURPOSE

The purpose of this investigation was to identify potential environmental risk factors for the diagnosis of equine deep stromal abscesses (DSA) in the subtropical climate at the University of Florida Veterinary Medical Center (UFVMC).

METHODS

Cases included were selected from the UFVMC medical record and imaging database, and included all cases of equine DSA diagnosed during the period from December 1991 to December 2013 in patients residing in north central Florida. Patient date of diagnosis and atmospheric data was obtained for north central Florida for the corresponding time period. Univariate and multivariate general linear models were generated testing effects and interactions between environmental conditions.

RESULTS

When year, sulfur dioxide (SO2 ) and wind were analyzed in the presence of each other, a one-mile per hour increase in wind (P = 0.005) significantly increased the number of DSA cases by 1.63 cases per year. When the influence of temperature was evaluated in conjunction with year and nitrogen dioxide (NO2 ), the number of cases decreased by 0.1534 per year for every degree increase in temperature (°C) (P = 0.039).

CONCLUSIONS

Wind speed is the first significant atmospheric risk factor to be identified for DSA formation in the horse. The importance of environmental variance in the incidence of DSA indicates that the pathogenesis of DSA formation may be multifactorial, interdependent and provides support in some horses for the micropuncture hypothesis of DSA formation related to the involvement of environmental conditions causing precorneal tear film and epithelial damage.

Land use regression models for crustal and traffic-related PM2.5 constituents in four areas of the SAPALDIA study

Many studies have documented adverse health effects of long-term exposure to fine particulate matter (PM2.5), but there is still limited knowledge regarding the causal relationship between specific sources of PM2.5 and such health effects. The spatial variability of PM2.5 constituents and sources, as a exposure assessment strategy for investigating source contributions to health effects, has been little explored so far. Between 2011 and 2012, three measurement campaigns of PM and nitrogen dioxide (NO2) were performed in 80 sites across four areas of the Swiss Study on Air Pollution and Lung and heart Diseases in Adults (SAPALDIA). Reflectance analysis and energy dispersive X-ray fluorescence (XRF) were performed on PM2.5 filter samples to estimate light absorbance and trace element concentrations, respectively. Three air pollution source factors were identified using principal-component factor analysis: vehicular, crustal, and long-range transport. Land use regression (LUR) models were developed for temporally-adjusted scores of each factor, combining the four study areas. Model performance was assessed using two cross-validation methods. Model explained variance was high for the vehicular factor (R(2)=0.76), moderate for the crustal factor (R(2)=0.46), and low for the long-range transport factor (R(2)=0.19). The cross-validation methods suggested that models for the vehicular and crustal factors moderately accounted for both the between and within-area variability, and therefore can be applied to the four study areas to estimate long-term exposures within the SAPALDIA study population. The combination of source apportionment techniques and LUR modelling may help in identifying air pollution sources and disentangling their contribution to observed health effects in epidemiologic studies.

Transferability and generalizability of regression models of ultrafine particles in urban neighborhoods in the Boston area

Land use regression (LUR) models have been used to assess air pollutant exposure, but limited evidence exists on whether location-specific LUR models are applicable to other locations (transferability) or general models are applicable to smaller areas (generalizability). We tested transferability and generalizability of spatial-temporal LUR models of hourly particle number concentration (PNC) for Boston-area (MA, U.S.A.) urban neighborhoods near Interstate 93. Four neighborhood-specific regression models and one Boston-area model were developed from mobile monitoring measurements (34-46 days/neighborhood over one year each). Transferability was tested by applying each neighborhood-specific model to the other neighborhoods; generalizability was tested by applying the Boston-area model to each neighborhood. Both the transferability and generalizability of models were tested with and without neighborhood-specific calibration. Important PNC predictors (adjusted-R(2) = 0.24-0.43) included wind speed and direction, temperature, highway traffic volume, and distance from the highway edge. Direct model transferability was poor (R(2) < 0.17). Locally-calibrated transferred models (R(2) = 0.19-0.40) and the Boston-area model (adjusted-R(2) = 0.26, range: 0.13-0.30) performed similarly to neighborhood-specific models; however, some coefficients of locally calibrated transferred models were uninterpretable. Our results show that transferability of neighborhood-specific LUR models of hourly PNC was limited, but that a general model performed acceptably in multiple areas when calibrated with local data.

Applying land use regression model to estimate spatial variation of PM₂.₅ in Beijing, China

Fine particulate matter (PM2.5) is the major air pollutant in Beijing, posing serious threats to human health. Land use regression (LUR) has been widely used in predicting spatiotemporal variation of ambient air-pollutant concentrations, though restricted to the European and North American context. We aimed to estimate spatiotemporal variations of PM2.5 by building separate LUR models in Beijing. Hourly routine PM2.5 measurements were collected at 35 sites from 4th March 2013 to 5th March 2014. Seventy-seven predictor variables were generated in GIS, including street network, land cover, population density, catering services distribution, bus stop density, intersection density, and others. Eight LUR models were developed on annual, seasonal, peak/non-peak, and incremental concentration subsets. The annual mean concentration across all sites is 90.7 μg/m(3) (SD = 13.7). PM2.5 shows more temporal variation than spatial variation, indicating the necessity of building different models to capture spatiotemporal trends. The adjusted R (2) of these models range between 0.43 and 0.65. Most LUR models are driven by significant predictors including major road length, vegetation, and water land use. Annual outdoor exposure in Beijing is as high as 96.5 μg/m(3). This is among the first LUR studies implemented in a seriously air-polluted Chinese context, which generally produce acceptable results and reliable spatial air-pollution maps. Apart from the models for winter and incremental concentration, LUR models are driven by similar variables, suggesting that the spatial variations of PM2.5 remain steady for most of the time. Temporal variations are explained by the intercepts, and spatial variations in the measurements determine the strength of variable coefficients in our models.