Search

Objective: Health Impact Assessments (HIAs) for air pollutant mixtures are challenging because risk estimates are primarily derived from single-pollutant models. Combining risk estimates from multiple pollutants requires new approaches, as a simple addition of single pollutant risk estimates from correlated air pollutants may result in double counting. We investigated approaches applying concentration-response functions (CRFs) from single- and two-pollutant models in HIAs, focusing on long-term exposure to particulate matter with a diameter less than 2.5 μm (PM 2.5 ) and nitrogen dioxide (NO 2 ) and their associations with all-cause mortality., Methods: A systematic literature search of MEDLINE and EMBASE identified cohort studies employing single- and two-pollutant models of long-term exposure to PM 2.5 and NO 2 with all-cause mortality. Pooled CRFs were calculated through random-effects meta-analyses of risk estimates from single- and two-pollutant models. Coefficient differences were calculated by comparing single- and two-pollutant model estimates. Four approaches to estimating population-attributable fractions (PAFs) were compared: PM 2.5 or NO 2 single-pollutant models to represent the mixture, the sum of single-pollutant models, the sum of two-pollutant models and the sum of single-pollutant models from a larger body of evidence adjusted by coefficient difference., Results: Seventeen papers reported both single and two-pollutant estimates. Pooled hazard ratios (HRs) for mortality from single- and two-pollutant models were 1.053 (95% confidence interval: 1.034-1.071) and 1.035 (1.014-1.057), respectively, for a 5 μg/m 3 increase in PM 2.5 . HRs for a 10 μg/m 3 increase in NO 2 were 1.032 (1.014-1.049) and 1.024 (1.000-1.049) for single- and two-pollutant models, respectively. The average coefficient difference between single- and two-pollutant models was 0.017 for PM 2.5 and 0.007 for NO 2 . Combined PAFs for the PM 2.5 -NO 2 mixture using joint HRs from single- and two-pollutant model CRFs were 0.09 and 0.06, respectively., Conclusion: Utilizing CRFs from two-pollutant models or applying the coefficient difference to a more extensive evidence base seems to mitigate the potential overestimation of mixture health impacts from adding single-pollutant CRFs., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 The Authors. Published by Elsevier Inc. All rights reserved.)

Annual average land-use regression (LUR) models have been widely used to assess spatial patterns of air pollution exposures. However, they fail to capture diurnal variability in air pollution and consequently might result in biased dynamic exposure assessments. In this study we aimed to model average hourly concentrations for two major pollutants, NO 2 and PM 2.5 , for the Netherlands using the LUR algorithm. We modelled the spatial variation of average hourly concentrations for the years 2016-2019 combined, for two seasons, and for two weekday types. Two modelling approaches were used, supervised linear regression (SLR) and random forest (RF). The potential predictors included population, road, land use, satellite retrievals, and chemical transport model pollution estimates variables with different buffer sizes. We also temporally adjusted hourly concentrations from a 2019 annual model using the hourly monitoring data, to compare its performance with the hourly modelling approach. The results showed that hourly NO 2 models performed overall well (5-fold cross validation R 2  = 0.50-0.78), while the PM 2.5 performed moderately (5-fold cross validation R 2  = 0.24-0.62). Both for NO 2 and PM 2.5 the warm season models performed worse than the cold season ones, and the weekends' worse than weekdays'. The performance of the RF and SLR models was similar for both pollutants. For both SLR and RF, variables with larger buffer sizes representing variation in background concentrations, were selected more often in the weekend models compared to the weekdays, and in the warm season compared to the cold one. Temporal adjustment of annual average models performed overall worse than both modelling approaches (NO 2 hourly R 2  = 0.35-0.70; PM 2.5 hourly R 2  = 0.01-0.15). The difference in model performance and selection of variables across hours, seasons, and weekday types documents the benefit to develop independent hourly models when matching it to hourly time activity data., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 The Authors. Published by Elsevier Inc. All rights reserved.)

Background: Air pollution contributes to a large disease burden and some populations are disproportionately exposed. We aimed to evaluate ethnic and socioeconomic differences in exposure to air pollution in the Netherlands., Methods: We did a nationwide, cross-sectional analysis of all residents of the Netherlands on Jan 1, 2019. Sociodemographic information was centralised by Statistics Netherlands and mainly originated from the National Population Register, the tax register, and education registers. Concentrations of NO 2 , PM 2·5 , PM 10 , and elemental carbon, modelled by the National Institute for Public Health and the Environment, were linked to the individual-level demographic data. We assessed differences in air pollution exposures across the 40 largest minority ethnic groups. Evaluation of how ethnicity intersected with socioeconomic position in relation to exposures was done for the ten largest ethnic groups, plus Chinese and Indian groups, in both urban and rural areas using multivariable linear regression analyses., Findings: The total study population consisted of 17 251 511 individuals. Minority ethnic groups were consistently exposed to higher levels of air pollution than the ethnic Dutch population. The magnitude of inequalities varied between the minority ethnic groups, with 3-44% higher exposures to NO 2 and 1-9% higher exposures to PM 2·5 compared with the ethnic Dutch group. Average exposures were highest for the lowest socioeconomic group. Ethnic inequalities in exposure remained after adjustment for socioeconomic position and were of similar magnitude in urban and rural areas., Interpretation: The variability in air pollution exposure across ethnic and socioeconomic subgroups in the Netherlands indicates environmental injustice at the intersection of social characteristics. The health consequences of the observed inequalities and the underlying processes driving them warrant further investigation., Funding: The Gravitation programme of the Dutch Ministry of Education, Culture, and Science, the Netherlands Organization for Scientific Research, the Netherlands Organisation for Health Research and Development, and Amsterdam University Medical Center., Competing Interests: Declaration of interests We declare no competing interests., (Copyright © 2024 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. Published by Elsevier Ltd.. All rights reserved.)

Hyperlocal air quality maps are becoming increasingly common, as they provide useful insights into the spatial variation and sources of air pollutants. In this study, we produced several high-resolution concentration maps to assess the spatial differences of three traffic-related pollutants, Nitrogen dioxide (NO 2 ), Black Carbon (BC) and Ultrafine Particles (UFP), in Amsterdam, the Netherlands, and Copenhagen, Denmark. All maps were based on a mixed-effect model approach by using state-of-the-art mobile measurements conducted by Google Street View (GSV) cars, during October 2018 - March 2020, and Land-use Regression (LUR) models based on several land-use and traffic predictor variables. We then explored the concentration ratio between the different normalised pollutants to understand possible contributing sources to the observed hyperlocal variations. The maps developed in this work reflect, (i) expected elevated pollution concentrations along busy roads, and (ii) similar concentration patterns on specific road types, e.g., motorways, for both cities. In the ratio maps, we observed a clear pattern of elevated concentrations of UFP near the airport in both cities, compared to BC and NO 2 . This is the first study to produce hyperlocal maps for BC and UFP using high-quality mobile measurements. These maps are important for policymakers and health-effect studies, trying to disentangle individual effects of key air pollutants of interest (e.g., UFP)., Competing Interests: Declaration of Competing Interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Jules Kerckhoffs reports financial support was provided by Environmental Defense Fund. Jules Kerckhoffs reports financial support was provided by EXPOSOME-NL. Jules Kerckhoffs reports financial support was provided by EXPANSE. Jibran Khan reports financial support was provided by Danish Big Data Centre for Environment and Health., (Copyright © 2022 The Authors. Published by Elsevier Ltd.. All rights reserved.)

Background: Emerging evidence suggests that air pollution may contribute to childhood asthma development. We estimated the burden of incident childhood asthma that may be attributable to outdoor nitrogen dioxide (NO 2 ), particulate matter ≤2.5 µm in diameter (PM 2.5 ) and black carbon (BC) in Europe., Methods: We combined country-level childhood incidence rates and pooled exposure-response functions with childhood (age 1-14 years) population counts, and exposure estimates at 1 540 386 1 km×1 km cells, across 18 European countries and 63 442 419 children. Annual average pollutant concentrations were obtained from a validated and harmonised European land-use regression model. We investigated two exposure reduction scenarios. For the first, we used recommended annual World Health Organization (WHO) air quality guideline values. For the second, we used the minimum air pollution levels recorded across 41 studies in the underlying meta-analysis., Results: NO 2 ranged from 1.4 to 70.0 µg·m -3 , with a mean of 11.8 µg·m -3 . PM 2.5 ranged from 2.0 to 41.1 µg·m -3 , with a mean of 11.6 µg·m -3 . BC ranged from 0.003 to 3.7×10 -5  m -1 , with a mean of 1.0×10 -5  m -1 . Compliance with the NO 2 and PM 2.5 WHO guidelines was estimated to prevent 2434 (0.4%) and 66 567 (11%) incident cases, respectively. Meeting the minimum air pollution levels for NO 2 (1.5 µg·m -3 ), PM 2.5 (0.4 µg·m -3 ) and BC (0.4×10 -5  m -1 ) was estimated to prevent 135 257 (23%), 191 883 (33%) and 89 191 (15%) incident cases, respectively., Conclusions: A significant proportion of childhood asthma cases may be attributable to outdoor air pollution and these cases could be prevented. Our estimates underline an urgent need to reduce children's exposure to air pollution., Competing Interests: Conflict of interest: H. Khreis has nothing to disclose. Conflict of interest: M. Cirach has nothing to disclose. Conflict of interest: N. Mueller has nothing to disclose. Conflict of interest: K. de Hoogh has nothing to disclose. Conflict of interest: G. Hoek has nothing to disclose. Conflict of interest: M.J. Nieuwenhuijsen has nothing to disclose. Conflict of interest: D. Rojas-Rueda has nothing to disclose., (Copyright ©ERS 2019.)

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 (NO 2 ) 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 NO 2 ), 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 (NO 2 ) 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 R 2 of 0.59 and a mean EV R 2 of 0.53. Generalized boosted machine, random forest and bagging performed best (CV R 2 ~0.63; EV R 2 0.58-0.61), while backward stepwise linear regression, support vector regression and artificial neural network performed less well (CV R 2 0.48-0.57; EV R 2 0.39-0.46). Most of the PM 2.5 model predictions at ESCAPE sites were highly correlated (R 2  > 0.85, with the exception of predictions from the artificial neural network). For NO 2 , the models performed even more similarly across different algorithms, with CV R 2 s ranging from 0.57 to 0.62, and EV R 2 s ranging from 0.49 to 0.51. The predicted concentrations from all algorithms at ESCAPE sites were highly correlated (R 2  > 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 NO 2 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., (Copyright © 2019 The Authors. Published by Elsevier Ltd.. All rights reserved.)

Background: Methods for estimating air pollutant exposures for epidemiological studies are becoming more complex in an effort to minimise exposure error and its associated bias. While land use regression (LUR) modelling is now an established method, there has been little comparison between LUR and other recent, more complex estimation methods. Our aim was to develop a LUR model to estimate intra-city exposures to nitrogen dioxide (NO 2 ) for a Sydney cohort, and to compare those with estimates from a national satellite-based LUR model (Sat-LUR) and a regional Bayesian Maximum Entropy (BME) model., Methods: Satellite-based LUR and BME estimates were obtained using existing models. We used methods consistent with the European Study of Cohorts for Air Pollution Effects (ESCAPE) methodology to develop LUR models for NO 2 and NOx. We deployed 46 Ogawa passive samplers across western Sydney during 2013/2014 and acquired data on land use, population density, and traffic volumes for the study area. Annual average NO 2 concentrations for 2013 were estimated for 947 addresses in the study area using the three models: standard LUR, Sat-LUR and a BME model. Agreement between the estimates from the three models was assessed using interclass correlation coefficient (ICC), Bland-Altman methods and correlation analysis (CC)., Results: The NO 2 LUR model predicted 84% of spatial variability in annual mean NO 2 (RMSE: 1.2 ppb; cross-validated R 2 : 0.82) with predictors of major roads, population and dwelling density, heavy traffic and commercial land use. A separate model was developed that captured 92% of variability in NOx (RMSE 2.3 ppb; cross-validated R 2 : 0.90). The annual average NO 2 concentrations were 7.31 ppb (SD: 1.91), 7.01 ppb (SD: 1.92) and 7.90 ppb (SD: 1.85), for the LUR, Sat-LUR and BME models respectively. Comparing the standard LUR with Sat-LUR NO 2 cohort estimates, the mean estimates from the LUR were 4% higher than the Sat-LUR estimates, and the ICC was 0.73. The Pearson's correlation coefficients (CC) for the LUR vs Sat-LUR values were r = 0.73 (log-transformed data) and r = 0.69 (untransformed data). Comparison of the NO 2 cohort estimates from the LUR model with the BME blended model indicated that the LUR mean estimates were 8% lower than the BME estimates. The ICC for the LUR vs BME estimates was 0.73. The CC for the logged LUR vs BME estimates was r = 0.73 and for the unlogged estimates was r = 0.69., Conclusions: Our LUR models explained a high degree of spatial variability in annual mean NO 2 and NOx in western Sydney. The results indicate very good agreement between the intra-city LUR, national-scale sat-LUR, and regional BME models for estimating NO 2 for a cohort of children residing in Sydney, despite the different data inputs and differences in spatial scales of the models, providing confidence in their use in epidemiological studies., (Copyright © 2019 Elsevier Inc. All rights reserved.)

Background: The evidence from observational epidemiological studies of a link between long-term air pollution exposure and diabetes prevalence and incidence is currently mixed. Some studies found the strongest associations of diabetes with fine particles, other studies with nitrogen dioxide and some studies found no associations., Objectives: Our aim was to investigate associations between long-term exposure to multiple air pollutants and diabetes prevalence in a large national survey in the Netherlands., Methods: We performed a cross-sectional analysis using the 2012 Dutch national health survey to investigate the associations between the 2009 annual average concentrations of multiple air pollutants (PM 10 , PM 2.5 , PM 10-2.5 , PM 2.5 absorbance, OP DTT , OP ESR and NO 2 ) and diabetes prevalence, among 289,703 adults. Air pollution exposure was assessed by land use regression models. Diabetes was defined based on a combined measure of self-reported physician diagnosis and medication prescription from an external database. Using logistic regression, we adjusted for potential confounders, including neighborhood- and individual socio-economic status and lifestyle-related risk factors such as smoking habits, alcohol consumption, physical activity and BMI., Results: After adjustment for potential confounders, all pollutants (except PM 2.5 ) were associated with diabetes prevalence. In two-pollutant models, NO 2 and OP DTT remained associated with increased diabetes prevalence. For NO 2 and OP DTT , single-pollutant ORs per interquartile range were 1.07 (95% CI: 1.05, 1.09) and 1.08 (95% CI: 1.05, 1.10), respectively. Stratified analysis showed no consistent effect modification by any of the included known diabetes risk factors., Conclusions: Long-term residential air pollution exposure was associated with diabetes prevalence in a large health survey in the Netherlands, strengthening the evidence of air pollution being an important diabetes risk factor. Most consistent associations were observed for NO 2 and oxidative potential of PM 2.5 measured by the DTT assay. The finding of an association with the oxidative potential of fine particles but not with PM 2.5 , suggests that particle composition may be important for a potential effect on diabetes., (Copyright © 2017 The Authors. Published by Elsevier Ltd.. All rights reserved.)

Both dispersion modeling (DM) and land-use regression modeling (LUR) are often used for assessment of long-term air pollution exposure in epidemiological studies, but seldom in combination. We developed a hybrid DM-LUR model using 93 biweekly observations of NO x at 31 sites in greater Stockholm (Sweden). The DM was based on spatially resolved topographic, physiographic and emission data, and hourly meteorological data from a diagnostic wind model. Other data were from land use, meteorology and routine monitoring of NO x . We built a linear regression model for NO x , using a stepwise forward selection of covariates. The resulting model predicted observed NO x (R 2 =0.89) better than the DM without covariates (R 2 =0.68, P-interaction <0.001) and with minimal apparent bias. The model included (in descending order of importance) DM, traffic intensity on the nearest street, population (number of inhabitants) within 100 m radius, global radiation (direct sunlight plus diffuse or scattered light) and urban contribution to NO x levels (routine urban NO x , less routine rural NO x ). Our results indicate that there is a potential for improving estimates of air pollutant concentrations based on DM, by incorporating further spatial characteristics of the immediate surroundings, possibly accounting for imperfections in the emission data.

Background: Recent reports have suggested that air pollution mixtures represented by nitrogen dioxide (NO 2 ) may have effects on human health, which are independent from those of particulate matter mass. We evaluate the association between NO 2 and daily mortality among elderly using one- and multipollutant models., Methods: This study was a daily time series of non-accidental and cause-specific mortality among the elderly living in São Paulo, Brazil, between 2000 and 2011. Effects of NO 2 , particulate matter smaller than 10µm (PM10), carbon monoxide (CO) and ozone (O 3 ) were estimated in Poisson generalized additive models. The single lag effect at lags 0 and 1 days and the cumulative effect from 0 to lag 10 days were evaluated in one-, two-, three- and four-pollutant models. The cumulative risk index (CRI) recently proposed to analyze associations with health of multiple correlated pollutants was additionally estimated for each multipollutant model., Results: An association between NO 2 , PM10, CO and O 3 exposures and non-accidental and cause-specific deaths was found in one-pollutant models. NO 2 effects remained significant in multipollutant models for non-accidental and circulatory deaths. The estimated CRIs suggested that circulatory deaths were mainly associated with NO 2 , and respiratory deaths mainly with CO and O 3 , regardless the lag. For non-accidental deaths, multipollutant models were associated with the highest CRI, with the main pollutants depending on the chosen lag., Conclusions: The results suggest that air pollution mixtures represented by NO 2 have an effect on non-accidental and circulatory mortality, which is independent from PM10, CO and O 3 . The CRI was always larger than the risks associated with single pollutants., (Copyright © 2017 Elsevier Inc. All rights reserved.)

Satellite-derived (SAT) and chemical transport model (CTM) estimates of PM 2.5 and NO 2 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 NO 2 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 NO 2 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 (adjR 2 : 0.33-0.38). For NO 2 CTM improved prediction modestly (adjR 2 : 0.58) compared to models without SAT and CTM (adjR 2 : 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 NO 2 significantly improved the performance of high spatial resolution LUR models at the European scale for use in large epidemiological studies., (Copyright © 2016 Elsevier Inc. All rights reserved.)

Background: Air pollution episodes are associated with increased cardiopulmonary hospital admissions. Cohort studies showed associations of spatial variation in traffic-related air pollution with respiratory and cardiovascular mortality. Much less is known in particular about associations with cardiovascular morbidity. We explored the relation between spatial variation in nitrogen dioxide (NO 2 ) concentrations and cardiopulmonary hospital admissions., Methods: This ecological study was based on hospital admissions data (2001-2004) from the National Medical Registration and general population data for the West of the Netherlands (population 4.04 million). At the 4-digit postcode area level (n=683) associations between modeled annual average outdoor NO 2 concentrations and hospital admissions for respiratory and cardiovascular causes were evaluated by linear regression with the log of the postcode-specific percentage of subjects that have been admitted at least once during the study period as the dependent variable. All analyses were adjusted for differences in composition of the population of the postcode areas (age, sex, income)., Results: At the postcode level, positive associations were found between outdoor NO 2 concentrations and hospital admission rates for asthma, chronic obstructive pulmonary disease (COPD), all cardiovascular causes, ischemic heart disease and stroke (e.g. adjusted relative risk (95% confidence interval) for the second to fourth quartile relative to the first quartile of exposure were 1.87 (1.46-2.40), 2.34 (1.83-3.01) and 2.81 (2.16-3.65) for asthma; 1.44 (1.19-1.74), 1.50 (1.24-1.82) and 1.60 (1.31-1.96) for COPD). Associations remained after additional (indirect) adjustment for smoking (COPD admission rate) and degree of urbanization., Conclusions: Our study suggests an increased risk of hospitalization for respiratory and cardiovascular causes in areas with higher levels of NO 2 . Our findings add to the currently limited evidence of a long-term effect of air pollution on hospitalization. The ecological design of our study is a limitation and more studies with individual data are needed to confirm our findings., (Copyright © 2016 Elsevier Inc. All rights reserved.)

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., (Copyright © 2016 Elsevier Ltd. All rights reserved.)

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., Competing Interests: The authors report no conflicts of interest.

Background: There is limited knowledge about the extent to which estimates of air pollution effects on health are affected by the choice for a specific exposure model., Objectives: We aimed to evaluate the correlation between long-term air pollution exposure estimates using two commonly used exposure modeling techniques [dispersion and land use regression (LUR) models] and, in addition, to compare the estimates of the association between long-term exposure to air pollution and lung function in children using these exposure modeling techniques., Methods: We used data of 1,058 participants of a Dutch birth cohort study with measured forced expiratory volume in 1 sec (FEV1), forced vital capacity (FVC), and peak expiratory flow (PEF) measurements at 8 years of age. For each child, annual average outdoor air pollution exposure [nitrogen dioxide (NO2), mass concentration of particulate matter with diameters ≤ 2.5 and ≤ 10 μm (PM2.5, PM10), and PM2.5 soot] was estimated for the current addresses of the participants by a dispersion and a LUR model. Associations between exposures to air pollution and lung function parameters were estimated using linear regression analysis with confounder adjustment., Results: Correlations between LUR- and dispersion-modeled pollution concentrations were high for NO2, PM2.5, and PM2.5 soot (R = 0.86-0.90) but low for PM10 (R = 0.57). Associations with lung function were similar for air pollutant exposures estimated using LUR and dispersion modeling, except for associations of PM2.5 with FEV1 and FVC, which were stronger but less precise for exposures based on LUR compared with dispersion model., Conclusions: Predictions from LUR and dispersion models correlated very well for PM2.5, NO2, and PM2.5 soot but not for PM10. Health effect estimates did not depend on the type of model used to estimate exposure in a population of Dutch children.

Background: Long-term exposure to air pollution has been associated with mortality in urban cohort studies. Few studies have investigated this association in large-scale population registries, including non-urban populations., Objectives: The aim of the study was to evaluate the associations between long-term exposure to air pollution and nonaccidental and cause-specific mortality in the Netherlands based on existing national databases., Methods: We used existing Dutch national databases on mortality, individual characteristics, residence history, neighborhood characteristics, and national air pollution maps based on land use regression (LUR) techniques for particulates with an aerodynamic diameter ≤ 10 μm (PM10) and nitrogen dioxide (NO2). Using these databases, we established a cohort of 7.1 million individuals ≥ 30 years of age. We followed the cohort for 7 years (2004-2011). We applied Cox proportional hazard models adjusting for potential individual and area-specific confounders., Results: After adjustment for individual and area-specific confounders, for each 10-μg/m3 increase, PM10 and NO2 were associated with nonaccidental mortality [hazard ratio (HR) = 1.08; 95% CI: 1.07, 1.09 and HR = 1.03; 95% CI: 1.02, 1.03, respectively], respiratory mortality (HR = 1.13; 95% CI: 1.10, 1.17 and HR = 1.02; 95% CI: 1.01, 1.03, respectively), and lung cancer mortality (HR = 1.26; 95% CI: 1.21, 1.30 and HR = 1.10 95% CI: 1.09, 1.11, respectively). Furthermore, PM10 was associated with circulatory disease mortality (HR = 1.06; 95% CI: 1.04, 1.08), but NO2 was not (HR = 1.00; 95% CI: 0.99, 1.01). PM10 associations were robust to adjustment for NO2; NO2 associations remained for nonaccidental mortality and lung cancer mortality after adjustment for PM10., Conclusions: Long-term exposure to PM10 and NO2 was associated with nonaccidental and cause-specific mortality in the Dutch population of ≥ 30 years of age.

Background: Land use regression (LUR) models have been developed mostly to explain intraurban variations in air pollution based on often small local monitoring campaigns. Transferability of LUR models from city to city has been investigated, but little is known about the performance of models based on large numbers of monitoring sites covering a large area., Objectives: We aimed to develop European and regional LUR models and to examine their transferability to areas not used for model development., Methods: We evaluated LUR models for nitrogen dioxide (NO2) and particulate matter (PM; PM2.5, PM2.5 absorbance) by combining standardized measurement data from 17 (PM) and 23 (NO2) ESCAPE (European Study of Cohorts for Air Pollution Effects) study areas across 14 European countries for PM and NO2. Models were evaluated with cross-validation (CV) and hold-out validation (HV). We investigated the transferability of the models by successively excluding each study area from model building., Results: The European model explained 56% of the concentration variability across all sites for NO2, 86% for PM2.5, and 70% for PM2.5 absorbance. The HV R2s were only slightly lower than the model R2 (NO2, 54%; PM2.5, 80%; PM2.5 absorbance, 70%). The European NO2, PM2.5, and PM2.5 absorbance models explained a median of 59%, 48%, and 70% of within-area variability in individual areas. The transferred models predicted a modest-to-large fraction of variability in areas that were excluded from model building (median R2: NO2, 59%; PM2.5, 42%; PM2.5 absorbance, 67%)., Conclusions: Using a large data set from 23 European study areas, we were able to develop LUR models for NO2 and PM metrics that predicted measurements made at independent sites and areas reasonably well. This finding is useful for assessing exposure in health studies conducted in areas where no measurements were conducted.

Epidemiological studies often use nitrogen dioxide (NO2) or proximity to roads to characterize exposure to more health-relevant pollutants (e.g., fine particles or black carbon aerosol) in vehicle exhaust. Due to the introduction of diesel-soot filters, particle-to-NO2 ratios may have decreased, but little information is available about these ratios over time. Our study aim was to evaluate the change in particle-to-NO2 ratios between 1999 and 2009. We compared data collected during measurement campaigns in 1999 and 2009 from Munich, the Netherlands, and Stockholm. Traffic-impacted and urban and regional background sites were studied during each campaign. The same pollutants were measured in each campaign (mass concentration of particles ≤ 2.5 microns in diameter (PM2.5), PM2.5 absorbance (a marker for black carbon), and NO2) using the same methods except for NO2: Palmes tubes were used in 1999 and Ogawa badges in 2009. Both NO2 methods were validated against side-by-side chemiluminescence measurements. The levels of PM2.5 absorbance and PM2.5 were significantly lower (p<0.05) in 2009 at traffic and urban background sites in Stockholm, but only slightly lower in Munich and the Netherlands. In contrast, NO2 levels were not significantly different in the three areas in 2009 compared to 1999. Statistically-significant decreases between 1999 and 2009 pollutant ratios in the Netherlands (PM2.5 absorbance/NO2 and PM2.5/NO2) and in Stockholm (PM2.5 absorbance/NO2 and PM2.5 absorbance/PM2.5) were observed. Smaller decreases in these ratios were observed in Munich. The contrast between traffic and urban background locations was larger in 2009 than 1999 for NO2, whereas it was smaller for PM2.5 absorbance and PM2.5. The lower PM2.5 absorbance/NO2 and PM2.5/NO2 ratios in 2009 is likely explained by the introduction of oxidative particle filters for diesel vehicles that reduce particles but produce NO2. The changed contrasts may affect estimates of health effects related to NO2 as a marker of proximity to roads., (Copyright © 2014 Elsevier B.V. All rights reserved.)

Studies have linked air pollution exposure to cardiovascular health effects, but it is not clear which components drive these effects. We examined the associations between air pollution exposure and circulating white blood cell (WBC) counts in humans. To investigate independent contributions of particulate matter (PM) characteristics, we exposed 31 healthy volunteers at five locations with high contrast and reduced correlations amongst pollutant components: two traffic sites, an underground train station, a farm and an urban background site. Each volunteer visited at least three sites and was exposed for 5 h with intermittent exercise. Exposure measurements on-site included PM mass and number concentration, oxidative potential (OP), elemental- and organic carbon, metals, O3 and NO2. Total and differential WBC counts were performed on blood collected before and 2 and 18 h post-exposure (PE). Changes in total WBC counts (2 and 18 h PE), number of neutrophils (2 h PE) and monocytes (18 h PE) were positively associated with PM characteristics that were high at the underground site. These time-dependent changes reflect an inflammatory response, but the characteristic driving this effect could not be isolated. Negative associations were observed for NO2 with lymphocytes and eosinophils. These associations were robust and did not change after adjustment for a large suite of PM characteristics, suggesting an independent effect of NO2. We conclude that short-term air pollution exposure at real-world locations can induce changes in WBC counts in healthy subjects. Future studies should indicate if air pollution exposure-induced changes in blood cell counts results in adverse cardiovascular effects in susceptible individuals.

Background: Socioeconomic variables are associated with mortality and morbidity in a variety of diseases at both the individual and neighborhood level. Investigating whether low socioeconomic status populations are exposed to higher air pollution has been an important objective for the scientific community during the last decade. The goal of this study was to analyze the associations between outdoor nitrogen dioxide (NO2) concentrations in an area of Asturias (Spain) and two socioeconomic indexes-one based on occupation and the other on educational level at the census-tract level., Methods: A map of NO2 concentration was obtained from a land-use regression model. To obtain a census-tract average value, NO2 was estimated at the centroids of all 50 × 50 grids within a census tract. Standard socioeconomic variables were used from the Census of Population and Housing 2001. We analyzed the association between NO2 concentration and socioeconomic indicators for the entire area and stratified for more urban and more rural areas., Results: A positive linear relationship was found between the levels of education and NO2 exposure in the urban area and the overall study area, but no association was found in the rural area. A positive association between socioeconomic index based upon occupation and NO2 concentration was found in urban areas; however, this association was reversed in the rural and overall study areas., Conclusions: The strength and direction of the association between socioeconomic status and NO2 concentration depended on the socioeconomic indicator used and the characteristics of the study area (urban, rural). More research is needed with different scenarios to clarify the uncertain relationship among socioeconomic indexes, particularly in non-urban areas, where little has been documented on this topic.

Land use regression (LUR) models typically investigate within-urban variability in air pollution. Recent improvements in data quality and availability, including satellite-derived pollutant measurements, support fine-scale LUR modeling for larger areas. Here, we describe NO2 and PM10 LUR models for Western Europe (years: 2005-2007) based on >1500 EuroAirnet monitoring sites covering background, industrial, and traffic environments. Predictor variables include land use characteristics, population density, and length of major and minor roads in zones from 0.1 km to 10 km, altitude, and distance to sea. We explore models with and without satellite-based NO2 and PM2.5 as predictor variables, and we compare two available land cover data sets (global; European). Model performance (adjusted R(2)) is 0.48-0.58 for NO2 and 0.22-0.50 for PM10. Inclusion of satellite data improved model performance (adjusted R(2)) by, on average, 0.05 for NO2 and 0.11 for PM10. Models were applied on a 100 m grid across Western Europe; to support future research, these data sets are publicly available.

Land use regression (LUR) models have become popular to explain the spatial variation of air pollution concentrations. Independent evaluation is important. We developed LUR models for nitrogen dioxide (NO(2)) using measurements conducted at 144 sampling sites in The Netherlands. Sites were randomly divided into training data sets with a size of 24, 36, 48, 72, 96, 108, and 120 sites. LUR models were evaluated using (1) internal "leave-one-out-cross-validation (LOOCV)" within the training data sets and (2) external "hold-out" validation (HV) against independent test data sets. In addition, we calculated Mean Square Error based validation R(2)s. The mean adjusted model and LOOCV R(2) slightly decreased from 0.87 to 0.82 and 0.83 to 0.79, respectively, with an increasing number of training sites. In contrast, the mean HV R(2) was lowest (0.60) with the smallest training sets and increased to 0.74 with the largest training sets. Predicted concentrations were more accurate in sites with out of range values for prediction variables after changing these values to the minimum or maximum of the range observed in the corresponding training data set. LUR models for NO(2) perform less well, when evaluated against independent measurements, when they are based on relatively small training sets. In our specific application, models based on as few as 24 training sites, however, achieved acceptable hold out validation R(2)s of, on average, 0.60.

Objectives: Land use regression (LUR) modelling is a popular method to estimate outdoor air pollution concentrations at the home and/or work addresses of individual subjects in epidemiological studies. Typically, such models are constructed using measurements from dedicated monitoring campaigns lasting up to 1 year. It is unknown to what extent such models can adequately predict concentrations in earlier or later time periods. We tested the stability of measured and modelled spatial contrasts in outdoor nitrogen dioxide (NO(2)) pollution across the Netherlands over 8 years., Methods: NO(2) measurements were conducted at 40 locations in the Netherlands in 1999-2000. In 2007, NO(2) was again measured at 144 locations, of which 35 were the same as in 1999-2000. This enabled us to compare measurements as well as model predictions between the two time periods., Results: NO(2) measurements conducted in 2007 agreed well with NO(2) measurements taken in 1999-2000 at the same locations (R(2)=0.86). LUR models from 1999-2000 and 2007 explained 85% and 86% of observed spatial variance, respectively. The 2007 LUR model explained 77% of spatial variability in the 1999-2000 measurements and the 1999-2000 model explained 81% of variability in the 2007 measurements., Conclusion: We found good agreement between measured spatial contrasts in outdoor NO(2) in 1999-2000 and 2007. LUR models predicted spatial contrast 8 years in the past (2007 model) and 8 years in the future (1999-2000 model) well. This supports the use of LUR models in epidemiological studies with health data available for a later or earlier timepoint.

Background: Air pollution may promote type 2 diabetes by increasing adipose inflammation and insulin resistance. This study examined the relation between long-term exposure to traffic-related air pollution and type 2 diabetes prevalence among 50- to 75-year-old subjects living in Westfriesland, the Netherlands., Methods: Participants were recruited in a cross-sectional diabetes screening-study conducted between 1998 and 2000. Exposure to traffic-related air pollution was characterized at the participants' home-address. Indicators of exposure were land use regression modeled nitrogen dioxide (NO2) concentration, distance to the nearest main road, traffic flow at the nearest main road and traffic in a 250 m circular buffer. Crude and age-, gender- and neighborhood income adjusted associations were examined by logistic regression., Results: 8,018 participants were included, of whom 619 (8%) subjects had type 2 diabetes. Smoothed plots of exposure versus type 2 diabetes supported some association with traffic in a 250 m buffer (the highest three quartiles compared to the lowest also showed increased prevalence, though non-significant and not increasing with increasing quartile), but not with the other exposure metrics. Modeled NO2-concentration, distance to the nearest main road and traffic flow at the nearest main road were not associated with diabetes. Exposure-response relations seemed somewhat more pronounced for women than for men (non-significant)., Conclusions: We did not find consistent associations between type 2 diabetes prevalence and exposure to traffic-related air pollution, though there were some indications for a relation with traffic in a 250 m buffer.

Background: In epidemiological studies, small-scale spatial variation in air quality is estimated using land-use regression (LUR) and dispersion models. An important issue of exposure modeling is the predictive performance of the model at unmeasured locations., Objective: In this study, we aimed to evaluate the performance of two LUR models (large area and city specific) and a dispersion model in estimating small-scale variations in nitrogen dioxide (NO₂) concentrations., Methods: Two LUR models were developed based on independent NO₂ monitoring campaigns performed in Amsterdam and in a larger area including Amsterdam, the Netherlands, in 2006 and 2007, respectively. The measurement data of the other campaign were used to evaluate each model. Predictions from both LUR models and the calculation of air pollution from road traffic (CAR) dispersion model were compared against NO₂ measurements obtained from Amsterdam., Results and Conclusion: The large-area and the city-specific LUR models provided good predictions of NO₂ concentrations [percentage of explained variation (R²) = 87% and 72%, respectively]. The models explained less variability of the concentrations in the other sampling campaign, probably related to differences in site selection, and illustrated the need to select sampling sites representative of the locations to which the model will be applied. More complete traffic information contributed more to a better model fit than did detailed land-use data. Dispersion-model estimates for NO₂ concentrations were within the range of both LUR estimates.

Land use regression (LUR) has been successfully used to assess the intraurban variability of air pollution. In the INMA (Environment and Childhood) Study, ambient nitrogen oxides (NO(x) and NO(2)) and aromatic hydrocarbons (BTEX) were measured at 57 sampling sites in Sabadell (northeast Spain). Multiple regression models were developed to predict residential outdoor concentrations in a cohortof pregnantwomen (n = 657), using geographic data as predictor variables. The models accounted for 68 and 69% of the variance in NO(x) and NO(2) levels, respectively, with four predictor variables (altitude, land coverage, and two road length indicators). These percentages of explained variability could be further improved by replacing the two road length indicators with an ordinal indicator (road type). To our knowledge, this is the first study using LUR to assess the intraurban variability of BTEX in Europe, with a model including altitude and source-proximity variables that explained 74% of the variance in BTEX levels. These models will be used to study the association between prenatal exposure to air pollution and adverse pregnancy outcomes and early childhhod effects in the cohort.

Objectives: NO2)is a major urban air pollutant. Previously reported associations between ambient NO2)and children's respiratory health have been inconsistent, and independent effects of correlated pollutants hard to assess. The authors examined effects of NO2 on a spectrum of 11 respiratory symptoms, controlling for PM10 and SO2, using a large pooled dataset., Methods: Cross sectional studies were conducted in Russia, Austria, Italy, Switzerland, and the Netherlands, during 1993-99, contributing in total 23 955 children. Study-specific odds ratios for associations with ambient NO2 are estimated using logistic regressions with area-level random effects. Heterogeneity between study-specific results, and mean estimates (allowing for heterogeneity) are calculated., Results: Long term average NO2 concentrations were unrelated to prevalences of bronchitis or asthma. Associations were found for sensitivity to inhaled allergens and allergy to pets, with mean odds ratios around 1.14 per 10 microg/m3 NO2. SO2 had little confounding effect, but an initial association between NO2 and morning cough was reduced after controlling for PM10. Associations with reported allergy were not reduced by adjustment for the other pollutants. Odds ratios for allergic symptoms tended to be higher for the 9-12 year old children compared with the 6-8 year old children., Conclusions: Evidence for associations between NO2 and respiratory symptoms was robust only for inhalation allergies. NO2 most likely is acting as an indicator of traffic related air pollutants, though its direct effect cannot be ruled out. This remains important, as policies to reduce traffic related air pollution will not result in rapid reductions.

Stochastic modeling was used to predict nitrogen dioxide and fine particles [particles collected with an upper 50% cut point of 2.5 microm aerodynamic diameter (PM2.5)] levels at 1,669 addresses of the participants of two ongoing birth cohort studies conducted in Munich, Germany. Alternatively, the Gaussian multisource dispersion model IMMIS(net/em) was used to estimate the annual mean values for NO2 and total suspended particles (TSP) for the 40 measurement sites and for all study subjects. The aim of this study was to compare the measured NO2 and PM2.5 levels with the levels predicted by the two modeling approaches (for the 40 measurement sites) and to compare the results of the stochastic and dispersion modeling for all study infants (1,669 sites). NO2 and PM2.5 concentrations obtained by the stochastic models were in the same range as the measured concentrations, whereas the NO2 and TSP levels estimated by dispersion modeling were higher than the measured values. However, the correlation between stochastic- and dispersion-modeled concentrations was strong for both pollutants: At the 40 measurement sites, for NO2, r = 0.83, and for PM, r = 0.79; at the 1,669 cohort sites, for NO2, r = 0.83 and for PM, r = 0.79. Both models yield similar results regarding exposure estimate of the study cohort to traffic-related air pollution, when classified into tertiles; that is, 70% of the study subjects were classified into the same category. In conclusion, despite different assumptions and procedures used for the stochastic and dispersion modeling, both models yield similar results regarding exposure estimation of the study cohort to traffic-related air pollutants.

In order to estimate the spatial variation within well-defined study areas, nitrogen dioxide was measured with diffusion samplers (Palmes tube) in 40-42 sites each in Germany (Munich), the Netherlands and Sweden (Stockholm County). Each site was measured over four 2-week periods during 1 year (spring 1999 to summer 2000). In each country, one reference site was measured during all periods and the results were used to adjust for seasonal variability, to improve the estimates of the annual average. Comparisons between the chemiluminescence method (European reference method) and Palmes tube measurement indicated a good agreement in Germany (with a ratio of 1.0 for Palmes tube/chemiluminescence) but underestimation for Palmes tube measurement in the Netherlands and Sweden (0.8 for both countries). The r2 values were between 0.86 and 0.90 for all three countries. The annual average values for NO2 for different sampling sites were between 15.9 and 50.6 (mean 28.8 microg/m3) in Germany, between 12.1 and 50.8 (mean 28.9 microg/m3) in the Netherlands and between 6.1 and 44.7 (mean 18.5 microg/m3) in Sweden. Comparing spatial variation between similar sites in the three countries, we did not find any significant differences between annual average levels for urban traffic sites. In Sweden, annual average levels in urban background and suburban backgrounds sites were about 8 microg/m3 lower than comparable sites in Germany and the Netherlands. Comparing site types within each country only urban traffic sites and suburban background sites differed in Germany. In the Netherlands and Sweden, the urban traffic sites differed from all other sites and in Sweden also the urban background sites differed from the other background sites. The observed contribution from local traffic was similar in the Netherlands and Sweden (10 and 8 microg/m3, corresponding to 26-27% of the NO2 concentration found in the urban traffic sites). In Germany, the contribution from local traffic was only 3 microg/m3, corresponding to 9% of the NO2 concentration found in the urban traffic sites. The spatial variation was substantially larger for NO2 than the variation for PM2.5 and similar to PM2.5 absorbance, measured in the same locations.

Recent prospective cohort studies have suggested that long-term exposure to low levels of particulate matter (PM) air pollution is associated with increased mortality due to, especially, cardio-pulmonary disease. Exposure to ambient air pollution was estimated mostly as city average concentrations, assuming homogenous exposure within the city. We used an ongoing cohort study - The Netherlands Cohort Study (NLCS) on diet and cancer - to investigate the relationship between traffic-related air pollution and mortality. The baseline data collection took place in 1986. A study was conducted to develop methods for exposure assessment and evaluate the contrast in exposure to air pollution within the cohort. Assessment of long-term exposure to two traffic-related air pollutants, Black Smoke (BS) and Nitrogen Dioxide (NO(2)), consisted of separate estimation of regional background, urban background, and local traffic contributions at the home address. Interpolation of concentration data from a routine monitoring network was used to estimate the regional background concentration. A regression model relating degree of urbanization to air pollution was used to allow for differences between different towns/neighborhoods of cities. Distance to major roads was calculated to characterize local traffic contributions, using a Geographic Information System (GIS). Interpolation resulted in reasonably precise regional background estimation when distant sites were not used and distance squared was used as the weight. Cross-validation showed that prediction errors were about 15% of the range in regional background concentration. Urban and local scales contributed significantly to the contrast within the cohort. Prediction errors for estimating the urban background were about 25% of the range in background concentrations. When the developed model was applied to the study cohort, there was substantial contrast in estimated exposure to BS and NO(2). About 90% of the study population lived 10 years or more at its 1986 home address - supporting the use of the estimated concentration at the 1986 address as a relevant exposure variable.

A small, personal monitoring study was performed in a subpopulation (14 families) of a case-control study on the relationship between indoor nitrogen-dioxide exposure and respiratory diseases of schoolchildren. Mothers, schoolchildren and pre-schoolchildren were asked to carry duplicate Palmes diffusion tubes during one week. Simultaneously nitrogen-dioxide concentrations were measured in the kitchen, living room, bedroom, outdoors and--for a few participants--at school and at work. Information on time activity patterns was gathered by means of a self administered diary. Several models for estimating exposure were constructed and tested against measured exposure. The personal exposure of the participants could well be explained by models containing indoor concentrations. Models with time-weighted average concentrations did not explain personal exposure better than models containing indoor concentrations. A calculated time-weighted average exposure was found to underestimate measured personal exposure by an average 20%, probably because the average concentration in a location does not necessarily reflect the actual exposure in that location. Personal exposure of mothers and children was very similar and highly correlated, indicating that the personal exposure of the mother might be a reasonable estimate for the exposure of the child.

The influence of indoor nitrogen-dioxide exposure on respiratory symptoms of schoolchildren was investigated in a case-control study. The election method used was useful in obtaining symptomatic children, but insufficient in defining cases and controls without additional information. No relationship between indoor NO2 and respiratory symptoms was found. Bias may have been present, especially because of the high mobility of the study population. Attempts to estimate historical exposure were inaccurate. Therefore the results do not exclude that an association between indoor NO2 and respiratory symptoms exists.

Low-cost urban air quality sensor networks are increasingly used to study the spatio-temporal variability in air pollutant concentrations. Recently installed low-cost urban sensors, however, are more prone to result in erroneous data than conventional monitors, e.g., leading to outliers. Commonly applied outlier detection methods are unsuitable for air pollutant measurements that have large spatial and temporal variations as occur in urban areas. We present a novel outlier detection method based upon a spatio-temporal classification, focusing on hourly NO2 concentrations. We divide a full year’s observations into 16 spatio-temporal classes, reflecting urban background vs. urban traffic stations, weekdays vs. weekends, and four periods per day. For each spatio-temporal class, we detect outliers using the mean and standard deviation of the normal distribution underlying the truncated normal distribution of the NO2 observations. Applying this method to a low-cost air quality sensor network in the city of Eindhoven, the Netherlands, we found 0.1-0.5% of outliers. Outliers could reflect measurement errors or unusual high air pollution events. Additional evaluation using expert knowledge is needed to decide on treatment of the identified outliers. We conclude that our method is able to detect outliers while maintaining the spatio-temporal variability of air pollutant concentrations in urban areas. [ABSTRACT FROM AUTHOR]

Numerous studies from Europe and elsewhere suggest that current levels of air pollution trigger a broad range of acute health effects among susceptible people of all ages. There is less information on the health effects related to long-term exposure to air pollution. Several prospective cohort studies have documented that long-term exposure to particulate matter was associated with increased mortality rates. Most of these studies were conducted in the United States. European cohort studies have found associations between nitrogen dioxide (marker for traffic emissions) and increased mortality. Effect estimates in the cohort studies were substantially larger than those observed in the time series studies, which assess the short-term effects of air pollution. Evidence of impacts of long-term exposure to air pollution on lung function (development), respiratory disease and possibly development of atherosclerosis and birth outcomes is increasing. Uncertainties remain about the components responsible for the observed associations, sensitive subgroups and transferability of effect estimates across populations. In the US studies effects of PM2.5 on mortality were higher for subjects with lower education. The US cohort studies have played an important role in the current proposal for new air quality guidelines in Europe. [ABSTRACT FROM AUTHOR]

• A systematic review was conducted on associations between long-term exposure to traffic-related air pollution and non-accidental mortality. • A new exposure framework was developed to determine whether a study was sufficiently specific to TRAP, which includes studies beyond the near-roadway environment. • The overall confidence in the evidence for a positive association between long-term exposure to TRAP and non-accidental mortality was high. The health effects of traffic-related air pollution (TRAP) continue to be of important public health interest across the globe. Following its 2010 review, the Health Effects Institute appointed a new expert Panel to systematically evaluate the epidemiological evidence regarding the associations between long-term exposure to TRAP and selected health outcomes. This paper describes the main findings of the systematic review on non-accidental mortality. The Panel used a systematic approach to conduct the review. An extensive search was conducted of literature published between 1980 and 2019. A new exposure framework was developed to determine whether a study was sufficiently specific to TRAP, which included studies beyond the near-roadway environment. We performed random-effects meta-analysis when at least three estimates were available of an association between a specific exposure and outcome. We evaluated confidence in the evidence using a modified Office of Health Assessment and Translation (OHAT) approach, supplemented with a broader narrative synthesis. Thirty-six cohort studies were included. Virtually all studies adjusted for a large number of individual and area-level covariates—including smoking, body mass index, and individual and area-level socioeconomic status—and were judged at a low or moderate risk for bias. Most studies were conducted in North America and Europe, and a few were based in Asia and Australia. The meta-analytic summary estimates for nitrogen dioxide, elemental carbon and fine particulate matter—pollutants with more than 10 studies—were 1.04 (95% CI 1.01, 1.06), 1.02 (1.00, 1.04) and 1.03 (1.01, 1.05) per 10, 1 and 5 µg/m3, respectively. Effect estimates are interpreted as the relative risk of mortality when the exposure differs with the selected increment. The confidence in the evidence for these pollutants was judged as high, because of upgrades for monotonic exposure–response and consistency across populations. The consistent findings across geographical regions, exposure assessment methods and confounder adjustment resulted in a high confidence rating using a narrative approach as well. The overall confidence in the evidence for a positive association between long-term exposure to TRAP and non-accidental mortality was high. [ABSTRACT FROM AUTHOR]