Yinkang Duan

GIS Approaches for the Estimation of Residential-Level Ambient PM Concentrations

Spatial estimations are increasingly used to estimate geocoded ambient particulate matter (PM) concentrations in epidemiologic studies because measures of daily PM concentrations are unavailable in most U.S. locations. This study was conducted to a) assess the feasibility of large-scale kriging estimations of daily residential-level ambient PM concentrations, b) perform and compare cross-validations of different kriging models, c) contrast three popular kriging approaches, and d ) calculate SE of the kriging estimations. We used PM data for PM with aerodynamic diameter ≤10 μm (PM10) and aerodynamic diameter ≤ 2.5 μm (PM2.5) from the U.S. Environmental Protection Agency for the year 2000. Kriging estimations were performed at 94,135 geocoded addresses of Women’s Health Initiative study participants using the ArcView geographic information system. We developed a semiautomated program to enable large-scale daily kriging estimation and assessed validity of semivariogram models using prediction error (PE), standardized prediction error (SPE), root mean square standardized (RMSS), and SE of the estimated PM. National- and regional-scale kriging performed satisfactorily, with the former slightly better. The average PE, SPE, and RMSS of daily PM10 semivariograms using regular ordinary kriging with a spherical model were 0.0629, −0.0011, and 1.255 μg/m3, respectively; the average SE of the estimated residential-level PM10 was 27.36 μg/m3. The values for PM2.5 were 0.049, 0.0085, 1.389, and 4.13 μg/m3, respectively. Lognormal ordinary kriging yielded a smaller average SE and effectively eliminated out-of-range predicted values compared to regular ordinary kriging. Semiautomated daily kriging estimations and semivariogram cross-validations are feasible on a national scale. Lognormal ordinary kriging with a spherical model is valid for estimating daily ambient PM at geocoded residential addresses.

Association of criteria pollutants with plasma hemostatic/inflammatory markers: a population-based study

To elucidate the health effects of air pollution, the short-term association of criteria pollutants (particles <10 μm in diameter [PM10], O3, CO, NO2, and SO2) with hemostatic and inflammatory markers were examined using a population-based sample of 10,208 middle-age males and females of the biracial cohort of Atherosclerosis Risk in Communities (ARIC) study. For each participant, we calculated the following pollutant exposures 1–3 days prior to the randomly allocated cohort examination date: PM10, CO, NO2, and SO2 as 24-h averages, and O3 as an 8-h average of the hourly measures. The hemostatic/inflammatory factors included fibrinogen, factor VIII-C, von Willebrand factor (vWF), albumin, and white blood cell count (WBC). Linear regression models were used to adjust for cardiovascular disease (CVD) risk factors, demographic and socioeconomic variables, and relevant meteorological variables. One standard deviation (SD) increment of PM10 (12.8 μg/m3) was significantly (P<0.05) associated with 3.93% higher of vWF among diabetics and 0.006 g/dl lower of serum albumin among persons with a history of CVD. One SD increment of CO (0.60 p.p.m.) was significantly (P<0.01) associated with 0.018 g/dl lower of serum albumin. Significant curvilinear associations, indicative of threshold effects, for PM10 with factor VIII-C, O3 with fibrinogen and vWF, and SO2 with factor VIII-C, WBC, and serum albumin were found. This population-based study suggest that the hemostasis/inmflammation markers analyzed, which are linked to higher risk of CHD, are associated adversely with environmentally relevant ambient pollutants, with the strongest associations in the upper range of the pollutant distributions, and in persons with a positive history of diabetes and CHD.

Metabolic syndrome, its preeminent clusters, incident coronary heart disease and all-cause mortality--results of prospective analysis for the Atherosclerosis Risk in Communities study.

Objective.  To investigate the prospective association between Metabolic Syndrome (MetS) and coronary heart disease (CHD) and all‐cause mortality.

Metabolic Syndrome Clusters and the Risk of Incident Stroke. The Atherosclerosis Risk in Communities (ARIC) Study

Background and Purpose— Little is known about the metabolic syndrome (MetS) and the risk of incident stroke. This study is designed to identify particular clusters of MetS components that carry the highest risk of incident stroke. Methods— We analyzed the public use data from the population-based Atherosclerosis Risk in Communities study. At baseline, 14 993 stroke-free middle-aged individuals were followed-up over 9 years for incident stroke. MetS components were defined according to the National Heart, Lung, and Blood Institute/American Heart Association criteria. Incident stroke was identified using a standardized incident events identification and classification protocol. Proportional hazard models were used to assess the RRs and 95% CIs of ischemic stroke associated with MetS and its different clusters. Results— At baseline, the prevalence of MetS was 39%. The mean age was 54, with 26% blacks and 55% females. The hazard ratio of incident ischemic stroke associated with MetS among women (hazard ratio, 2.41; 95% CI, 1.69 to 3.49) and men (hazard ratio, 2.11; 95% CI, 1.56–2.85) was similar. There was a dose–response relationship between the numbers of MetS components and the risk of incidence stroke. Persons with either elevated blood pressure or elevated fasting glucose in the clusters to form a MetS had the highest risk for incident stroke (hazard ratio, 2.74–4.16 comparing to the reference group) than MetS without these 2 components (hazard ratio, ≤2.00 comparing to the reference group). Conclusions— The data support the need to target MetS, especially MetS, with these 2 highest risk components (elevated blood pressure or elevated fasting glucose) in the clusters.

Ambient particulate air pollution and ectopy - The environmental epidemiology of arrhythmogenesis in women's health initiative study, 1999-2004

The relationships between ambient PM2.5 and PM10 and arrhythmia and the effect modification by cigarette smoking were investigated. Data from U.S. Environmental Protection Agency (EPA) air quality monitors and an established national-scale, log-normal kriging method were used to spatially estimate daily mean concentrations of PM at addresses of 57,422 individuals from 59 examination sites in 24 U.S. states in 1999–2004. The acute and subacute exposures were estimated as mean, geocoded address-specific PM concentrations on the day of, 0–2 d before, and averaged over 30 d before the electrocardiogram (ECG) (Lag0; Lag1; Lag2; Lag1–30). At the time of standard 12-lead resting ECG, the mean age (SD) of participants was 67.5 (6.9) yr (84% non-Hispanic White; 6% current smoker; 15% with coronary heart disease; 5% with ectopy). After the identification of significant effect modifiers, two-stage random-effects models were used to calculate center-pooled odds ratios and 95% confidence intervals (OR, 95% CI) of arrhythmia per 10 μg/m3 increase in PM concentrations. Among current smokers, Lag0 and Lag1 PM concentrations were significantly associated ventricular ectopy (VE)—the OR (95% CI) for VE among current smokers was 2 (1.32–3.3) and 1.32 (1.07–1.65) at Lag1 PM2.5 and PM10, respectively. The interactions between current smoking and acute exposures (Lag0; Lag1; Lag2) were significant in relationship to VE. Acute exposures were not significantly associated with supraventricular ectopy (SVE), or with VE among nonsmokers. Subacute (Lag1–30) exposures were not significantly associated with arrhythmia. Acute PM2.5 and PM10 exposure is directly associated with the odds of VE among smokers, suggesting that they are more vulnerable to the arrhythmogenic effects of PM.

Associations Between Air Pollution and Peak Expiratory Flow Among Patients with Persistent Asthma

Responses of patients with persistent asthma to ambient air pollution may be different from those of general populations. For example, asthma medications may modify the effects of ambient air pollutants on peak expiratory flow (PEF). Few studies examined the association between air pollution and PEF in patients with persistent asthma on well-defined medication regimens using asthma clinical trial data. Airway obstruction effects of ambient air pollutants, using 14,919 person-days of daily self-measured peak expiratory flow (PEF), were assessed from 154 patients with persistent asthma during the 16 wk of active treatment in the Salmeterol Off Corticosteroids Study trial. The three therapies were an inhaled corticosteroid, an inhaled long-acting β-agonist, and placebo. The participants were nonsmokers aged 12 through 63 yr, recruited from 6 university-based ambulatory care centers from February 1997 to January 1999. Air pollution data were derived from the U.S. Environmental Protection Agency Aerometric Information Retrieval System. An increase of 10 ppb of ambient daily mean concentrations of NO2 was associated with a decrease in PEF of 1.53 L/min (95% confidence interval [CI] –2.93 to –0.14) in models adjusted for age, gender, race/ethnicity, asthma clinical center, season, week, daily average temperature, and daily average relative humidity. The strongest association between NO2 and PEF was observed among the patients treated with salmeterol. Negative associations were also found between PEF and SO2 and between PEF and PM10, respectively. The results show that the two medication regimens protected against the effects of PM10. However, salmeterol increased the sensitivity to NO2 and triamcinalone enhanced the sensitivity to SO2.

Is Age-Related Macular Degeneration Associated with Stroke Among Elderly Americans?§

Objective: To investigate whether age-related macular degeneration (AMD) is associated with the development of ischemic and hemorrhagic stroke among elderly Americans. Design: Population-based cohort study. Participants: The five percent random sample of 2000-2003 Medicare enrollees was obtained. The cohort (n=1,519,086) consisted of enrollees who were aged 65 or older at the first two-year (January 1, 2000 to December 31, 2001). Methods: Baseline demographic variables and chronic conditions (AMD and type, history of myocardial infarction (MI), stroke, hypertension, and diabetes) were defined based on the occurrence of relevant ICD-9 codes in relevant diagnosis fields of the baseline Medicare Data. We excluded 215,900 persons who had a diagnosis of MI or stroke during baseline period to form a cohort of 1,303,186 individuals who were free of major cardio-cerebral vascular disease (CVD) at baseline. Main Outcome Measures: In two years of follow-up (January 1, 2002 to December 31, 2003), a total of 89,501 incident stroke cases were identified, including 80,018 ischemic, 7048 hemorrhagic, and 2,435 stroke cases of both types. Results: Baseline mean age was 75 years (Standard Divination=7.7), with 60% women and 88% whites. The prevalence of AMD was 10.6%, with 19.7% being neovascular AMD and 80.3% being non-neovascular AMD. Baseline age, gender, race, hypertension, and diabetes adjusted 2-year incident odds ratios and 95% confidence internal of stroke associated with AMD were 1.31 (1.26, 1.36) for neovascular AMD, 1.18 (1.15, 1.21) for non-neovascular AMD, and 1.21 (1.18, 1.23) for either neovascular or non-neovascular AMD. Conclusion: The findings are suggestive of an association between AMD, especially neovascular AMD, and incident stroke, independent of demographic factors and co-morbidity. These findings, if confirmed by other studies that control for smoking and other lifestyle covariables not measured in this study, suggest the possibility of shared common antecedents between stroke and AMD.

Interaction of Ambient Air Pollution With Asthma Medication on Exhaled Nitric Oxide Among Asthmatics

ABSTRACT The interaction between ambient air pollution and asthma medication remains unclear. The authors compared airway inflammation response to air pollution among asthmatics. Increases of 10 ppb of nitrogen dioxide (NO2) and of 10 μg/m3 of particulate matter < 10 micron in diameter (PM10) daily concentrations were associated with an increase in exhaled nitric oxide (eNO) of 0.13 ppb (95% confidence interval = 0.06, 0.19) and of 0.07 ppb (95% confidence interval = 0.02, 0.12), respectively, in models adjusted for important covariates. The results show that the medication could not counteract airway inflammation effects of air pollution. Specifically, the patients on triamcinolone decreased the sensitivity to PM10 but increased the sensitivity to NO2. The patients on salmeterol were more vulnerable to both NO2 and PM10. This study indicates that the current pollution levels may still enhance airway inflammation among patients with persistent asthma even when they are on asthma medications.

National Kriging Exposure Estimation: Liao et Al. Respond

Szpiro et al. suggest that our findings Liao et al. (2006) do not adequately support using national-scale, log-normal ordinary kriging to estimate daily mean concentrations of PM10 (particulate matter with aerodynamic diameter ≤ 10 μm) at unmonitored locations in the contiguous United States. They posit that the absence of the cross-validation SE prevents evaluating the validity of kriging estimation, as we implemented in this context, and the comparability of both regional-versus national-scale kriging and manually modified versus semiautomated, default-calculated semivariograms. Little literature is available on the use of kriging methods to estimate daily air pollution data for large population–based multi-center epidemiologic studies. The four studies cited by Szpiro et al. (Dockery et al. 1993; Jerrett et al. 2005; Miller et al. 2007; Pope et al. 1995) all used cohort analyses for which only long-term average exposures are required, and only one of those (Jerrett et al. 2005) actually involved interpolation methods at all, although the study was restricted to a single city. In contrast, our objective was to create an interpolated daily pollutant concentration database for a multisite population-based epidemiologic study. The cross-validation mean square error (MSE) mentioned by Szpiro et al. is also termed the “root-mean-square (RMS) prediction error,” which is the empirical SE based on the mean square of the predictions, as opposed to SE, the mathematical formula for the RMS prediction error. RMS and SE, both are available from ArcView (ESRI Inc., Redlands, CA), are often considered jointly as an alternative measure (to RMSS) of the validity of spatial analysis. The average RMS and SE from 366 daily PM10 spherical model cross-validations based on year 2000 PM10 data were 19.48 and 16.19 μg/m3, respectively, from the log-normal regular kriging, and 26.43 and 25.60 μg/m3, respectively from a ordinary kriging. The validity of the model is supported by RMSS alone (≈1), by the similarity of RMS and SE, and by SPE (≈ 0). Additionally, the average daily SD of PM10 measured at the monitor locations was 27.20 μg/m3. Comparing SD with the kriging-RMS provides a measure of the reduction in error due to interpolation. If RMS is less than the SD, then the kriging approach has some benefit, compared with using long-run averages. From both ordinary and log-normal kriging, especially for the latter, we see a notable reduction in RMS compared with SD. Meanwhile, substantial variability remains, suggesting that kriging error should be taken into account when using the kriged values. Szpiro et al. also implicitly criticize our use of daily kriging when the objective was to interpolate daily data. Spatial–temporal models have potentially greater power than a 1-day-at-a-time spatial analysis but are not easy to apply in practice, with large datasets and many missing values. Regional kriging could be superior to national kriging if the spatial dependence parameters (range, sill, and nugget) vary substantially from region to region, in which case a national kriging model could result in misspecified covariances. However, regional kriging also uses fewer data points to estimate those parameters and could result in greater errors. We would welcome theoretical or empirical studies that could cast further light on this trade-off. However, as far as our article (Liao et al. 2006) is concerned, our main purpose was to note that the national kriging method appears to be competitive when assessed by overall RMS error. We compared the results of regional- and national-scale kriging on a small set (17%) of days when the largest number of monitors (≥ 400) were reporting data—a scenario heavily favoring regional spatial interpolation strategies. On the remaining days when only 120–400 monitors were reporting data, regional kriging was inherently problematic given the restricted availability of monitors within regions. Szpiro et al. suggest that the problems of interpolation near the boundary could be solved by “overlapping,” but this is only one of the issues encountered using regional-kriging: One would still need to decide how to consistently define the regions, considering the number of available data points that change substantially from day to day, to achieve a meaningful reduction in RMS error. Based on the 12 “optimal” days in 2000, the average RMS and SE were 12.68 and 12.82 μg/m3, respectively, from the national scale kriging, compared with 12.22 and 12.49 μg/m3, respectively, from regional-scale kriging (Liao et al. 2006). These results, together with RMSS and SPE we reported, support our conclusion that national kriging performs comparably to regional kriging even when restricted to optimal days. Szpiro et al. correctly note that it is possible to improve the RMSS values by manual adjustment. However, typically we found that when one of the validity measures (RMSS, PE, or SPE) was improved by manual adjustment, other measures became worse. It is difficult to manually adjust models to improve all cross-validation parameters simultaneously. Manually adjusting daily semivariogrms is not feasible when kriging over 10 years. Moreover, the predicted SE at unmeasured locations was uniformly lower in automatically fit models. Szpiro et al. are correct that cross-validation may not be representative of the performance at participant address locations, although it is unclear what alternative methods they would like us to use. The ability to do semi-automatic cross-validations was a major attraction of ArcView and, despite limitations, is the best tool we know for validating spatial predictions. The semiautomated kriging approach presents considerable advantages in estimating daily residential-level pollutant concentrations in large cohorts over long periods. Our proposed method (Liao et al. 2006) used log-normal kriging based on a spherical model to interpolate daily data on a national scale, and the weighted least squares method of parameter estimation without manual adjustment. We believe that the cross-validation statistics, presented in our article and amplified here, provide adequate support for these recommendations against reasonable alternatives that we considered.

GIS APPROACHES FOR ESTIMATION OF RESIDENTIAL-LEVEL AMBIENT PM CONCENTRATIONS

Spatial estimations are increasingly used to estimate geocoded ambient particulate matter (PM) concentrations in epidemiologic studies because measures of daily PM concentrations are unavailable in most U.S. locations. This study was conducted to a) assess the feasibility of large-scale kriging estimations of daily residential-level ambient PM concentrations, b) perform and compare cross-validations of different kriging models, c) contrast three popular kriging approaches, and d) calculate SE of the kriging estimations. We used PM data for PM with aerodynamic diameter </= 10 microm (PM10) and aerodynamic diameter </= 2.5 microm (PM2.5) from the U.S. Environmental Protection Agency for the year 2000. Kriging estimations were performed at 94,135 geocoded addresses of Womens Health Initiative study participants using the ArcView geographic information system. We developed a semiautomated program to enable large-scale daily kriging estimation and assessed validity of semivariogram models using prediction error (PE) , standardized prediction error (SPE) , root mean square standardized (RMSS) , and SE of the estimated PM. National- and regional-scale kriging performed satisfactorily, with the former slightly better. The average PE, SPE, and RMSS of daily PM10 semivariograms using regular ordinary kriging with a spherical model were 0.0629, -0.0011, and 1.255 microg/m3, respectively ; the average SE of the estimated residential-level PM10 was 27.36 microg/m3. The values for PM2.5 were 0.049, 0.0085, 1.389, and 4.13 microg/m3, respectively. Lognormal ordinary kriging yielded a smaller average SE and effectively eliminated out-of-range predicted values compared to regular ordinary kriging. Semiautomated daily kriging estimations and semivariogram cross-validations are feasible on a national scale. Lognormal ordinary kriging with a spherical model is valid for estimating daily ambient PM at geocoded residential addresses.