Abstract:This report describes the results of a time-series analysis of the effect of short-term exposure to particulate matter with an aerodynamic diameter < or = 10 pm (PM10) on mortality in metropolitan Chennai, India (formerly Madras). This was one of three sites in India chosen by HEI as part of its Public Health and Air Pollution in Asia (PAPA) initiative. The study involved integration and analysis of retrospective data for the years 2002 through 2004. The data were obtained from relevant government agencies in charge of routine data collection. Data on meteorologic confounders (including temperature, relative humidity, and dew point) were available on all days of the study period. Data on mortality were also available on all days, but information on cause-of-death (including accidental deaths) could not be reliably ascertained. Hence, only all-cause daily mortality was used as the major outcome for the time-series analyses. Data on PM10, nitrogen dioxide (NO2), and sulfur dioxide (SO2) were limited to a much smaller number of days, but spanned the full study period. Data limitations resulting from low sensitivity of gaseous pollutant measurements led to using only PM10 in the main analysis. Of the eight operational ambient air quality monitor (AQM) stations in the city, seven met the selection criteria set forth in the common protocol developed for the three PAPA studies in India. In addition, all raw data used in the analysis were subjected to additional quality assurance (QA) and quality control (QC) criteria to ensure the validity of the measurements. Two salient features of the PM10 data set in Chennai were a high percentage of missing readings and a low correlation among daily data recorded by the AQMs. The latter resulted partly because each AQM had a small footprint (approximate area over which the air pollutant measurements recorded in the AQM are considered valid), and partly because of differences in source profiles among the 10 zones within the city. The zones were defined by the Chennai Corporation based on population density. Alternative exposure series were developed to control for these data features. We first developed exposure series based on data from single AQMs and multiple AQMs. Because neither was found to satisfactorily represent population exposures, we subsequently developed an exposure series that disaggregated pollutant data to individual zones within the city boundary. The zonal series, despite some uncertainties, was found to best represent population exposures among other available choices. The core model was thus a zonal model developed using disaggregated mortality and pollutant data from individual zones. We used quasi-Poisson generalized additive models (GAMs) with smooth functions of time, temperature, and relative humidity modeled using penalized splines. The degrees of freedom (df) for these confounders were selected to maximize the precision with which the relative risk for PM10 was estimated. This is a deviation from the traditional approaches to degrees of freedom selection, which usually aim to optimize overall model fit. Our approach led to the use of 8 df/year for time, 6 df/year for temperature, and 5 df/year for relative humidity. The core model estimated a 0.44% (95% confidence interval [CI] = 0.17 to 0.71) increase in daily all-cause mortality per 10-pg/m3 increase in daily average PM10 concentrations. Extensive sensitivity analyses compared models constructed using alternative exposure series and contributions of model parameters to the core model with regard to confounder degrees of freedom, alternative lags for exposure and meteorologic confounders, inclusion of outliers, seasonality, inclusion of multiple pollutants, and stratification by sex and age. The sensitivity analyses showed that our estimates were robust to a range of specifications and were also comparable to estimates reported in previous time-series studies: PAPA, the National Morbidity, Mortality, and Air Pollution Study (NMMAPS), Air Pollution and Health: A European Approach (APHEA), and Air Pollution and Health: A European and North American Approach (APHENA). While the approaches developed in previous studies served as the basis for our model development, the present study has new refinements that have allowed us to address specific data limitations (such as missing measurements and small footprints of air pollution monitors). The methods developed in the study may allow better use of routine data for time-series analysis in a broad range of settings where similar exposure and data-related issues prevail. We hope that the estimates derived in this study, although somewhat tentative, will facilitate local environmental management initiatives and spur future studies.

Mortality Associated with Ambient PM2.5 Exposure in India: Results from the Million Death Study

Daily nonaccidental mortality associated with short-term PM2.5 exposures in Delhi, India

Estimating the effect of annual PM 2·5 exposure on mortality in India: a difference-in-differences approach

Mortality Burden Attributable to PM1 in Zhejiang Province, China.

Air Pollution and Mortality in India: Investigating the Nexus of Ambient and Household Pollution Across Life Stages

Increasing Mortality Caused by Chronic Obstructive Pulmonary Disease (COPD) in Relation with Exposure to Ambient Fine Particulate Matters: an Analysis in Southeastern China

Long-Term Exposure to PM2.5 and Mortality: A Cohort Study in China

Association Between Short-Term Exposure to Ambient Air Pollution and Daily Mortality: a Time-Series Study in Eastern China.

Ambient fine particulate matter and daily mortality: a comparative analysis of observed and estimated exposure in 347 cities

Association between ambient fine particulate matters and chronic obstructive pulmonary disease (COPD) mortality: an analysis in Southeastern China

Estimates of Global Mortality Burden Associated with Short-Term Exposure to Fine Particulate Matter (PM2·5)

Coarse Particulate Air Pollution and Daily Mortality: A Global Study in 205 Cities.

Fine Particulate Air Pollution and Daily Mortality. A Nationwide Analysis in 272 Chinese Cities.

Ambient air pollution and daily mortality in ten cities of India: a causal modelling study

Long-term exposure to particulate matter and all-cause and cause-specific mortality in an analysis of multiple Asian cohorts

Child Survival and Early Lifetime Exposures to Ambient Fine Particulate Matter in India: A Retrospective Cohort Study

Ambient Particulate Air Pollution and Daily Mortality in 652 Cities

Part 1. Short-term effects of air pollution on mortality: results from a time-series analysis in Chennai, India

Health Risk and Disease Burden Attributable to Long-Term Global Fine-Mode Particles.

Mortality Risk and Burden From a Spectrum of Causes in Relation to Size-Fractionated Particulate Matters: Time Series Analysis

Low concentrations of fine particle air pollution and mortality in the Canadian Community Health Survey cohort