Differentiating the effects of fine and coarse particles on daily mortality in Shanghai, China

Abstract

The findings on health effects of ambient fine particles (PM2.5) and coarse particles (PM10-2.5) remain inconsistent. In China, PM2.5 and PM10-2.5 are not the criteria air pollutants, and their monitoring data are scarce. There have been no epidemiological studies of health effects of PM2.5 and PM10-2.5 simultaneously in China. We conducted a time series study to examine the acute effects of PM2.5 and PM10-2.5 on daily mortality in Shanghai, China from Mar. 4, 2004 to Dec. 31, 2005. We used the generalized additive model (GAM) with penalized splines to analyze the mortality, air pollution and covariate data. The average concentrations of PM2.5 and PM10-2.5 were 56.4 microg/m3 and 52.3 microg/m3 in our study period, and PM2.5 constituted around 53.0% of the PM10 mass. Compared with the Global Air Quality Guidelines set by World Health Organization (10 microg/m3 for annual mean) and U.S. National Ambient Air Quality Standards (15 microg/m3 for annual mean), the PM2.5 level in Shanghai was much higher. We found that PM2.5 was associated with the death rates from all causes and from cardiorespiratory diseases in Shanghai. We did not find a significant effect of PM10-2.5 on mortality outcomes. A10 microg/m3 increase in the 2-day moving average (lag01) concentration of PM2.5 corresponded to 0.36% (95% CI 0.11%, 0.61%), 0.41% (95% CI 0.01%, 0.82%) and 0.95% (95% CI 0.16%, 1.73%) increase of total, cardiovascular and respiratory mortality. For PM10-2.5, the effects were attenuated and less precise. Our analyses provide the first statistically significant evidence in China that PM2.5 has an adverse effect on population health and strengthen the rationale for further limiting levels of PM2.5 in outdoor air in Shanghai.

Figure 1

Percent increase in numbers of deaths due to all, cardiovascular, and respiratory causes per 10 ug/m³ increase in particle concentration for different lag days. (a: total mortality; b: cardiovascular mortality; c: respiratory mortality)* * Current-day temperature and relative humidity (lag=0) were used; 3 df were applied to temperature and relative humidity respectively.

Figure 1

Percent increase in numbers of deaths due to all, cardiovascular, and respiratory causes per 10 ug/m³ increase in particle concentration for different lag days. (a: total mortality; b: cardiovascular mortality; c: respiratory mortality)* * Current-day temperature and relative humidity (lag=0) were used; 3 df were applied to temperature and relative humidity respectively.

Figure 1

Percent increase in numbers of deaths due to all, cardiovascular, and respiratory causes per 10 ug/m³ increase in particle concentration for different lag days. (a: total mortality; b: cardiovascular mortality; c: respiratory mortality)* * Current-day temperature and relative humidity (lag=0) were used; 3 df were applied to temperature and relative humidity respectively.

Figure 2

Smoothing plots of fine particles against mortality risk (df=3). X-axis is the 2-day average (lag 01) PM_2.5 concentrations (µg/m³). The estimated mean percentage of change in daily mortality is shown by the solid line, and the dotted lines represent twice the point-wise standard error* (a: total mortality; b: cardiovascular mortality; c: respiratory mortality) * Current-day temperature and relative humidity (lag=0) were used; 3 df were applied to temperature and relative humidity respectively.

Figure 2

Smoothing plots of fine particles against mortality risk (df=3). X-axis is the 2-day average (lag 01) PM_2.5 concentrations (µg/m³). The estimated mean percentage of change in daily mortality is shown by the solid line, and the dotted lines represent twice the point-wise standard error* (a: total mortality; b: cardiovascular mortality; c: respiratory mortality) * Current-day temperature and relative humidity (lag=0) were used; 3 df were applied to temperature and relative humidity respectively.

Figure 2

Smoothing plots of fine particles against mortality risk (df=3). X-axis is the 2-day average (lag 01) PM_2.5 concentrations (µg/m³). The estimated mean percentage of change in daily mortality is shown by the solid line, and the dotted lines represent twice the point-wise standard error* (a: total mortality; b: cardiovascular mortality; c: respiratory mortality) * Current-day temperature and relative humidity (lag=0) were used; 3 df were applied to temperature and relative humidity respectively.

Figure 3

Effects of varying the df for time trend and weather conditions on the percent increase of total mortality associated with a 10µg/m³ increase of 2-day average (lag01) PM_2.5 and PM_10-2.5 (a: time trend; b: mean temperature; c: relative humidity)* * The df value used in the core models; current-day temperature and relative humidity (lag=0) were used.

Figure 3

Effects of varying the df for time trend and weather conditions on the percent increase of total mortality associated with a 10µg/m³ increase of 2-day average (lag01) PM_2.5 and PM_10-2.5 (a: time trend; b: mean temperature; c: relative humidity)* * The df value used in the core models; current-day temperature and relative humidity (lag=0) were used.

Figure 3

Effects of varying the df for time trend and weather conditions on the percent increase of total mortality associated with a 10µg/m³ increase of 2-day average (lag01) PM_2.5 and PM_10-2.5 (a: time trend; b: mean temperature; c: relative humidity)* * The df value used in the core models; current-day temperature and relative humidity (lag=0) were used.

Figure 4

Percent increase in total mortality per 10 ug/m³ increase in 2-day average (lag01) particle concentration before and after excluding days with extremely high and low concentrations* * Model 1: whole period analysis; model 2: analysis excluding days with extremely high and low concentrations; current-day temperature and relative humidity (lag=0) were used; 3 df were applied to temperature and relative humidity respectively.