Characteristics of Non-accidental Deaths and PM2.5 Concentration
During the study period, a total of 336,379 non-accidental deaths occurred in Shanghai, with a mean of 307 daily deaths. Some 53.1% of these deaths were men and the average ages at death were 74.9 and 79.3 years for men and women, respectively (Table 1).
Table 1. Characteristics of non-accidental deaths in Shanghai, China, 2012-2014.
Variables
|
Men
|
Women
|
Total numbers of non-accidental deaths during the study period (%)
|
178,786 (53.1%)
|
157,593 (46.9%)
|
Average population (million) during the study period (%)
|
12.53 (52.1%)
|
11.52 (47.9%)
|
Average daily death during the study period (SD)
|
163 (27)
|
144 (27)
|
Average age (years) at death (SD)
|
74.9 (13.0)
|
79.3 (11.7)
|
Population-standardized smoking rate of the deceased people (SD)
|
29.71% (15.25%)
|
0.92% (0.71%)
|
The detailed results of PM2.5 concentrations and weather conditions during the study period were published previously [13, 19]. Overall, the average daily concentration of PM2.5 in Shanghai was 55.0 µg/m3, with a similar seasonal trend as the daily non-accidental deaths (i.e. high values in cold seasons and low values in warm seasons) [19].
Sex Difference in Effect of PM2.5 on Non-accidental Mortality
When considering the nonlinear associations of PM2.5, temperature, and smoking with non-accidental death, and adjusting weather types and day of week, older people and men showed a higher risk for non-accidental death (RR = 1.025 per year increase in age, 95% confidence interval (CI): 1.021 – 1.029, and RR = 0.892 for women, 95% CI: 0.802 – 0.993, respectively) (Table 2).
Table 2. Risk ratios of sex and its interaction with PM2.5 for non-accidental mortality.
|
Risk ratio (RR)
|
95% confidence interval (CI)
|
P-value
|
Lower limit
|
Upper limit
|
Age
|
1.025
|
1.021
|
1.029
|
<0.001
|
Female
|
0.892
|
0.802
|
0.993
|
0.036
|
Female × PM2.5 (per 10 µg/m3)
|
0.998
|
0.996
|
1.000
|
0.010
|
SWT
|
|
|
|
|
Warm humid
|
0.996
|
0.980
|
1.012
|
0.596
|
Cold dry
|
0.979
|
0.956
|
1.003
|
0.090
|
Cool dry
|
1.006
|
0.988
|
1.025
|
0.486
|
Cool humid
|
1.044
|
1.025
|
1.064
|
<0.001
|
Cold humid
|
1.010
|
0.986
|
1.035
|
0.409
|
Day of Week
|
|
|
|
|
Monday
|
1.017
|
1.004
|
1.030
|
0.009
|
Tuesday
|
1.011
|
0.998
|
1.024
|
0.110
|
Wednesday
|
1.013
|
1.000
|
1.026
|
0.050
|
Thursday
|
1.006
|
0.993
|
1.019
|
0.358
|
Friday
|
1.004
|
0.991
|
1.017
|
0.598
|
Saturday
|
0.997
|
0.984
|
1.010
|
0.624
|
SWT, synoptic weather type.
The predicted numbers of daily non-accidental deaths by day and PM2.5 concentration, and by temperature and PM2.5 concentration are shown in the 3D perspective plots in Figure 1. Daily non-accidental deaths fluctuated with season and increased with PM.2.5 concentration (Figure 1a), and a U-shape exposure-response relationship was observed between temperature and daily non-accidental deaths (Figure 1b). The predicted values by the CGAM indicate that 66.2% of the deviance can be explained by the model.
A low magnitude but statistically significant interaction was found between sex and PM2.5 concentration, i.e. compared with men, the risk ratio for non-accidental death in relation to increasing PM2.5 concentration was smaller in women (Figure 2). Per 10 µg/m3 increase in PM2.5 concentration, the risk ratio was 0.2% smaller (RR = 0.998, 95% CI: 0.996 – 1.000) for women, compared to men (Table 2). When including the nonlinear interaction term between smoking rate and PM2.5 concentration in the model, although the conditional mortality rate for women was higher than that of men, the increased risk for non-accidental death per 10 µg/m3 increase in PM2.5 concentration for women compared to men was even lower (RR = 0.983, 95% CI: 0.979 – 0.988, P < 0.001).
Lag Struscture of Interaction between Sex and PM2.5 Concentration
In our sensitivity analysis investigating the interaction effect of sex after applying up to 14 lag days of PM2.5 concentration, we got consistent results. During the two lag weeks’ exposure to PM2.5, in general, the increased risk for non-accidental death is about 0.2% – 0.4% lower in women than in men, per 10 µg/m3 increase in PM2.5 concentration (Table 3). The results for the weighted moving average of PM2.5 concentrations up to 14 days indicate that the risk ratio of non-accidental deaths for cumulative effects of PM2.5 was 0.3% – 0.6% lower in women compared to men, and the results are all statistically significant and consistent throughout the 14-day time window (Table 3).
Table 3. Risk ratio of the interaction between sex and PM2.5 (per 10 µg/m3) for non-accidental death, after applying different lag times for exposure.
Lag
|
Single-day
|
Lag
|
Weighted moving average
|
RR
|
95% CI
|
P-value
|
RR
|
95% CI
|
P-value
|
Lag 1
|
0.998
|
0.997 – 1.000
|
0.014
|
Lag 0-1
|
0.997
|
0.995 – 0.999
|
0.001
|
Lag 2
|
0.999
|
0.997 – 1.000
|
0.135
|
Lag 0-2
|
0.997
|
0.995 -0.999
|
0.001
|
Lag 3
|
0.997
|
0.995 – 0.998
|
0.000
|
Lag 0-3
|
0.997
|
0.995 – 0.999
|
0.004
|
Lag 4
|
0.996
|
0.995 – 0.998
|
0.000
|
Lag 0-4
|
0.996
|
0.994 – 0.999
|
<0.001
|
Lag 5
|
0.997
|
0.996 – 0.999
|
0.000
|
Lag 0-5
|
0.996
|
0.994 – 0.998
|
0.001
|
Lag 6
|
0.997
|
0.995 – 0.999
|
0.000
|
Lag 0-6
|
0.996
|
0.993 – 0.998
|
0.001
|
Lag 7
|
0.997
|
0.995 – 0.998
|
0.000
|
Lag 0-7
|
0.995
|
0.993 – 0.998
|
0.001
|
Lag 8
|
0.997
|
0.996 – 0.999
|
0.000
|
Lag 0-8
|
0.995
|
0.993 -0.998
|
0.001
|
Lag 9
|
0.997
|
0.996 – 0.999
|
0.000
|
Lag 0-9
|
0.996
|
0.993 – 0.998
|
0.001
|
Lag 10
|
0.997
|
0.995 – 0.998
|
0.000
|
Lag 0-10
|
0.995
|
0.993 – 0.998
|
<0.001
|
Lag 11
|
0.999
|
0.997 – 1.000
|
0.082
|
Lag 0-11
|
0.995
|
0.992 – 0.997
|
<0.001
|
Lag 12
|
0.999
|
0.997 – 1.000
|
0.084
|
Lag 0-12
|
0.995
|
0.992 – 0.997
|
<0.001
|
Lag 13
|
0.998
|
0.997 – 1.000
|
0.008
|
Lag 0-13
|
0.995
|
0.992 – 0.997
|
<0.001
|
Lag 14
|
0.998
|
0.996 – 1.000
|
0.014
|
Lag 0-114
|
0.994
|
0.992 – 0.997
|
<0.001
|
RR: risk ratio; CI, confidence interval.
Subgroup analysis by age groups and specific causes of death
The reduced incremental effects of PM2.5 on non-accidental deaths in women were consistent in the subgroup analyses. Although not statistically significant in the aged 40 years or younger group (might be due to the small proportion, i.e. 1.3%, of the total deaths), the risk ratio was 0.4% – 1.0% lower for women, compared to men (Table 4) Regarding the case-specific deaths, in general, the risk ratio was 0.2% – 0.8% lower for women, compared to men. However, the reduction was not statistically significant for cerebrovascular deaths (Table 4).
Table 4. Risk ratio of the interaction between sex and PM2.5 (per 10 µg/m3) for non-accidental death, by age groups and specific causes of death.
Subgroup
|
Risk ratio
(RR)
|
95% confidence interval (CI)
|
P-value
|
Lower limit
|
Upper limit
|
Age (years)
|
|
|
|
|
<40
|
0.990
|
0.978
|
1.003
|
0.118
|
≥40, <60
|
0.991
|
0.985
|
0.996
|
0.001
|
≥60, <80
|
0.996
|
0.994
|
0.999
|
0.008
|
≥80
|
0.996
|
0.994
|
0.998
|
<0.001
|
Causes of death
|
|
|
|
|
Respiratory disease
|
0.992
|
0.987
|
0.997
|
0.001
|
Cardiovascular disease
|
0.993
|
0.990
|
0.997
|
<0.001
|
Cerebrovascular disease
|
0.998
|
0.994
|
1.001
|
0.164
|
Other circulatory system diseases
|
0.997
|
0.994
|
0.999
|
0.003
|