Potential gains in life expectancy by attaining daily ambient fine particulate matter pollution standards in mainland China: A modeling study based on nationwide data.

BACKGROUND: Ambient fine particulate matter pollution (PM2.5) is one leading cause of disease burden, but no study has quantified the association between daily PM2.5 exposure and life expectancy. We aimed to assess the potential benefits in life expectancy by attaining the daily PM2.5 standards in 72 cities of China during 2013-2016. METHODS AND
FINDINGS: We applied a two-stage approach for the analysis. At the first stage, we used a generalized additive model (GAM) with a Gaussian link to examine the city-specific short-term association between daily PM2.5 and years of life lost (YLL); at the second stage, a random-effects meta-analysis was used to generate the regional and national estimations. We further estimated the potential gains in life expectancy (PGLE) by assuming that ambient PM2.5 has met the Chinese National Ambient Air Quality Standard (NAAQS, 75 μg/m3) or the ambient air quality guideline (AQG) of the World Health Organization (WHO) (25 μg/m3). We also calculated the attributable fraction (AF), which denoted the proportion of YLL attributable to a higher-than-standards daily mean PM2.5 concentration. During the period from January 18, 2013 to December 31, 2016, we recorded 1,226,849 nonaccidental deaths in the study area. We observed significant associations between daily PM2.5 and YLL: each 10 μg/m3 increase in three-day-averaged (lag02) PM2.5 concentrations corresponded to an increment of 0.43 years of life lost (95% CI: 0.29-0.57). We estimated that 168,065.18 (95% CI: 114,144.91-221,985.45) and 68,684.95 (95% CI: 46,648.79-90,721.11) years of life lost can be avoided by achieving WHO's AQG and Chinese NAAQS in the study area, which corresponded to 0.14 (95% CI: 0.09-0.18) and 0.06 (95% CI: 0.04-0.07) years of gain in life expectancy for each death in these cities. We observed differential regional estimates across the 7 regions, with the highest gains in the Northwest region (0.28 years of gain [95% CI: 0.06-0.49]) and the lowest in the North region (0.08 [95% CI: 0.02-0.15]). Furthermore, using WHO's AQG and Chinese NAAQS as the references, we estimated that 1.00% (95% CI: 0.68%-1.32%) and 0.41% (95% CI: 0.28%-0.54%) of YLL could be attributable to the PM2.5 exposure at the national level. Findings from this study were mainly limited by the unavailability of data on individual PM2.5 exposure.
CONCLUSIONS: This study indicates that significantly longer life expectancy could be achieved by a reduction in the ambient PM2.5 concentrations. It also highlights the need to formulate a stricter ambient PM2.5 standard at both national and regional levels of China to protect the population's health.

Introduction

The health effects of fine particulate matter (particulate matter with an aerodynamic diameter less than or equal to 2.5 μm, PM2.5) have attracted increasing public concern over the past decade in China [1,2]. The population-weighted annual PM2.5 concentration in mainland China reached 54.3 μg/m3 in 2013 [3], which was much higher than that in 1990 (39 μg/m3) and far above the ambient air quality guidelines (AQGs, 25 μg/m3) recommended by the World Health Organization (WHO) [4,5]. Meanwhile, mounting evidence has linked the ambient PM2.5 exposure with excess premature deaths and years of life lost (YLL) [6,7]. Such findings have provided valuable information to estimate the disease burden of ambient PM2.5 [8,9].

Previous studies have examined the association between short-term and long-term exposure to ambient air pollution and mortality or YLL [10,11]. The short-term studies, usually based on the daily time-series data, evaluated the acute health effects of air pollution [12], while the long-term studies estimated health effects of chronic and cumulative air pollution exposures (usually with the average concentration of several years as the exposure indicator) [13]. The long-term health effect studies generally reported relatively larger effects than those in short-term analyses [14]. The exact biological mechanisms for the health effects of PM2.5 exposure remain unclear; previous studies suggested that oxidative stress, systemic inflammation, direct vascular endothelial impairment, and alterations in arterial tone might play important roles [15-17].

Considering the widely reported effects of air pollution exposure on premature mortality and increased years of life lost [7,18], it was reasonable to hypothesize that high levels of air pollution exposure could lead to losses in life expectancy; however, only a few studies have investigated this association, and most of those studies focused on the long-term air pollution exposure [19,20]. For example, one study from the United States and two studies from China reported long-term exposure to higher levels of particulate pollution was associated with reduced life expectancy [21-23]. However, to the best of our knowledge, the evidence is lacking on the effects of short-term (e.g., daily) PM2.5 exposure on life expectancy.

Furthermore, there is a need to estimate the potential benefits of reduction in daily ambient PM2.5 concentration by attaining the air quality standards. As such, we used potential gains in life expectancy (PGLE) to investigate the benefit on life expectancy by assuming the PM2.5 concentration was in compliance with certain ambient air quality standards. Compared with other indicators such as excess mortality and YLL, PGLE is a more informative indicator for epidemiological research [24]. Through directly quantifying the health benefits by attaining the air quality standards, PGLE is more relevant to air pollution controlling and formulation of air quality standards. Another advantage of PGLE is that it can be easily compared across different areas, while excess deaths and YLL are somewhat influenced by the age structure and size of the study population [25]. Although this limitation can be solved by several standardization techniques, the YLL was subject to one important issue of its sensitivity to competing risks of death [25,26].

In this study, we firstly examined the associations between daily PM2.5 and YLL after adjusting for potential confounders at both national and regional levels of mainland China from 2013 to 2016, based on which we estimated the PGLE by postulating that ambient PM2.5 concentrations were successfully controlled under the Chinese National Ambient Air Quality Standards (NAAQS), as well as WHO’s AQG and its Interim Targets (ITs).

Methods

Mortality data and YLL calculation

This is a nationwide modeling study based on a time-series analysis. The daily time-series mortality data on nonaccidental causes in 72 Chinese cities (S1 Table) for the period of January 18, 2013 through December 31, 2016 were selected for this study, and a total of 1,226,849 nonaccidental deaths were recorded. The data were extracted from the Disease Surveillance Points (DSP) System of China, which is operated by the National Center for Chronic and Noncommunicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention [27]. The data from the DSP System have been widely used in the assessment of health risk factors or disease burden and policy formulation [28,29]. These cities were selected based on the following process: (1) they were randomly selected using a multistage stratification approach that took the sociodemographic characteristics of the Chinese population into consideration; (2) the daily mortality counts in these cities were temporally stable without large fluctuations, and no change in the administrative divisions occurred during the study period; and (3) their air pollution and meteorological records were accessible during the study period. The completeness and accuracy of the death data in the DSP System were strictly checked by different administrative levels of the Chinese Center for Disease Control and Prevention network. Practitioners in the health facilities were responsible for checking the accuracy, completeness, and data quality of the death data, and they then reported that information to the DSP System. Staff in the district-level CDC reviewed all new information to ensure the data quality (i.e., to check that the ICD codes were maintained and to exclude the duplicate records and redundant information) in the system within 7 days, as well as returning the unclear or uncertain records back to the reporting health facilities. Then, practitioners in those health facilities asked the physicians to correct and confirm the data. Staff in district-level CDC also collected nonaccidental death information from the security department and civil affairs bureau (the other government departments collecting the death information for the purpose of residence) every month. Then, the staff of the provincial- or regional-level CDC would conduct a second round of checking and reviewing. Finally, data were sent to the national-level CDC to undergo a further round of review, which included the duplication, logic, data analysis, and investigation of misreported data.

The 72 cities in our study were divided into the following 7 regions: Northwest, North, Northeast, Central, East, Southwest, and South (Fig 1), and cities in the same region usually incorporated similar features in terms of geographical, meteorological, and cultural conditions.

Fig 1

Geographical distribution of the 72 cities across 7 regions of mainland China.

The location of the 72 cities are indicated by red dots.

The average life expectancy in China was 76.25 years in 2016. We used the life expectancy in the corresponding years to calculate the YLL for each death by matching age and sex to the Chinese national life table [30], which was obtained from WHO’s website, and then summed the YLL for all deaths on each day of the study period to compute the daily YLL of each city.

This study was based on one project aiming to examine the short-term health effects of air pollution in China, which has been approved by the Ethical Review Committee of Institute for Environmental Health and Related Product Safety, Chinese Center for Disease Control and Prevention. No individual consent was required because all data were analyzed at an aggregated level. The present study is reported as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (S1 STROBE Checklist). The data analyses were performed following a prospective analysis plan (S1 Text), and the model structure of this study is provided as a diagram in S1 Fig.

Air pollution and meteorological factors

Daily concentrations for ambient PM2.5 and other air pollutants (including sulfur dioxide [SO2], nitrogen dioxide [NO2], and ozone [O3]) were obtained from China’s National Real-time Publishing Platform for Daily Air Quality (http://106.37.208.233:20035), which delivered the real-time concentrations of ambient air pollutants that were measured by state-controlled air-monitoring stations [31]. The 24-hour mean concentrations of ambient PM2.5, SO2, and NO2 and the maximum 8-hour mean levels for O3 were averaged from all available monitoring data within each city.

In addition, daily meteorological data on mean temperature (°C) and relative humidity (%) were obtained from the National Meteorological Data Service Center of China (http://data.cma.cn), which is publicly accessible.

Statistical analysis

Descriptive analysis

For descriptive analysis, the number of cities and mean air pollutant concentrations, meteorological conditions, and mortality and YLL in the 7 regions during the study period were summarized. In addition, the Spearman correlation was performed to quantify the correlation between air pollutants and weather variables. Analyses were based on complete mortality records during the study period.

Analysis for the PM2.5–YLL association

We examined the national and regional short-term association between daily PM2.5 and YLL using two-stage models. At the first stage, we applied a generalized additive model (GAM) with a Gaussian link to explore the city-specific short-term association between daily PM2.5 and YLL. In the GAM model, daily mean concentration of PM2.5 in each city was incorporated as the independent variable while daily YLL was used as the dependent variable, and all the quantitative variables were treated as continuous variables. We controlled for public holidays and day of the week in the form of categorical variables, while long-term and seasonal trends, temperature, and relative humidity were adjusted using penalized smoothing splines [32]. A complete list of model parameters was provided as a supplemental table (S2 Table). We selected the model specifications and the degree of freedom (df) for the smoothers according to previous experiences of similar studies [33]. For example, we applied a df of 6 per year for long-term trends to filter out the information at time scales of about 2 months, a df of 6 for moving average temperature of the current day and previous 3 days (lag03) for the potential nonlinear relationship, and 3 df for the same day’s relative humidity. We explored the associations with different lag structures from the current day (lag0) up to 3 days before (lag3), and we also evaluated the effects of moving averages for the current day and the previous 1, 2, and 3 days (lag01, lag02, lag03). The statistical model can be specified as

At the second stage, we used a random-effects meta-analysis to generate the regional and national estimates. This approach provided a useful tool to pool risk estimates while interpreting within-city statistical error and between-city heterogeneity of the genuine risks [34].

Sensitivity analyses

We conducted a series of sensitivity analyses to check the robustness of the findings. Two-pollutant models were used to examine the associations between daily PM2.5 and YLL after adjusting for other air pollutants. Specifically, PM2.5 was included alone in the single-pollutant models, while PM2.5 and SO2 (or NO2, O3) were included simultaneously in the two-pollutant models. In addition, we observed that the Northwest and Southwest regions covered a rather large area, which may have wide variation in basic characteristics, and relatively fewer cities were included in these two regions. Considering the uncertainty and the complex geospatial correlation between the cities, we performed a spatial statistical model by adjusting for the longitude and latitude of the cities in the model using a penalized smoothing splines function [35]. Furthermore, we also used a mixed-effect GAM as a one-stage approach to examine the regional and national estimates, in which we included the variable of city as a random term.

In addition, we performed a meta-regression to evaluate whether the observed PM2.5–YLL relationship could be explained by some city-level variables: Gross Domestic Product (GDP), population density, GDP per capita, elevation, precipitation, poverty, education, annual PM2.5 concentration, annual CO concentration, annual O3 concentration, annual SO2 concentration, annual NO2 concentration, air pressure, annual temperature, and annual relative humidity. The potential interaction between annual PM2.5 and GDP was checked by including an interactive term of PM2.5 and GDP in the meta-regression model.

Estimating the avoidable YLL, PGLE, and attributable fraction

Based on the established associations between ambient PM2.5 and YLL, we further estimated the avoidable YLL by assuming the ambient PM2.5 had been controlled at specified concentrations as in China’s NAAQS or WHO’s AQG and its ITs. We further estimated the PGLE, which was the average years longer each deceased person would have lived if ambient PM2.5 were kept under a certain standard in the study area. We also calculated the attributable fraction (AF) that denoted the proportion of YLL due to a higher-than-standards daily PM2.5 concentration. The two indicators can be calculated using the following formulas: where avoidable YLL is the sum of estimated YLL that can be prevented in the study area if ambient PM2.5 were kept under a certain concentration, overall mortality count is the total mortality number during the study period, AF is the AF, and overall YLL is the sum of the YLL for all deaths that occurred during the study period. The reference levels of PM2.5 included WHO's AQG (25 μg/m3) and its ITs, including IT-1 (75 μg/m3, which was the same as the China’s NAAQS), IT-2 (50 μg/m3), and IT-3 (37.5 μg/m3).

Our main analyses were performed using R (version 3.5.1; R foundation for Statistical Computing, Vienna, Austria) with the “mgcv” and “metafor” packages. All statistical tests were two-sided, and values of p < 0.05 were considered statistically significant.

Results

Descriptive results

During the study period, a total of 1,226,849 nonaccidental deaths were recorded in the 72 cities across the 7 regions of China; 44.0% of the study population were females. The average age of death of the subjects included in this study was 71.72 ± 16.74 years. Table 1 summarizes the number of cities, daily mean air pollutant concentrations, meteorological conditions, daily mean mortality, and YLL of these regions. The daily mean concentrations of PM2.5, SO2, NO2, and O3 ranged from 49.29 to 95.90, 27.15 to 53.94, 26.80 to 48.99, and 66.75 to 100.69, respectively. The mean temperature ranged from 7.35°C to 21.84°C, and relative humidity ranged from 49.31% to 77.79%. Moreover, the daily mean YLL were 87.16, 123.44, 146.09, 145.04, 158.06, 121.37, and 179.78 years in the 7 regions of Northwest, North, Northeast, Central, East, Southwest, and South, respectively.

Table 1

Summary characteristics of the study cities by regions.

Variable	Northwest	North	Northeast	Central	East	Southwest	South	National
Number of cities	8	6	7	8	24	11	8	72
Mean concentration of air pollutants (μg/m3)
PM2.5	49.29	95.90	64.47	71.27	71.05	52.82	51.94	67.65
SO2	34.20	53.94	45.82	31.50	41.60	28.76	27.15	38.97
NO2	26.80	48.99	39.96	34.99	37.24	29.73	32.02	36.70
O3	77.72	100.69	87.33	66.75	92.76	72.28	76.04	85.14
Weather
Mean temperature (°C)	10.39	13.33	7.35	16.85	15.93	15.40	21.84	14.88
Relative humidity (%)	49.31	56.31	62.91	72.95	71.19	65.11	77.79	66.30
Daily mean mortality	5.32	9.68	10.20	9.77	12.39	8.06	12.89	10.34
Daily mean YLL (years)	87.16	123.44	146.09	145.04	158.06	121.37	179.78	141.88

Abbreviations: NO2, nitrogen dioxide; O3, ozone; PM2.5, particulate matter with an aerodynamic diameter less than or equal to 2.5 μm or fine particulate matter; SO2, sulfur dioxide; YLL, years of life lost.

The correlation analyses showed low to moderate correlation coefficients between air pollutants and weather variables. For example, PM2.5 had moderate positive correlations with NO2 (correlation coefficient = 0.50), had relatively lower correlations with SO2 and O3 (correlation coefficients of 0.26 and 0.29, respectively), and had a negative correlation with mean temperature and relative humidity (correlation coefficients of −0.15 and −0.02, respectively) (S3 Table).

The association between daily PM2.5 and YLL

S2 Fig, S3 Fig, and S4 Fig show the diagnostic graphs of the model, including the plot of the residuals, the plot of partial autocorrelation function (PACF), and Q-Q plot for 6 provincial capital cities. These results showed that there were no discernible autocorrelation and patterns in the residuals, suggesting that the models had acceptable goodness of fit.

In the single-pollutant models, we observed statistically significant associations between PM2.5 and YLL at both national and regional levels, especially in the lag02 models (Fig 2). At the national level, we estimated that each 10 μg/m3 increase in the PM2.5 concentrations of lag02 was associated with an increment of 0.43 (95% CI: 0.29–0.57) YLL (S4 Table). The plot of residuals at the national level suggested that these residuals were generally independent, and there were no obviously discernible autocorrelation and patterns (S5 Fig). The region-specific results showed that the associations varied by regions. For example, the Northwest region was found to have the highest association (β = 0.94, 95% CI: 0.21–1.68), while the North region had the lowest association, with a regression coefficient of 0.12 (95% CI: 0.03–0.22).

Fig 2

The associations between ambient PM2.5 and YLL in 7 regions of mainland China, 2013–2016.

PM2.5, particulate matter with an aerodynamic diameter less than or equal to 2.5 μm or fine particulate matter; YLL, years of life lost.

In the two-pollutant models, the significant associations between PM2.5 and YLL generally remained (S4 Table). For instance, at the national level, each 10 μg/m3 increase in lag02 PM2.5 concentration was associated with an increment of 0.41 (95% CI: 0.27–0.55), 0.32 (95% CI: 0.19–0.45), or 0.41 (95% CI: 0.27–0.55) in YLL after controlling for SO2, NO2, and O3, respectively. The spatial statistical models for the Northwest and Southwest regions, which additionally adjusted for the longitude and latitude of each city, also produced significant effect estimates, with the coefficients of 0.45 (95% CI: 0.31–0.59) for the Northwest region and 0.63 (95% CI: 0.52–0.74) for the Southwest region. While the one-stage mix-effect model also yielded a significant effect (the overall regression coefficient was 0.18 [95% CI: 0.09–0.27]), the estimate was relatively smaller (S5 Table).

In addition, we evaluated whether the observed PM2.5–YLL relationship could be explained by some city-level factors (S6 Table). The analysis showed that the associations between PM2.5 and YLL were relatively higher in cities with lower annual mean concentrations of PM2.5. Each IQR (39.40 μg/m3) increase in annual concentrations of PM2.5 was associated with a 0.59 decrease in the regression coefficient. Furthermore, we did not find a significant interactive effect of PM2.5 and GDP on the associations between PM2.5 and YLL (p = 0.89).

Avoidable YLL, PGLE, and the AF

Based on the established relationship between daily PM2.5 and YLL, we estimated the avoidable YLL and AF in different regions of China (Table 2). Specifically, we estimated that 68,684.95 (95% CI: 46,648.79–90,721.11) YLL can potentially be avoided by attaining China’s NAAQS (75 μg/m3) in the study area, and this number could rise to 168,065.18 (95% CI: 114,144.91–221,985.45) by meeting WHO's AQG (25 μg/m3). In general, we observed higher effect estimates when adopting stricter air quality standards. Heterogeneity of the estimates was observed across the 7 regions. For example, when adopting WHO’s AQG, the East region was found to have the largest avoidable YLL (46,572.93 YLL), while the North region had the lowest avoidable YLL (12,661.48 YLL).

Table 2

The avoidable YLL and AF by improving ambient PM2.5 to China’s and WHO’s standards in 72 cities of mainland China, 2013–2016.

Region	Avoidable YLL (95% CI)		AF (%, 95% CI)
	China’s Standard (75 μg/m3)	WHO’s Guideline (25 μg/m3)	China’s Standard (75 μg/m3)	WHO’s Guideline (25 μg/m3)
Northwest	4,241.55 (929.63–7,553.46)	15,911.15 (3,487.29–28,335.01)	0.45 (0.10–0.80)	1.69 (0.37–3.02)
North	6,290.23 (1,343.86–11,236.61)	12,661.48 (2,705.03–22,617.94)	0.32 (0.07–0.57)	0.64 (0.14–1.15)
Northeast	7,112.83 (825.47–13,400.18)	16,660.23 (1,933.49–31,386.97)	0.39 (0.04–0.73)	0.91 (0.11–1.71)
Central	7,179.33 (547.35–13,811.32)	19,483.29 (1,485.40–37,481.17)	0.43 (0.03–0.82)	1.16 (0.09–2.23)
East	18,068.94 (6,464.92–29,672.96)	46,572.93 (16,663.41–76,482.45)	0.32 (0.11–0.52)	0.81 (0.29–1.33)
Southwest	12,604.24 (8,889.43–16,319.04)	33,087.32 (23,335.60–42,839.04)	0.61 (0.43–0.79)	1.61 (1.14–2.09)
South	6,261.74 (2,680.24–9,843.24)	18,909.99 (8,094.13–29,725.86)	0.24 (0.10–0.38)	0.72 (0.31–1.13)
National	68,684.95 (46,648.79–90,721.11)	168,065.18 (114,144.91–221,985.45)	0.41 (0.28–0.54)	1.00 (0.68–1.32)

Based on the effects of moving averaged concentration of lag 0 to lag 2 (lag02) of daily PM2.5.

Abbreviations: AF, attributable fraction; PM2.5, particulate matter with an aerodynamic diameter less than or equal to 2.5 μm or fine particulate matter; WHO, World Health Organization; YLL, years of life lost.

We further estimated that 0.41% (95% CI: 0.28%–0.54%) and 1.00% (95% CI: 0.68%–1.32%) of the YLL could be attributable to the daily exposure of PM2.5 by using China’s NAAQS and WHO’s AQG as the reference (Table 2). In addition, different effect estimates were observed among these regions, with the largest being observed in the Northwest region (1.69% [95% CI: 0.37%–3.02%]) and the minimum in the South region (0.24% [95% CI: 0.10%–0.38%]). Fig 3 shows the regional and national estimates of the PGLE using different air quality standards. Overall, we estimated that 0.14 (95% CI: 0.09–0.18) and 0.06 (95% CI: 0.04–0.07) years in life expectancy can be potentially gained according to WHO’s AQG (25 μg/m3) and China’s standard (75 μg/m3), respectively. Among the 7 regions, the largest value of 0.28 (95% CI: 0.06–0.49) was observed in the Northwest region, and the minimum value of 0.08% (95% CI: 0.02–0.15) was found in the North region by using WHO’s AQG as the reference.

Fig 3

The estimated PGLE by attaining WHO’s AQGs and their ITs in 7 regions of mainland China, 2013–2016.

Based on the effects of moving averaged concentration of lag 0 to lag 2 (lag02) of daily PM2.5. AQG, air quality guideline; IT, Interim Target; PGLE, potential gains in life expectancy; PM2.5, particulate matter with an aerodynamic diameter less than or equal to 2.5 μm or fine particulate matter; WHO, World Health Organization.

Discussion

To our knowledge, this might be the first study to quantify the short-term association between ambient PM2.5 and life expectancy in China. Using a large data set coving 72 Chinese cities, we estimated that about 0.14 years in life expectancy could be prolonged based on the hypothetical situation that the daily ambient PM2.5 concentration was in compliance with WHO’s ambient AQG.

Previous studies have well-documented the health effects of ambient air pollutants using a series of health outcomes such as premature mortality, excess morbidity, and YLL, which provided crucial information to measure the harmful effects of ambient air pollutants [36-38]. A few studies further examined the effects of long-term air pollution exposure on life expectancy [39-41]; however, little has been done to address the association of short-term PM2.5 exposure with life expectancy, and no studies, to our knowledge, have quantified the potential benefits in life expectancy due to short-term air quality improvement [42,43]. Such evidence will be helpful for policy-making, risk management, and resource allocation.

A few studies have reported the association between long-term exposure to ambient particulate matter pollution and life expectancy. For example, one study reported that a reduction of 10 μg/m3 in annual PM2.5 concentration could increase the life expectancy by about 0.61 years in the United States [21]. Another study similarly reported that an increase of 10 μg/m3 in long-term PM10 exposure was associated with a decrease of 0.64 years in life expectancy in China, and it may save 3.7 billion life-years in the whole country if the concentrations of PM10 reached the Class I standard of 40 μg/m3 [23]. In the present study, we estimated that 0.14 years in life expectancy can be potentially gained by reaching WHO’s AQG on daily PM2.5 concentrations in China. This finding was in line with previous observations that the short-term health effects of PM2.5 were relatively smaller than those from long-term exposure [44], and this may be due to the cumulative effects of prolonged exposures [45]. Nevertheless, findings from this study provided valuable evidence for the potential benefits in life expectancy of improved daily air quality, indicating that exposure to higher levels of air pollution even for a short time could reduce life expectancy.

The underlying biological mechanisms linking short-term PM2.5 exposure to life expectancy included a range of pathophysiological pathways. For example, one reason was that short-term PM2.5 exposure could lead to increased mortality and morbidity of cardiopulmonary diseases through formation of atherosclerotic plaque, systemic oxidative stress, and inflammation [46,47]. This explanation was supported by an intervention study that reduction of particle exposure by indoor air filtration could improve microvascular function in the elderly [48].

We observed a larger potential health benefit when using WHO’s AQG (25 μg/m3) as the reference than using China’s NAAQS (75 μg/m3), indicating that a stricter ambient quality standard would lead to more health benefits and therefore should be considered in future revision of China’s air quality standards.

We observed some evidence for spatial heterogeneity in the association between PM2.5 and YLL across different regions. This finding was in line with previous studies [31,49]. Generally, we found relatively weaker associations in the North, East, and South regions, whereas the associations were stronger in the Northwest and Southwest regions. The underlying reasons remained unclear. One possible underlying reason might be the differences in emission sources and chemical constituents of ambient PM2.5 among the different regions. The PM2.5 in the Northwest and Southwest regions may be more hazardous than that in other regions; most of the ambient fine particles were related to biomass combustion, which was more toxic than other sources [50]. Our meta-regression analysis showed that the areas with higher annual concentrations of PM2.5 tended to have a lower PM2.5–YLL association, indicating a better adaptation to the local environmental conditions in the areas with higher levels of air pollution. It was possible that people living in highly polluted areas have higher self-protection awareness, which could lead to taking better protective actions such as wearing masks, reducing outdoor activities, and use of air purifiers [31]. Moreover, considering that the cities with a higher PM2.5 concentration may also be wealthier and have better healthcare access, we cannot rule out the possibility that there may be some protective effect of economic development level. We therefore included the interactive term of PM2.5 and GDP in the meta-regression model and did not find a significant interactive effect. Additionally, in light of previous studies that reported varying effects of PM2.5 constituents on human health, we suspect that the differences in chemical components of PM2.5 in different areas may be a potential explanation [51,52].

The observed associations between PM2.5 and YLL were generally robust in the sensitivity analyses. In particular, the associations remained consistent in the two-pollutant models with adjustment for other air pollutants, indicating that the associations were not confounded by these air pollutants. However, we observed a relatively smaller estimate when adjusting for NO2, which could be partly explained by the moderate positive correlation between PM2.5 and NO2 (r = 0.50). It was also possible that PM2.5 and NO2 shared similar emission sources and biological pathways in their health effects [53,54]. Furthermore, spatial autocorrelation might be one issue in this analysis; however, this concern should be minimal because the cities were sparsely distributed in different areas, and our spatial model controlling for longitude and latitude of the cities yielded a consistent result.

This study applied a novel, to our knowledge, indicator, namely PGLE, to measure the potential health benefits by controlling air pollution to a certain level. This indicator estimated the average years a person would have lived longer through air quality improvement. This measurement took into consideration of the age of the deceased and the population size of the study area, making it comparable across different areas [55].

A few limitations should be noted for this study. This was an ecological time-series study, which used the city-averaged concentrations of ambient air pollutants as the exposure measurement. It might have led to ecological fallacy and thus limited our ability of causal inference. However, it is not feasible to measure every participants’ exposure directly for such a large-scale study, and this strategy has been widely used in previous time-series studies [56,57]. Relatively fewer cities were included in some regions such as the Northwest and Southwest regions, which might have limited the representativeness of these two regions; however, our sensitivity analyses based on a spatial statistical model produced consistent results, suggesting that the issue did not affect the result estimate to a great extent.

The findings from this study have some important implications for both public health and environment management. We suggest applying this indicator in future efforts. For example, the PGLE can be applied to estimate the effects of other air pollutants on life expectancy, as well as for conducting studies in different populations. The average life expectancy in China was 76.25 years in 2016. The Chinese government released the Healthy China (HC 2030) blueprint in 2016 as a national strategy. One goal of this plan is to increase the average life expectancy to 79 years by 2030. To achieve that goal, a series of action plans were suggested such as health education, diet control, and sufficient physical exercise [58]. In this respect, our study provided some new evidence that the life expectancy can be prolonged by controlling the concentrations of air pollution, and we suggest that this finding should be considered in future policy-making.

In conclusion, this study indicates that ambient PM2.5 might be a risk factor for YLL that should not be neglected, and significantly longer life expectancy could be achieved by a reduction in the pollution level.

The checklist of STROBE guidelines.

STROBE, Strengthening the Reporting of Observational Studies in Epidemiology.

(DOCX)

Click here for additional data file.

The list of 72 cities in our study.

(DOCX)

Click here for additional data file.

The list of model parameters in this study.

(DOCX)

Click here for additional data file.

Spearman correlation between air pollutants and meteorological factors in 72 cities of mainland China, 2013–2016.

(DOCX)

Click here for additional data file.

Regional-specific estimates of absolute change in YLL associated with each 10 μg/m3 increase in PM2.5 in single- and two-pollutant models in 72 cities of mainland China, 2013–2016.

(DOCX)

Click here for additional data file.

Sensitivity analyses for the absolute change in YLL associated with each 10 μg/m3 increase in PM2.5 in different models.

(DOCX)

Click here for additional data file.

Change in PM2.5–YLL relationship per IQR increase in city-level variables.

(DOCX)

Click here for additional data file.

Prospective analysis plan of the current study.

(DOCX)

Click here for additional data file.

Diagram of the model structure.

(TIF)

Click here for additional data file.

The plot of the residuals for 6 provincial capital cities.

(TIF)

Click here for additional data file.

The plot of PACF for 6 provincial capital cities.

PACF, partial autocorrelation function.

(TIF)

Click here for additional data file.

The normal Q-Q plot for 6 provincial capital cities.

(TIF)

Click here for additional data file.

The plot of the residuals for all cities.

(TIF)

Click here for additional data file.

22 Aug 2019

Dear Dr. Lin,

Thank you very much for submitting your manuscript "Potential gains in life expectancy by attaining daily ambient fine particulate matter pollution standards in mainland China: a nationwide analysis" (PMEDICINE-D-19-01976) for consideration at PLOS Medicine.

Your paper was evaluated by a senior editor and discussed among all the editors here. It was also discussed with an academic editor with relevant expertise, and sent to three independent reviewers, including a statistical reviewer. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of these reviews, I am afraid that we will not be able to accept the manuscript for publication in the journal in its current form, but we would like to consider a revised version that addresses the reviewers' and editors' comments. Obviously we cannot make any decision about publication until we have seen the revised manuscript and your response, and we plan to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We expect to receive your revised manuscript by Sep 12 2019 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see http://journals.plos.org/plosmedicine/s/submission-guidelines#loc-methods.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

We look forward to receiving your revised manuscript.

Sincerely,

Thomas McBride, PhD

Senior Editor

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

1- Did your study have a prospective protocol or analysis plan? Please state this (either way) early in the Methods section.

a) If a prospective analysis plan (from your funding proposal, IRB or other ethics committee submission, study protocol, or other planning document written before analyzing the data) was used in designing the study, please include the relevant prospectively written document with your revised manuscript as a Supporting Information file to be published alongside your study, and cite it in the Methods section. A legend for this file should be included at the end of your manuscript.

b) If no such document exists, please make sure that the Methods section transparently describes when analyses were planned, and when/why any data-driven changes to analyses took place.

c) In either case, changes in the analysis—including those made in response to peer review comments—should be identified as such in the Methods section of the paper, with rationale.

2- Thank you for providing information on how readers can access the primary data. In addition to the url for the air pollutant concentrations, please include the url or email contact information for obtaining meteorological data from the National Weather Data Sharing System of China. Additionally, while we understand that you cannot share the daily mortality time series data, please provide contact information (url or email) you used to obtain this data from the Death Surveillance Points System of China.

3- Please place the study design in your title following the colon. We suggest: “Potential gains in life expectancy by attaining daily ambient fine particulate matter pollution standards in mainland China: a modelling study based on nationwide data”

4- Please structure your abstract using the PLOS Medicine headings (Background, Methods and Findings, Conclusions).

5- Please provide the context of why the study is important in the Abstract Background. The final sentence should clearly state the study question.

6- In the Abstract Methods and Findings, please include the total population of the 72 cities, and the exact date range (include day and month) of the study.

7- It would also be helpful to include the average life expectancy and other demographic information (average age, % female) of the population included.

8- In the last sentence of the Abstract Methods and Findings section, please describe the main limitation(s) of the study's methodology.

9- Is it worth specifying in the Abstract Conclusions that these estimates and conclusions are based on short-term air pollution exposure?

10- At this stage, we ask that you include a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract. Please see our author guidelines for more information: https://journals.plos.org/plosmedicine/s/revising-your-manuscript#loc-author-summary

11- Please provide the name(s) of the institutional review board(s) that provided ethical approval, and whether informed consent was provided or waived (if waived, by whom).

12- Please include a list of the 72 cities included, perhaps as a supplemental text referenced from the Methods.

13- Please provide a diagram that shows the model structure, including how the disease natural history is represented, the process and determinants of disease acquisition, and how the putative intervention could affect the system.

14- Please provide a complete list of model parameters, including clear and precise descriptions of the meaning of each parameter, together with the values or ranges for each, with justification or the primary source cited, and important caveats about the use of these values noted.

15- Please provide a diagram that shows the model structure.

16- In the Discussion, please include a paragraph just preceding the Conclusion that discusses implications and next steps for research, clinical practice, and/or public policy.

Comments from the reviewers:

Reviewer #1: The study examined the associations between ambient PM2.5 and years of life lost (YLL) using four-year daily death data (1,226,849 non-accidental deaths) in 72 cities in China. These authors thus estimated the potential gains in life expectancy (PGLE) under the Chinese Ambient Air Quality Standards and WHO's air quality guidelines. They found that ambient PM2.5 reduction was significantly associated with decreased YLL and suggested that WHO's AQG should be selected as reference in China.

This paper addresses important public health issues related to the air pollution. Both the research aim and results in the study are important and interesting. However, I have some considerations in methods section:

1) The research divided the 72 cities into the seven regions (Figure 1) and assessed the association between PM2.5 and death at seven regional level. Authors state that same region has a similar features (meteorological, culture condition etc.). However, the northwest and southwest cover a rather big area which may have big variation for the features and the two big regions only include a few cities data. Considering this issue, I worry about the research results in regional level which may not be good representative for the two regions. Authors may consider using Bayesian spatial geo-statistical model to assess the association as the model can incorpate the spatial dependence and uncertainly?

2) These authors used two-stage Bayesian hierarchical statistical models (Generalized linear model with a Gaussian link and Bayesian hierarchical random-effects meta-analysis) to conduct data analysis. I would suggest authors providing more details of modeling such as model formula, goodness-of-fit of model which will help readers to understand the methods.

3) Have these authors considered the other factors such as smoking and social economic factors at city level at meta-analysis?

4) Line 138 GAM or GLM?

Reviewer #2: This study estimate the potential benefits on life expectancy by the daily ambient PM2.5 concentration meets the Chinese and WHO air quality standard respectively in both national and regional levels of China. The result provides a useful information for policy making on quantifying the beneficial effects of air pollution reduction. This manuscript is suitable to be published in PLOS Medicine after the following revision is conducted.

General comments:

Dataset includes concentrations of PM2.5, SO2, and NO2 and meteorological parameters such as temperature and relative humidity. However, very few discussion has been made using those air pollutants and meteorological parameters. Only a simple correlation results are shown in Table S1, so the meaning to use the above dataset in this paper is not clear. If this paper includes those data more deeply analysis and discussion is needed.

The 72 cities are divided into 7 regions in this study. However, there is almost no discussion on the comparison and interpretation about the different results of PGLE, Avoidable YLL and AF among those regions.

The English of this manuscript needs to be revised by native speaker.

Specific comments:

Line 50: "air pollution" is no necessary shown as Keywords.

Line 64-65: Data sources needs to be cited for annual PM2.5 concentration.

Line 103: What does the "randomly select and extract" mean?

Line 105: How to check the data set needs to be described more detailed.

Line 195-196: How to interpret this correlation results? What does the slightly positive or negative correlation mean?

Table 1: The difference of those variables among the 7 regions needs to be interpreted and discussed in main text. I think population is also important factor for health risk assessment. Why there is no data and discussion on population in different regions?

Line 211: After control NO2 the increment of YYL is smaller than control SO2 and O3. The reason needs to be interpreted.

Line 229-231: The description of "reducing reference standard" is not clear.

Line 246-250: The reason and indication from this different results of YLL in 7 regions needs to be interpreted.

Reviewer #3: Thank you for the opportunity to review this paper. Lin and colleagues conducted a nationwide study in China evaluating the association between daily PM2.5, years of life lost and further modelled potential gains in life due to achieving targeted national and WHO standards of PM2.5 targets. They have found across China, although with variation in levels of potential gains in life expectancy across regions, that overall significant gains in average life expectancy could be made by further reducing concentration levels. The 2-stage modelling approach is commonly used in ecological studies with the first stage estimating association using a generalised linear model (link function in this case was Gaussian) between PM 2.5 exposure and outcome of YLL. The second stage utilised a random-effects meta-analysis to tool cities and regions to obtain a national average. The authors also explored 3 day lag structure and moving averages, confounding by temperature and humidity. The study is well-conducted and presented. The only major issue I would have liked to have seen more thought given to in the analysis is incorporation other population level covariates which may affect this association - education, urbanicity, smoking, per capita income, unemployment come to mind.

1) Abstract methods does not mention the method on using adopting a two-stage model to combine city-specific results to estimate the nation-wide mean potential years of life expectancy gained

2) Abstract methods - would be helpful to indicate what is the recommended level of ambient PM 2.5 by the Chinese National air quality threshold and the WHO threshold.

3) Abstract conclusions - it's unclear what the authors mean by "PM2.5 might be a non-ignored risk factor" as national and international policy is currently targeting reduction in PM2.5 to reduce premature deaths.

4) Intro lines 77-81. It would be helpful for the reader to clearly delineate between short-term exposure and long-term exposure PM2.5 and what the suggested mechanism is for reduction of life expectancy and how they may differ from one another (i.e. some literature suggests long-term exposure has bigger effects)

5) Intro lines 89-92. YLL is still very commonly used and has been extended by the WHO recommended DALY measure combining mortality and morbidity. Influence by age size of the population can be solved by age-direct standardisation approaches.

see https://academic.oup.com/ije/article/48/4/1367/5281229

To me, the main advantage of PGLE isn't the influence of age structure and size as standardisation techniques allow for comparisons across difference areas, but rather the fact that YLL is extremely sensitive to competing risks (that is the exposure of interest contributes to several causes of death). This has not been mentioned in the text.

6) Methods Line 103 - What is the rationale here for random selection? I would have thought that it would be more advantageous to consider stratified sampling strategy to have a equal coverage across regions. Was this a pragmatic decision due to the number of cities which needed to be reduced or had a priori rationale.

7) Methods line 167 on two-stage Bayesian hierarchical model. This is an appropriate approach, where in stage 1 GLM with Gaussian link to estimate city specific association between PM2.5 and YLL, then using second stage to include a random-effect terms to pool cities to create regional and national estimates.

My question related to whether authors also consider a fully specified one-stage models (random study intercept or fixed study-specific intercept; random exposure effect; and fixed study-specific effects for covariate), investigating interaction - and if using a 1-stage this would have impacted the results instead of a 2-stage model.

8) Figure 2 - northwest region (n=8) missing reference line for 0.

9) Figure 3 needs better labelling y axis label missing and x-axis labels shared common labels between vertical panels are not intuitive. Also some bars are missing what I assume are the 95% CIs

Any attachments provided with reviews can be seen via the following link:

[LINK]

26 Sep 2019

Submitted filename: 1 Response_20190926v6.docx

Click here for additional data file.

7 Nov 2019

Dear Dr. Lin,

Thank you very much for submitting your revised manuscript "Potential gains in life expectancy by attaining daily ambient fine particulate matter pollution standards in mainland China: a modelling study based on nationwide data" (PMEDICINE-D-19-01976R1) for consideration at PLOS Medicine.

Your revision was evaluated by a senior editor and discussed among all the editors here. It was also sent to the original reviewers. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of these reviews, I am afraid that we still will not be able to accept the manuscript for publication in the journal in its current form, but we would like to consider a further revised version that addresses the reviewers' and editors' remaining comments. Obviously we cannot make any decision about publication until we have seen the revised manuscript and your response, and we plan to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We expect to receive your revised manuscript by Nov 14 2019 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see http://journals.plos.org/plosmedicine/s/submission-guidelines#loc-methods.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

We look forward to receiving your revised manuscript.

Sincerely,

Thomas McBride, PhD

Senior Editor

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

1- Thank you for including an Author Summary. The last point of the What did the researchers do and find? Section (“By adopting the WHO’s guideline as a reference, we estimated that about 1.00% of the years of life lost could be avoided in the study area.”) could use a bit more context or explanation. How is this point different from the point that precedes it? Over what timeframe is this estimate? Perhaps the point can be removed.

2- The Abstract Methods and Findings should also include the results for the Chinese NAAQS projection, quoted around line 376.

3- Line 480 do you mean “neglected” instead of “non-ignored”?

4- Please make sure all refrerences use the "Vancouver" style for reference formatting, and see our website for other reference guidelines https://journals.plos.org/plosmedicine/s/submission-guidelines#loc-references

Comments from the reviewers:

Reviewer #1: I think the manuscript has been improved. However, I have little consideration for the spatial model. These authors stated "A spatial statistical model by adjusting for the longitude and latitude of the cities in the model using a penalized smoothing splines function." I would suggest mapping these residuals from the model and check if these residuals are independence or appear spatial auto-correlation. Some discussion for why need to control the spatial auto-correlation in spatial model is needed.

Reviewer #2: I am appreciated to have a chance to review this paper again.

The authors have answered my comments sufficiently and revised the manuscript accordingling.

I have no more comment.

There are many comments about the models from other reviewers who are experts on this field. I am not sure whether the reply and revision is sufficient or not.

If the answer and revision about the method can be acceptable, this paper is suitable to be publish.

Reviewer #3: Thank you for addressing issues raised by in the reviews and incorporating additional analyses using the suggested methods. With these additional analyses, it's clear now that the primary results are robust (after numerous sensitivity/stress tests). There were some differences (i.e. use of one-stage model resulted in smaller effect sizes but the result retained significance), but the authors were able to detail and interpret these findings appropriately.

For me, the key strength now of the analyses is the separate meta-regression analyses exploring population level confounders. The interesting highlight of the confounders was the annual concentration being higher in areas had smaller effects between PM2.5 and life years. The authors suggest this could be self-awareness in areas with higher annual levels of pollution. I agree this could play a role but the final suggestion I think would be worthwhile to check is if there an interaction between annual PM2.5 and GDP with more industrialised areas being also wealthier. If so, this could partially explain as well whether there is some protection afforded by being located in an area with better health care access and care. This is a minor point though, so for that reason I recommend proceeding to acceptance of the very interesting and well-conducted analyses.

Any attachments provided with reviews can be seen via the following link:

[LINK]

14 Nov 2019

Submitted filename: 2 Response_20191114_R2.docx

Click here for additional data file.

11 Dec 2019

Dear Dr. Lin,

Thank you very much for re-submitting your manuscript "Potential gains in life expectancy by attaining daily ambient fine particulate matter pollution standards in mainland China: a modelling study based on nationwide data" (PMEDICINE-D-19-01976R2) for review by PLOS Medicine.

I have discussed the paper with my colleagues and the academic editor and it was also seen again by one of the original reviewers. I am pleased to say that provided the remaining editorial and production issues are dealt with we are planning to accept the paper for publication in the journal.

The remaining issues that need to be addressed are listed at the end of this email. Any accompanying reviewer attachments can be seen via the link below. Please take these into account before resubmitting your manuscript:

[LINK]

Our publications team (plosmedicine@plos.org) will be in touch shortly about the production requirements for your paper, and the link and deadline for resubmission. DO NOT RESUBMIT BEFORE YOU'VE RECEIVED THE PRODUCTION REQUIREMENTS.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

In revising the manuscript for further consideration here, please ensure you address the specific points made by each reviewer and the editors. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments and the changes you have made in the manuscript. Please submit a clean version of the paper as the main article file. A version with changes marked must also be uploaded as a marked up manuscript file.

Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. If you haven't already, we ask that you provide a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract.

We expect to receive your revised manuscript within 1 week. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

We ask every co-author listed on the manuscript to fill in a contributing author statement. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

If you have any questions in the meantime, please contact me or the journal staff on plosmedicine@plos.org.

We look forward to receiving the revised manuscript by Dec 18 2019 11:59PM.

Sincerely,

Clare Stone, PhD

Managing Editor

PLOS Medicine

plosmedicine.org

------------------------------------------------------------

Requests from Editors:

Line 162 – please remove data contact from main text.

Comments from Reviewers:

Reviewer #1: I think the paper has been further improved. I am happy with the current version.

Any attachments provided with reviews can be seen via the following link:

[LINK]

20 Dec 2019

Dear Prof. Lin,

On behalf of my colleagues and the academic editor, Dr. Jonathan A. Patz, I am delighted to inform you that your manuscript entitled "Potential gains in life expectancy by attaining daily ambient fine particulate matter pollution standards in mainland China: a modelling study based on nationwide data" (PMEDICINE-D-19-01976R3) has been accepted for publication in PLOS Medicine.

PRODUCTION PROCESS

Before publication you will see the copyedited word document (in around 1-2 weeks from now) and a PDF galley proof shortly after that. The copyeditor will be in touch shortly before sending you the copyedited Word document. We will make some revisions at the copyediting stage to conform to our general style, and for clarification. When you receive this version you should check and revise it very carefully, including figures, tables, references, and supporting information, because corrections at the next stage (proofs) will be strictly limited to (1) errors in author names or affiliations, (2) errors of scientific fact that would cause misunderstandings to readers, and (3) printer's (introduced) errors.

If you are likely to be away when either this document or the proof is sent, please ensure we have contact information of a second person, as we will need you to respond quickly at each point.

PRESS

A selection of our articles each week are press released by the journal. You will be contacted nearer the time if we are press releasing your article in order to approve the content and check the contact information for journalists is correct. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact.

PROFILE INFORMATION

Now that your manuscript has been accepted, please log into EM and update your profile. Go to https://www.editorialmanager.com/pmedicine, log in, and click on the "Update My Information" link at the top of the page. Please update your user information to ensure an efficient production and billing process.

Thank you again for submitting the manuscript to PLOS Medicine. We look forward to publishing it.

Best wishes,

Clare Stone, PhD

Managing Editor

PLOS Medicine

plosmedicine.org