Introduction
Great health concern has been placed on the severe air pollution in China. Particular matters as the dominant air pollutants in Chinese cities (e.g. PM
1, PM
2.5 and PM
10), have already been recognized as the Group I carcinogenic factor to lung cancer diseases in the world [
1]. As reported by the State of Global Air 2020, particular matter air pollution has led to the mortality of around 500,000 infants across the world [
2]. Despite considerable efforts on the estimates of particular matter effects, especially for PM
2.5 and PM
10 [
3‐
5], however, whether finer particular matter has the greater effect on human health has not been well understood in China and across the world.
Several potential mechanisms have been proposed to explain the varying effects of size-fractioned particular matters. Biologically, particular matters including PM
1, PM
2.5 and PM
10, can exert adverse effects on the physical health of human beings by the way of aggregating genetic damage [
1]. With regards to the difference in health effects, firstly, there is high ratio of surface area to volume in finer than in coarser particular matters. This enables finer particular matter to more easily approach the deeper places in lung, such as lung alveoli [
6,
7]. Secondly, the proportion of toxic chemical composition is usually higher in finer than in coarser particular matters. Such physicochemical property makes finer particular matter more easily cause detrimental effects on lung function and epigenetic alteration [
8,
9].
Empirically, few attempts have examined the effects of PMs with different particle sizes. In general, the argument that smaller particular matters have greater effects on human health, is still debated. Many studies tend to support this argument, especially for research investigating the effects of PM
1, PM
2.5 and PM
10 [
10‐
12]. Particularly, a time-series study performed in 65 Chinese cities between 2014 and 2017 indicated that the association with cardiovascular disease was stronger for PM
1 than for PM
2.5 and PM
10 [
7]. Similarly, as reported in the 33 Communities Chinese Health Study, the odds ratio of cardiovascular disease associated with a 10 μg/m
3 increase in PMs was 1.12 (95% CI:1.05, 1.20) for PM
1, which was higher than 1.06 (95% CI:1.01, 1.11) of PM
2.5 [
13]. By contrast, some studies report the greater effects of particular matters with larger particle sizes [
14,
15] or the insignificant effects of some size-fractioned particular matters [
16,
17]. For example, a case–crossover study performed in Barcelona of Spain suggested that the effect on cardiovascular mortality during Non-Saharan dust days was smaller for PM
1 than for PM
2.5 and PM
10 [
18].
Apart from inconsistent findings above, more efforts are required due to the three reasons. Firstly, of research investigating the varying effects of particular matters with different particle sizes, most are single- or several-site studies [
12,
19‐
21], while nationwide studies are quite limited [
7,
22]. Hence, findings from previous studies are still not sufficient to conclude the greater effects of finer particular matters than those of coarser particular matters. Secondly, few studies pay attention to PM
1 which is the dominant component of severe PM
2.5 air pollution in Chinese cities [
23]. This is partly resulted from the unavailable data on PM
1, especially at the national scale [
7,
10]. Thirdly, it remains unknown whether smaller particular matter has the larger effect on lung cancer which has become the second-order of cancer incidences for the female in China [
24], although numerous studies have suggested the effects of PMs (especially for PM
2.5 and PM
10) on lung cancer diseases [
3,
25‐
27].
To fill the aforementioned gaps, this work used data collected from 436 Chinse counties between 2014 and 2016 to examine whether finer particular matter has the greater effect on the incidence rate of female lung cancer in China where particular matter air pollution is much more severe than developed countries. To answer the research question, three regression models were developed with different controls of time, location and socioeconomic covariates. We further investigated whether the findings are sensitive to the controls of smoking and drinking behaviors as well as additional air pollutant. Moreover, we tried to answer whether urban-rural division modifies the association of the incidence rate of female lung cancer with each of three particular matters (i.e. PM1, PM2.5 and PM10).
Discussions
Many previous studies focus on PM2.5 and PM10 air pollutions in Chinese cities, while little attention has been placed on PM1. Currently, the strategies and standard already have been set for PM2.5 and PM10 in China, while the standard for PM1 is still missing. The nationwide or large-scale studies on health effects of PM1 benefit the establishment of PM1 standard and prevention measures all over China and even the world, while such studies are quite limited. Moreover, it remains unknown whether particular matters with smaller sizes are more harmful to human health, especially in China where particular matter air pollutions are still much more severe than western countries. As a response, this work as one of the earliest and largest nationwide studies in China, investigated the effects of PM1, PM2.5 and PM10 on the basis of data acquired from 436 counties/districts of China.
We observed the greatest effect of PM
1, followed by PM
2.5 and PM
10. This is consistent with many previous Chinese studies [
7,
10,
11]. In particular, a time-series study of eleven cities in Zhejiang Province of China suggested that the relative risk of all-cause mortality was higher for PM
1 than for PM
2.5 and PM
10, when there was the same increase in each of the three particular matters [
11]. Similarly, a short-term study conducted in two metropolitan cities of China (Guangzhou and Shenzhen) indicated that PM
1 was more closely associated with emergency department visits from all cause than PM
2.5 and PM
10 [
12]. The finding of more harmful effect of PM
1 was further supported in other prior studies [
13,
21,
49,
50]. Biologically, the high ratio of surface area to volume as well as the large percentage of toxic chemical composition in PM
1 may be responsible for the greater effect of PM
1. Our finding supports the argument that finer particular matters are more harmful than those with coarse particle sizes. Our finding also highlights that the regulation of particular matter air pollution in China should not only focus on PM
2.5 and PM
10, but also PM
1. Standards, measures and strategies targeting the alleviation of PM
1 should be highly prioritized because of the differences in chemical components and emission sources between PM
1 and other particular matters in China, especially the control of emissions from automobile transport and industry which are usually identified as the main sources of PM
1 air pollution.
We found that urban-rural division (rural group as the reference) positively modified the effect of PM
1, PM
2.5 and PM
10. The finding of urban-rural modifying role is in line with those of some previous studies [
43,
51,
52]. As suggested in a nationwide study of China [
27], when there was a 10 μg/m
3 increase in PM
2.5, the relative risk of lung cancer incidence was 1.06 (95% CI: 1.04, 1.08) in urban areas, which is higher than the value of rural areas at 1.04 (95% CI: 1.00, 1.08). A Chinese study also reported that the effect of PM
10 on the incidence of lung cancer was larger for urban inhabitants than for rural populations [
53]. From a biological perspective, the difference in air pollution effects may be attributed to the varieties of material resources, biological factors and psychological stress among different socioeconomic groups. In China, the difference in particular matter effect between urban and rural groups may partly result from the difference in smoking behaviours (e.g. smoking prevalence and smoking strength), which has been discussed in our prior research [
52]. The finding from the present study enhances the notion of the modifying role of socioeconomic factors on the effect of air pollution.
There are several limitations in the present study. Firstly, similar to most ecological research in relation to air pollution [
54,
55], the well-recognized inherent errors in the estimates of particular matter exposures may be produced in our work. We operationalized PM
1 (PM
2.5 and PM
10) exposure as the county (district)-aggregated mean concentrations of particular matter. Such operationalization did not take individual mobility into account and thus may produce exposure misclassification errors [
56,
57]. Secondly, the estimates of PM
1, PM
2.5 and PM
10 effects may be sensitive to the control of other potential socioeconomic and behavior covariates. In this work, our model construction is partly limited to the availability of socioeconomic and health data, with the number of variables by six and three, respectively. Hence, it may not be sufficient to control cofounders in relation to lung cancer diseases in the present study.
Thirdly, our work on the estimate of particular matter effect is an ecological study in nature, which may suffer from problems such as ecological fallacy [
58]. However, ecological studies have their strengths of large sample size as well as big spatial coverage. Ecological studies, in combination with studies using individual-level data, may benefit an in-depth and more scientific understanding of the differential effects of PM
1, PM
2.5 and PM
10 air pollution. Fourthly, the findings of PMs effects may be sensitive to the control of smoking and drinking factors, primarily resulted from the lack of smoking and drinking data for 436 registries. The sensitivity analysis, which derived city-level smoking and drinking data from the CHARLS, showed the robustness of PMs effects to the control of these behavior covariates. However, it is still not sufficient to well consider the effects of smoking and drinking factors. Attention should be paid if data these two covariates are available in the future.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.