3.1 Article selection and description
We initially identified 1977articles. Ultimately, we included 38 publications on long-term PM exposure in our meta-analysis (Figure 1). They are listed by region and pollutants and then chronologically in Table 1.
There is a large number of studies now that have evaluated mortality and morbidity effects of long-term exposure to PM2.5 (n=24) or PM10 (n=18), but few studies on PM1 (n=2). The vast majority uses the cohort study design (n=35), with only 3 articles use a case-control study design (two in patient population and one in infants). For the outcome, 23 publications were on cardiovascular diseases, 13 were on respiratory diseases, and 2 were on both. Studies have been conducted in a wide range of countries, though the majority has been conducted in North America and Europe. But because the study is about low-and middle-income countries, studies from Europe and the Americas were excluded. There is an increasing number of studies from Asia, but currently there are no studies from Africaand South America.
Table 1
Characteristics of the Selected Studies on Long-Term Exposure to PM
First Author (Year of Publication) | Area | Species | Study Period | Study Design | Samplesize | Age | Diseases | Outcome |
Zhang 2011 | Northern China | PM10 | 1998-2009 | COH | 12584 | 35-103years | CVD | mortality |
Dong 2012 | Shenyang, China | PM10 | 1998-2009 | COH | 9941 | ≥25years | RSD | mortality |
Zhang 2014 | four cities in northern China | PM10 | 1998-2009 | COH | 39054 | All | CVD | mortality |
Zhou 2014 | 25 cities | PM10 | 1990-1991 | COH | 71431 | ≥40years | CVD and RSD | mortality |
Tseng 2015 | Taiwan, China | PM2.5 | 1989-2008 | COH | 43227 | All | CVD | mortality |
Hwang 2015 | Taiwan, China | PM10 | 2001-2007 | CAS | 1087 | All | CVD | morbidity |
Yin 2015 | 44 counties or cities in China | PM2.5 | 2000-2005 | COH | 186399 | 40-79years | CVD | mortality |
Lai 2015 | Taiwan, China | PM2.5 | 2005-2012 | COH | 106678 | All | RSD | morbidity |
Jin 2015 | Lanzhou, China | PM10 | 2010-2012 | COH | 8969 | Infants | CVD | morbidity |
Liu 2016 | Shanghai, China | PM10 | 2011-2012 | COH | 3358 | Infants | RSD | morbidity |
Peng 2016 | Shanghai, China | PM2.5 | 2003-2013 | COH | 4444 | All | RSD | mortality |
Chen 2016 | Northern China | PM10 | 1998-2009 | COH | 39054 | All | RSD | mortality |
Wong 2016 | Hongkong, China | PM2.5 | 1998-2001 | COH | 66820 | ≥65years | Lung cancer | mortality |
Deng 2016 | Changsha, China | PM10 | 2011-2012 | COH | 2598 | 3-6years | Allergic rhinitis | morbidity |
Zhang 2016 | Wuhan, China | PM2.5,PM10 | 2012-2013 | COH | 105988 | Infants | CVD | morbidity |
Chen 2017 | Four cities in China | PM10 | 1999-2009 | COH | 39054 | All | RSD | mortality |
Yin 2017 | 45areas in China | PM2.5 | 1990-1991 | COH | 189793 | ≥40years | CVD and RSD | mortality |
Ren 2018 | Beijing, China | PM10 | 2009-2012 | COH | 30669 | Infants | CHD | morbidity |
Jiang 2018 | Changsha, China | PM10 | 2011-2012 | COH | 2598 | 3-6years | Pneumonia | morbidity |
Yang 2018 | Hongkong, China | PM2.5 | 1998-2011 | COH | 61386 | ≥65years | CVD | mortality |
Huang 2019a | Taiwan, China | PM2.5, PM10 | 2007-2014 | CAS | 5474 | Infants | CVD | morbidity |
Huang 2019b | 15provinces in China | PM2.5 | 1992-2008 | COH | 117575 | ≥18years | Stroke | morbidity |
Huang 2019c | China | PM2.5 | 2014-2015 | COH | 59456 | ≥18years | Hypertension | morbidity |
Chen 2019 | China | PM1, PM2.5,PM10 | 2007-2008 | CAS | 12291 | All | Stroke | mortality |
Mao 2019 | Henan, China | PM2.5, PM10 | 2015-2017 | COH | 39259 | 18-79years | CVD | morbidity |
Sun 2019 | Hongkong, China | PM2.5 | 1998-2001 | COH | 58643 | ≥65years | RSD | mortality |
Yang 2019 | Northeastern China | PM1, PM2.5 | 2006-2008 | COH | 24845 | 18-74years | CVD | mortality |
Bo 2019 | Hongkong, China | PM2.5 | 2001-2014 | COH | 134978 | ≥18years | Incident Hypertension | morbidity |
Hystad 2020 | 21 HIMCsand LMICs | PM2.5 | 2003-2018 | COH | 157436 | 35-70years | CVD | morbidity and mortality |
Ruchiraset 2020 | Thailand | PM10 | 2003-2014 | COH | 41085 | All | Pneumonia | morbidity |
Yin 2020 | Foshan, China | PM2.5 | 2015-2019 | COH | 63213 | All | CHD | morbidity |
Li 2020 | China | PM2.5 | 1992-2015 | COH | 118551 | ≥18years | Lung cancer | morbidity and mortality |
Liang 2020 | China | PM2.5 | 2000-2015 | COH | 127840 | ≥18years | CVD | morbidity and mortality |
Yang 2020 a | China | PM2.5 | 2000-2015 | COH | 116821 | ≥18years | CVD | mortality |
Yang 2020 b | Foshan, China | PM2.5 | 2015-2019 | COH | 61884 | All | CHD | morbidity |
Lin 2021 | Taiwan, China | PM2.5 | 2005-2011 | COH | 140911 | Infants | RSD | morbidity |
Yang 2021 | northern China | PM2.5 | - | COH | 38140 | All | Stroke | mortality |
Paoin 2021 | Thailand | PM10 | 2005-2013 | COH | 25532 | All | CVD | morbidity |
COH, Cohort study; CAS, Case-control study; CVD, Cardiovasculardiseases; RSD, Respiratory diseases; CHD, Congenital Heart Disease |
3.2 Impacts of PM2.5 on morbidity and mortality
After a systematic search and review of English literature, 22 cohort studies and 2 case-control studies assessing chronic morbidity and mortality effects attributable to PM2.5 exposure were identified and included in the meta-analysis, which were all published after 2014. Most of the studies were conducted in China.
PM 2.5 on cardiovascular diseases
As can be seen from the forestplot below, with per 10 µg/m3 increase in PM2.5 concentrations, cardiovascular morbidity was increased by 13% (95% CI: 1.07, 1.2), P<0.001, I2=91.2% (Figure 2). Cardiovascular mortality was increased by13% (95% CI: 1.08, 1.19), P<0.001, I2=79.9% (Figure 3).
Subgroup analysis
We did a subgroup analysis on the morbidity of PM2.5 on cardiovascular diseases (table 2), including study design (cohort study and case-control study), sex (male and famale) and disease (stroke, ischemic heart disease, congenital heart disease, tetralogy of fallot and hypertension)).
For males, the morbidity increased by 8% (95% CI :1.06,1.1), P<0.001, I2=98.2% for 10 µg/m3 increase in PM2.5 concentrations; but in famales, it’s 14% (95% CI:1.12, 1.17), P<0.001, I2=96.4%. There are also differences in the morbidity of different diseases. For 10 µg/m3 increase in PM2.5 concentrations, stroke morbidity increase 9% (95% CI:1.06,1.12), I2=78%, P<0.001.
Table2. Subgroup analysis of PM 2.5 on the morbidity of cardiovascular diseases
Characteristics | Morbidity | |
n | HR(95%CI) | I² | P | P-interaction |
Type of study | |
Cohort | 9 | 1.07(1.06, 1.09) | 99.1 | <0.001 | 0.264 |
Case-control | 1 | 1.15(1.02, 1.3) | - | 0.024 | |
Sex | |
Male | 3 | 1.08(1.06, 1.1) | 98.2 | <0.001 | <0.001 |
Famale | 3 | 1.14(1.12, 1.17) | 96.4 | <0.001 | |
Disease | |
Stroke | 4 | 1.09(1.06, 1.12) | 78 | <0.001 | <0.001 |
CHD | 3 | 1.01(0.96, 1.08) | 75.2 | 0.635 | |
TF | 2 | 1.07(0.92, 1.24) | 0 | 0.379 | |
Hypertension | 2 | 0.78(0.76, 0.81) | 99.6 | <0.001 | |
CHD, Congenital Heart Disease; TF, Tetralogy of Fallot |
Publication bias
Egger test was performed on the literature about the influence of PM2.5 on cardiovascular diseases. For morbidity, P = 0.767, there was no publication bias.
PM 2.5 on respiratory diseases
Respiratory diseases have reveieved less study as compared to cardiobvascular diseases. We performed a meta-analysis of respiratory morbidity and mortality, and subgroup analysis of mortality for different diseases like COPD, Tuberculosis, and lung cancer as well as for ages older than 65 years and those less than 65 years.
As can be seen from the forest plot, with every 10 µg/m3 increase in PM2.5 concentrations, respiratory morbidity has no statistical significance (Figure 4). Respiratory mortality increased by 10% (95% CI: 1.02, 1.18), P<0.05, I2 = 67.5% (Figure 5).
Subgroup analysis
Meta-analysis shows that with every 10 µg/m3 increase in PM2.5 concentrations, the mortality of COPD increased by 12% (95% CI: 1.11, 1.14), P<0.001, while that of tuberculosis increased by 22% (95% CI: 1.09, 1.36), P =0.001 and that of lung cancer by 12% (95% CI: 1.09, 1.16), P<0.001 (Table 3).
For people over 65 years, respiratory diseases mortality had no significance; for people less than 65 years, it increased by 12% (95% CI: 1.11 1.14), P<0.001. The result is the same with those of the cardiovascular diseases. This may be due to the fact that people under the age of 65 spend more time outdoors and are more exposed to PM2.5. The elder’s mortality was more related to other diseases.
Table 3
Subgroup analysis of PM2.5 on mortality of respiratory diseases
Characteristic | Mortality |
n | HR(95%CI) | I² | P | P-interaction |
Disease | |
COPD | 3 | 1.12(1.11,1.14) | 0 | <0.001 | 0.36 |
Tuberculosis | 3 | 1.22(1.09,1.36) | 0 | 0.001 | |
Lung cancer | 4 | 1.12(1.09,1.16) | 0 | <0.001 | |
Age | |
≥65 | 3 | 1.02(0.95, 1.09) | 0 | 0.606 | 0.005 |
<65 | 4 | 1.12(1.11, 1.14) | 66.20 | <0.001 | |
COPD, Chronic Obstructive Pulmonary Disease |
3.2 Impacts of PM10 on morbidity and mortality
PM 10 On cardiovascular diseases
PM10 was positively associated with cardiovascular diseases at increased levels. We can see from the forest plot, with every 10 µg/m3 increase in PM10 concentrations, cardiovascular morbidity increased by 8% (95% CI: 1.00, 1.16), P<0.001, I2 = 84.1% (Figure 6). Cardiovascular mortality increased by 17% (95% CI: 1.00, 1.37), P<0.001, I2 = 99.7% (Figure 7).
Subgroup analysis
The subgroup analysis for the impact of PM10 on cardiovascular diseases including morbidity by different diseases (CHD, VSD, ASD, TF), population (non-infant and infant), and mortality by sex (male and female) (Table 4) was conducted.
The meta-analysis of different diseases shows that with every 10 µg/m3 increase in PM10 concentrations, the morbidity of CHD increased by 2% (95% CI : 0.99, 1.06), I2 = 87.5%, P<0.001, while that of VSD increased by 3% (95% CI : 1.02, 1.04), I2 = 98.6%, P<0.001. But the morbidity of ASD and TF have no statistical significance, and no difference in the morbidity among different diseases.
We made an analysis of mortality by sex. For male, the mortality increased by 6% (95% CI : 1.04, 1.07), I2 = 98.4%, P<0.001 for 10 µg/m3 increase in PM10 concentrations; for famale, it’s 15% (95% CI: 1.13, 1.17), I2 = 99.5%, P<0.001.
Table 4
Subgroup analysis of PM10 on morbidity and mortality of cardiovascular diseases
Characteristic | n | HR(95%CI) | I² | P | P-interaction |
Morbidity | | | | | |
Disease | | | | | |
CHD | 3 | 1.02(0.99,1.06) | 87.5 | <0.001 | 0.386 |
VSD | 3 | 1.03(1.02,1.04) | 98.6 | <0.001 | |
ASD | 2 | 1.02(1.01,1.03) | 9.7 | 0.293 | |
TF | 3 | 1.01(0.99,1.03) | 52.3 | 0.123 | |
Population | | | | |
Infant | 5 | 1.04(1.01,1.07) | 87.6 | <0.001 | 0.02 |
Non-infant | 2 | 1.11(1.06,1.17) | - | 0.863 | |
Mortality | | | | | |
Sex | | | | | |
Male | 3 | 1.06(1.04, 1.07) | 98.4 | <0.001 | <0.001 |
Famale | 3 | 1.15(1.13, 1.17) | 99.5 | <0.001 | |
CHD, Congenital Heart Disease; VSD, Ventricular Septal Defect; ASD, Atrial Septal Defect; TF, Tetralogy of Fallot |
PM 10 On Respiratory diseases
As the forest plot shows, with every 10 µg/m3 increase in PM10 concentrations, respiratory morbidity increased by 38% (95% CI: 1.00, 1.88), P<0.001 (Figure 8), while respiratory mortality increased by 28% (95% CI: 1.10, 1.49), P<0.001 (Figure 9).
Subgroup analysis
There were few studies on the chronic effects of PM10 on respiratory diseases, but we still analysed the major diseases, and the results are shown in Table 5. The subgroup analysis including morbidity was conducted on two different countries (China and Thailand) and for three different diseases (pneumonia, AR and asthma). The mortality was analysed for COPD and lung cancer.
There was no significant difference in morbidity between China and Thailand. The morbidity of pneumonia increased by 1% (95% CI: 1.00, 1.01), P<0.011. The analysis showed a significant difference between pneumonia and AR and asthma, but since there was only one study for AR and asthma, this result need to be confirmed with those of more related studies. The mortality of COPD increased by 2% (95% CI: 0.99, 1.06), I2 = 95.3%, P<0.001. The mortality of lung cancer increased by 4% (95%CI: 1.02, 10.6), I2 = 99.3%, P<0.001.
Table 5
Subgroup analysis of PM10 on morbidity and mortality of respiratory diseases
Characteristic | n | HR(95%CI) | I² | P | P-interaction |
Morbidity | | | | | |
Country | | | | | |
China | 3 | 0.94(0.76,1.16) | 77.8 | 0.011 | 0.496 |
Thailand | 1 | 1.01(1.00,1.02) | 0 | - | |
Disease | | | | | |
Pneumonia | 3 | 1.01(1.00,1.01) | 0 | 0.358 | 0.013 |
AR | 1 | 3.34(1.42,7.88) | 0 | - | |
Asthma | 1 | 0.85(0.62,1.16) | 0 | - | |
Mortality | | | | | |
Disease | | | | | |
COPD | 2 | 1.02(0.99,1.06) | 95.3 | <0.001 | 0.443 |
Lung cancer | 2 | 1.04(1.02,1.06) | 99.3 | <0.001 | |
AR, Allergic Rhinitis; COPD, Chronic Obstructive Pulmonary Disease |
3.2 Impacts of PM1 on cardiovascular diseases
There were only two studies on the chronic effects of PM1 on the cardiovascular diseases, and both analyses were sex-specific. Therefore, we analysed the results of the two studies by sex classification (Figure 10).
From the forest plot we can see that for males, the effect on the cardiovascular diseases increased by 10% (95%CI: 1.01, 1.21) for every 10µg/m3 increase in PM1 concentration, while the results were not significant for females. There was also no correlation between males and females.