| First author and year by domaina . | Population, year coverage, and country/region . | Study design . | Exposuresb . | Summary of results . | Percentage of results in expected direction . | Quality assessment (ROBIS) . |
|---|---|---|---|---|---|---|
| Air pollution | ||||||
| Alexeeff et al.34 | General population, till 31 December 2019, country not restricted: Europe, Canada, USA, UK, Australia, China, South Korea, and Israel | 42 CS | PM2.5 | Per 10 μg/m3 increase in long-term exposure, the pooled RRs and 95% CIs were 1.23 (1.15–1.31) for IHD mortality, 1.08 (0.99–1.18) for incident acute MI, 1.24 (1.13–1.36) for CeVD mortality, and 1.13 (1.11–1.15) for incident stroke. | 83% | High risk of bias |
| Atkinson et al. 2016 | General population, till October 2015, country not restricted: USA, UK, China (Taiwan), and France | Total 22 studies, relevant 8 studies: 8 CS | O3 | HRs expressed per 10 ppb increase in O3. For long-term annual O3 concentrations, the standardized effect estimates (HRs and 95% CIs) were 1.01 (0.99–1.03) for CVD mortality, 1.02 (1.00–1.04) for IHD mortality, and 1.01 (0.97–1.05) for stroke mortality. For long-term annual O3 concentrations, the random-effects summary estimates (HR and 95% CI) were 0.98 (0.93–1.04) for CVD mortality and 1.00 (0.92–1.09) for IHD mortality. For the warm season/peak O3, random-effects summary estimates were 1.01 (1.00 to 1.02). | 42% | High risk of bias |
| Atkinson et al.35 | General population, 1996–October 2016 (Medline, EMBASE), 1970–October 2016 (Web of Science), 1966–October 2016 (PubMed), country not restricted: Europe, North America, China (mainland and Taiwan), and Japan | Total 48 studies, relevant 22 studies: 22 CS | NO2 (annual or multi-year averages) | Per 10 µg/m3 increase in long-term exposure, the pooled HRs and 95% CIs were 1.03 (1.02–1.05) for CVD mortality, 1.05 (1.03–1.06) for CHD mortality, and 1.01 (0.98–1.03) for CeVD mortality. | 91% | High risk of bias |
| Chen et al. 2008 | Adults, 1 January 1950–31 December 2007, country not restricted: USA, Norway, France, the Netherlands, Canada, and Germany | Total 32 studies, relevant 17 studies: 14 CS, 3 CCS | O3, SO2, NO/NO2, black smoke, PM10, PM2.5, CO, benzene, and polycyclic aromatic hydrocarbons | RR per 10 µg/m3 increase. For PM2.5, the pooled RRs and 95% CIs were 1.14 (1.09–1.18) for CVD mortality and 1.16 (0.96–1.40) for CHD mortality. For other particulate and gaseous pollutants, the paucity of data precludes drawing conclusions. | 79% | High risk of bias |
| Chen et al. 2020 | General population, till 9 October 2018, country not restricted: Europe, Canada, UK, USA, Israel, New Zealand, South Korea, Japan, and China (mainland, Taiwan, and Hong Kong) | 67 CS | PM2.5 and PM10 | Per 10 µg/m3 increase in long-term PM2.5 exposure, the pooled RRs and 95% CIs were 1.11 (1.09–1.14) for CVD mortality, 1.16 (1.10–1.21) for IHD mortality, and 1.11 (1.04–1.18) for stroke mortality. The estimates of PM10 were 1.04 (0.99–1.10) for CVD mortality, 1.06 (1.01–1.10) for IHD mortality, and 1.01 (0.83–1.21) for stroke mortality. The certainty of evidence was high for PM2.5 and CVD mortality and was moderate for PM10 and CVD mortality, as measured by GRADE framework. | 81% | Low risk of bias |
| Chen et al.36 | General population, till October 2019, country not restricted: South Korea, UK, Denmark, Sweden, and Canada | Total 18 studies, relevant 6 studies: 6 CS | PM2.5, PM10, NO2, SO2, O3, and CO | Per 10 μg/m3 increase in long-term exposure, the pooled HRs, and 95% CIs of AF were 1.116 (1.031–1.207) for PM2.5, 1.034 (1.032–1.035) for PM10, 1.017 (1.001–1.033) for NO2, 1.005 (1.004–1.007) for SO2, 1.017 (1.013–1.022) for CO, and 1.007 (0.927–1.094) for O3. | 83% | Low risk of bias |
| Faustini et al. 2014 | General population, January 2004–January 2013, country not restricted: identified countries: Japan, China, Canada, USA, UK, Italy, Germany, Sweden, the Netherlands, and Norway | Total 23 studies, relevant 17 studies: 2 CCS, 15 CS | NO2 | RR per 10 µg/m3 increase. The pooled RRs and 95% CIs for CVD mortality were 1.13 (1.09–1.18) for NO2 and 1.20 (1.09–1.31) for PM2.5. | 94% | High risk of bias |
| Hak-Kan et al. 2013 | Chinese population, till 30 June 2012, 80 major Chinese cities in Mainland China, Hong Kong, and Taiwan | Total 48 studies, relevant 3 studies: 3 CS | PM10, NO2, SO2, and O3 | RR per 10 µg/m3 increase. In one cohort study examining PM10 and NO2, the corresponding RRs and 95% CIs were 1.0155 (1.0151–1.0160) and 1.0246 (1.0231–1.0263) for CVD mortality and 1.0149 (1.0145–1.0153) and 1.0244 (1.0227–1.0262) for CeVD mortality. In another cohort study examining SO2 and CVD, the corresponding RR was 1.032 (1.023–1.040). | 100% | High risk of bias |
| Hoek et al. 2013 | Not specified (adults), till January 2013, country not restricted: identified countries: USA, Germany, the Netherlands, Switzerland, Canada, China, New Zealand, Japan, Italy, France, and Denmark | Total 67 studies, relevant 34 studies: 34 CS | Long-term exposure to fine particulate matter (PM2.5, PM10, NO2, elemental carbon, and coarse particles) | RR per 10 µg/m3 increase. For PM2.5, the pooled RR and 95% CI was 1.11 (1.05–1.16) for CVD mortality. There was no consistent evidence that long-term exposure to coarse PM or elemental carbon is associated with CVD mortality. Several studies found positive associations between NO2 exposure and fatal MI, but not non-fatal MI. The evidence for an association between air pollution and CeVD mortality was inconsistent. | 88% | High risk of bias |
| Huang et al. 2021 | General population, till 2020.02.29, country not restricted: USA, Canada, Norway, the Netherlands, UK, Italy, Denmark, France, Spain, Japan, China, South Korea, Australia, Sweden, Norway, Germany, Austria, Switzerland, France, Italy, Spain, Greece, and Finland | 32 CS | NO2 | Per 10 ppb increase in annual NO2 concentration, the pooled HR and 95% CI was 1.11 (1.07–1.16) for cardiovascular mortality. | 71% | High risk of bias |
| Jadambaa et al.37 | Mongolian population (adults and children), till April 2014, Mongolia | Total 59 studies, relevant 2 studies: 2 CSS | NO2 and PM2.5 | Two studies found an increased risk of CVD with increased exposure to NO2 and PM2.5. | 100% | Unclear |
| Jaganathan et al.38 | General population, 1 January 1948–6 March 2018, country restricted to low- and middle-income countries, Mexico (Mexico City), Brazil (São José dos Campos, Cuiabá, and Várzea Grande), China, and India (Varanasi) | Total 17 studies, relevant 12 studies: 8 LS, 2 CSS, 1 CCR, 1 CS | PM2.5 (annual average or average measure of more than 3 days) | Eight out of nine studies (91%) reported significant effects on CVD mortality. Per 10 µg/m3 increase in long-term exposure, the effect estimates of CVD mortality ranged from 0.24 to 6.11%. All four studies reported significant effects of long-term exposure on CVD hospitalization. Few studies have evaluated this association in LMICs. No studies were found in North and Sub-Saharan Africa. | 92% | Low risk of bias |
| Kan et al. 2005 | General population, 1990–2002, China and worldwide | Total 26 studies, relevant 7 studies: 7 CS | Effects of particulate air pollution. PM10 was selected as the indicator particulate matter. | RR per 10 µg/m3 increase. For PM10, the pooled RRs and 95% CIs were 1.0095 (1.0060–1.0130) for CHA, 1.013 (1.007–1.019) for CHA based on four European studies, and 1.008 (1.004–1.011) for CHA based on three US and Canadian studies. | 100% | High risk of bias |
| Karimi et al. 2019 | Iran population, January 1980–January 2018, country restricted to Iran | Total 38 studies, relevant 28 studies: 27 CSS, 1 ES | O3, PM2.5, PM10, NO2, NOx, SO2, and CO measured by environmental protection organization and air quality control centre | Per 10 µg/m3 increase in all air pollutants, the pooled increased risk (95% CI) in CVD mortality was 0.5% (0.4–0.6%). The estimate for PM2.5 and PM10 was 0.7% (0.4–1%). | NA | Unclear |
| Liu et al. 2018 | General population, adults, January 1974–July 2017, country not restricted: USA, UK, Italy, Canada, China (mainland and Hong Kong), Europe, New Zealand, and Japan | 16 CS | PM2.5 and PM10 | Per 10 μg/m3 increase in long-term exposure, the pooled HRs and 95% CIs of CVD mortality were 1.12 (1.08–1.16) for PM2.5, 1.02 (0.89–1.16) for PM10, and 1.10 (1.06–1.14) for combined. In subgroup analyses, there is no difference in the association stratified by categories of WHO PM levels or smoking status. The estimates of PM2.5 were 1.19 (1.11–1.27) for studies with ≥ 11 years of follow-up, higher than those <11 years: 1.07 (1.04–1.11). | 88% | Low risk of bias |
| Lu et al. 2015 | Chinese population (adults only), 1990–2013, Mainland China, Hong Kong, and Taiwan | Total 59 studies, relevant 2 studies: 2 CS | PM10 and PM2.5 | RR per 10 μg/m3 increase. For the annual average concentration of PM10, the RR and 95% CI was 1.23 (1.19–1.26) for CVD mortality in one study and 1.55 (1.51–1.60) for CVD mortality in another study. | 100% | High risk of bias |
| Luben et al. 2017 | Adults, till 15 June 2017, country not restricted: USA, China (mainland and Taiwan), the Netherlands, Canada, South Korea, Spain, and Italy | Total 24 studies, relevant, 3 studies: 2 CS, 1 LS | Ambient black carbon | There are generally modest, positive associations of long-term exposure to black carbon and elemental carbon with cardiovascular hospital admissions and mortality. | 100% | High risk of bias |
| Niu et al. 2021 | General population, till 1 February 2020, country not restricted: China, Europe, England, Japan (Shizuoka), USA (California), Ghana, India, Mexico, Russia, and South Africa | Total 68 studies, relevant 13 studies: 13 CS | PM2.5, PM10, and NO2 | Per 10 μg/m3 increase in long-term exposure, the pooled HRs and 95% CIs of stroke incidence were 1.081 (0.971–1.023) for PM2.5, 1.033 (0.907–1.175) for PM10, and 1.005 (0.977–1.034) for NO2; the HRs and 95%CI of stroke mortality were 1.047 (0.995–1.101) for NO2. | 82% | High risk of bias |
| Prueitt et al.39 | General population, 1 January 2006–4 November 2013, country not restricted: USA, UK, Canada (Toronto), and China (Liaoning) | Total 25 studies, relevant 11 studies: 8 CS, 2 CSS, 1 ES | O3 | For long-term O3 exposure and CVD morbidity, studies were rare and reports were inconsistent. For CVD mortality, of 10 high-quality studies, 5 reported positive association, and the other 5 reported null or negative association. | 17% | High risk of bias |
| Scheers et al. 2015 | General population, till 20 July 2015, country not restricted: Japan, China, UK, the Netherlands, Switzerland, Greece, USA, Canada, Finland, Norway, Sweden, Denmark, Germany, Austria, Italy, Greece, and France | Total 20 studies, relevant 20 studies: 14 CS, 6 ES | PM10 or PM2.5 | HR per 10 μg/m3 increase. For PM10, the pooled HRs and 95% CIs were 1.061 (1.018–1.105) for overall stroke events and 1.080 (0.992–1.177) for stroke mortality. For PM2.5, the pooled HRs and 95% CIs were 1.064 (1.021–1.109) for overall stroke events and 1.125 (1.007–1.256) for stroke mortality. | 50% | High risk of bias |
| Shin et al. 2014 | Not specified (adults), from 1990, country not restricted: USA and UK | Total 20 studies, relevant 4 studies: 4 CS | PM2.5 | RR per 10 μg/m3 increase. In the frequentist meta-analysis, the pooled RR and 95% CI for long-term exposure to PM2.5 was 1.06 (1.00–1.13) for non-fatal strokes. The Bayesian meta-analysis found a posterior mean 1.08 (0.96–1.26) from a normal prior and 1.05 (1.02–1.10) from a gamma prior. | 100% | High risk of bias |
| Stieb et al.40 | General population, till 25 February 2020, country not restricted: Canada, USA, UK, Europe, China (mainland, Hong Kong, and Taipei), Australia, South Korea (Seoul), and Japan (Shizuoka) | 49 CS | NO2 | Per 10 ppb increase in long-term exposure, the pooled HRs and 95% CIs were 1.139 (0.997–1.301) for CVD mortality, 1.128 (1.076–1.182) for IHD mortality, and 1.167 (0.936–1.456) for CeVD mortality. After excluding studies with probably high or high risk of bias, the pooled HRs and 95% CIs were 1.058 (1.026–1.091) for CVD mortality, 1.111 (1.079–1.144) for IHD mortality, and 1.014 (0.997–1.032) for CeVD mortality. | 74% | Low risk of bias |
| Wang et al. 2020 | Older adults aged ≥ 55 years, till January 2020, country not restricted: USA (Steubenville, Eastern Massachusetts, Boston), Germany (Erfurt), Finland (Helsinki), the Netherlands (Amsterdam), UK (Scotland: Aberdeen), and China (Beijing and Taipei) | Total 19 studies, relevant 10 studies: 10 LS | Concentration of PM2.5 | Per 10 mg/m3 increase in long-term exposure, the pooled estimates and 95% CIs of HRV were −0.92% (−2.14 to 0.31%) for SDNN, −1.96% (−3.48 to −0.44%) for RMSSD in time-domain measurements, −2.78% (−4.02 to −1.55%) for LF, and −1.61% (−4.02 to 0.80%) for HF in frequency domain measurements. | 68% | High risk of bias |
| Yang et al. 201941 | General population, till 25 April 2018, country not restricted: Europe, UK, Canada, USA, South Korea, China, Ghana, India, Mexico, Russia, South Africa, and Japan | 35 CS | PM2.5, PM10, O3, and NO2 | Per 10 μg/m3 increase in long-term PM2.5 exposure, the pooled RRs and 95% CIs were 1.11 (1.07–1.15) for CVD events, 1.12 (1.05–1.19) for stroke incidence, 1.12 (1.08–1.16) for stroke events, 1.19 (1.09–1.30) for IHD incidence, and 1.14 (1.08–1.21) for IHD events. The estimates of CVD mortality were 1.11 (1.07–1.15) for PM2.5, 1.09 (1.02–1.16) for PM10, 1.23 (1.15–1.31) for NO2, and 1.03 (1.02–1.05) for O3. The estimates of NO2 and IHD events were 1.05 (1.04–1.06). No significant associations were found between PM10 and CVD, stroke and IHD incidence. | 87% | High risk of bias |
| Yuan et al. 2019 | General population, 1980–December 2018, country not restricted: Europe, USA, China (Hong Kong), Ghana, India, Mexico, Russia, South Africa, UK, Sweden (Gothenburg), and Italy | 16 CS | PM2.5 | Per 5 μg/m3 increase in long-term exposure, the pooled HRs and 95% CIs were 1.11 (1.05–1.17) for stroke incidence and 1.11 (1.05–1.17) for stroke mortality. In subgroup analysis, the estimates of stroke incidence were 1.09 (1.05–1.14) for North America (5 CS), 1.07 (1.05–1.10) for Europe (4 CS), and 2.31 (0.49–10.95) for Asia (2 CS). The associations were insignificant in both sex and significant in both ischaemic and haemorrhagic stroke. The estimates of stroke incidence were 1.08 (1.03–1.13) for never smokers, 1.11 (1.01–1.22) for former smokers, and 1.08 (0.94–1.25) for current smokers. | 95% | Low risk of bias |
| Zhao et al. 2017 | General population, 1990–2016, country not restricted: USA, Israel, Japan, UK, China, Italy, Norway, Greece, Canada, Denmark, France, South Korea, Iran, Germany, Finland, Sweden, Spain, and the Netherlands | Total 48 studies, relevant 48 studies: 25 CS, 23 LS | PM10, PM2.5, SO2, NO2, CO, and O3 | HR per 10 μg/m3 increase. For CHD mortality, the pooled HRs and 95% CIs were 1.12 (1.04–1.20) for PM10, 1.17 (1.12–1.22) for PM2.5, 1.03 (1.00–1.07) for SO2, 1.04 (1.01–1.06) for NO2, 1.04 (0.98–1.10) for CO, and 1.06 (1.01–1.11) for O3 (10 mg/m3 increase). For CHD incidence, the pooled HRs and 95% CIs were 1.01 (1.00–1.02) for PM10, 1.02 (1.00–1.03) for PM2.5, 1.01 (1.00–1.02) for SO2, 1.04 (1.03–1.06) for NO2, 1.01 (0.97–1.04) for O3, and 1.03 (1.00–1.05) for CO (10 mg/m3 increase). | NA | High risk of bias |
| Zhao et al. 2021 | General population, time and country not restricted: China, Norway, UK, the Netherlands, China (Hong Kong), and Canada (Ontario) | 7 CS | PM2.5 acquired through satellite-based model (5 studies) and outdoor-automated monitoring stations (2 studies) | Per 1.4–10 μg/m3 increase in long-term PM2.5 exposure, the pooled HRs and 95% CIs of haemorrhagic stroke were 1.16 (1.03–1.30) for total, 1.41 (0.92–2.15) for current smoker, and 1.04 (0.74–1.46) for never and former smoker. | 71% | Low risk of bias |
| Zhu et al. 2021 | General population, till 2 August 2020, country not restricted: Canada, Denmark, the Netherlands, China, USA, South Korea, Israel, and UK (London) | 12 CS | PM2.5 | Per 10 μg/m3 increase in long-term PM2.5 exposure, the pooled HRs and 95% CIs were 1.10 (1.02–1.18) for MI incidence and 1.07 (1.04–1.09) for post-MI mortality. | 75% | Unclear |
| Zou et al.42 | General population, till September 2019, country not restricted: USA, South Korea, UK, Canada, Sweden, Israel, Italy, the Netherlands, Switzerland, and Finland | 27 CS | PM2.5 and PM10 | Per 10 μg/m3 increase in long-term exposure, the pooled RRs and 95% CIs of MI were 1.18 (1.11–1.26) for PM2.5 and 1.03 (1.00–1.05) for PM10. | 91% | Unclear |
| Physical activity environment | ||||||
| Gascon et al. 2016 | Adults, till 14 November 2014, country not restricted: USA, UK, New Zealand, Lithuania, and Canada | Total 12 studies, relevant 8 studies; 4 ES, 2 CS, 1 CSS | Residential natural outdoor environments, particularly green and blue spaces | For each 10% increase of greenness, the RR and 95% CI was 0.993 (0.985–1.001) for CVD mortality. For high vs. low categories of greenness, the RR and 95% CI was 0.96 (0.94–0.97) for CVD mortality. | 75% | Unclear |
| Twohig-Bennett et al.43 | General population, till January 2017, country not restricted: USA, UK, and Lithuania | Total 143 studies, relevant 4 studies: 3 CS, 1 ES | Greenspace measured by residential NVDI, distance to the nearest greenspace, and proportion of city area covered by green land | Comparing higher to lower greenspace exposure, the pooled ORs and 95% CIs were 0.82 (0.61–1.11) for stroke (3 studies), 0.84 (0.76–0.93) for CVD mortality (2 studies), and 0.92 (0.78–1.07) for CHD (2 studies). | 86% | High risk of bias |
| Yuan et al. 2020 | Older adults (mostly ≥ 60 years), 1 January 2000–1 July 2020, country not restricted: Japan, Canada, USA, Finland, China, Rome, Australia, the Netherlands, Lithuania, Brazil, Israel, South Korea, Iran, and UK | Total 22 studies, relevant 17 studies:12 CS, 5 CSS | Greenspace measured by NDVI (mostly), percent of greenspace coverage, distance to the nearest green space, park visitation and length of stay, and loss of trees from emerald ash bore disease | Of 8 studies in total CVD, 7 found beneficial effects of green space, and the other study showed a lower risk of CVD with higher percentage of tree canopy, but not total green space. Evidence for stroke and MI was less consistent. Only cohort studies measuring NDVI and mortality were included in meta-analysis. Per 0.1 unit increase in NDVI, the pooled HRs and 95% CIs were 0.99 (0.89–1.09) for CVD mortality, 0.96 (0.88–1.05) for IHD mortality, and 0.77 (0.59–1.00) for stroke mortality. | 67% | Low risk of bias |
| Urbanization | ||||||
| Angkurawaranon et al.44 | Southeast Asian populations, till April 2013, SE Asia countries: Brunei Darussalam, Cambodia, Indonesia, Laos PDR, Malaysia, Myanmar, Philippines, Singapore, Thailand, Timor Leste, and Vietnam | Total 37 studies, relevant 7 studies: 7 CSS | Urban exposure | For urban exposure, the pooled ORs and 95% CIs were 1.01 (0.56–1.82) for stroke, 1.19 (0.35–4.07) for non-specific heart disease in the elderly, 2.48 (1.20–5.11) for CHD, and 0.31 (0.13–0.76) for RHD. | 56% | Unclear |
| Residential noisec | ||||||
| Babisch et al.45 | Not specified (adults), time and country not restricted: identified countries: UK, the Netherlands, Canada, Denmark, Germany, Sweden, and Japan | 5 CS, 4 CCS, 5 CSS | Road traffic noise. L Aeq16hr, L Aeq24hr, L DEN, LDay, LNight | Relative risk per increase of the traffic noise level of 10 dB. For road traffic noise, the pooled OR and 95% CI was 1.08 (1.04–1.13) for CHD. | 71% | High risk of bias |
| Banerjee et al. 2014 | Adult population, 1980–2010, country not restricted: the Netherlands, UK, Germany, Serbia, Sweden, Austria, Italy, Lithuania, Portugal, Switzerland, France, Slovakia, and Hungary | 14 CSS | Transportation noise exposure | (No information on unit) For traffic noise (all sources), the pooled RRs and 95% CIs were 1.04 (0.96–1.12) for CVD, 1.01 (0.89–1.14) for MI, 1.08 (0.80–1.36) for AP, and 1.00 (0.73–1.26) for IHD. The estimates for air traffic noise exposure were 1.00 (0.91–1.09) for CVD, 1.04 (0.80–1.28) for AP, 1.02 (0.89–1.14) for MI, and 0.96 (0.80–1.12) for IHD. The pooled RR for road traffic noise was 1.03 (0.97–1.09) for CVD, 1.23 (0.38–2.09) for AP, 0.85 (−0.58 to 2.29) for MI, and 1.35 (0.78–1.92) for IHD. | 73% | High risk of bias |
| Cai et al. 2021 | Adults, general population, 1 January 2000–5 October 2020, country not restricted: Denmark (Copenhagen and Aarhus), France (Paris, Lyon, and Toulouse), Switzerland, Sweden (Gothenburg), Spain (Barcelona), the Netherlands, UK (London), and Canada (Vancouver) | Total 12 studies, relevant 10 studies: 8 CS, 1 CSS, 2 ES | Residential traffic noise from road, rail, and aircraft, measured or modelled: mostly Lden, LAeq24hr, LAeq16hr, LDN, Lday, Lnight | For road traffic, per 10 dB increase in Lden, the pooled HRs and 95% CIs were 1.01 (0.98–1.05) for CVD mortality, 1.03 (0.99–1.08) for IHD mortality, and 1.05 (0.97–1.14) for stroke mortality. For aircraft traffic, the estimates based on three studies were 1.17 (1.10–1.25) for CVD mortality, 1.03 (0.82–1.29) for IHD mortality, and 1.06 (0.93–1.20) for stroke mortality. For rail traffic, the estimates were 0.98 (0.94–1.01) for CVD mortality (1 study) and 1.02 (0.91–1.14) for IHD mortality (2 studies). | 68% | Unclear |
| Dzhambov et al. 2016 | Adults, till 24 November 2015, country not restricted: the Netherlands, UK, Denmark, Germany, France, Switzerland, USA, Canada, Sweden, Greece, and Italy | 7 CS, 2 CSS, ES 4 | Traffic noise | RR per 10 dB noise increase. For road traffic noise, the pooled RR and 95% CI was 1.03 (0.87–1.22). For air traffic noise, the pooled RR was 1.05 (1.00–1.10). | 72% | High risk of bias |
| Khosravipour et al. 2020 | General population, time and country not restricted: till 29 November 2019, UK, Germany, Sweden, Lithuania, Denmark, and the Netherlands | 7 CS, 5 CCS, 1 CSS | Road traffic noise | Comparing highest to lowest category of noise exposure (results from categorical analysis), the pooled RR and 95% CI of MI were 1.03 (0.93–1.13). Per 10 dB increment (results from exposure–response analysis and transformed from categorical analysis), the pooled estimate was 1.02 (1.00–1.05). In subgroup analysis, pooled estimates were significant for CCS and CSS, but not for CS. Estimates for the exposure–response analyses were 1.03 (1.00–1.05) after excluding two conference papers and 1.02 (1.01–1.03) after further excluding the studies with only results from categorical analysis. | 57% | Low risk of bias |
| van Kempen et al. 2002 | Adults, 1970–1999, country not restricted: Iran, Belgium, Germany, Canada, India, Finland, Italy, the Netherlands, Russia, USA, Poland, Japan, Israel, China, France, South Africa, China (Taiwan), and UK. | Total 43 studies, relevant 10 studies: 6 CSS, 2CCS, 2 CS | Community noise exposure (road and air traffic) assessed by calculations, personal dosimeter, or sound level meter | RR per 5 dB(A) noise increase. For road traffic noise, the pooled RRs and 95% CIs were 1.09 (1.05–1.13) for IHD, 0.99 (0.84–1.16) for AP, and 1.03 (0.99–1.09) for MI. For air traffic noise, the pooled RR was 1.03 (0.90–1.18) for AP. | 25% | High risk of bias |
| van Kempen, et al.46 | European, 2000–October 2014, European countries | Total 61 studies, relevant 32 studies: 14 CSS, 5 ES, 8 CS, 5 CCS | Noise from road, rail, and air traffic and wind turbines: LDEN | Road, rail, and air traffic noise in relation to prevalence, incidence, and mortality of IHD and stroke were analysed, respectively. Number of studies for each analysis is small. Per 10 dB increase in exposure, the pooled RR and 95% CI of IHD was 1.08 (1.01–1.15) for road traffic. Estimates for other associations were of low quality or from <3 studies, and mostly insignificant. | NA | Low risk of bias |
| Vienneau et al.47 | Not specified (general population), January 1994–January 2014, country not restricted: Germany, UK, the Netherlands, Sweden, Switzerland, Denmark, Canada, and USA | 3 CCS, 5 CS, 2 LS | Transportation noise exposure | RR per 10 dB increase in exposure. The pooled RR and 95% CI for IHD was 1.06 (1.03–1.09). | 75% | High risk of bias |
| Weihofen et al.48 | General population, till 31 August 2017, country not restricted: USA, France (Paris, Lyon Saint, and Toulouse), Canada (Vancouver), UK (London), Switzerland, Germany (Berlin and Frankfurt), the Netherlands (Amsterdam), Sweden (Stockholm), Greece (Athens), and Italy (Milan) | 3 CSS, 1 ES, 4 CS, 1 CCS | Aircraft noise: LAeq, LDay, LNight, LDN, Lden, LDENAEI | Per 10 dB increase in Lden, the pooled RR and 95% CI of stroke was 1.013 (0.998–1.028) from seven studies. | 71% | Low risk of bias |
| Ambient temperature | ||||||
| Bunker et al.49 | Elderly (65+), 1 January 1975–24 July 2015, country not restricted: USA, Bangladesh, China (mainland, Taiwan, and Hong Kong), Portugal, UK, Denmark, Australia, Russia, Italy, Hungary, Brazil, Vietnam, Sweden, Thailand, Norway, South Korea, and Germany | Total 60 studies, relevant 47 studies; 47 LS | Ambient hot and cold temperature | Per 1°C temperature change, for heat, the pooled percentage changes and 95% CIs were 3.79 (3.40–4.18) for CVD mortality, 1.62 (0.24–3.03) for IHD mortality, 1.40 (0.06–2.75) for CeVD mortality, 0.33 (−0.09 to 0.75) for IS, −0.66 (−2.13 to 0.84) for ICH, −0.17(−0.96 to 0.63) for CeVD, −0.16(−2.05 to 1.77) for MI, and 0.30(−0.12 to 0.81) for CVD. For cold, the estimates were 1.84 (0.85–2.84) for CVD mortality, 0.45 (−0.01 to 0.91) for IHD mortality, 1.21 (0.66–1.77) for CeVD mortality, 3.63 (−3.94 to 11.8) for IS, 1.49 (1.04–1.94) for ICH, −0.46 (−1.12 to 0.2) for CeVD, 0.66 (−0.14 to 1.48) for MI, −0.80 (−2.21 to 0.64) for AP, −0.67 (−2.15 to 0.83) for HF, and −0.28 (−1.39 to 0.84) for CVD. | 73% | High risk of bias |
| Kofi Amegah et al. 2016 | Sub-Saharan African populations, till December 2014, Sub-Saharan Africa | Total 23 studies, relevant 5 studies: 4 LS, 1 CSS | Temperature | One study found that low temperature was associated with increased risk of CVD. Two studies found associations of low and high temperatures with CVD mortality. One study found no association between mean monthly temperature and CVD mortality. One study found 5°C change in the monthly mean temperature to be associated with decreased risk of hospitalization for venous thromboembolism, stroke, and acute MI. | 56% | High risk of bias |
| Ma et al. 2020 | Chinese population, January 2010–January 2020, country restricted to China | Total 175 studies, relevant 19 studies: 19 LS | (i) Every 1°C temperature increase/decrease beyond certain reference points (ii) Comparison between extreme temperatures and reference normal temperatures | Pooled RRs and 95% CIs of CVD were 1.089 (1.062–1.116) and 1.171 (1.125–1.218), respectively, for hot and cold temperatures as compared with normal temperatures. | 100% | High risk of bias |
| Moghadamnia et al.50 | General population, January 2000–31 December 2015, country not restricted: China (mainland, Taiwan, and Hong Kong), Australia, Thailand, Philippines, South Korea, Germany, and Spain | 26 LS | Ambient temperature | RR per 1°C change of temperature. For CVD mortality, the RRs and 95% CIs were 1.055 (1.050–1.060) for cold exposure and 1.013 (1.011–1.015) for heat exposure. Coefficient per 1°C change in mean annual temperature. For CVD mortality, the pooled estimates were 0.026 (−0.019 to 0.072) for cold exposure and 0.008 (−0.015 to 0.031) for heat exposure. | 96% | High risk of bias |
| Odame et al. 2018 | Rural population, till April 2018, country not restricted: Bangladesh (MATLAB), Czech Republic, and China (Naidong and Jiangzi in Tibet) | All 14 studies, relevant 3 studies: 3 LS | Daily mean temperature | Per 1°C increase, the pooled RR and 95% CI of CVD mortality was 1.111 (1.045–1.181). The associations were significant in subgroup analyses of both developing and developed countries. | 100% | High risk of bias |
| Turner et al. 2012 | Not specified (general population), time and country not restricted: South Korea (Incheon), USA, UK (London and Scotland), Europe, Australia (Brisbane), and Thailand (Muang) | Total 21 studies, relevant 18 studies: 18 LS | Effects of ambient temperature. Maximum, minimum, and mean daily temperature | RR per 1°C increase in temperature. The pooled RRs and 95% CI were 0.999 (0.982–1.016) for CVD morbidities, 0.990 (0.887–1.105) for stroke, and 1.010 (0.930–1.097) for ACS/MI. | 43% | High risk of bias |
| Wang et al. 2016 | Adults, till 16 October 2015, country not restricted: Japan, UK, Russia, Spain, China, Portugal, Italy, South Korea, China (Taiwan), Mozambique, USA, France, and Australia | 7 CS, 13 LS, 1 CSS | Ambient temperature | OR per 1°C increase in mean ambient temperature. The pooled ORs and 95% CIs were 0.97 (0.94–1.00) for ICH, 1.00 (0.99–1.01) for IS, and 1.00 (0.98–1.01) for SAH. The pooled estimates for ambient minimum and maximum temperature and IS were OR 0.99 (0.96–1.01) and 0.98 (0.94–1.02), respectively. | 19% | Unclear |
| Zafeiratou et al. 2021 | General population, 1990–2020, country not restricted: Serbia (Belgrade), China (Hong Kong), UK (London), USA (New England), Finland (Helsinki), Switzerland, and South Korea | Total 34 studies, relevant 6 studies: 4 ES, 2 CS | Mean annual temperature or variability, seasonal temperature or variability, annual temperature categories, and mean annual degrees above/below minimum mortality temperature | In temporal comparisons within the same area, increased cardiovascular mortality was associated with both increased and decreased temperature. Stronger association was found with cold rather than hot temperature. In geographical comparison in just one study, people living in areas with higher temperature were found a lower rate of IHD mortality, though no dose–response. | 5 studies on CVD mortality: 60% | High risk of bias |
| Multiple domains | ||||||
| Rugel et al. 2020 | Urban residents, 1 January 2003–November 2019, country not restricted: the Netherlands, Sweden (Skåne), UK (London), Canada, Denmark (Copenhagen and Aarhus), Germany (Bochum, Essen, and Mülheim/Ruhr), South Korea (Seoul, Ulsan, and Cheonan), France, Spain, Norway, Greece, Italy (Rome and Verona), Switzerland, USA (California), and China (Shenyang, Anshan, Jinzhou, and Taiwan) | Total 51 studies, relevant 21 studies: 15 CS, 5 CSS, 1 ES | Traffic-related air pollution; natural spaces, neighbourhood walkability; noise | Based on the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) system for assessing the quality of the synthesized evidence base, evidence was sufficient for higher noise exposure with increased CVD. And evidence was sufficient for no association between traffic-related air pollution and CVD. | NA | High risk of bias |
| First author and year by domaina . | Population, year coverage, and country/region . | Study design . | Exposuresb . | Summary of results . | Percentage of results in expected direction . | Quality assessment (ROBIS) . |
|---|---|---|---|---|---|---|
| Air pollution | ||||||
| Alexeeff et al.34 | General population, till 31 December 2019, country not restricted: Europe, Canada, USA, UK, Australia, China, South Korea, and Israel | 42 CS | PM2.5 | Per 10 μg/m3 increase in long-term exposure, the pooled RRs and 95% CIs were 1.23 (1.15–1.31) for IHD mortality, 1.08 (0.99–1.18) for incident acute MI, 1.24 (1.13–1.36) for CeVD mortality, and 1.13 (1.11–1.15) for incident stroke. | 83% | High risk of bias |
| Atkinson et al. 2016 | General population, till October 2015, country not restricted: USA, UK, China (Taiwan), and France | Total 22 studies, relevant 8 studies: 8 CS | O3 | HRs expressed per 10 ppb increase in O3. For long-term annual O3 concentrations, the standardized effect estimates (HRs and 95% CIs) were 1.01 (0.99–1.03) for CVD mortality, 1.02 (1.00–1.04) for IHD mortality, and 1.01 (0.97–1.05) for stroke mortality. For long-term annual O3 concentrations, the random-effects summary estimates (HR and 95% CI) were 0.98 (0.93–1.04) for CVD mortality and 1.00 (0.92–1.09) for IHD mortality. For the warm season/peak O3, random-effects summary estimates were 1.01 (1.00 to 1.02). | 42% | High risk of bias |
| Atkinson et al.35 | General population, 1996–October 2016 (Medline, EMBASE), 1970–October 2016 (Web of Science), 1966–October 2016 (PubMed), country not restricted: Europe, North America, China (mainland and Taiwan), and Japan | Total 48 studies, relevant 22 studies: 22 CS | NO2 (annual or multi-year averages) | Per 10 µg/m3 increase in long-term exposure, the pooled HRs and 95% CIs were 1.03 (1.02–1.05) for CVD mortality, 1.05 (1.03–1.06) for CHD mortality, and 1.01 (0.98–1.03) for CeVD mortality. | 91% | High risk of bias |
| Chen et al. 2008 | Adults, 1 January 1950–31 December 2007, country not restricted: USA, Norway, France, the Netherlands, Canada, and Germany | Total 32 studies, relevant 17 studies: 14 CS, 3 CCS | O3, SO2, NO/NO2, black smoke, PM10, PM2.5, CO, benzene, and polycyclic aromatic hydrocarbons | RR per 10 µg/m3 increase. For PM2.5, the pooled RRs and 95% CIs were 1.14 (1.09–1.18) for CVD mortality and 1.16 (0.96–1.40) for CHD mortality. For other particulate and gaseous pollutants, the paucity of data precludes drawing conclusions. | 79% | High risk of bias |
| Chen et al. 2020 | General population, till 9 October 2018, country not restricted: Europe, Canada, UK, USA, Israel, New Zealand, South Korea, Japan, and China (mainland, Taiwan, and Hong Kong) | 67 CS | PM2.5 and PM10 | Per 10 µg/m3 increase in long-term PM2.5 exposure, the pooled RRs and 95% CIs were 1.11 (1.09–1.14) for CVD mortality, 1.16 (1.10–1.21) for IHD mortality, and 1.11 (1.04–1.18) for stroke mortality. The estimates of PM10 were 1.04 (0.99–1.10) for CVD mortality, 1.06 (1.01–1.10) for IHD mortality, and 1.01 (0.83–1.21) for stroke mortality. The certainty of evidence was high for PM2.5 and CVD mortality and was moderate for PM10 and CVD mortality, as measured by GRADE framework. | 81% | Low risk of bias |
| Chen et al.36 | General population, till October 2019, country not restricted: South Korea, UK, Denmark, Sweden, and Canada | Total 18 studies, relevant 6 studies: 6 CS | PM2.5, PM10, NO2, SO2, O3, and CO | Per 10 μg/m3 increase in long-term exposure, the pooled HRs, and 95% CIs of AF were 1.116 (1.031–1.207) for PM2.5, 1.034 (1.032–1.035) for PM10, 1.017 (1.001–1.033) for NO2, 1.005 (1.004–1.007) for SO2, 1.017 (1.013–1.022) for CO, and 1.007 (0.927–1.094) for O3. | 83% | Low risk of bias |
| Faustini et al. 2014 | General population, January 2004–January 2013, country not restricted: identified countries: Japan, China, Canada, USA, UK, Italy, Germany, Sweden, the Netherlands, and Norway | Total 23 studies, relevant 17 studies: 2 CCS, 15 CS | NO2 | RR per 10 µg/m3 increase. The pooled RRs and 95% CIs for CVD mortality were 1.13 (1.09–1.18) for NO2 and 1.20 (1.09–1.31) for PM2.5. | 94% | High risk of bias |
| Hak-Kan et al. 2013 | Chinese population, till 30 June 2012, 80 major Chinese cities in Mainland China, Hong Kong, and Taiwan | Total 48 studies, relevant 3 studies: 3 CS | PM10, NO2, SO2, and O3 | RR per 10 µg/m3 increase. In one cohort study examining PM10 and NO2, the corresponding RRs and 95% CIs were 1.0155 (1.0151–1.0160) and 1.0246 (1.0231–1.0263) for CVD mortality and 1.0149 (1.0145–1.0153) and 1.0244 (1.0227–1.0262) for CeVD mortality. In another cohort study examining SO2 and CVD, the corresponding RR was 1.032 (1.023–1.040). | 100% | High risk of bias |
| Hoek et al. 2013 | Not specified (adults), till January 2013, country not restricted: identified countries: USA, Germany, the Netherlands, Switzerland, Canada, China, New Zealand, Japan, Italy, France, and Denmark | Total 67 studies, relevant 34 studies: 34 CS | Long-term exposure to fine particulate matter (PM2.5, PM10, NO2, elemental carbon, and coarse particles) | RR per 10 µg/m3 increase. For PM2.5, the pooled RR and 95% CI was 1.11 (1.05–1.16) for CVD mortality. There was no consistent evidence that long-term exposure to coarse PM or elemental carbon is associated with CVD mortality. Several studies found positive associations between NO2 exposure and fatal MI, but not non-fatal MI. The evidence for an association between air pollution and CeVD mortality was inconsistent. | 88% | High risk of bias |
| Huang et al. 2021 | General population, till 2020.02.29, country not restricted: USA, Canada, Norway, the Netherlands, UK, Italy, Denmark, France, Spain, Japan, China, South Korea, Australia, Sweden, Norway, Germany, Austria, Switzerland, France, Italy, Spain, Greece, and Finland | 32 CS | NO2 | Per 10 ppb increase in annual NO2 concentration, the pooled HR and 95% CI was 1.11 (1.07–1.16) for cardiovascular mortality. | 71% | High risk of bias |
| Jadambaa et al.37 | Mongolian population (adults and children), till April 2014, Mongolia | Total 59 studies, relevant 2 studies: 2 CSS | NO2 and PM2.5 | Two studies found an increased risk of CVD with increased exposure to NO2 and PM2.5. | 100% | Unclear |
| Jaganathan et al.38 | General population, 1 January 1948–6 March 2018, country restricted to low- and middle-income countries, Mexico (Mexico City), Brazil (São José dos Campos, Cuiabá, and Várzea Grande), China, and India (Varanasi) | Total 17 studies, relevant 12 studies: 8 LS, 2 CSS, 1 CCR, 1 CS | PM2.5 (annual average or average measure of more than 3 days) | Eight out of nine studies (91%) reported significant effects on CVD mortality. Per 10 µg/m3 increase in long-term exposure, the effect estimates of CVD mortality ranged from 0.24 to 6.11%. All four studies reported significant effects of long-term exposure on CVD hospitalization. Few studies have evaluated this association in LMICs. No studies were found in North and Sub-Saharan Africa. | 92% | Low risk of bias |
| Kan et al. 2005 | General population, 1990–2002, China and worldwide | Total 26 studies, relevant 7 studies: 7 CS | Effects of particulate air pollution. PM10 was selected as the indicator particulate matter. | RR per 10 µg/m3 increase. For PM10, the pooled RRs and 95% CIs were 1.0095 (1.0060–1.0130) for CHA, 1.013 (1.007–1.019) for CHA based on four European studies, and 1.008 (1.004–1.011) for CHA based on three US and Canadian studies. | 100% | High risk of bias |
| Karimi et al. 2019 | Iran population, January 1980–January 2018, country restricted to Iran | Total 38 studies, relevant 28 studies: 27 CSS, 1 ES | O3, PM2.5, PM10, NO2, NOx, SO2, and CO measured by environmental protection organization and air quality control centre | Per 10 µg/m3 increase in all air pollutants, the pooled increased risk (95% CI) in CVD mortality was 0.5% (0.4–0.6%). The estimate for PM2.5 and PM10 was 0.7% (0.4–1%). | NA | Unclear |
| Liu et al. 2018 | General population, adults, January 1974–July 2017, country not restricted: USA, UK, Italy, Canada, China (mainland and Hong Kong), Europe, New Zealand, and Japan | 16 CS | PM2.5 and PM10 | Per 10 μg/m3 increase in long-term exposure, the pooled HRs and 95% CIs of CVD mortality were 1.12 (1.08–1.16) for PM2.5, 1.02 (0.89–1.16) for PM10, and 1.10 (1.06–1.14) for combined. In subgroup analyses, there is no difference in the association stratified by categories of WHO PM levels or smoking status. The estimates of PM2.5 were 1.19 (1.11–1.27) for studies with ≥ 11 years of follow-up, higher than those <11 years: 1.07 (1.04–1.11). | 88% | Low risk of bias |
| Lu et al. 2015 | Chinese population (adults only), 1990–2013, Mainland China, Hong Kong, and Taiwan | Total 59 studies, relevant 2 studies: 2 CS | PM10 and PM2.5 | RR per 10 μg/m3 increase. For the annual average concentration of PM10, the RR and 95% CI was 1.23 (1.19–1.26) for CVD mortality in one study and 1.55 (1.51–1.60) for CVD mortality in another study. | 100% | High risk of bias |
| Luben et al. 2017 | Adults, till 15 June 2017, country not restricted: USA, China (mainland and Taiwan), the Netherlands, Canada, South Korea, Spain, and Italy | Total 24 studies, relevant, 3 studies: 2 CS, 1 LS | Ambient black carbon | There are generally modest, positive associations of long-term exposure to black carbon and elemental carbon with cardiovascular hospital admissions and mortality. | 100% | High risk of bias |
| Niu et al. 2021 | General population, till 1 February 2020, country not restricted: China, Europe, England, Japan (Shizuoka), USA (California), Ghana, India, Mexico, Russia, and South Africa | Total 68 studies, relevant 13 studies: 13 CS | PM2.5, PM10, and NO2 | Per 10 μg/m3 increase in long-term exposure, the pooled HRs and 95% CIs of stroke incidence were 1.081 (0.971–1.023) for PM2.5, 1.033 (0.907–1.175) for PM10, and 1.005 (0.977–1.034) for NO2; the HRs and 95%CI of stroke mortality were 1.047 (0.995–1.101) for NO2. | 82% | High risk of bias |
| Prueitt et al.39 | General population, 1 January 2006–4 November 2013, country not restricted: USA, UK, Canada (Toronto), and China (Liaoning) | Total 25 studies, relevant 11 studies: 8 CS, 2 CSS, 1 ES | O3 | For long-term O3 exposure and CVD morbidity, studies were rare and reports were inconsistent. For CVD mortality, of 10 high-quality studies, 5 reported positive association, and the other 5 reported null or negative association. | 17% | High risk of bias |
| Scheers et al. 2015 | General population, till 20 July 2015, country not restricted: Japan, China, UK, the Netherlands, Switzerland, Greece, USA, Canada, Finland, Norway, Sweden, Denmark, Germany, Austria, Italy, Greece, and France | Total 20 studies, relevant 20 studies: 14 CS, 6 ES | PM10 or PM2.5 | HR per 10 μg/m3 increase. For PM10, the pooled HRs and 95% CIs were 1.061 (1.018–1.105) for overall stroke events and 1.080 (0.992–1.177) for stroke mortality. For PM2.5, the pooled HRs and 95% CIs were 1.064 (1.021–1.109) for overall stroke events and 1.125 (1.007–1.256) for stroke mortality. | 50% | High risk of bias |
| Shin et al. 2014 | Not specified (adults), from 1990, country not restricted: USA and UK | Total 20 studies, relevant 4 studies: 4 CS | PM2.5 | RR per 10 μg/m3 increase. In the frequentist meta-analysis, the pooled RR and 95% CI for long-term exposure to PM2.5 was 1.06 (1.00–1.13) for non-fatal strokes. The Bayesian meta-analysis found a posterior mean 1.08 (0.96–1.26) from a normal prior and 1.05 (1.02–1.10) from a gamma prior. | 100% | High risk of bias |
| Stieb et al.40 | General population, till 25 February 2020, country not restricted: Canada, USA, UK, Europe, China (mainland, Hong Kong, and Taipei), Australia, South Korea (Seoul), and Japan (Shizuoka) | 49 CS | NO2 | Per 10 ppb increase in long-term exposure, the pooled HRs and 95% CIs were 1.139 (0.997–1.301) for CVD mortality, 1.128 (1.076–1.182) for IHD mortality, and 1.167 (0.936–1.456) for CeVD mortality. After excluding studies with probably high or high risk of bias, the pooled HRs and 95% CIs were 1.058 (1.026–1.091) for CVD mortality, 1.111 (1.079–1.144) for IHD mortality, and 1.014 (0.997–1.032) for CeVD mortality. | 74% | Low risk of bias |
| Wang et al. 2020 | Older adults aged ≥ 55 years, till January 2020, country not restricted: USA (Steubenville, Eastern Massachusetts, Boston), Germany (Erfurt), Finland (Helsinki), the Netherlands (Amsterdam), UK (Scotland: Aberdeen), and China (Beijing and Taipei) | Total 19 studies, relevant 10 studies: 10 LS | Concentration of PM2.5 | Per 10 mg/m3 increase in long-term exposure, the pooled estimates and 95% CIs of HRV were −0.92% (−2.14 to 0.31%) for SDNN, −1.96% (−3.48 to −0.44%) for RMSSD in time-domain measurements, −2.78% (−4.02 to −1.55%) for LF, and −1.61% (−4.02 to 0.80%) for HF in frequency domain measurements. | 68% | High risk of bias |
| Yang et al. 201941 | General population, till 25 April 2018, country not restricted: Europe, UK, Canada, USA, South Korea, China, Ghana, India, Mexico, Russia, South Africa, and Japan | 35 CS | PM2.5, PM10, O3, and NO2 | Per 10 μg/m3 increase in long-term PM2.5 exposure, the pooled RRs and 95% CIs were 1.11 (1.07–1.15) for CVD events, 1.12 (1.05–1.19) for stroke incidence, 1.12 (1.08–1.16) for stroke events, 1.19 (1.09–1.30) for IHD incidence, and 1.14 (1.08–1.21) for IHD events. The estimates of CVD mortality were 1.11 (1.07–1.15) for PM2.5, 1.09 (1.02–1.16) for PM10, 1.23 (1.15–1.31) for NO2, and 1.03 (1.02–1.05) for O3. The estimates of NO2 and IHD events were 1.05 (1.04–1.06). No significant associations were found between PM10 and CVD, stroke and IHD incidence. | 87% | High risk of bias |
| Yuan et al. 2019 | General population, 1980–December 2018, country not restricted: Europe, USA, China (Hong Kong), Ghana, India, Mexico, Russia, South Africa, UK, Sweden (Gothenburg), and Italy | 16 CS | PM2.5 | Per 5 μg/m3 increase in long-term exposure, the pooled HRs and 95% CIs were 1.11 (1.05–1.17) for stroke incidence and 1.11 (1.05–1.17) for stroke mortality. In subgroup analysis, the estimates of stroke incidence were 1.09 (1.05–1.14) for North America (5 CS), 1.07 (1.05–1.10) for Europe (4 CS), and 2.31 (0.49–10.95) for Asia (2 CS). The associations were insignificant in both sex and significant in both ischaemic and haemorrhagic stroke. The estimates of stroke incidence were 1.08 (1.03–1.13) for never smokers, 1.11 (1.01–1.22) for former smokers, and 1.08 (0.94–1.25) for current smokers. | 95% | Low risk of bias |
| Zhao et al. 2017 | General population, 1990–2016, country not restricted: USA, Israel, Japan, UK, China, Italy, Norway, Greece, Canada, Denmark, France, South Korea, Iran, Germany, Finland, Sweden, Spain, and the Netherlands | Total 48 studies, relevant 48 studies: 25 CS, 23 LS | PM10, PM2.5, SO2, NO2, CO, and O3 | HR per 10 μg/m3 increase. For CHD mortality, the pooled HRs and 95% CIs were 1.12 (1.04–1.20) for PM10, 1.17 (1.12–1.22) for PM2.5, 1.03 (1.00–1.07) for SO2, 1.04 (1.01–1.06) for NO2, 1.04 (0.98–1.10) for CO, and 1.06 (1.01–1.11) for O3 (10 mg/m3 increase). For CHD incidence, the pooled HRs and 95% CIs were 1.01 (1.00–1.02) for PM10, 1.02 (1.00–1.03) for PM2.5, 1.01 (1.00–1.02) for SO2, 1.04 (1.03–1.06) for NO2, 1.01 (0.97–1.04) for O3, and 1.03 (1.00–1.05) for CO (10 mg/m3 increase). | NA | High risk of bias |
| Zhao et al. 2021 | General population, time and country not restricted: China, Norway, UK, the Netherlands, China (Hong Kong), and Canada (Ontario) | 7 CS | PM2.5 acquired through satellite-based model (5 studies) and outdoor-automated monitoring stations (2 studies) | Per 1.4–10 μg/m3 increase in long-term PM2.5 exposure, the pooled HRs and 95% CIs of haemorrhagic stroke were 1.16 (1.03–1.30) for total, 1.41 (0.92–2.15) for current smoker, and 1.04 (0.74–1.46) for never and former smoker. | 71% | Low risk of bias |
| Zhu et al. 2021 | General population, till 2 August 2020, country not restricted: Canada, Denmark, the Netherlands, China, USA, South Korea, Israel, and UK (London) | 12 CS | PM2.5 | Per 10 μg/m3 increase in long-term PM2.5 exposure, the pooled HRs and 95% CIs were 1.10 (1.02–1.18) for MI incidence and 1.07 (1.04–1.09) for post-MI mortality. | 75% | Unclear |
| Zou et al.42 | General population, till September 2019, country not restricted: USA, South Korea, UK, Canada, Sweden, Israel, Italy, the Netherlands, Switzerland, and Finland | 27 CS | PM2.5 and PM10 | Per 10 μg/m3 increase in long-term exposure, the pooled RRs and 95% CIs of MI were 1.18 (1.11–1.26) for PM2.5 and 1.03 (1.00–1.05) for PM10. | 91% | Unclear |
| Physical activity environment | ||||||
| Gascon et al. 2016 | Adults, till 14 November 2014, country not restricted: USA, UK, New Zealand, Lithuania, and Canada | Total 12 studies, relevant 8 studies; 4 ES, 2 CS, 1 CSS | Residential natural outdoor environments, particularly green and blue spaces | For each 10% increase of greenness, the RR and 95% CI was 0.993 (0.985–1.001) for CVD mortality. For high vs. low categories of greenness, the RR and 95% CI was 0.96 (0.94–0.97) for CVD mortality. | 75% | Unclear |
| Twohig-Bennett et al.43 | General population, till January 2017, country not restricted: USA, UK, and Lithuania | Total 143 studies, relevant 4 studies: 3 CS, 1 ES | Greenspace measured by residential NVDI, distance to the nearest greenspace, and proportion of city area covered by green land | Comparing higher to lower greenspace exposure, the pooled ORs and 95% CIs were 0.82 (0.61–1.11) for stroke (3 studies), 0.84 (0.76–0.93) for CVD mortality (2 studies), and 0.92 (0.78–1.07) for CHD (2 studies). | 86% | High risk of bias |
| Yuan et al. 2020 | Older adults (mostly ≥ 60 years), 1 January 2000–1 July 2020, country not restricted: Japan, Canada, USA, Finland, China, Rome, Australia, the Netherlands, Lithuania, Brazil, Israel, South Korea, Iran, and UK | Total 22 studies, relevant 17 studies:12 CS, 5 CSS | Greenspace measured by NDVI (mostly), percent of greenspace coverage, distance to the nearest green space, park visitation and length of stay, and loss of trees from emerald ash bore disease | Of 8 studies in total CVD, 7 found beneficial effects of green space, and the other study showed a lower risk of CVD with higher percentage of tree canopy, but not total green space. Evidence for stroke and MI was less consistent. Only cohort studies measuring NDVI and mortality were included in meta-analysis. Per 0.1 unit increase in NDVI, the pooled HRs and 95% CIs were 0.99 (0.89–1.09) for CVD mortality, 0.96 (0.88–1.05) for IHD mortality, and 0.77 (0.59–1.00) for stroke mortality. | 67% | Low risk of bias |
| Urbanization | ||||||
| Angkurawaranon et al.44 | Southeast Asian populations, till April 2013, SE Asia countries: Brunei Darussalam, Cambodia, Indonesia, Laos PDR, Malaysia, Myanmar, Philippines, Singapore, Thailand, Timor Leste, and Vietnam | Total 37 studies, relevant 7 studies: 7 CSS | Urban exposure | For urban exposure, the pooled ORs and 95% CIs were 1.01 (0.56–1.82) for stroke, 1.19 (0.35–4.07) for non-specific heart disease in the elderly, 2.48 (1.20–5.11) for CHD, and 0.31 (0.13–0.76) for RHD. | 56% | Unclear |
| Residential noisec | ||||||
| Babisch et al.45 | Not specified (adults), time and country not restricted: identified countries: UK, the Netherlands, Canada, Denmark, Germany, Sweden, and Japan | 5 CS, 4 CCS, 5 CSS | Road traffic noise. L Aeq16hr, L Aeq24hr, L DEN, LDay, LNight | Relative risk per increase of the traffic noise level of 10 dB. For road traffic noise, the pooled OR and 95% CI was 1.08 (1.04–1.13) for CHD. | 71% | High risk of bias |
| Banerjee et al. 2014 | Adult population, 1980–2010, country not restricted: the Netherlands, UK, Germany, Serbia, Sweden, Austria, Italy, Lithuania, Portugal, Switzerland, France, Slovakia, and Hungary | 14 CSS | Transportation noise exposure | (No information on unit) For traffic noise (all sources), the pooled RRs and 95% CIs were 1.04 (0.96–1.12) for CVD, 1.01 (0.89–1.14) for MI, 1.08 (0.80–1.36) for AP, and 1.00 (0.73–1.26) for IHD. The estimates for air traffic noise exposure were 1.00 (0.91–1.09) for CVD, 1.04 (0.80–1.28) for AP, 1.02 (0.89–1.14) for MI, and 0.96 (0.80–1.12) for IHD. The pooled RR for road traffic noise was 1.03 (0.97–1.09) for CVD, 1.23 (0.38–2.09) for AP, 0.85 (−0.58 to 2.29) for MI, and 1.35 (0.78–1.92) for IHD. | 73% | High risk of bias |
| Cai et al. 2021 | Adults, general population, 1 January 2000–5 October 2020, country not restricted: Denmark (Copenhagen and Aarhus), France (Paris, Lyon, and Toulouse), Switzerland, Sweden (Gothenburg), Spain (Barcelona), the Netherlands, UK (London), and Canada (Vancouver) | Total 12 studies, relevant 10 studies: 8 CS, 1 CSS, 2 ES | Residential traffic noise from road, rail, and aircraft, measured or modelled: mostly Lden, LAeq24hr, LAeq16hr, LDN, Lday, Lnight | For road traffic, per 10 dB increase in Lden, the pooled HRs and 95% CIs were 1.01 (0.98–1.05) for CVD mortality, 1.03 (0.99–1.08) for IHD mortality, and 1.05 (0.97–1.14) for stroke mortality. For aircraft traffic, the estimates based on three studies were 1.17 (1.10–1.25) for CVD mortality, 1.03 (0.82–1.29) for IHD mortality, and 1.06 (0.93–1.20) for stroke mortality. For rail traffic, the estimates were 0.98 (0.94–1.01) for CVD mortality (1 study) and 1.02 (0.91–1.14) for IHD mortality (2 studies). | 68% | Unclear |
| Dzhambov et al. 2016 | Adults, till 24 November 2015, country not restricted: the Netherlands, UK, Denmark, Germany, France, Switzerland, USA, Canada, Sweden, Greece, and Italy | 7 CS, 2 CSS, ES 4 | Traffic noise | RR per 10 dB noise increase. For road traffic noise, the pooled RR and 95% CI was 1.03 (0.87–1.22). For air traffic noise, the pooled RR was 1.05 (1.00–1.10). | 72% | High risk of bias |
| Khosravipour et al. 2020 | General population, time and country not restricted: till 29 November 2019, UK, Germany, Sweden, Lithuania, Denmark, and the Netherlands | 7 CS, 5 CCS, 1 CSS | Road traffic noise | Comparing highest to lowest category of noise exposure (results from categorical analysis), the pooled RR and 95% CI of MI were 1.03 (0.93–1.13). Per 10 dB increment (results from exposure–response analysis and transformed from categorical analysis), the pooled estimate was 1.02 (1.00–1.05). In subgroup analysis, pooled estimates were significant for CCS and CSS, but not for CS. Estimates for the exposure–response analyses were 1.03 (1.00–1.05) after excluding two conference papers and 1.02 (1.01–1.03) after further excluding the studies with only results from categorical analysis. | 57% | Low risk of bias |
| van Kempen et al. 2002 | Adults, 1970–1999, country not restricted: Iran, Belgium, Germany, Canada, India, Finland, Italy, the Netherlands, Russia, USA, Poland, Japan, Israel, China, France, South Africa, China (Taiwan), and UK. | Total 43 studies, relevant 10 studies: 6 CSS, 2CCS, 2 CS | Community noise exposure (road and air traffic) assessed by calculations, personal dosimeter, or sound level meter | RR per 5 dB(A) noise increase. For road traffic noise, the pooled RRs and 95% CIs were 1.09 (1.05–1.13) for IHD, 0.99 (0.84–1.16) for AP, and 1.03 (0.99–1.09) for MI. For air traffic noise, the pooled RR was 1.03 (0.90–1.18) for AP. | 25% | High risk of bias |
| van Kempen, et al.46 | European, 2000–October 2014, European countries | Total 61 studies, relevant 32 studies: 14 CSS, 5 ES, 8 CS, 5 CCS | Noise from road, rail, and air traffic and wind turbines: LDEN | Road, rail, and air traffic noise in relation to prevalence, incidence, and mortality of IHD and stroke were analysed, respectively. Number of studies for each analysis is small. Per 10 dB increase in exposure, the pooled RR and 95% CI of IHD was 1.08 (1.01–1.15) for road traffic. Estimates for other associations were of low quality or from <3 studies, and mostly insignificant. | NA | Low risk of bias |
| Vienneau et al.47 | Not specified (general population), January 1994–January 2014, country not restricted: Germany, UK, the Netherlands, Sweden, Switzerland, Denmark, Canada, and USA | 3 CCS, 5 CS, 2 LS | Transportation noise exposure | RR per 10 dB increase in exposure. The pooled RR and 95% CI for IHD was 1.06 (1.03–1.09). | 75% | High risk of bias |
| Weihofen et al.48 | General population, till 31 August 2017, country not restricted: USA, France (Paris, Lyon Saint, and Toulouse), Canada (Vancouver), UK (London), Switzerland, Germany (Berlin and Frankfurt), the Netherlands (Amsterdam), Sweden (Stockholm), Greece (Athens), and Italy (Milan) | 3 CSS, 1 ES, 4 CS, 1 CCS | Aircraft noise: LAeq, LDay, LNight, LDN, Lden, LDENAEI | Per 10 dB increase in Lden, the pooled RR and 95% CI of stroke was 1.013 (0.998–1.028) from seven studies. | 71% | Low risk of bias |
| Ambient temperature | ||||||
| Bunker et al.49 | Elderly (65+), 1 January 1975–24 July 2015, country not restricted: USA, Bangladesh, China (mainland, Taiwan, and Hong Kong), Portugal, UK, Denmark, Australia, Russia, Italy, Hungary, Brazil, Vietnam, Sweden, Thailand, Norway, South Korea, and Germany | Total 60 studies, relevant 47 studies; 47 LS | Ambient hot and cold temperature | Per 1°C temperature change, for heat, the pooled percentage changes and 95% CIs were 3.79 (3.40–4.18) for CVD mortality, 1.62 (0.24–3.03) for IHD mortality, 1.40 (0.06–2.75) for CeVD mortality, 0.33 (−0.09 to 0.75) for IS, −0.66 (−2.13 to 0.84) for ICH, −0.17(−0.96 to 0.63) for CeVD, −0.16(−2.05 to 1.77) for MI, and 0.30(−0.12 to 0.81) for CVD. For cold, the estimates were 1.84 (0.85–2.84) for CVD mortality, 0.45 (−0.01 to 0.91) for IHD mortality, 1.21 (0.66–1.77) for CeVD mortality, 3.63 (−3.94 to 11.8) for IS, 1.49 (1.04–1.94) for ICH, −0.46 (−1.12 to 0.2) for CeVD, 0.66 (−0.14 to 1.48) for MI, −0.80 (−2.21 to 0.64) for AP, −0.67 (−2.15 to 0.83) for HF, and −0.28 (−1.39 to 0.84) for CVD. | 73% | High risk of bias |
| Kofi Amegah et al. 2016 | Sub-Saharan African populations, till December 2014, Sub-Saharan Africa | Total 23 studies, relevant 5 studies: 4 LS, 1 CSS | Temperature | One study found that low temperature was associated with increased risk of CVD. Two studies found associations of low and high temperatures with CVD mortality. One study found no association between mean monthly temperature and CVD mortality. One study found 5°C change in the monthly mean temperature to be associated with decreased risk of hospitalization for venous thromboembolism, stroke, and acute MI. | 56% | High risk of bias |
| Ma et al. 2020 | Chinese population, January 2010–January 2020, country restricted to China | Total 175 studies, relevant 19 studies: 19 LS | (i) Every 1°C temperature increase/decrease beyond certain reference points (ii) Comparison between extreme temperatures and reference normal temperatures | Pooled RRs and 95% CIs of CVD were 1.089 (1.062–1.116) and 1.171 (1.125–1.218), respectively, for hot and cold temperatures as compared with normal temperatures. | 100% | High risk of bias |
| Moghadamnia et al.50 | General population, January 2000–31 December 2015, country not restricted: China (mainland, Taiwan, and Hong Kong), Australia, Thailand, Philippines, South Korea, Germany, and Spain | 26 LS | Ambient temperature | RR per 1°C change of temperature. For CVD mortality, the RRs and 95% CIs were 1.055 (1.050–1.060) for cold exposure and 1.013 (1.011–1.015) for heat exposure. Coefficient per 1°C change in mean annual temperature. For CVD mortality, the pooled estimates were 0.026 (−0.019 to 0.072) for cold exposure and 0.008 (−0.015 to 0.031) for heat exposure. | 96% | High risk of bias |
| Odame et al. 2018 | Rural population, till April 2018, country not restricted: Bangladesh (MATLAB), Czech Republic, and China (Naidong and Jiangzi in Tibet) | All 14 studies, relevant 3 studies: 3 LS | Daily mean temperature | Per 1°C increase, the pooled RR and 95% CI of CVD mortality was 1.111 (1.045–1.181). The associations were significant in subgroup analyses of both developing and developed countries. | 100% | High risk of bias |
| Turner et al. 2012 | Not specified (general population), time and country not restricted: South Korea (Incheon), USA, UK (London and Scotland), Europe, Australia (Brisbane), and Thailand (Muang) | Total 21 studies, relevant 18 studies: 18 LS | Effects of ambient temperature. Maximum, minimum, and mean daily temperature | RR per 1°C increase in temperature. The pooled RRs and 95% CI were 0.999 (0.982–1.016) for CVD morbidities, 0.990 (0.887–1.105) for stroke, and 1.010 (0.930–1.097) for ACS/MI. | 43% | High risk of bias |
| Wang et al. 2016 | Adults, till 16 October 2015, country not restricted: Japan, UK, Russia, Spain, China, Portugal, Italy, South Korea, China (Taiwan), Mozambique, USA, France, and Australia | 7 CS, 13 LS, 1 CSS | Ambient temperature | OR per 1°C increase in mean ambient temperature. The pooled ORs and 95% CIs were 0.97 (0.94–1.00) for ICH, 1.00 (0.99–1.01) for IS, and 1.00 (0.98–1.01) for SAH. The pooled estimates for ambient minimum and maximum temperature and IS were OR 0.99 (0.96–1.01) and 0.98 (0.94–1.02), respectively. | 19% | Unclear |
| Zafeiratou et al. 2021 | General population, 1990–2020, country not restricted: Serbia (Belgrade), China (Hong Kong), UK (London), USA (New England), Finland (Helsinki), Switzerland, and South Korea | Total 34 studies, relevant 6 studies: 4 ES, 2 CS | Mean annual temperature or variability, seasonal temperature or variability, annual temperature categories, and mean annual degrees above/below minimum mortality temperature | In temporal comparisons within the same area, increased cardiovascular mortality was associated with both increased and decreased temperature. Stronger association was found with cold rather than hot temperature. In geographical comparison in just one study, people living in areas with higher temperature were found a lower rate of IHD mortality, though no dose–response. | 5 studies on CVD mortality: 60% | High risk of bias |
| Multiple domains | ||||||
| Rugel et al. 2020 | Urban residents, 1 January 2003–November 2019, country not restricted: the Netherlands, Sweden (Skåne), UK (London), Canada, Denmark (Copenhagen and Aarhus), Germany (Bochum, Essen, and Mülheim/Ruhr), South Korea (Seoul, Ulsan, and Cheonan), France, Spain, Norway, Greece, Italy (Rome and Verona), Switzerland, USA (California), and China (Shenyang, Anshan, Jinzhou, and Taiwan) | Total 51 studies, relevant 21 studies: 15 CS, 5 CSS, 1 ES | Traffic-related air pollution; natural spaces, neighbourhood walkability; noise | Based on the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) system for assessing the quality of the synthesized evidence base, evidence was sufficient for higher noise exposure with increased CVD. And evidence was sufficient for no association between traffic-related air pollution and CVD. | NA | High risk of bias |
ES, ecological studies; CS, cohort studies; CSS, cross-sectional studies; CCS, case-control studies; CCR, case-crossover studies; SAS, small-area studies; longitudinal study, LS (e.g. panel study and time-series); RR, relative risk; HR, hazard ratio; MI, myocardial infarction; AP, angina pectoris; CHA, cardiovascular hospital admission; ACS, acute coronary syndrome; CVD, cardiovascular disease; CHD, coronary heart disease; AF, atrial fibrillation; IHD, ischaemic heart disease; CeVD, cerebrovascular disease; RHD, rheumatic heart disease; HRV, heart rate variability; SDNN, the standard deviation of NN intervals; RMSSD, root mean square of successive RR interval differences; LF, low-frequency bands; HF, high-frequency bands; RE, risk estimate; events include both morbidity and mortality; ICH, intracerebral haemorrhage; IS, ischaemic stroke; SAH, subarachnoid haemorrhage; HF, heart failure. The significant pooled results are in bold.
A bibliography of all included reviews is presented in Supplementary material online, Appendix S3.
There were no systematic reviews found in the domains of food environment and light pollution.
LAeq16hr, annual non-weighted 16 h average noise level during the day; LAeq24hr, the annual non-weighted day–night average noise level; LDEN, the annual weighted (day + 0 dB, evening + 5 dB, and night + 10 dB) day–evening–night average noise level; LAeq, A-weighted average of an energy-equivalent continuous sound level over a period of time (A-weighting: in noise research, typically, the A filter is used which adjusts for deep and high frequencies, as these are perceived as less loud); LDay, LAeq for the day (usually 7:00 a.m.–7:00 p.m.) for all day periods of a year; LNight, LAeq for the night (usually from 11:00 p.m. to 7:00 a.m.) for all night periods of a year; LDN, all 24 h LAeq periods of a year with additional 10 dB for nighttime noise annoyance (usually from 11:00 p.m. to 7:00 a.m.); LDENAEI, see LDEN, but in addition, weighted average exposure on municipal level.
| First author and year by domaina . | Population, year coverage, and country/region . | Study design . | Exposuresb . | Summary of results . | Percentage of results in expected direction . | Quality assessment (ROBIS) . |
|---|---|---|---|---|---|---|
| Air pollution | ||||||
| Alexeeff et al.34 | General population, till 31 December 2019, country not restricted: Europe, Canada, USA, UK, Australia, China, South Korea, and Israel | 42 CS | PM2.5 | Per 10 μg/m3 increase in long-term exposure, the pooled RRs and 95% CIs were 1.23 (1.15–1.31) for IHD mortality, 1.08 (0.99–1.18) for incident acute MI, 1.24 (1.13–1.36) for CeVD mortality, and 1.13 (1.11–1.15) for incident stroke. | 83% | High risk of bias |
| Atkinson et al. 2016 | General population, till October 2015, country not restricted: USA, UK, China (Taiwan), and France | Total 22 studies, relevant 8 studies: 8 CS | O3 | HRs expressed per 10 ppb increase in O3. For long-term annual O3 concentrations, the standardized effect estimates (HRs and 95% CIs) were 1.01 (0.99–1.03) for CVD mortality, 1.02 (1.00–1.04) for IHD mortality, and 1.01 (0.97–1.05) for stroke mortality. For long-term annual O3 concentrations, the random-effects summary estimates (HR and 95% CI) were 0.98 (0.93–1.04) for CVD mortality and 1.00 (0.92–1.09) for IHD mortality. For the warm season/peak O3, random-effects summary estimates were 1.01 (1.00 to 1.02). | 42% | High risk of bias |
| Atkinson et al.35 | General population, 1996–October 2016 (Medline, EMBASE), 1970–October 2016 (Web of Science), 1966–October 2016 (PubMed), country not restricted: Europe, North America, China (mainland and Taiwan), and Japan | Total 48 studies, relevant 22 studies: 22 CS | NO2 (annual or multi-year averages) | Per 10 µg/m3 increase in long-term exposure, the pooled HRs and 95% CIs were 1.03 (1.02–1.05) for CVD mortality, 1.05 (1.03–1.06) for CHD mortality, and 1.01 (0.98–1.03) for CeVD mortality. | 91% | High risk of bias |
| Chen et al. 2008 | Adults, 1 January 1950–31 December 2007, country not restricted: USA, Norway, France, the Netherlands, Canada, and Germany | Total 32 studies, relevant 17 studies: 14 CS, 3 CCS | O3, SO2, NO/NO2, black smoke, PM10, PM2.5, CO, benzene, and polycyclic aromatic hydrocarbons | RR per 10 µg/m3 increase. For PM2.5, the pooled RRs and 95% CIs were 1.14 (1.09–1.18) for CVD mortality and 1.16 (0.96–1.40) for CHD mortality. For other particulate and gaseous pollutants, the paucity of data precludes drawing conclusions. | 79% | High risk of bias |
| Chen et al. 2020 | General population, till 9 October 2018, country not restricted: Europe, Canada, UK, USA, Israel, New Zealand, South Korea, Japan, and China (mainland, Taiwan, and Hong Kong) | 67 CS | PM2.5 and PM10 | Per 10 µg/m3 increase in long-term PM2.5 exposure, the pooled RRs and 95% CIs were 1.11 (1.09–1.14) for CVD mortality, 1.16 (1.10–1.21) for IHD mortality, and 1.11 (1.04–1.18) for stroke mortality. The estimates of PM10 were 1.04 (0.99–1.10) for CVD mortality, 1.06 (1.01–1.10) for IHD mortality, and 1.01 (0.83–1.21) for stroke mortality. The certainty of evidence was high for PM2.5 and CVD mortality and was moderate for PM10 and CVD mortality, as measured by GRADE framework. | 81% | Low risk of bias |
| Chen et al.36 | General population, till October 2019, country not restricted: South Korea, UK, Denmark, Sweden, and Canada | Total 18 studies, relevant 6 studies: 6 CS | PM2.5, PM10, NO2, SO2, O3, and CO | Per 10 μg/m3 increase in long-term exposure, the pooled HRs, and 95% CIs of AF were 1.116 (1.031–1.207) for PM2.5, 1.034 (1.032–1.035) for PM10, 1.017 (1.001–1.033) for NO2, 1.005 (1.004–1.007) for SO2, 1.017 (1.013–1.022) for CO, and 1.007 (0.927–1.094) for O3. | 83% | Low risk of bias |
| Faustini et al. 2014 | General population, January 2004–January 2013, country not restricted: identified countries: Japan, China, Canada, USA, UK, Italy, Germany, Sweden, the Netherlands, and Norway | Total 23 studies, relevant 17 studies: 2 CCS, 15 CS | NO2 | RR per 10 µg/m3 increase. The pooled RRs and 95% CIs for CVD mortality were 1.13 (1.09–1.18) for NO2 and 1.20 (1.09–1.31) for PM2.5. | 94% | High risk of bias |
| Hak-Kan et al. 2013 | Chinese population, till 30 June 2012, 80 major Chinese cities in Mainland China, Hong Kong, and Taiwan | Total 48 studies, relevant 3 studies: 3 CS | PM10, NO2, SO2, and O3 | RR per 10 µg/m3 increase. In one cohort study examining PM10 and NO2, the corresponding RRs and 95% CIs were 1.0155 (1.0151–1.0160) and 1.0246 (1.0231–1.0263) for CVD mortality and 1.0149 (1.0145–1.0153) and 1.0244 (1.0227–1.0262) for CeVD mortality. In another cohort study examining SO2 and CVD, the corresponding RR was 1.032 (1.023–1.040). | 100% | High risk of bias |
| Hoek et al. 2013 | Not specified (adults), till January 2013, country not restricted: identified countries: USA, Germany, the Netherlands, Switzerland, Canada, China, New Zealand, Japan, Italy, France, and Denmark | Total 67 studies, relevant 34 studies: 34 CS | Long-term exposure to fine particulate matter (PM2.5, PM10, NO2, elemental carbon, and coarse particles) | RR per 10 µg/m3 increase. For PM2.5, the pooled RR and 95% CI was 1.11 (1.05–1.16) for CVD mortality. There was no consistent evidence that long-term exposure to coarse PM or elemental carbon is associated with CVD mortality. Several studies found positive associations between NO2 exposure and fatal MI, but not non-fatal MI. The evidence for an association between air pollution and CeVD mortality was inconsistent. | 88% | High risk of bias |
| Huang et al. 2021 | General population, till 2020.02.29, country not restricted: USA, Canada, Norway, the Netherlands, UK, Italy, Denmark, France, Spain, Japan, China, South Korea, Australia, Sweden, Norway, Germany, Austria, Switzerland, France, Italy, Spain, Greece, and Finland | 32 CS | NO2 | Per 10 ppb increase in annual NO2 concentration, the pooled HR and 95% CI was 1.11 (1.07–1.16) for cardiovascular mortality. | 71% | High risk of bias |
| Jadambaa et al.37 | Mongolian population (adults and children), till April 2014, Mongolia | Total 59 studies, relevant 2 studies: 2 CSS | NO2 and PM2.5 | Two studies found an increased risk of CVD with increased exposure to NO2 and PM2.5. | 100% | Unclear |
| Jaganathan et al.38 | General population, 1 January 1948–6 March 2018, country restricted to low- and middle-income countries, Mexico (Mexico City), Brazil (São José dos Campos, Cuiabá, and Várzea Grande), China, and India (Varanasi) | Total 17 studies, relevant 12 studies: 8 LS, 2 CSS, 1 CCR, 1 CS | PM2.5 (annual average or average measure of more than 3 days) | Eight out of nine studies (91%) reported significant effects on CVD mortality. Per 10 µg/m3 increase in long-term exposure, the effect estimates of CVD mortality ranged from 0.24 to 6.11%. All four studies reported significant effects of long-term exposure on CVD hospitalization. Few studies have evaluated this association in LMICs. No studies were found in North and Sub-Saharan Africa. | 92% | Low risk of bias |
| Kan et al. 2005 | General population, 1990–2002, China and worldwide | Total 26 studies, relevant 7 studies: 7 CS | Effects of particulate air pollution. PM10 was selected as the indicator particulate matter. | RR per 10 µg/m3 increase. For PM10, the pooled RRs and 95% CIs were 1.0095 (1.0060–1.0130) for CHA, 1.013 (1.007–1.019) for CHA based on four European studies, and 1.008 (1.004–1.011) for CHA based on three US and Canadian studies. | 100% | High risk of bias |
| Karimi et al. 2019 | Iran population, January 1980–January 2018, country restricted to Iran | Total 38 studies, relevant 28 studies: 27 CSS, 1 ES | O3, PM2.5, PM10, NO2, NOx, SO2, and CO measured by environmental protection organization and air quality control centre | Per 10 µg/m3 increase in all air pollutants, the pooled increased risk (95% CI) in CVD mortality was 0.5% (0.4–0.6%). The estimate for PM2.5 and PM10 was 0.7% (0.4–1%). | NA | Unclear |
| Liu et al. 2018 | General population, adults, January 1974–July 2017, country not restricted: USA, UK, Italy, Canada, China (mainland and Hong Kong), Europe, New Zealand, and Japan | 16 CS | PM2.5 and PM10 | Per 10 μg/m3 increase in long-term exposure, the pooled HRs and 95% CIs of CVD mortality were 1.12 (1.08–1.16) for PM2.5, 1.02 (0.89–1.16) for PM10, and 1.10 (1.06–1.14) for combined. In subgroup analyses, there is no difference in the association stratified by categories of WHO PM levels or smoking status. The estimates of PM2.5 were 1.19 (1.11–1.27) for studies with ≥ 11 years of follow-up, higher than those <11 years: 1.07 (1.04–1.11). | 88% | Low risk of bias |
| Lu et al. 2015 | Chinese population (adults only), 1990–2013, Mainland China, Hong Kong, and Taiwan | Total 59 studies, relevant 2 studies: 2 CS | PM10 and PM2.5 | RR per 10 μg/m3 increase. For the annual average concentration of PM10, the RR and 95% CI was 1.23 (1.19–1.26) for CVD mortality in one study and 1.55 (1.51–1.60) for CVD mortality in another study. | 100% | High risk of bias |
| Luben et al. 2017 | Adults, till 15 June 2017, country not restricted: USA, China (mainland and Taiwan), the Netherlands, Canada, South Korea, Spain, and Italy | Total 24 studies, relevant, 3 studies: 2 CS, 1 LS | Ambient black carbon | There are generally modest, positive associations of long-term exposure to black carbon and elemental carbon with cardiovascular hospital admissions and mortality. | 100% | High risk of bias |
| Niu et al. 2021 | General population, till 1 February 2020, country not restricted: China, Europe, England, Japan (Shizuoka), USA (California), Ghana, India, Mexico, Russia, and South Africa | Total 68 studies, relevant 13 studies: 13 CS | PM2.5, PM10, and NO2 | Per 10 μg/m3 increase in long-term exposure, the pooled HRs and 95% CIs of stroke incidence were 1.081 (0.971–1.023) for PM2.5, 1.033 (0.907–1.175) for PM10, and 1.005 (0.977–1.034) for NO2; the HRs and 95%CI of stroke mortality were 1.047 (0.995–1.101) for NO2. | 82% | High risk of bias |
| Prueitt et al.39 | General population, 1 January 2006–4 November 2013, country not restricted: USA, UK, Canada (Toronto), and China (Liaoning) | Total 25 studies, relevant 11 studies: 8 CS, 2 CSS, 1 ES | O3 | For long-term O3 exposure and CVD morbidity, studies were rare and reports were inconsistent. For CVD mortality, of 10 high-quality studies, 5 reported positive association, and the other 5 reported null or negative association. | 17% | High risk of bias |
| Scheers et al. 2015 | General population, till 20 July 2015, country not restricted: Japan, China, UK, the Netherlands, Switzerland, Greece, USA, Canada, Finland, Norway, Sweden, Denmark, Germany, Austria, Italy, Greece, and France | Total 20 studies, relevant 20 studies: 14 CS, 6 ES | PM10 or PM2.5 | HR per 10 μg/m3 increase. For PM10, the pooled HRs and 95% CIs were 1.061 (1.018–1.105) for overall stroke events and 1.080 (0.992–1.177) for stroke mortality. For PM2.5, the pooled HRs and 95% CIs were 1.064 (1.021–1.109) for overall stroke events and 1.125 (1.007–1.256) for stroke mortality. | 50% | High risk of bias |
| Shin et al. 2014 | Not specified (adults), from 1990, country not restricted: USA and UK | Total 20 studies, relevant 4 studies: 4 CS | PM2.5 | RR per 10 μg/m3 increase. In the frequentist meta-analysis, the pooled RR and 95% CI for long-term exposure to PM2.5 was 1.06 (1.00–1.13) for non-fatal strokes. The Bayesian meta-analysis found a posterior mean 1.08 (0.96–1.26) from a normal prior and 1.05 (1.02–1.10) from a gamma prior. | 100% | High risk of bias |
| Stieb et al.40 | General population, till 25 February 2020, country not restricted: Canada, USA, UK, Europe, China (mainland, Hong Kong, and Taipei), Australia, South Korea (Seoul), and Japan (Shizuoka) | 49 CS | NO2 | Per 10 ppb increase in long-term exposure, the pooled HRs and 95% CIs were 1.139 (0.997–1.301) for CVD mortality, 1.128 (1.076–1.182) for IHD mortality, and 1.167 (0.936–1.456) for CeVD mortality. After excluding studies with probably high or high risk of bias, the pooled HRs and 95% CIs were 1.058 (1.026–1.091) for CVD mortality, 1.111 (1.079–1.144) for IHD mortality, and 1.014 (0.997–1.032) for CeVD mortality. | 74% | Low risk of bias |
| Wang et al. 2020 | Older adults aged ≥ 55 years, till January 2020, country not restricted: USA (Steubenville, Eastern Massachusetts, Boston), Germany (Erfurt), Finland (Helsinki), the Netherlands (Amsterdam), UK (Scotland: Aberdeen), and China (Beijing and Taipei) | Total 19 studies, relevant 10 studies: 10 LS | Concentration of PM2.5 | Per 10 mg/m3 increase in long-term exposure, the pooled estimates and 95% CIs of HRV were −0.92% (−2.14 to 0.31%) for SDNN, −1.96% (−3.48 to −0.44%) for RMSSD in time-domain measurements, −2.78% (−4.02 to −1.55%) for LF, and −1.61% (−4.02 to 0.80%) for HF in frequency domain measurements. | 68% | High risk of bias |
| Yang et al. 201941 | General population, till 25 April 2018, country not restricted: Europe, UK, Canada, USA, South Korea, China, Ghana, India, Mexico, Russia, South Africa, and Japan | 35 CS | PM2.5, PM10, O3, and NO2 | Per 10 μg/m3 increase in long-term PM2.5 exposure, the pooled RRs and 95% CIs were 1.11 (1.07–1.15) for CVD events, 1.12 (1.05–1.19) for stroke incidence, 1.12 (1.08–1.16) for stroke events, 1.19 (1.09–1.30) for IHD incidence, and 1.14 (1.08–1.21) for IHD events. The estimates of CVD mortality were 1.11 (1.07–1.15) for PM2.5, 1.09 (1.02–1.16) for PM10, 1.23 (1.15–1.31) for NO2, and 1.03 (1.02–1.05) for O3. The estimates of NO2 and IHD events were 1.05 (1.04–1.06). No significant associations were found between PM10 and CVD, stroke and IHD incidence. | 87% | High risk of bias |
| Yuan et al. 2019 | General population, 1980–December 2018, country not restricted: Europe, USA, China (Hong Kong), Ghana, India, Mexico, Russia, South Africa, UK, Sweden (Gothenburg), and Italy | 16 CS | PM2.5 | Per 5 μg/m3 increase in long-term exposure, the pooled HRs and 95% CIs were 1.11 (1.05–1.17) for stroke incidence and 1.11 (1.05–1.17) for stroke mortality. In subgroup analysis, the estimates of stroke incidence were 1.09 (1.05–1.14) for North America (5 CS), 1.07 (1.05–1.10) for Europe (4 CS), and 2.31 (0.49–10.95) for Asia (2 CS). The associations were insignificant in both sex and significant in both ischaemic and haemorrhagic stroke. The estimates of stroke incidence were 1.08 (1.03–1.13) for never smokers, 1.11 (1.01–1.22) for former smokers, and 1.08 (0.94–1.25) for current smokers. | 95% | Low risk of bias |
| Zhao et al. 2017 | General population, 1990–2016, country not restricted: USA, Israel, Japan, UK, China, Italy, Norway, Greece, Canada, Denmark, France, South Korea, Iran, Germany, Finland, Sweden, Spain, and the Netherlands | Total 48 studies, relevant 48 studies: 25 CS, 23 LS | PM10, PM2.5, SO2, NO2, CO, and O3 | HR per 10 μg/m3 increase. For CHD mortality, the pooled HRs and 95% CIs were 1.12 (1.04–1.20) for PM10, 1.17 (1.12–1.22) for PM2.5, 1.03 (1.00–1.07) for SO2, 1.04 (1.01–1.06) for NO2, 1.04 (0.98–1.10) for CO, and 1.06 (1.01–1.11) for O3 (10 mg/m3 increase). For CHD incidence, the pooled HRs and 95% CIs were 1.01 (1.00–1.02) for PM10, 1.02 (1.00–1.03) for PM2.5, 1.01 (1.00–1.02) for SO2, 1.04 (1.03–1.06) for NO2, 1.01 (0.97–1.04) for O3, and 1.03 (1.00–1.05) for CO (10 mg/m3 increase). | NA | High risk of bias |
| Zhao et al. 2021 | General population, time and country not restricted: China, Norway, UK, the Netherlands, China (Hong Kong), and Canada (Ontario) | 7 CS | PM2.5 acquired through satellite-based model (5 studies) and outdoor-automated monitoring stations (2 studies) | Per 1.4–10 μg/m3 increase in long-term PM2.5 exposure, the pooled HRs and 95% CIs of haemorrhagic stroke were 1.16 (1.03–1.30) for total, 1.41 (0.92–2.15) for current smoker, and 1.04 (0.74–1.46) for never and former smoker. | 71% | Low risk of bias |
| Zhu et al. 2021 | General population, till 2 August 2020, country not restricted: Canada, Denmark, the Netherlands, China, USA, South Korea, Israel, and UK (London) | 12 CS | PM2.5 | Per 10 μg/m3 increase in long-term PM2.5 exposure, the pooled HRs and 95% CIs were 1.10 (1.02–1.18) for MI incidence and 1.07 (1.04–1.09) for post-MI mortality. | 75% | Unclear |
| Zou et al.42 | General population, till September 2019, country not restricted: USA, South Korea, UK, Canada, Sweden, Israel, Italy, the Netherlands, Switzerland, and Finland | 27 CS | PM2.5 and PM10 | Per 10 μg/m3 increase in long-term exposure, the pooled RRs and 95% CIs of MI were 1.18 (1.11–1.26) for PM2.5 and 1.03 (1.00–1.05) for PM10. | 91% | Unclear |
| Physical activity environment | ||||||
| Gascon et al. 2016 | Adults, till 14 November 2014, country not restricted: USA, UK, New Zealand, Lithuania, and Canada | Total 12 studies, relevant 8 studies; 4 ES, 2 CS, 1 CSS | Residential natural outdoor environments, particularly green and blue spaces | For each 10% increase of greenness, the RR and 95% CI was 0.993 (0.985–1.001) for CVD mortality. For high vs. low categories of greenness, the RR and 95% CI was 0.96 (0.94–0.97) for CVD mortality. | 75% | Unclear |
| Twohig-Bennett et al.43 | General population, till January 2017, country not restricted: USA, UK, and Lithuania | Total 143 studies, relevant 4 studies: 3 CS, 1 ES | Greenspace measured by residential NVDI, distance to the nearest greenspace, and proportion of city area covered by green land | Comparing higher to lower greenspace exposure, the pooled ORs and 95% CIs were 0.82 (0.61–1.11) for stroke (3 studies), 0.84 (0.76–0.93) for CVD mortality (2 studies), and 0.92 (0.78–1.07) for CHD (2 studies). | 86% | High risk of bias |
| Yuan et al. 2020 | Older adults (mostly ≥ 60 years), 1 January 2000–1 July 2020, country not restricted: Japan, Canada, USA, Finland, China, Rome, Australia, the Netherlands, Lithuania, Brazil, Israel, South Korea, Iran, and UK | Total 22 studies, relevant 17 studies:12 CS, 5 CSS | Greenspace measured by NDVI (mostly), percent of greenspace coverage, distance to the nearest green space, park visitation and length of stay, and loss of trees from emerald ash bore disease | Of 8 studies in total CVD, 7 found beneficial effects of green space, and the other study showed a lower risk of CVD with higher percentage of tree canopy, but not total green space. Evidence for stroke and MI was less consistent. Only cohort studies measuring NDVI and mortality were included in meta-analysis. Per 0.1 unit increase in NDVI, the pooled HRs and 95% CIs were 0.99 (0.89–1.09) for CVD mortality, 0.96 (0.88–1.05) for IHD mortality, and 0.77 (0.59–1.00) for stroke mortality. | 67% | Low risk of bias |
| Urbanization | ||||||
| Angkurawaranon et al.44 | Southeast Asian populations, till April 2013, SE Asia countries: Brunei Darussalam, Cambodia, Indonesia, Laos PDR, Malaysia, Myanmar, Philippines, Singapore, Thailand, Timor Leste, and Vietnam | Total 37 studies, relevant 7 studies: 7 CSS | Urban exposure | For urban exposure, the pooled ORs and 95% CIs were 1.01 (0.56–1.82) for stroke, 1.19 (0.35–4.07) for non-specific heart disease in the elderly, 2.48 (1.20–5.11) for CHD, and 0.31 (0.13–0.76) for RHD. | 56% | Unclear |
| Residential noisec | ||||||
| Babisch et al.45 | Not specified (adults), time and country not restricted: identified countries: UK, the Netherlands, Canada, Denmark, Germany, Sweden, and Japan | 5 CS, 4 CCS, 5 CSS | Road traffic noise. L Aeq16hr, L Aeq24hr, L DEN, LDay, LNight | Relative risk per increase of the traffic noise level of 10 dB. For road traffic noise, the pooled OR and 95% CI was 1.08 (1.04–1.13) for CHD. | 71% | High risk of bias |
| Banerjee et al. 2014 | Adult population, 1980–2010, country not restricted: the Netherlands, UK, Germany, Serbia, Sweden, Austria, Italy, Lithuania, Portugal, Switzerland, France, Slovakia, and Hungary | 14 CSS | Transportation noise exposure | (No information on unit) For traffic noise (all sources), the pooled RRs and 95% CIs were 1.04 (0.96–1.12) for CVD, 1.01 (0.89–1.14) for MI, 1.08 (0.80–1.36) for AP, and 1.00 (0.73–1.26) for IHD. The estimates for air traffic noise exposure were 1.00 (0.91–1.09) for CVD, 1.04 (0.80–1.28) for AP, 1.02 (0.89–1.14) for MI, and 0.96 (0.80–1.12) for IHD. The pooled RR for road traffic noise was 1.03 (0.97–1.09) for CVD, 1.23 (0.38–2.09) for AP, 0.85 (−0.58 to 2.29) for MI, and 1.35 (0.78–1.92) for IHD. | 73% | High risk of bias |
| Cai et al. 2021 | Adults, general population, 1 January 2000–5 October 2020, country not restricted: Denmark (Copenhagen and Aarhus), France (Paris, Lyon, and Toulouse), Switzerland, Sweden (Gothenburg), Spain (Barcelona), the Netherlands, UK (London), and Canada (Vancouver) | Total 12 studies, relevant 10 studies: 8 CS, 1 CSS, 2 ES | Residential traffic noise from road, rail, and aircraft, measured or modelled: mostly Lden, LAeq24hr, LAeq16hr, LDN, Lday, Lnight | For road traffic, per 10 dB increase in Lden, the pooled HRs and 95% CIs were 1.01 (0.98–1.05) for CVD mortality, 1.03 (0.99–1.08) for IHD mortality, and 1.05 (0.97–1.14) for stroke mortality. For aircraft traffic, the estimates based on three studies were 1.17 (1.10–1.25) for CVD mortality, 1.03 (0.82–1.29) for IHD mortality, and 1.06 (0.93–1.20) for stroke mortality. For rail traffic, the estimates were 0.98 (0.94–1.01) for CVD mortality (1 study) and 1.02 (0.91–1.14) for IHD mortality (2 studies). | 68% | Unclear |
| Dzhambov et al. 2016 | Adults, till 24 November 2015, country not restricted: the Netherlands, UK, Denmark, Germany, France, Switzerland, USA, Canada, Sweden, Greece, and Italy | 7 CS, 2 CSS, ES 4 | Traffic noise | RR per 10 dB noise increase. For road traffic noise, the pooled RR and 95% CI was 1.03 (0.87–1.22). For air traffic noise, the pooled RR was 1.05 (1.00–1.10). | 72% | High risk of bias |
| Khosravipour et al. 2020 | General population, time and country not restricted: till 29 November 2019, UK, Germany, Sweden, Lithuania, Denmark, and the Netherlands | 7 CS, 5 CCS, 1 CSS | Road traffic noise | Comparing highest to lowest category of noise exposure (results from categorical analysis), the pooled RR and 95% CI of MI were 1.03 (0.93–1.13). Per 10 dB increment (results from exposure–response analysis and transformed from categorical analysis), the pooled estimate was 1.02 (1.00–1.05). In subgroup analysis, pooled estimates were significant for CCS and CSS, but not for CS. Estimates for the exposure–response analyses were 1.03 (1.00–1.05) after excluding two conference papers and 1.02 (1.01–1.03) after further excluding the studies with only results from categorical analysis. | 57% | Low risk of bias |
| van Kempen et al. 2002 | Adults, 1970–1999, country not restricted: Iran, Belgium, Germany, Canada, India, Finland, Italy, the Netherlands, Russia, USA, Poland, Japan, Israel, China, France, South Africa, China (Taiwan), and UK. | Total 43 studies, relevant 10 studies: 6 CSS, 2CCS, 2 CS | Community noise exposure (road and air traffic) assessed by calculations, personal dosimeter, or sound level meter | RR per 5 dB(A) noise increase. For road traffic noise, the pooled RRs and 95% CIs were 1.09 (1.05–1.13) for IHD, 0.99 (0.84–1.16) for AP, and 1.03 (0.99–1.09) for MI. For air traffic noise, the pooled RR was 1.03 (0.90–1.18) for AP. | 25% | High risk of bias |
| van Kempen, et al.46 | European, 2000–October 2014, European countries | Total 61 studies, relevant 32 studies: 14 CSS, 5 ES, 8 CS, 5 CCS | Noise from road, rail, and air traffic and wind turbines: LDEN | Road, rail, and air traffic noise in relation to prevalence, incidence, and mortality of IHD and stroke were analysed, respectively. Number of studies for each analysis is small. Per 10 dB increase in exposure, the pooled RR and 95% CI of IHD was 1.08 (1.01–1.15) for road traffic. Estimates for other associations were of low quality or from <3 studies, and mostly insignificant. | NA | Low risk of bias |
| Vienneau et al.47 | Not specified (general population), January 1994–January 2014, country not restricted: Germany, UK, the Netherlands, Sweden, Switzerland, Denmark, Canada, and USA | 3 CCS, 5 CS, 2 LS | Transportation noise exposure | RR per 10 dB increase in exposure. The pooled RR and 95% CI for IHD was 1.06 (1.03–1.09). | 75% | High risk of bias |
| Weihofen et al.48 | General population, till 31 August 2017, country not restricted: USA, France (Paris, Lyon Saint, and Toulouse), Canada (Vancouver), UK (London), Switzerland, Germany (Berlin and Frankfurt), the Netherlands (Amsterdam), Sweden (Stockholm), Greece (Athens), and Italy (Milan) | 3 CSS, 1 ES, 4 CS, 1 CCS | Aircraft noise: LAeq, LDay, LNight, LDN, Lden, LDENAEI | Per 10 dB increase in Lden, the pooled RR and 95% CI of stroke was 1.013 (0.998–1.028) from seven studies. | 71% | Low risk of bias |
| Ambient temperature | ||||||
| Bunker et al.49 | Elderly (65+), 1 January 1975–24 July 2015, country not restricted: USA, Bangladesh, China (mainland, Taiwan, and Hong Kong), Portugal, UK, Denmark, Australia, Russia, Italy, Hungary, Brazil, Vietnam, Sweden, Thailand, Norway, South Korea, and Germany | Total 60 studies, relevant 47 studies; 47 LS | Ambient hot and cold temperature | Per 1°C temperature change, for heat, the pooled percentage changes and 95% CIs were 3.79 (3.40–4.18) for CVD mortality, 1.62 (0.24–3.03) for IHD mortality, 1.40 (0.06–2.75) for CeVD mortality, 0.33 (−0.09 to 0.75) for IS, −0.66 (−2.13 to 0.84) for ICH, −0.17(−0.96 to 0.63) for CeVD, −0.16(−2.05 to 1.77) for MI, and 0.30(−0.12 to 0.81) for CVD. For cold, the estimates were 1.84 (0.85–2.84) for CVD mortality, 0.45 (−0.01 to 0.91) for IHD mortality, 1.21 (0.66–1.77) for CeVD mortality, 3.63 (−3.94 to 11.8) for IS, 1.49 (1.04–1.94) for ICH, −0.46 (−1.12 to 0.2) for CeVD, 0.66 (−0.14 to 1.48) for MI, −0.80 (−2.21 to 0.64) for AP, −0.67 (−2.15 to 0.83) for HF, and −0.28 (−1.39 to 0.84) for CVD. | 73% | High risk of bias |
| Kofi Amegah et al. 2016 | Sub-Saharan African populations, till December 2014, Sub-Saharan Africa | Total 23 studies, relevant 5 studies: 4 LS, 1 CSS | Temperature | One study found that low temperature was associated with increased risk of CVD. Two studies found associations of low and high temperatures with CVD mortality. One study found no association between mean monthly temperature and CVD mortality. One study found 5°C change in the monthly mean temperature to be associated with decreased risk of hospitalization for venous thromboembolism, stroke, and acute MI. | 56% | High risk of bias |
| Ma et al. 2020 | Chinese population, January 2010–January 2020, country restricted to China | Total 175 studies, relevant 19 studies: 19 LS | (i) Every 1°C temperature increase/decrease beyond certain reference points (ii) Comparison between extreme temperatures and reference normal temperatures | Pooled RRs and 95% CIs of CVD were 1.089 (1.062–1.116) and 1.171 (1.125–1.218), respectively, for hot and cold temperatures as compared with normal temperatures. | 100% | High risk of bias |
| Moghadamnia et al.50 | General population, January 2000–31 December 2015, country not restricted: China (mainland, Taiwan, and Hong Kong), Australia, Thailand, Philippines, South Korea, Germany, and Spain | 26 LS | Ambient temperature | RR per 1°C change of temperature. For CVD mortality, the RRs and 95% CIs were 1.055 (1.050–1.060) for cold exposure and 1.013 (1.011–1.015) for heat exposure. Coefficient per 1°C change in mean annual temperature. For CVD mortality, the pooled estimates were 0.026 (−0.019 to 0.072) for cold exposure and 0.008 (−0.015 to 0.031) for heat exposure. | 96% | High risk of bias |
| Odame et al. 2018 | Rural population, till April 2018, country not restricted: Bangladesh (MATLAB), Czech Republic, and China (Naidong and Jiangzi in Tibet) | All 14 studies, relevant 3 studies: 3 LS | Daily mean temperature | Per 1°C increase, the pooled RR and 95% CI of CVD mortality was 1.111 (1.045–1.181). The associations were significant in subgroup analyses of both developing and developed countries. | 100% | High risk of bias |
| Turner et al. 2012 | Not specified (general population), time and country not restricted: South Korea (Incheon), USA, UK (London and Scotland), Europe, Australia (Brisbane), and Thailand (Muang) | Total 21 studies, relevant 18 studies: 18 LS | Effects of ambient temperature. Maximum, minimum, and mean daily temperature | RR per 1°C increase in temperature. The pooled RRs and 95% CI were 0.999 (0.982–1.016) for CVD morbidities, 0.990 (0.887–1.105) for stroke, and 1.010 (0.930–1.097) for ACS/MI. | 43% | High risk of bias |
| Wang et al. 2016 | Adults, till 16 October 2015, country not restricted: Japan, UK, Russia, Spain, China, Portugal, Italy, South Korea, China (Taiwan), Mozambique, USA, France, and Australia | 7 CS, 13 LS, 1 CSS | Ambient temperature | OR per 1°C increase in mean ambient temperature. The pooled ORs and 95% CIs were 0.97 (0.94–1.00) for ICH, 1.00 (0.99–1.01) for IS, and 1.00 (0.98–1.01) for SAH. The pooled estimates for ambient minimum and maximum temperature and IS were OR 0.99 (0.96–1.01) and 0.98 (0.94–1.02), respectively. | 19% | Unclear |
| Zafeiratou et al. 2021 | General population, 1990–2020, country not restricted: Serbia (Belgrade), China (Hong Kong), UK (London), USA (New England), Finland (Helsinki), Switzerland, and South Korea | Total 34 studies, relevant 6 studies: 4 ES, 2 CS | Mean annual temperature or variability, seasonal temperature or variability, annual temperature categories, and mean annual degrees above/below minimum mortality temperature | In temporal comparisons within the same area, increased cardiovascular mortality was associated with both increased and decreased temperature. Stronger association was found with cold rather than hot temperature. In geographical comparison in just one study, people living in areas with higher temperature were found a lower rate of IHD mortality, though no dose–response. | 5 studies on CVD mortality: 60% | High risk of bias |
| Multiple domains | ||||||
| Rugel et al. 2020 | Urban residents, 1 January 2003–November 2019, country not restricted: the Netherlands, Sweden (Skåne), UK (London), Canada, Denmark (Copenhagen and Aarhus), Germany (Bochum, Essen, and Mülheim/Ruhr), South Korea (Seoul, Ulsan, and Cheonan), France, Spain, Norway, Greece, Italy (Rome and Verona), Switzerland, USA (California), and China (Shenyang, Anshan, Jinzhou, and Taiwan) | Total 51 studies, relevant 21 studies: 15 CS, 5 CSS, 1 ES | Traffic-related air pollution; natural spaces, neighbourhood walkability; noise | Based on the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) system for assessing the quality of the synthesized evidence base, evidence was sufficient for higher noise exposure with increased CVD. And evidence was sufficient for no association between traffic-related air pollution and CVD. | NA | High risk of bias |
| First author and year by domaina . | Population, year coverage, and country/region . | Study design . | Exposuresb . | Summary of results . | Percentage of results in expected direction . | Quality assessment (ROBIS) . |
|---|---|---|---|---|---|---|
| Air pollution | ||||||
| Alexeeff et al.34 | General population, till 31 December 2019, country not restricted: Europe, Canada, USA, UK, Australia, China, South Korea, and Israel | 42 CS | PM2.5 | Per 10 μg/m3 increase in long-term exposure, the pooled RRs and 95% CIs were 1.23 (1.15–1.31) for IHD mortality, 1.08 (0.99–1.18) for incident acute MI, 1.24 (1.13–1.36) for CeVD mortality, and 1.13 (1.11–1.15) for incident stroke. | 83% | High risk of bias |
| Atkinson et al. 2016 | General population, till October 2015, country not restricted: USA, UK, China (Taiwan), and France | Total 22 studies, relevant 8 studies: 8 CS | O3 | HRs expressed per 10 ppb increase in O3. For long-term annual O3 concentrations, the standardized effect estimates (HRs and 95% CIs) were 1.01 (0.99–1.03) for CVD mortality, 1.02 (1.00–1.04) for IHD mortality, and 1.01 (0.97–1.05) for stroke mortality. For long-term annual O3 concentrations, the random-effects summary estimates (HR and 95% CI) were 0.98 (0.93–1.04) for CVD mortality and 1.00 (0.92–1.09) for IHD mortality. For the warm season/peak O3, random-effects summary estimates were 1.01 (1.00 to 1.02). | 42% | High risk of bias |
| Atkinson et al.35 | General population, 1996–October 2016 (Medline, EMBASE), 1970–October 2016 (Web of Science), 1966–October 2016 (PubMed), country not restricted: Europe, North America, China (mainland and Taiwan), and Japan | Total 48 studies, relevant 22 studies: 22 CS | NO2 (annual or multi-year averages) | Per 10 µg/m3 increase in long-term exposure, the pooled HRs and 95% CIs were 1.03 (1.02–1.05) for CVD mortality, 1.05 (1.03–1.06) for CHD mortality, and 1.01 (0.98–1.03) for CeVD mortality. | 91% | High risk of bias |
| Chen et al. 2008 | Adults, 1 January 1950–31 December 2007, country not restricted: USA, Norway, France, the Netherlands, Canada, and Germany | Total 32 studies, relevant 17 studies: 14 CS, 3 CCS | O3, SO2, NO/NO2, black smoke, PM10, PM2.5, CO, benzene, and polycyclic aromatic hydrocarbons | RR per 10 µg/m3 increase. For PM2.5, the pooled RRs and 95% CIs were 1.14 (1.09–1.18) for CVD mortality and 1.16 (0.96–1.40) for CHD mortality. For other particulate and gaseous pollutants, the paucity of data precludes drawing conclusions. | 79% | High risk of bias |
| Chen et al. 2020 | General population, till 9 October 2018, country not restricted: Europe, Canada, UK, USA, Israel, New Zealand, South Korea, Japan, and China (mainland, Taiwan, and Hong Kong) | 67 CS | PM2.5 and PM10 | Per 10 µg/m3 increase in long-term PM2.5 exposure, the pooled RRs and 95% CIs were 1.11 (1.09–1.14) for CVD mortality, 1.16 (1.10–1.21) for IHD mortality, and 1.11 (1.04–1.18) for stroke mortality. The estimates of PM10 were 1.04 (0.99–1.10) for CVD mortality, 1.06 (1.01–1.10) for IHD mortality, and 1.01 (0.83–1.21) for stroke mortality. The certainty of evidence was high for PM2.5 and CVD mortality and was moderate for PM10 and CVD mortality, as measured by GRADE framework. | 81% | Low risk of bias |
| Chen et al.36 | General population, till October 2019, country not restricted: South Korea, UK, Denmark, Sweden, and Canada | Total 18 studies, relevant 6 studies: 6 CS | PM2.5, PM10, NO2, SO2, O3, and CO | Per 10 μg/m3 increase in long-term exposure, the pooled HRs, and 95% CIs of AF were 1.116 (1.031–1.207) for PM2.5, 1.034 (1.032–1.035) for PM10, 1.017 (1.001–1.033) for NO2, 1.005 (1.004–1.007) for SO2, 1.017 (1.013–1.022) for CO, and 1.007 (0.927–1.094) for O3. | 83% | Low risk of bias |
| Faustini et al. 2014 | General population, January 2004–January 2013, country not restricted: identified countries: Japan, China, Canada, USA, UK, Italy, Germany, Sweden, the Netherlands, and Norway | Total 23 studies, relevant 17 studies: 2 CCS, 15 CS | NO2 | RR per 10 µg/m3 increase. The pooled RRs and 95% CIs for CVD mortality were 1.13 (1.09–1.18) for NO2 and 1.20 (1.09–1.31) for PM2.5. | 94% | High risk of bias |
| Hak-Kan et al. 2013 | Chinese population, till 30 June 2012, 80 major Chinese cities in Mainland China, Hong Kong, and Taiwan | Total 48 studies, relevant 3 studies: 3 CS | PM10, NO2, SO2, and O3 | RR per 10 µg/m3 increase. In one cohort study examining PM10 and NO2, the corresponding RRs and 95% CIs were 1.0155 (1.0151–1.0160) and 1.0246 (1.0231–1.0263) for CVD mortality and 1.0149 (1.0145–1.0153) and 1.0244 (1.0227–1.0262) for CeVD mortality. In another cohort study examining SO2 and CVD, the corresponding RR was 1.032 (1.023–1.040). | 100% | High risk of bias |
| Hoek et al. 2013 | Not specified (adults), till January 2013, country not restricted: identified countries: USA, Germany, the Netherlands, Switzerland, Canada, China, New Zealand, Japan, Italy, France, and Denmark | Total 67 studies, relevant 34 studies: 34 CS | Long-term exposure to fine particulate matter (PM2.5, PM10, NO2, elemental carbon, and coarse particles) | RR per 10 µg/m3 increase. For PM2.5, the pooled RR and 95% CI was 1.11 (1.05–1.16) for CVD mortality. There was no consistent evidence that long-term exposure to coarse PM or elemental carbon is associated with CVD mortality. Several studies found positive associations between NO2 exposure and fatal MI, but not non-fatal MI. The evidence for an association between air pollution and CeVD mortality was inconsistent. | 88% | High risk of bias |
| Huang et al. 2021 | General population, till 2020.02.29, country not restricted: USA, Canada, Norway, the Netherlands, UK, Italy, Denmark, France, Spain, Japan, China, South Korea, Australia, Sweden, Norway, Germany, Austria, Switzerland, France, Italy, Spain, Greece, and Finland | 32 CS | NO2 | Per 10 ppb increase in annual NO2 concentration, the pooled HR and 95% CI was 1.11 (1.07–1.16) for cardiovascular mortality. | 71% | High risk of bias |
| Jadambaa et al.37 | Mongolian population (adults and children), till April 2014, Mongolia | Total 59 studies, relevant 2 studies: 2 CSS | NO2 and PM2.5 | Two studies found an increased risk of CVD with increased exposure to NO2 and PM2.5. | 100% | Unclear |
| Jaganathan et al.38 | General population, 1 January 1948–6 March 2018, country restricted to low- and middle-income countries, Mexico (Mexico City), Brazil (São José dos Campos, Cuiabá, and Várzea Grande), China, and India (Varanasi) | Total 17 studies, relevant 12 studies: 8 LS, 2 CSS, 1 CCR, 1 CS | PM2.5 (annual average or average measure of more than 3 days) | Eight out of nine studies (91%) reported significant effects on CVD mortality. Per 10 µg/m3 increase in long-term exposure, the effect estimates of CVD mortality ranged from 0.24 to 6.11%. All four studies reported significant effects of long-term exposure on CVD hospitalization. Few studies have evaluated this association in LMICs. No studies were found in North and Sub-Saharan Africa. | 92% | Low risk of bias |
| Kan et al. 2005 | General population, 1990–2002, China and worldwide | Total 26 studies, relevant 7 studies: 7 CS | Effects of particulate air pollution. PM10 was selected as the indicator particulate matter. | RR per 10 µg/m3 increase. For PM10, the pooled RRs and 95% CIs were 1.0095 (1.0060–1.0130) for CHA, 1.013 (1.007–1.019) for CHA based on four European studies, and 1.008 (1.004–1.011) for CHA based on three US and Canadian studies. | 100% | High risk of bias |
| Karimi et al. 2019 | Iran population, January 1980–January 2018, country restricted to Iran | Total 38 studies, relevant 28 studies: 27 CSS, 1 ES | O3, PM2.5, PM10, NO2, NOx, SO2, and CO measured by environmental protection organization and air quality control centre | Per 10 µg/m3 increase in all air pollutants, the pooled increased risk (95% CI) in CVD mortality was 0.5% (0.4–0.6%). The estimate for PM2.5 and PM10 was 0.7% (0.4–1%). | NA | Unclear |
| Liu et al. 2018 | General population, adults, January 1974–July 2017, country not restricted: USA, UK, Italy, Canada, China (mainland and Hong Kong), Europe, New Zealand, and Japan | 16 CS | PM2.5 and PM10 | Per 10 μg/m3 increase in long-term exposure, the pooled HRs and 95% CIs of CVD mortality were 1.12 (1.08–1.16) for PM2.5, 1.02 (0.89–1.16) for PM10, and 1.10 (1.06–1.14) for combined. In subgroup analyses, there is no difference in the association stratified by categories of WHO PM levels or smoking status. The estimates of PM2.5 were 1.19 (1.11–1.27) for studies with ≥ 11 years of follow-up, higher than those <11 years: 1.07 (1.04–1.11). | 88% | Low risk of bias |
| Lu et al. 2015 | Chinese population (adults only), 1990–2013, Mainland China, Hong Kong, and Taiwan | Total 59 studies, relevant 2 studies: 2 CS | PM10 and PM2.5 | RR per 10 μg/m3 increase. For the annual average concentration of PM10, the RR and 95% CI was 1.23 (1.19–1.26) for CVD mortality in one study and 1.55 (1.51–1.60) for CVD mortality in another study. | 100% | High risk of bias |
| Luben et al. 2017 | Adults, till 15 June 2017, country not restricted: USA, China (mainland and Taiwan), the Netherlands, Canada, South Korea, Spain, and Italy | Total 24 studies, relevant, 3 studies: 2 CS, 1 LS | Ambient black carbon | There are generally modest, positive associations of long-term exposure to black carbon and elemental carbon with cardiovascular hospital admissions and mortality. | 100% | High risk of bias |
| Niu et al. 2021 | General population, till 1 February 2020, country not restricted: China, Europe, England, Japan (Shizuoka), USA (California), Ghana, India, Mexico, Russia, and South Africa | Total 68 studies, relevant 13 studies: 13 CS | PM2.5, PM10, and NO2 | Per 10 μg/m3 increase in long-term exposure, the pooled HRs and 95% CIs of stroke incidence were 1.081 (0.971–1.023) for PM2.5, 1.033 (0.907–1.175) for PM10, and 1.005 (0.977–1.034) for NO2; the HRs and 95%CI of stroke mortality were 1.047 (0.995–1.101) for NO2. | 82% | High risk of bias |
| Prueitt et al.39 | General population, 1 January 2006–4 November 2013, country not restricted: USA, UK, Canada (Toronto), and China (Liaoning) | Total 25 studies, relevant 11 studies: 8 CS, 2 CSS, 1 ES | O3 | For long-term O3 exposure and CVD morbidity, studies were rare and reports were inconsistent. For CVD mortality, of 10 high-quality studies, 5 reported positive association, and the other 5 reported null or negative association. | 17% | High risk of bias |
| Scheers et al. 2015 | General population, till 20 July 2015, country not restricted: Japan, China, UK, the Netherlands, Switzerland, Greece, USA, Canada, Finland, Norway, Sweden, Denmark, Germany, Austria, Italy, Greece, and France | Total 20 studies, relevant 20 studies: 14 CS, 6 ES | PM10 or PM2.5 | HR per 10 μg/m3 increase. For PM10, the pooled HRs and 95% CIs were 1.061 (1.018–1.105) for overall stroke events and 1.080 (0.992–1.177) for stroke mortality. For PM2.5, the pooled HRs and 95% CIs were 1.064 (1.021–1.109) for overall stroke events and 1.125 (1.007–1.256) for stroke mortality. | 50% | High risk of bias |
| Shin et al. 2014 | Not specified (adults), from 1990, country not restricted: USA and UK | Total 20 studies, relevant 4 studies: 4 CS | PM2.5 | RR per 10 μg/m3 increase. In the frequentist meta-analysis, the pooled RR and 95% CI for long-term exposure to PM2.5 was 1.06 (1.00–1.13) for non-fatal strokes. The Bayesian meta-analysis found a posterior mean 1.08 (0.96–1.26) from a normal prior and 1.05 (1.02–1.10) from a gamma prior. | 100% | High risk of bias |
| Stieb et al.40 | General population, till 25 February 2020, country not restricted: Canada, USA, UK, Europe, China (mainland, Hong Kong, and Taipei), Australia, South Korea (Seoul), and Japan (Shizuoka) | 49 CS | NO2 | Per 10 ppb increase in long-term exposure, the pooled HRs and 95% CIs were 1.139 (0.997–1.301) for CVD mortality, 1.128 (1.076–1.182) for IHD mortality, and 1.167 (0.936–1.456) for CeVD mortality. After excluding studies with probably high or high risk of bias, the pooled HRs and 95% CIs were 1.058 (1.026–1.091) for CVD mortality, 1.111 (1.079–1.144) for IHD mortality, and 1.014 (0.997–1.032) for CeVD mortality. | 74% | Low risk of bias |
| Wang et al. 2020 | Older adults aged ≥ 55 years, till January 2020, country not restricted: USA (Steubenville, Eastern Massachusetts, Boston), Germany (Erfurt), Finland (Helsinki), the Netherlands (Amsterdam), UK (Scotland: Aberdeen), and China (Beijing and Taipei) | Total 19 studies, relevant 10 studies: 10 LS | Concentration of PM2.5 | Per 10 mg/m3 increase in long-term exposure, the pooled estimates and 95% CIs of HRV were −0.92% (−2.14 to 0.31%) for SDNN, −1.96% (−3.48 to −0.44%) for RMSSD in time-domain measurements, −2.78% (−4.02 to −1.55%) for LF, and −1.61% (−4.02 to 0.80%) for HF in frequency domain measurements. | 68% | High risk of bias |
| Yang et al. 201941 | General population, till 25 April 2018, country not restricted: Europe, UK, Canada, USA, South Korea, China, Ghana, India, Mexico, Russia, South Africa, and Japan | 35 CS | PM2.5, PM10, O3, and NO2 | Per 10 μg/m3 increase in long-term PM2.5 exposure, the pooled RRs and 95% CIs were 1.11 (1.07–1.15) for CVD events, 1.12 (1.05–1.19) for stroke incidence, 1.12 (1.08–1.16) for stroke events, 1.19 (1.09–1.30) for IHD incidence, and 1.14 (1.08–1.21) for IHD events. The estimates of CVD mortality were 1.11 (1.07–1.15) for PM2.5, 1.09 (1.02–1.16) for PM10, 1.23 (1.15–1.31) for NO2, and 1.03 (1.02–1.05) for O3. The estimates of NO2 and IHD events were 1.05 (1.04–1.06). No significant associations were found between PM10 and CVD, stroke and IHD incidence. | 87% | High risk of bias |
| Yuan et al. 2019 | General population, 1980–December 2018, country not restricted: Europe, USA, China (Hong Kong), Ghana, India, Mexico, Russia, South Africa, UK, Sweden (Gothenburg), and Italy | 16 CS | PM2.5 | Per 5 μg/m3 increase in long-term exposure, the pooled HRs and 95% CIs were 1.11 (1.05–1.17) for stroke incidence and 1.11 (1.05–1.17) for stroke mortality. In subgroup analysis, the estimates of stroke incidence were 1.09 (1.05–1.14) for North America (5 CS), 1.07 (1.05–1.10) for Europe (4 CS), and 2.31 (0.49–10.95) for Asia (2 CS). The associations were insignificant in both sex and significant in both ischaemic and haemorrhagic stroke. The estimates of stroke incidence were 1.08 (1.03–1.13) for never smokers, 1.11 (1.01–1.22) for former smokers, and 1.08 (0.94–1.25) for current smokers. | 95% | Low risk of bias |
| Zhao et al. 2017 | General population, 1990–2016, country not restricted: USA, Israel, Japan, UK, China, Italy, Norway, Greece, Canada, Denmark, France, South Korea, Iran, Germany, Finland, Sweden, Spain, and the Netherlands | Total 48 studies, relevant 48 studies: 25 CS, 23 LS | PM10, PM2.5, SO2, NO2, CO, and O3 | HR per 10 μg/m3 increase. For CHD mortality, the pooled HRs and 95% CIs were 1.12 (1.04–1.20) for PM10, 1.17 (1.12–1.22) for PM2.5, 1.03 (1.00–1.07) for SO2, 1.04 (1.01–1.06) for NO2, 1.04 (0.98–1.10) for CO, and 1.06 (1.01–1.11) for O3 (10 mg/m3 increase). For CHD incidence, the pooled HRs and 95% CIs were 1.01 (1.00–1.02) for PM10, 1.02 (1.00–1.03) for PM2.5, 1.01 (1.00–1.02) for SO2, 1.04 (1.03–1.06) for NO2, 1.01 (0.97–1.04) for O3, and 1.03 (1.00–1.05) for CO (10 mg/m3 increase). | NA | High risk of bias |
| Zhao et al. 2021 | General population, time and country not restricted: China, Norway, UK, the Netherlands, China (Hong Kong), and Canada (Ontario) | 7 CS | PM2.5 acquired through satellite-based model (5 studies) and outdoor-automated monitoring stations (2 studies) | Per 1.4–10 μg/m3 increase in long-term PM2.5 exposure, the pooled HRs and 95% CIs of haemorrhagic stroke were 1.16 (1.03–1.30) for total, 1.41 (0.92–2.15) for current smoker, and 1.04 (0.74–1.46) for never and former smoker. | 71% | Low risk of bias |
| Zhu et al. 2021 | General population, till 2 August 2020, country not restricted: Canada, Denmark, the Netherlands, China, USA, South Korea, Israel, and UK (London) | 12 CS | PM2.5 | Per 10 μg/m3 increase in long-term PM2.5 exposure, the pooled HRs and 95% CIs were 1.10 (1.02–1.18) for MI incidence and 1.07 (1.04–1.09) for post-MI mortality. | 75% | Unclear |
| Zou et al.42 | General population, till September 2019, country not restricted: USA, South Korea, UK, Canada, Sweden, Israel, Italy, the Netherlands, Switzerland, and Finland | 27 CS | PM2.5 and PM10 | Per 10 μg/m3 increase in long-term exposure, the pooled RRs and 95% CIs of MI were 1.18 (1.11–1.26) for PM2.5 and 1.03 (1.00–1.05) for PM10. | 91% | Unclear |
| Physical activity environment | ||||||
| Gascon et al. 2016 | Adults, till 14 November 2014, country not restricted: USA, UK, New Zealand, Lithuania, and Canada | Total 12 studies, relevant 8 studies; 4 ES, 2 CS, 1 CSS | Residential natural outdoor environments, particularly green and blue spaces | For each 10% increase of greenness, the RR and 95% CI was 0.993 (0.985–1.001) for CVD mortality. For high vs. low categories of greenness, the RR and 95% CI was 0.96 (0.94–0.97) for CVD mortality. | 75% | Unclear |
| Twohig-Bennett et al.43 | General population, till January 2017, country not restricted: USA, UK, and Lithuania | Total 143 studies, relevant 4 studies: 3 CS, 1 ES | Greenspace measured by residential NVDI, distance to the nearest greenspace, and proportion of city area covered by green land | Comparing higher to lower greenspace exposure, the pooled ORs and 95% CIs were 0.82 (0.61–1.11) for stroke (3 studies), 0.84 (0.76–0.93) for CVD mortality (2 studies), and 0.92 (0.78–1.07) for CHD (2 studies). | 86% | High risk of bias |
| Yuan et al. 2020 | Older adults (mostly ≥ 60 years), 1 January 2000–1 July 2020, country not restricted: Japan, Canada, USA, Finland, China, Rome, Australia, the Netherlands, Lithuania, Brazil, Israel, South Korea, Iran, and UK | Total 22 studies, relevant 17 studies:12 CS, 5 CSS | Greenspace measured by NDVI (mostly), percent of greenspace coverage, distance to the nearest green space, park visitation and length of stay, and loss of trees from emerald ash bore disease | Of 8 studies in total CVD, 7 found beneficial effects of green space, and the other study showed a lower risk of CVD with higher percentage of tree canopy, but not total green space. Evidence for stroke and MI was less consistent. Only cohort studies measuring NDVI and mortality were included in meta-analysis. Per 0.1 unit increase in NDVI, the pooled HRs and 95% CIs were 0.99 (0.89–1.09) for CVD mortality, 0.96 (0.88–1.05) for IHD mortality, and 0.77 (0.59–1.00) for stroke mortality. | 67% | Low risk of bias |
| Urbanization | ||||||
| Angkurawaranon et al.44 | Southeast Asian populations, till April 2013, SE Asia countries: Brunei Darussalam, Cambodia, Indonesia, Laos PDR, Malaysia, Myanmar, Philippines, Singapore, Thailand, Timor Leste, and Vietnam | Total 37 studies, relevant 7 studies: 7 CSS | Urban exposure | For urban exposure, the pooled ORs and 95% CIs were 1.01 (0.56–1.82) for stroke, 1.19 (0.35–4.07) for non-specific heart disease in the elderly, 2.48 (1.20–5.11) for CHD, and 0.31 (0.13–0.76) for RHD. | 56% | Unclear |
| Residential noisec | ||||||
| Babisch et al.45 | Not specified (adults), time and country not restricted: identified countries: UK, the Netherlands, Canada, Denmark, Germany, Sweden, and Japan | 5 CS, 4 CCS, 5 CSS | Road traffic noise. L Aeq16hr, L Aeq24hr, L DEN, LDay, LNight | Relative risk per increase of the traffic noise level of 10 dB. For road traffic noise, the pooled OR and 95% CI was 1.08 (1.04–1.13) for CHD. | 71% | High risk of bias |
| Banerjee et al. 2014 | Adult population, 1980–2010, country not restricted: the Netherlands, UK, Germany, Serbia, Sweden, Austria, Italy, Lithuania, Portugal, Switzerland, France, Slovakia, and Hungary | 14 CSS | Transportation noise exposure | (No information on unit) For traffic noise (all sources), the pooled RRs and 95% CIs were 1.04 (0.96–1.12) for CVD, 1.01 (0.89–1.14) for MI, 1.08 (0.80–1.36) for AP, and 1.00 (0.73–1.26) for IHD. The estimates for air traffic noise exposure were 1.00 (0.91–1.09) for CVD, 1.04 (0.80–1.28) for AP, 1.02 (0.89–1.14) for MI, and 0.96 (0.80–1.12) for IHD. The pooled RR for road traffic noise was 1.03 (0.97–1.09) for CVD, 1.23 (0.38–2.09) for AP, 0.85 (−0.58 to 2.29) for MI, and 1.35 (0.78–1.92) for IHD. | 73% | High risk of bias |
| Cai et al. 2021 | Adults, general population, 1 January 2000–5 October 2020, country not restricted: Denmark (Copenhagen and Aarhus), France (Paris, Lyon, and Toulouse), Switzerland, Sweden (Gothenburg), Spain (Barcelona), the Netherlands, UK (London), and Canada (Vancouver) | Total 12 studies, relevant 10 studies: 8 CS, 1 CSS, 2 ES | Residential traffic noise from road, rail, and aircraft, measured or modelled: mostly Lden, LAeq24hr, LAeq16hr, LDN, Lday, Lnight | For road traffic, per 10 dB increase in Lden, the pooled HRs and 95% CIs were 1.01 (0.98–1.05) for CVD mortality, 1.03 (0.99–1.08) for IHD mortality, and 1.05 (0.97–1.14) for stroke mortality. For aircraft traffic, the estimates based on three studies were 1.17 (1.10–1.25) for CVD mortality, 1.03 (0.82–1.29) for IHD mortality, and 1.06 (0.93–1.20) for stroke mortality. For rail traffic, the estimates were 0.98 (0.94–1.01) for CVD mortality (1 study) and 1.02 (0.91–1.14) for IHD mortality (2 studies). | 68% | Unclear |
| Dzhambov et al. 2016 | Adults, till 24 November 2015, country not restricted: the Netherlands, UK, Denmark, Germany, France, Switzerland, USA, Canada, Sweden, Greece, and Italy | 7 CS, 2 CSS, ES 4 | Traffic noise | RR per 10 dB noise increase. For road traffic noise, the pooled RR and 95% CI was 1.03 (0.87–1.22). For air traffic noise, the pooled RR was 1.05 (1.00–1.10). | 72% | High risk of bias |
| Khosravipour et al. 2020 | General population, time and country not restricted: till 29 November 2019, UK, Germany, Sweden, Lithuania, Denmark, and the Netherlands | 7 CS, 5 CCS, 1 CSS | Road traffic noise | Comparing highest to lowest category of noise exposure (results from categorical analysis), the pooled RR and 95% CI of MI were 1.03 (0.93–1.13). Per 10 dB increment (results from exposure–response analysis and transformed from categorical analysis), the pooled estimate was 1.02 (1.00–1.05). In subgroup analysis, pooled estimates were significant for CCS and CSS, but not for CS. Estimates for the exposure–response analyses were 1.03 (1.00–1.05) after excluding two conference papers and 1.02 (1.01–1.03) after further excluding the studies with only results from categorical analysis. | 57% | Low risk of bias |
| van Kempen et al. 2002 | Adults, 1970–1999, country not restricted: Iran, Belgium, Germany, Canada, India, Finland, Italy, the Netherlands, Russia, USA, Poland, Japan, Israel, China, France, South Africa, China (Taiwan), and UK. | Total 43 studies, relevant 10 studies: 6 CSS, 2CCS, 2 CS | Community noise exposure (road and air traffic) assessed by calculations, personal dosimeter, or sound level meter | RR per 5 dB(A) noise increase. For road traffic noise, the pooled RRs and 95% CIs were 1.09 (1.05–1.13) for IHD, 0.99 (0.84–1.16) for AP, and 1.03 (0.99–1.09) for MI. For air traffic noise, the pooled RR was 1.03 (0.90–1.18) for AP. | 25% | High risk of bias |
| van Kempen, et al.46 | European, 2000–October 2014, European countries | Total 61 studies, relevant 32 studies: 14 CSS, 5 ES, 8 CS, 5 CCS | Noise from road, rail, and air traffic and wind turbines: LDEN | Road, rail, and air traffic noise in relation to prevalence, incidence, and mortality of IHD and stroke were analysed, respectively. Number of studies for each analysis is small. Per 10 dB increase in exposure, the pooled RR and 95% CI of IHD was 1.08 (1.01–1.15) for road traffic. Estimates for other associations were of low quality or from <3 studies, and mostly insignificant. | NA | Low risk of bias |
| Vienneau et al.47 | Not specified (general population), January 1994–January 2014, country not restricted: Germany, UK, the Netherlands, Sweden, Switzerland, Denmark, Canada, and USA | 3 CCS, 5 CS, 2 LS | Transportation noise exposure | RR per 10 dB increase in exposure. The pooled RR and 95% CI for IHD was 1.06 (1.03–1.09). | 75% | High risk of bias |
| Weihofen et al.48 | General population, till 31 August 2017, country not restricted: USA, France (Paris, Lyon Saint, and Toulouse), Canada (Vancouver), UK (London), Switzerland, Germany (Berlin and Frankfurt), the Netherlands (Amsterdam), Sweden (Stockholm), Greece (Athens), and Italy (Milan) | 3 CSS, 1 ES, 4 CS, 1 CCS | Aircraft noise: LAeq, LDay, LNight, LDN, Lden, LDENAEI | Per 10 dB increase in Lden, the pooled RR and 95% CI of stroke was 1.013 (0.998–1.028) from seven studies. | 71% | Low risk of bias |
| Ambient temperature | ||||||
| Bunker et al.49 | Elderly (65+), 1 January 1975–24 July 2015, country not restricted: USA, Bangladesh, China (mainland, Taiwan, and Hong Kong), Portugal, UK, Denmark, Australia, Russia, Italy, Hungary, Brazil, Vietnam, Sweden, Thailand, Norway, South Korea, and Germany | Total 60 studies, relevant 47 studies; 47 LS | Ambient hot and cold temperature | Per 1°C temperature change, for heat, the pooled percentage changes and 95% CIs were 3.79 (3.40–4.18) for CVD mortality, 1.62 (0.24–3.03) for IHD mortality, 1.40 (0.06–2.75) for CeVD mortality, 0.33 (−0.09 to 0.75) for IS, −0.66 (−2.13 to 0.84) for ICH, −0.17(−0.96 to 0.63) for CeVD, −0.16(−2.05 to 1.77) for MI, and 0.30(−0.12 to 0.81) for CVD. For cold, the estimates were 1.84 (0.85–2.84) for CVD mortality, 0.45 (−0.01 to 0.91) for IHD mortality, 1.21 (0.66–1.77) for CeVD mortality, 3.63 (−3.94 to 11.8) for IS, 1.49 (1.04–1.94) for ICH, −0.46 (−1.12 to 0.2) for CeVD, 0.66 (−0.14 to 1.48) for MI, −0.80 (−2.21 to 0.64) for AP, −0.67 (−2.15 to 0.83) for HF, and −0.28 (−1.39 to 0.84) for CVD. | 73% | High risk of bias |
| Kofi Amegah et al. 2016 | Sub-Saharan African populations, till December 2014, Sub-Saharan Africa | Total 23 studies, relevant 5 studies: 4 LS, 1 CSS | Temperature | One study found that low temperature was associated with increased risk of CVD. Two studies found associations of low and high temperatures with CVD mortality. One study found no association between mean monthly temperature and CVD mortality. One study found 5°C change in the monthly mean temperature to be associated with decreased risk of hospitalization for venous thromboembolism, stroke, and acute MI. | 56% | High risk of bias |
| Ma et al. 2020 | Chinese population, January 2010–January 2020, country restricted to China | Total 175 studies, relevant 19 studies: 19 LS | (i) Every 1°C temperature increase/decrease beyond certain reference points (ii) Comparison between extreme temperatures and reference normal temperatures | Pooled RRs and 95% CIs of CVD were 1.089 (1.062–1.116) and 1.171 (1.125–1.218), respectively, for hot and cold temperatures as compared with normal temperatures. | 100% | High risk of bias |
| Moghadamnia et al.50 | General population, January 2000–31 December 2015, country not restricted: China (mainland, Taiwan, and Hong Kong), Australia, Thailand, Philippines, South Korea, Germany, and Spain | 26 LS | Ambient temperature | RR per 1°C change of temperature. For CVD mortality, the RRs and 95% CIs were 1.055 (1.050–1.060) for cold exposure and 1.013 (1.011–1.015) for heat exposure. Coefficient per 1°C change in mean annual temperature. For CVD mortality, the pooled estimates were 0.026 (−0.019 to 0.072) for cold exposure and 0.008 (−0.015 to 0.031) for heat exposure. | 96% | High risk of bias |
| Odame et al. 2018 | Rural population, till April 2018, country not restricted: Bangladesh (MATLAB), Czech Republic, and China (Naidong and Jiangzi in Tibet) | All 14 studies, relevant 3 studies: 3 LS | Daily mean temperature | Per 1°C increase, the pooled RR and 95% CI of CVD mortality was 1.111 (1.045–1.181). The associations were significant in subgroup analyses of both developing and developed countries. | 100% | High risk of bias |
| Turner et al. 2012 | Not specified (general population), time and country not restricted: South Korea (Incheon), USA, UK (London and Scotland), Europe, Australia (Brisbane), and Thailand (Muang) | Total 21 studies, relevant 18 studies: 18 LS | Effects of ambient temperature. Maximum, minimum, and mean daily temperature | RR per 1°C increase in temperature. The pooled RRs and 95% CI were 0.999 (0.982–1.016) for CVD morbidities, 0.990 (0.887–1.105) for stroke, and 1.010 (0.930–1.097) for ACS/MI. | 43% | High risk of bias |
| Wang et al. 2016 | Adults, till 16 October 2015, country not restricted: Japan, UK, Russia, Spain, China, Portugal, Italy, South Korea, China (Taiwan), Mozambique, USA, France, and Australia | 7 CS, 13 LS, 1 CSS | Ambient temperature | OR per 1°C increase in mean ambient temperature. The pooled ORs and 95% CIs were 0.97 (0.94–1.00) for ICH, 1.00 (0.99–1.01) for IS, and 1.00 (0.98–1.01) for SAH. The pooled estimates for ambient minimum and maximum temperature and IS were OR 0.99 (0.96–1.01) and 0.98 (0.94–1.02), respectively. | 19% | Unclear |
| Zafeiratou et al. 2021 | General population, 1990–2020, country not restricted: Serbia (Belgrade), China (Hong Kong), UK (London), USA (New England), Finland (Helsinki), Switzerland, and South Korea | Total 34 studies, relevant 6 studies: 4 ES, 2 CS | Mean annual temperature or variability, seasonal temperature or variability, annual temperature categories, and mean annual degrees above/below minimum mortality temperature | In temporal comparisons within the same area, increased cardiovascular mortality was associated with both increased and decreased temperature. Stronger association was found with cold rather than hot temperature. In geographical comparison in just one study, people living in areas with higher temperature were found a lower rate of IHD mortality, though no dose–response. | 5 studies on CVD mortality: 60% | High risk of bias |
| Multiple domains | ||||||
| Rugel et al. 2020 | Urban residents, 1 January 2003–November 2019, country not restricted: the Netherlands, Sweden (Skåne), UK (London), Canada, Denmark (Copenhagen and Aarhus), Germany (Bochum, Essen, and Mülheim/Ruhr), South Korea (Seoul, Ulsan, and Cheonan), France, Spain, Norway, Greece, Italy (Rome and Verona), Switzerland, USA (California), and China (Shenyang, Anshan, Jinzhou, and Taiwan) | Total 51 studies, relevant 21 studies: 15 CS, 5 CSS, 1 ES | Traffic-related air pollution; natural spaces, neighbourhood walkability; noise | Based on the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) system for assessing the quality of the synthesized evidence base, evidence was sufficient for higher noise exposure with increased CVD. And evidence was sufficient for no association between traffic-related air pollution and CVD. | NA | High risk of bias |
ES, ecological studies; CS, cohort studies; CSS, cross-sectional studies; CCS, case-control studies; CCR, case-crossover studies; SAS, small-area studies; longitudinal study, LS (e.g. panel study and time-series); RR, relative risk; HR, hazard ratio; MI, myocardial infarction; AP, angina pectoris; CHA, cardiovascular hospital admission; ACS, acute coronary syndrome; CVD, cardiovascular disease; CHD, coronary heart disease; AF, atrial fibrillation; IHD, ischaemic heart disease; CeVD, cerebrovascular disease; RHD, rheumatic heart disease; HRV, heart rate variability; SDNN, the standard deviation of NN intervals; RMSSD, root mean square of successive RR interval differences; LF, low-frequency bands; HF, high-frequency bands; RE, risk estimate; events include both morbidity and mortality; ICH, intracerebral haemorrhage; IS, ischaemic stroke; SAH, subarachnoid haemorrhage; HF, heart failure. The significant pooled results are in bold.
A bibliography of all included reviews is presented in Supplementary material online, Appendix S3.
There were no systematic reviews found in the domains of food environment and light pollution.
LAeq16hr, annual non-weighted 16 h average noise level during the day; LAeq24hr, the annual non-weighted day–night average noise level; LDEN, the annual weighted (day + 0 dB, evening + 5 dB, and night + 10 dB) day–evening–night average noise level; LAeq, A-weighted average of an energy-equivalent continuous sound level over a period of time (A-weighting: in noise research, typically, the A filter is used which adjusts for deep and high frequencies, as these are perceived as less loud); LDay, LAeq for the day (usually 7:00 a.m.–7:00 p.m.) for all day periods of a year; LNight, LAeq for the night (usually from 11:00 p.m. to 7:00 a.m.) for all night periods of a year; LDN, all 24 h LAeq periods of a year with additional 10 dB for nighttime noise annoyance (usually from 11:00 p.m. to 7:00 a.m.); LDENAEI, see LDEN, but in addition, weighted average exposure on municipal level.
This PDF is available to Subscribers Only
View Article Abstract & Purchase OptionsFor full access to this pdf, sign in to an existing account, or purchase an annual subscription.
