Air pollution and venous thrombosis: a meta-analysis

Exposure to air pollution has been linked to cardiovascular and respiratory disorders. However, the effect of air pollution on venous thrombotic disorders is uncertain. We performed a meta-analysis to assess the association between air pollution and venous thrombosis. PubMed, Embase, EBM Reviews, Healthstar, Global Health, Nursing Database, and Web of Science were searched for citations on air pollutants (carbon monoxide, sulfur dioxide, nitrogen dioxide, ozone, and particulate matters) and venous thrombosis. Using a random-effects model, overall risk estimates were derived for each increment of 10 μg/m3 of pollutant concentration. Of the 485 in-depth reviewed studies, 8 citations, involving approximately 700,000 events, fulfilled the inclusion criteria. All the main air pollutants analyzed were not associated with an increased risk of venous thrombosis (OR = 1.005, 95% CI = 0.998–1.012 for PM2.5; OR = 0.995, 95% CI = 0.984–1.007 for PM10; OR = 1.006, 95% CI = 0.994–1.019 for NO2). Based on exposure period and thrombosis location, additional subgroup analyses provided results comparable with those of the overall analyses. There was no evidence of publication bias. Therefore, this meta analysis does not suggest the possible role of air pollution as risk factor for venous thrombosis in general population.

Scientific RepoRts | 6:32794 | DOI: 10.1038/srep32794 Databases. We searched Pubmed, Embase, EBM Reviews, Healthstar, Global Health, Nursing Database, and Web of Science using the terms related to the type of exposure (air pollution, ozone, carbon monoxide, nitrogen dioxide, sulfur dioxide, PM 10 , and PM 2.5 ) and to the type of outcome (venous thrombosis, venous thromboembolism, deep vein thrombosis, pulmonary embolism, and pulmonary thromboembolism). The full search criteria are provided in the Supplemental Material. Furthermore, we supplemented citations by cross checking the reference lists of eligible studies and relevant reviews to identify additional published and unpublished data.

Study selection and data extraction.
All studies in human beings that presented original data and were published in full text, or meeting abstract were eligible for inclusion, with no restrictions on publication date, language, or ethnicity. We excluded animal studies, ex vivo and toxicological studies, commentaries and editorials, case reports, and studies with no original data. If a citation is lacking of enough quantitative data and these essential data could not be obtained from the correspondent author, the study was excluded. Two investigators (L.T. and Q.Y.W.) screened all citations for potentially eligible studies and extracted data independently. Disagreements were adjudicated by a third investigator (B.H.).
Study quality assessment. The Newcastle-Ottawa Scale, with some modifications was adapted to judge study quality, according to validated scales in previous studies and the Cochrane Collaboration 22,23 . We evaluated 10 items as follows: (1) VT diagnosis: We considered the diagnosis to be validated if it was coded according to the International Classification of Diseases or based on objective investigations (color Doppler ultrasonography or vein angiography for deep vein thrombosis, CT angiography or ventilation/perfusion lung scan for pulmonary embolism). For cases reported in registries, we considered the diagnosis as validated. (2) Pollutant measurement: Good quality was considered if measurement was performed at least daily. (3) Study area: Complete evaluation of study area was referred to as a study where the unit of analysis for the exposure matched that of the outcome. (4) Total population: Good quality was considered if the study population was not limited to some special groups. (5) Multiple lags: Multiple lags were defined as studies that evaluated pollutant levels in a distributed lag model beyond lag of 1 day (in short-term analyses) or 1 year (in long-term analyses). (6) Repeated events: Good quality was considered if repeated events were controlled. (7)-(9) Temperature, Time trends, and Season: Good quality was considered if adjustment has been made. (10) Other factors: Good quality was considered if additional confounding factors were adjusted, such as humidity, day of week, smoking, body-mass index, and cancer. For each item, 1 was given to good quality, while 0 was given to low quality. Data synthesis. Adjusted odds ratios (OR) for venous thrombosis were pooled for a standardized increment in pollutant concentration of 10 μ g/m 3 . This level is the one that is used most frequently. If the increment in pollutant concentration was not 10 μ g/m 3 , standardized risk estimates were calculated using the following formula: OR (standardized) = OR Increment (10)/Increment (original) . Most studies have verified a linear relation between air pollutants increases and venous thrombosis risk. Due to the significant heterogeneity between studies, we estimated RRs and 95% CIs using a random-effects model. To further assess heterogeneity, subgroup analyses were performed. Statistical heterogeneity across the studies was examined using the standard I 2 statistic and Q-test. Because this test has limited power when the number of studies is small, the presence of heterogeneity was considered at a significance level of 0.10. Publication bias was assessed using Egger's regression test and Begg's test 24 . All tests were 2-sided and statistical significance was defined as P ≤ 0.05. Analyses were performed with Stata SE/MP 11.0 (StataCORP, College Station, TX, USA).

Results
Included studies. Our initial search yielded 1,250 potential literature citations ( Fig. 1). After screening abstracts and in-depth review, 21 citations were selected for further evaluation. Of these, additional data were requested from the authors of 3 citations but obtained for only 2. The remaining one citation with insufficient data was excluded. Altogether 8 citations fulfilled the inclusion criteria [25][26][27][28][29][30][31][32] . Interrater agreement for study selection was high (κ = 0.93). The 8 eligible studies included 2 time-series studies, 3 case-crossover studies, 2 prospective cohort studies, and 1 case-control study. Although different methods and study designs were employed, it is reasonable to analyze these studies together because there is a common exposure (air pollution) 33 .
The characteristics of included studies are shown in Table 1. The number of patients or events per study ranged between 302 and 605,242. The study population was predominantly the general population, with the exception of one study that focused on post-menopausal women 28 . Because only two studies (Milojevic 2014 andDales 2010) investigated the association between CO/SO 2 /O 3 and venous thrombosis 26,30 , these pollutants were excluded from further analysis. One study evaluated both short-term and long-term effects of air pollution 28 , while one study assessed deep vein thrombosis and pulmonary embolism, respectively 30 . Thus, data from the two studies were extracted separately for subgroup analyses.

Discussion
This meta-analysis is the first to assess the association between exposure to major air pollutants and venous thrombosis risk. According to the overall analysis and the subgroup analysis, no significant association was observed between the three air pollutants and venous thrombosis. These results are really unexpected, since a considerable number of experimental studies conducted in animals and in humans have suggested that air pollution may increase the susceptibility to venous thrombosis 34 .
Animal studies demonstrated that exposure to particulate air pollutants could lead to the activation of platelets, the increase in hemostasis factors, histamine release, and the heightened thrombus formation [35][36][37] . Experiments in humans showed a similar phenomenon. Elevated plasma viscosity was associated with acute changes in PM concentrations, as observed in 3256 randomly selected participants 38 . Healthy adults exposed to traffic-related pollutants had increased plasma levels of homocysteine, an established risk factor for arterial and venous thrombosis 39 . Ambient PM 10 levels have also been related to platelet aggregation among healthy persons, even 1 to 4 days after exposure 40 . Studies conducted in healthy persons as well as in individuals with venous thrombosis found that prothrombin time was shortened in relation to high levels of ambient PM 10 32,41 . Nevertheless, our work was consistent with evidence from another recent animal study 42 . Using a murine model, the experimental study showed that acute exposure to PM triggered primary hemostasis activation without substantial secondary hemostasis activation, resulted in arterial but not venous thrombosis. Potential explanations for the discrepancy among the reported studies are described as follows. First, the protocol details greatly vary in regard to the exposure duration, pollutants concentrations, PM compositions, animal models, and preexisting susceptibility. Second, the procoagulant and thrombotic effects of air pollutants were relatively mild, even at concentrations that were much higher than those of the real daily air pollutants. The procoagulant effects observed in experimental studies were too weak to give rise to a pathological venous thrombosis in healthy human beings. Potential limitations of this study should be considered. First, compared with the investigations on the associations between air pollutants and cardiovascular diseases, the number of available studies in this meta-analysis is small (8 citations). This number is relatively small even though all the major databases were searched with no other restrictions. However, the combined 95% CIs were narrow, which was attributed to the large absolute number of participants and events. Moreover, all of the included studies were of high quality (Table S1). For example, all these studies used multi-pollutant models. Therefore, the findings in this work should be solid. Second, most of the included studies were primarily focused on particulate matters. It is not clear whether other types of air pollution (CO, SO 2 , and O 3 ) will have adverse effects on venous thrombosis. Third, to date, the studies reported in literature and those considered in the meta-analysis, are characterized by intrinsic bias. This intrinsic bias includes "exposure bias" due to the difficulty to estimate the personal exposure of each subject enrolled 43 , and "population bias" due to the limited information regarding some peculiar aspects of the study population 44 . Fourth, our conclusion was different from a recently published paper 45 , in which a relationship between air pollution and venous thrombosis was considered. However, in that study, only a systematic review, but no pooled analysis was conducted. In our study, the pooled analyses showed no association in various subgroups (different pollutants, long term or short term effects, for PE only). In addition, we excluded one study 46 without enough data and another study comparing unprovoked PE with provoked PE 44 . The two excluded studies were of case-control design, with a sample size of only 2,613 or 105. Even if these two studies were included, we believed that they would not change our conclusion, because they would have a very low weight in the pooled analysis.
It is worth noting that all of the available studies were conducted in Western countries, where the median PM 2.5 concentration is less than 20 μ g/m 3 . In developing countries, however, the PM 2·5 concentrations in urban cities are likely to be up to 100-200 μ g/m 3 , such as Beijing, Wuhan, and most of the other large cities in Asia ( Figure S1). Will such a high level of air pollution have a marked effect on human blood coagulation and result in venous thrombosis tendency? Thus, relevant studies from these countries, including China, are warranted in the future.

Conclusions
In conclusion, despite being an important adverse factor for cardiovascular diseases, air pollution may not be a risk factor for venous thrombosis in the general population. More epidemiologic studies are urgently needed to establish the association between air pollution and venous thrombosis in middle-income and low-income countries.