|Home | About | Journals | Submit | Contact Us | Français|
Indoor exposure to fine particulate matter (PM2.5) from outdoor sources is a major health concern, especially in highly polluted developing countries, such as China. Few studies have evaluated the effectiveness of indoor air purification on the improvement of cardiopulmonary health in these areas.
To evaluate whether a short-term indoor air purifier intervention improves cardiopulmonary health.
We conducted a randomized double-blind crossover trial among 35 healthy college students in Shanghai, China in 2014. These students lived in dormitories that were randomized into 2 groups and alternated the use of true or sham air purifiers for 48 h with a 2-week washout interval. We measured 14 circulating biomarkers of inflammation, coagulation and vasoconstriction, lung function, blood pressure (BP), and fractional exhaled nitric oxide (FeNO). We applied linear mixed-effect models to evaluate the effect of the intervention on health outcome variables.
On average, air purification resulted in a 57% reduction in PM2.5 concentration from 96.2 to 41.3 μg/m3 within hours of operation. Air purification was significantly associated with decreases in geometric means of several circulating inflammatory and thrombogenic biomarkers, including 17.5% in monocyte chemoattractant protein-1, 68.1% in interleukin-1β, 32.8% in myeloperoxidase and 64.9% in soluble CD40 ligand. Further, systolic BP, diastolic BP, and FeNO were significantly decreased by 2.7%, 4.8%, and 17.0% in geometric mean, respectively. The impacts on lung function and vasoconstriction biomarkers were beneficial, but not statistically significant.
This intervention study demonstrated clear cardiopulmonary benefits of indoor air purification among young, healthy adults in a Chinese city with severe ambient particulate air pollution.
(Intervention Study on the Health Impact of Air Filters in Chinese Adults; NCT02239744)
Numerous studies have demonstrated that ambient fine particulate (particulate matter with an aerodynamic diameter <2.5 μm, PM2.5) air pollution is significantly associated with increased cardiopulmonary morbidity and mortality (1,2). Furthermore, PM2.5 may exacerbate cardiopulmonary symptoms, sometimes even within hours of exposure, and may result in serious adverse outcomes, such as chronic obstructive pulmonary disease (3), myocardial infarction (4), heart failure (5), fatal arrhythmias (6), sudden cardiac arrest (7), and stroke (8). The underlying biological mechanisms for these associations are not yet well understood, although there are hypotheses implicating inflammation, impaired lung function, oxidative stress, increased arterial blood pressure (BP), endothelial dysfunction, blood coagulation, arterial vasoconstriction, and altered cardiac autonomic function (1).
China has 1 of the highest levels of ambient PM2.5 in the world. It was estimated that ambient PM2.5 contributed to over 1.2 million deaths and a loss of 24 million healthy years in China and ranked fourth among all contributions to the health burden (9). Outdoor PM2.5 can penetrate indoors. In China, urban residents on average spend 87% of their time indoors and this percentage may be even higher for vulnerable subgroups, such as young children and the elderly (10). Therefore, it is critically important to identify ways to effectively reduce indoor exposure to PM2.5 of outdoor origin. Unlike the outdoor environment, a substantial reduction of PM2.5 can be achieved in the indoor environment simply by using air filters, cleaners or purifiers. Indeed, several previous studies suggest potential benefits of indoor use of air filters on cardiopulmonary health (11–14). However, these studies were conducted in less polluted countries and empirical evidence from China is lacking.
Therefore, we conducted a randomized double-blind crossover study to examine whether short-term use of air purifiers improves cardiopulmonary health among healthy young adults in Shanghai, the largest city in China. We chose circulating biomarkers and lung function as the primary endpoints because they have been consistently associated with air pollution in observational studies (15–17). The secondary endpoints included BP and indicators of respiratory inflammation.
We recruited 35 healthy college students on the basis of sample size calculations for the primary endpoints, using noninferiority tests in the software PASS 11 (NCSS, LLC. Kaysville, Utah) (18,19). We set the significance level (alpha) at 0.05 with 80% power, and estimated the noninferior margins and standard deviations of primary endpoints on the basis of data from 2 quasi-experimental air pollution studies among healthy college students in China (17,20). Our study participants were from 10 dormitory rooms (each about 20 m2) in 2 adjacent buildings with 3 or 4 participants per room. All participants and their roommates were nonsmokers. Their rooms were thoroughly cleaned before the intervention to ensure that there were no observable indoor sources of air pollution. All subjects declared that they had no clinically-diagnosed chronic cardiopulmonary diseases.
The study was conducted in several weekends of 2014 at the Fenglin campus of Fudan University, located in the central urban area of Shanghai. This intervention was designed as a randomized double-blind crossover study. To be specific, the 10 rooms were randomized into 2 groups of 5 rooms each. One group used an air purifier placed in the center of the room for 48 h, corresponding to 2 weekends, followed by a 2-week washout period, and then another 48 h of using a sham air purifier under the same conditions. The other group simply reversed the order in which the real and sham air purifiers were used. All rooms used the same qualified air purifiers (model FAP04, 3M Filtrete), with the only difference being removal of the filter gauze in the sham purifiers. The air pollution auto-sensing feature of air purifiers was disabled in both groups. All participants and research staff were blinded to the group assignment. We requested all participants to stay in their dormitory room with the windows/doors closed throughout each 48-h intervention period. We delivered foods and drinks to each room during the intervention period. All interventions started at 8 a.m. to avoid issues related to diurnal variation. We evaluated health endpoints and drew blood immediately after the completion of each 48-h intervention.
The Institutional Review Board of the School of Public Health, Fudan University approved the study protocol and all participants provided written informed consent before enrollment. This study was registered at www.clinicaltrials.gov (NCT02239744).
We measured indoor and outdoor PM2.5 concentrations in real time using the TSI SidePak AM510 Personal Aerosol Monitor (TSI Inc., St. Paul, Minnesota). This monitor was installed indoors at least 1 m away from the air purifier. In addition, we also installed an outdoor monitor on the rooftop of 1 dormitory building to represent the outdoor PM2.5 levels of all rooms. Before the intervention, all devices were factory calibrated to Arizona Test Dust, and externally calibrated to a neighboring government-controlled ambient monitor that measured PM2.5 using the method of tapered element oscillating microbalance. At the beginning of each intervention day, we calibrated the Sidepak using the manufacturer-supplied zero filter and applied clean grease to the impactor. We also placed a HOBO data logger (Onset Computer Corporation, Pocasset, Massachusetts) to monitor indoor temperature and relative humidity. We recorded these environmental data on an hourly basis and used the 48-h average as the uniform exposure level for the 3 or 4 subjects in each room.
At baseline, we collected basic demographic information such as age and sex. We also measured height and weight to calculate the body mass index (BMI). Immediately after each intervention period, we drew a blood sample and measured the following health indicators:
Peripheral blood samples (5 ml) were drawn by a nurse, separated into serum and plasma, and stored at −80 °C within 30 min. We measured the levels of 14 circulating biomarkers: (1) 8 biomarkers of inflammation, including C-reactive protein (CRP), fibrinogen, P-selectin, monocyte chemoattractant protein-1 (MCP-1), interleukin-1β, interleukin-6, tumor necrosis factor-α (TNF-α) and myeloperoxidase; (2) 4 biomarkers of coagulation, including soluble CD40 ligand (sCD40L), plasminogen activator inhibitor-1, tissue plasminogen activator and D-Dimer; and (3) 2 biomarkers of vasoconstriction, including endothelin-1 and angiotensin-converting enzyme. We measured CRP, fibrinogen, P-selectin, MCP-1, TNF-α, myeloperoxidase, and D-Dimer using the Millipore MILLIPLEX MAP human cytokine/chemokine kit (Millipore Corp., Billerica, Massachusetts). We measured the other biomarkers using enzyme-linked immunosorbent assays. All biomarker tests were performed according to the manufacturer’s instructions.
A respiratory physician measured forced vital capacity, forced expiratory volume in 1 s (FEV1), and peak expiratory flow of each participant using JAEGER Masterlab equipment (Würzburg, Germany) that meets American Thoracic Society criteria. The volume signal was calibrated at least once on a testing day with a 3.0 l syringe connected to the pneumotachograph, in accordance with the manufacturer’s recommendations. We instructed participants to perform at least 3 forced expiratory lung function maneuvers in order to obtain a minimum of 2 acceptable and reproducible values, and we recorded the best results.
After sitting in a quiet room for at least 5 min, participants had their left upper arm BP measured by trained technicians using a mercury sphygmomanometer at least 3 times with 2-min minimum intervals between measurements. The second and third sets of readings were averaged to obtain systolic BP and diastolic BP. Pulse pressure was calculated as the difference between systolic BP and diastolic BP. If the differences among the 3 measurements were more than 5 mm Hg, a new round of measurements were arranged.
FeNO is an established biomarker of respiratory inflammation, and has been widely used in epidemiological studies because of its high sensitivity, specificity, and noninvasive nature (21). We measured FeNO levels using a portable NIOX MINO machine (Aerocrine AB, Solna, Sweden) according to standardized procedures recommended by the American Thoracic Society and the European Respiratory Society.
Because the distributions of the health outcome variables were all skewed, we log-transformed them before statistical analysis. To examine potential effects due to the order of intervention, we compared health endpoints between the 2 groups with different treatment order using the 2-sample Wilcoxon rank-sum (Mann-Whitney) test (19). To account for the repeated measurement of health endpoints under the 2 experimental scenarios, we applied linear mixed-effect models to investigate the effect of air purification on outcome variables (12). This automatically allows each subject to serve as his/her own control over time, and also adjusts for between-subject covariates that do not change over time. The intervention was coded as a dummy variable (i.e., 1 for true-purified scenario and 0 for sham-purified scenario) and was analyzed as a fixed effect in the model. We incorporated random intercepts for subjects to account for intraindividual correlations between repeated measurements, as well as interindividual correlations of repeated measures in each room (14). We also controlled for the following variables as fixed-effect covariates: age, sex, BMI, indoor temperature, and indoor relative humidity. As the trial was completed in about half a month, we did not control for the temporal trends of the health measurements. Furthermore, ambient gaseous air pollutants did not confound the analyses, as their concentrations were exactly the same between the 2 scenarios due to this 1:1 crossover design. Similarly, the potentially lagging confounding effects of ambient air pollution and temperature may also be excluded from the final models. We calculated the effect of air purification as a percent change of the geometric mean and its 95% confidence interval (CI) in a health endpoint comparing the true-purified air scenario with the sham intervention.
As a sensitivity analysis, we replaced the dummy variable of the intervention with indoor PM2.5 concentrations in the above model to examine whether an empirical decrease in indoor PM2.5 could lead to a change in health indicators.
All statistical tests were 2-sided with alpha = 0.05. All analyses were conducted using the “lme4” package of R software (Version 2.15.3; R Development Core Team).
Study participants were 25 females and 10 males with a mean age of 23 ± 2 years and an average BMI of 22 kg/m2. All participants completed this study. According to the self-administrated questionnaire, they stayed indoors almost the entire time, and stayed within the central urban area of Shanghai during the washout period. Furthermore, all participants remained healthy throughout the study period.
Before the intervention, the average PM2.5 concentration was comparable between the 2 groups. PM2.5 concentration was markedly reduced within the first 4 h of using the true air purifiers and remained stable over the rest of the 48-h period (Figure 1); in contrast, PM2.5 concentration in the sham-purification group was barely reduced, suggesting that closing the windows and doors did not efficiently block the penetration of outdoor PM2.5.
Table 1 summarizes the indoor and outdoor air pollutant concentrations and meteorological parameters during the study periods. The average outdoor concentrations of PM2.5 were 103 μg/m3, which were much higher than those in North America and Western Europe. The indoor PM2.5 concentration in rooms with sham air purifier (96.2 μg/m3) was only slightly lower than outdoor levels. In contrast, the mean PM2.5 concentration in rooms with a true air purifier was greatly reduced to 41.3 μg/m3, 57% lower than those of the sham group.
There were appreciable decreases in the levels of circulating biomarkers, BP, and FeNO in the true-purified air scenario compared to the sham-purified air scenario (Table 2). However, the differences in lung function indicators between the 2 scenarios were not significant. The Wilcoxon rank-sum tests did not show any statistically significant differences in all health indicators between orders (p values ranging from 0.11 to 0.93), suggesting that there were no order effects or interactions between period and order.
In the mixed-effect model analysis, compared with participants in the sham purification group, those assigned to true air purification showed decreased levels of 4 blood biomarkers, BP, and FeNO, although nonsignificant improvement was also observed for lung function and several other blood biomarkers (Central illustration and Table 3).
All biomarkers of systematic inflammation, coagulation and vasoconstriction decreased in response to the air purification intervention, although not all were statistically significant. The intervention had significant effects on 3 of 8 inflammation markers, and 1 of 4 coagulation markers, and no significant effects on 2 vasoconstriction markers. The magnitude of the effects varied by biomarkers. For example, the intervention led to a significant geometric mean decrease of 17.5% (95% CI: 5.5% to 30.8%) in MCP-1, 68.1% (95% CI: 44.3% to 81.7%) in interleukin-1β, 32.8% (95% CI: 5.3% to 67.5%) in myeloperoxidase, and 64.9% (95% CI: 30.3% to 82.3%) in sCD40L. Systolic and diastolic BP was significantly decreased by 2.7% (95% CI: 0.4% to 5.1%) and 4.8% (95% CI: 1.2% to 8.5%) in geometric mean, respectively. However, pulse pressure was not altered with the introduction of air purifiers.
FeNO level was significantly decreased by 17.0 % (95% CI: 3.6% to 32.5%) in geometric mean in the air purification intervention group. There was some indication of improved lung function associated with this intervention, but no evidence of statistical significance was observed.
Overall, as indicated in Table 4, the sensitivity analysis showed positive associations of continuous exposure to indoor PM2.5 with circulating biomarkers, BP, and FeNO, and inverse, but nonsignificant associations with lung function. Inconsistent with the main analyses, lower indoor PM2.5 exposure was significantly associated with lower diastolic BP, but not systolic BP; furthermore, unlike the main analyses, PM2.5 was inversely associated with tissue plasminogen activator, but its association with MCP-1 did not reach statistical significance.
China has 1 of the worst air pollution levels in the world and the government and society are making significant efforts to reduce air pollution levels. Research on easy and effective means of reducing air pollution in China may therefore have substantial public health implications. In this crossover intervention study in Shanghai, China, we found that closing doors/windows barely reduced indoor PM2.5 concentration. In contrast, indoor use of air purifiers efficiently reduced indoor PM2.5 within hours of operation. Furthermore, this improvement in air quality resulted in improved cardiopulmonary function in 35 generally healthy college students. To the best of our knowledge, this is the first study to examine the impact of short-term purification of indoor air on clinical and biochemistry measures of cardiorespiratory health in areas with severe air pollution.
Short-term indoor air purification may have cardiovascular benefits. In the current study, we found a modest, but statistically significant decrease in BP after the intervention. This finding is in contrast to those from several previous air filtration studies in countries with cleaner air (11,12,14), suggesting that these cardiovascular benefits may be more easily achieved in regions with severe air pollution problems. Moreover, with indoor air purification, we also observed improvements in nearly all of 14 examined blood biomarkers of inflammation, coagulation, and vasoconstriction, particularly for proinflammatory proteins, such as MCP-1, interleukin-1β, and myeloperoxidase, but not CRP. This nonsignificant effect on CRP is consistent with 2 home-based air filtration studies in Copenhagen (12,13), but conflicts with another similar study in a woodsmoke-impacted community in Canada (11). In the intervention period, we found a significant decrease in circulating sCD40L, a surface adhesion molecule involved in both inflammatory and thrombogenic processes (22). A prior air filtration study also reported reduced expression of other adhesion molecules (13). However, the present study did not demonstrate a significant benefit for the remaining biomarkers, whereas these associations were reported in several observational panel studies (1,23,24). Consistent with another air filtration study among 37 Canadian residents (14), the nonsignificant effects on the majority of blood biomarkers may be explained by the relatively small sample size, making the effect estimates more imprecise.
Our data also suggest beneficial effects of air purification on respiratory health. For example, we found a 17.0% reduction in FeNO, an established biomarker of airway inflammation that can be incipiently induced by the inhalation of PM2.5 (25). The current study show modest, nonsignificant beneficial impacts on lung function, probably due to the short intervention period. Two previous intervention studies, both conducted in areas with lower ambient air pollution, reported conflicting results on indoor air purification and lung function. In a study in Denmark that reduced PM2.5 from 8 to 4 μg/m3, there was no significant association of 2 to 14 days of indoor air purification with lung function improvements (13). In the second study, 7 days of indoor air purification was associated with significantly improved lung function in young adults in Manitoba, Canada (14).
Taken together, results from this interventional study demonstrated clear, albeit modest, cardiopulmonary benefits of using an indoor air purifier in China. Because our study participants are healthy young adults, one could reasonably expect similar or even larger cardiopulmonary benefits of air purification among vulnerable populations, such as young children or older adults. Furthermore, the potential for additional benefits with a longer intervention can be expected and should be investigated. The use of air purifiers offers ordinary citizens a feasible and affordable way to reduce exposure to hazardous air pollution in a highly-polluted developing country, such as China, leading to significant public health benefits.
This study contributes to our understanding about potential biological mechanisms of PM2.5 and health by examining a short-term intervention of reducing PM2.5 exposure and a range of clinical and biochemical markers. Previous human studies have linked short-term PM2.5 exposure to a wide range of adverse cardiopulmonary endpoints (1,26,27). However, most of these were observational panel studies, which evaluated changes in health indicators in association with natural day-to-day variations in ambient PM2.5 concentrations. Residual confounding may occur due to a number of individual characteristics and environmental factors, such as weather conditions. Therefore, controlled-exposure interventional studies offer a viable alternative to investigate the causal relationship between ambient PM2.5 and health parameters (28), but this kind of evidence is largely lacking. Overall, our results supported the biological mechanisms proposed by previous observational studies (1).
Our study has strengths. First, this randomized double-blind crossover design facilitated causal inferences of air purification, reduction of air pollutants, and improved health indicators. Secondly, we measured more indicators of cardiorespiratory health than previous studies, allowing a more systematic assessment of potential health benefits of reducing indoor air pollution. Thirdly, the experimental environment was well controlled in that only PM2.5 was probably different between the 2 groups, and exposure measurement error was minimized. Finally, as this intervention study was completed over a 2-week period, we avoided potential temporal confounding due to seasonal changes and changes in participants behavior.
Our study also has limitations. First, the study included only 35 participants in 10 rooms. We might therefore have missed some potentially important, but modest differences due to the relatively small sample size. Secondly, because of the relatively small sample size and examination of multiple health endpoints, our study was considered exploratory in nature. Thirdly, we chose to enroll college students from school dormitories with a short intervention period in order to better control for potential confounding that might have been difficult to control in other study settings (e.g., indoor cooking, smoking, medication use, and individual health status). However, this study strategy might have limited the generalizability of our study results, and have led us to underestimate or miss some potential health effects (such as lung function) that may be observed with long-term air purification and/or in more vulnerable populations. Fourthly, we did not monitor indoor gaseous air pollutants, and were therefore unable to make a direct inference between PM2.5 reduction and the observed health benefits. Because the dormitory rooms were almost identical, except for the intervention, and because air purifiers could not eliminate gaseous pollutants, we believe that this is not a major concern,.
This intervention study demonstrated clear cardiopulmonary benefits of indoor air purification among young, healthy adults in a Chinese city with severe ambient particulate air pollution. Future studies should further evaluate the potential health benefits of long-term air purification among more vulnerable populations, such as children, older adults, or individuals with cardiopulmonary diseases.
China has 1 of the worst air pollution levels in the world. This randomized double-blind crossover study in China demonstrated that a 2-day intervention of indoor air purification could significantly reduce indoor PM2.5 concentration and improve cardiopulmonary health.
The study may offer a practical way to reduce the potential adverse health effects of air pollution in a country with 1 of the worst air pollution problems in the world. Findings from this study, if confirmed and extended to vulnerable populations, such as children, older adults, and patients with existing cardiovascular and pulmonary conditions, will have significant clinical and public health implications.
Disclosures: The authors have reported that they have no relationships relevant to the contents of this paper to disclose. The National Natural Science Foundation of China (81222036), China Medical Board Collaborating Program (13–152), Cyrus Tang Foundation (CTF2013001), and Public Welfare Research Program of National Health and Family Planning Commission of China (201402022) supported this study. The intramural research program of the U.S. National Institutes of Health, National Institute of Environmental Health Sciences supports Dr. H. Chen.