Preplanned Studies: Baseline Investigation on Residential PM_2.5 Pollution of General Living Scenarios — 12 Cities, China, 2018

Short-term and long-term exposure to outdoor airborne fine particulate matter (particles with aerodynamic diameter ≤2.5 μm; PM_2.5) can increase the morbidity and mortality of cardiovascular and respiratory diseases (1). Considering time-activity patterns, people spent 85%–90% of their daily time in households (2), and for indoor PM_2.5 pollution, some studies were reported from many countries with a focus on special conditions such as residential ventilation, biofuel combustion, environmental tobacco smoke, etc (3-4). With rapid urbanization, residential PM_2.5 pollution still remains to be studied in many cities of China, especially under daily general living circumstances.

To explore the representative levels, characteristics, and influencing factors of residential PM_2.5 pollution in China, the National Institute of Environmental Health (NIEH) of China CDC initiated a multicenter investigation for indoor air pollution in 2018. The study selected cities and resident families based on two-step random sampling. Twelve representative cities were selected considering various factors such as climatic and geographical locations. These cities were mainly from the Northeast (Harbin and Panjin), Northwest (Lanzhou and Xi’an), Southwest (Mianyang), North (Shijiazhuang), East (Wuxi, Ningbo and Qingdao), Central (Luoyang), and South (Nanning and Shenzhen) of China. The selected cities within the zone of temperate climate were Harbin, Panjin, Qingdao, Shijiazhuang, Lanzhou, Luoyang, and Xi’an, and the selected cities within the zone of subtropical climate included the other five cities. The sampling dates were in the cold season (December) and the warm season (June) in 2018. Samples from families of Xi’an and Mianyang were collected only in the cold season because of unexpected interruptions to the field investigation.

Resident families in each city were randomly selected from one district downwind of the city center and another district located upwind. At least 25 families in each district were identified as confirmed target households with the following inclusion criteria: 1) families lived in the house for more than 3 years without plans to move away in the next 3 years; 2) families included at least one infant; and 3) families were willing to participate in this investigation. Families including individuals engaged in occupations with high health risks caused by environmental pollution and with any individuals smoking regularly were excluded.

In order to collect representative data of residential air quality, samples from bedrooms and living rooms of each family were collected by local CDCs. To collect the data of residential pollution with many parameters such as PM_2.5 related to the unified general living scenario of each household, the sampling condition of each family was regulated with the following guidelines: 1) the doors and windows were pre-closed for 12 hours for indoor sampling; 2) air conditioners, fans, and other equipment that may interfere with airflow were all turned off during the sampling period; 3) the period between 9 AM to 10 AM was chosen as the preferred sampling time to avoid traffic peaks and indoor cooking; 4) behaviors like indoor smoking were prohibited and temporary visiting guests were avoided for sampling in all families under the general living circumstances; 5) the height of the measuring point was 1 to 1.5 meters above the ground; and 6) the measuring point was not less than 0.5 meters from the wall. Indoor air quality indicators in this study included temperature, humidity, PM_2.5, PM₁₀ (particles with aerodynamic diameter ≤10 μm; PM₁₀), etc. PM_2.5 and PM₁₀ were monitored by calibrated light-scattering dust meters. The average value of 10 monitoring values recorded on a 5 minute interval was consistently taken as the final representative concentration for a site in living rooms and bedrooms. Approximately 3% of the total rooms were monitored repetitively as the parallel sites, and the sampling method had acceptable repeatability. The family members were interviewed about the lifestyles and living conditions of households with questionnaires. Informed consent forms were signed before the investigation. The sample size of target families was calculated based on a cross-sectional study design. Finally, after accounting for data censoring and the lack of coordination of the families, a total of 642 families were identified to evaluate residential air pollution in 12 major cities in China.

In 2018, the population in all selected cities was exposed to a residential PM_2.5 average concentration of 79.34 μg/m³, and the PM_2.5 concentration distribution in each city was also shown (Table 1). A significant difference was found in PM_2.5 concentrations among 12 cities (F=72.13, p<0.001). The concentration of residential PM_2.5 in two seasons was significantly different in 7 representative cities including Harbin, Panjin, Qingdao, Shijiazhuang, Lanzhou, Luoyang, and Wuxi (p<0.05) (Figure 1).

Figure 1. 

Spatial and seasonal distribution of residential PM_2.5 of 12 cities in China, 2018 (μg/m³).

Residential PM_2.5 concentration in this study was not normally distributed, so we estimated the odds ratio and corresponding 95% confidence interval (95% CI) for each potential risk factor by using the generalized linear model (GLM) with Poisson connection function (Table 2). After adjusting for the influence of PM₁₀, the concentrations of residential PM_2.5 in warm climate zones were found to be likely higher than those in the cold climate zone (OR₂=1.4454, 95% CI: 1.3973–1.4951). Statistically significant associations were found with residential PM_2.5 and physical indicators including temperature (OR₂=0.9436, 95% CI: 0.9406–0.9465) and humidity (OR₂=0.9906, 95% CI: 0.9894–0.9918). For the residential environment, households more than 1 kilometer away from the road had a lower residential PM_2.5 concentration than those less than 1 kilometer (OR₂=0.7511, 95% CI: 0.7103–0.7937).

Architectural characteristics showed potential influences on residential PM_2.5. As to the layers of window glass, the PM_2.5 concentrations in families with more than two layers were higher than those of less than two layers (OR₂=1.2841, 95% CI: 1.1694–1.4066). In addition, the PM_2.5 concentrations in households between the 5th and 10th floors (OR₂=1.1102, 95% CI: 1.0788–1.1425) and above 10th floor (OR₂=1.1616, 95% CI: 1.1276–1.1965) were higher than those in households below the 5th floor. Compared with bungalows, the residential PM_2.5 concentrations were higher in buildings (OR₂=1.3860, 95% CI: 1.3028–1.4760).

Some family-related information and lifestyle habits may also influence residential PM_2.5. As for the influence of family economic status, compared with families with an annual total income of fewer than 100,000 RMB (roughly 14,300 USD), households with an annual income of 100,000 to 200,000 RMB (OR₂=0.8209, 95% CI: 0.8004–0.8419) and an income of more than 200,000 RMB (OR₂=0.8074, 95% CI: 0.7778–0.8379) had lower residential PM_2.5 concentrations. The study also found that the average living area of family members showed a positive correlation with residential PM_2.5 concentration (OR₂=1.0020, 95% CI: 1.0008–1.0032). As for lifestyle habits, the PM_2.5 concentrations of households that never use air purifiers were significantly higher than those families using air purifiers (OR₂=1.0856, 95% CI: 1.0551–1.1170). The frequency of using the range hood in the kitchen was also correlated with residential PM_2.5. The PM_2.5 concentrations in households using range hoods frequently (OR₂=0.8864, 95% CI: 0.8207–0.9588) and occasionally (OR₂=0.5241, 95% CI: 0.4366–0.6253) were lower than those that never used the range hoods. Moreover, households without carpets showed lower PM_2.5 concentrations than those with carpets (OR₂=0.8027, 95% CI: 0.7699–0.8372), and households that never grow plants had a higher PM_2.5 concentration than those grow plants (OR₂=1.0284, 95% CI: 1.0020–1.0555).