The purpose of this study was to investigate the relationship between the concentration of air pollutants and the occurrence of pneumonia. The main discovery of our research was that the risk of hospital visits for pneumonia increases with an IQR increment of atmospheric pollutant levels (PM2.5, PM10, SO2, NO2). The effects of air pollutants were slightly different in diverse subgroups.
The effect of atmospheric mainly arose in the cold period rather than the warm period. Younger children might be more sensitive to atmospheric pollutants. These findings could better explain the effect of air pollutants on human health.
Before this study, there have much research focus on the health effects of atmospheric pollution, especially PM. Many recent investigations found that the rise of the air pollutants levels had a positive relationship with the incidence and mortality of respiratory diseases such as asthma, upper respiratory infections, and pneumonia [24–27]. This research, like previous singular-source studies, discovered a connection in atmospheric pollutants and hospital visits with pneumonia. With an IQR increment of PM2.5, PM10, NO2, and SO2 levels, the odds ratio of ED visits with pneumonia for each air pollutant increased by 14.0%, 10.9%, 14.1%, and 4.5% in Kaohsiung City on lag three days [22]. Another study in Jinan also demonstrated the short-term effects of atmospheric particulates on pediatric patients with pneumonia, and PM2.5 and PM10 achieved the most significant impact, respectively, on the lag five days(OR = 1.06, 95%CI: 1,02-1.10) and lag two days (OR = 1.04, 95%CI: 1.00-1.08) [13]. Duan et al. [12] reported a significant relationship between increased PM2.5 concentrations and hospitalization for pneumonia in Shijiazhuang during 2013, and the effect of PM2.5 was highest on the day of admission(OR = 1.011, 95%CI: 1.005–1.017). Although studies have found a connection between atmospheric pollutants and pneumonia, the results on the hysteresis effect of pollutants were not the same. Besides, this study found a bimodal effect between atmospheric particulate matter and hospital visits for pneumonia. This result was similar to the discovery of Tecer et al. [28] which also found a bimodal effect of PM on respiratory disease. The reason why this phenomenon occurs might be due to the variety of the severity of the disease, lifestyle, and sensitivity to atmospheric particulates. Moreover, there is evidence that the socioeconomic status of a population could affect the exposure to atmospheric particulate matter, which in turn leads to different health burdens [29]. Sue et al. [30] once implied that there was a correlation between the chemical properties of air pollutants and the risk of hospitalization. The composition and sources of air pollutants in different regions were different, which may be the reason for the different lag effects of air pollution in different areas. Another possible reason might be the difference in the structure and characteristics of the population.
The mechanism of pneumonia caused by air pollution is not known adequately. Animal experiments have proved that exposure to atmospheric particulates would reduce the antibacterial ability of the lungs and aggravate the original pulmonary inflammation [31]. NO2 also could provoke changes in pulmonary immune function by causing damage to bronchial and alveolar epithelial cells [32]. PM could reduce the activity of pulmonary macrophages and epithelial cells, and play a role in promoting inflammation as well as oxidative stress [33]. Furthermore, Exposure to PM would inhibit pulmonary macrophages, which increases the susceptibility of the lungs to infection [34]. Besides, experiments had shown that exposure to atmospheric particles would cause the production of reactive oxygen species (ROS), then ROS would produce a series of cellular responses (Mitochondrial damage [35], inflammatory mediators release, apoptosis), and eventually lead to the occurrence of disease [36]. Except for atmospheric particulates, SO2 exposure also could induce lung mitochondrial function, resulting in cellular dysfunction and lung illness [37]. These studies implied that air pollution might cause pneumonia by inducing the abnormal immune function of the lungs and producing oxidative damage.
Gender and age differences had always been the focus of environmental epidemiological research. In this study, we found that the impact of atmospheric particulate matter on men and women was both statistically significant, but the difference between them was not significant. This result was consistent with some previous studies. An epidemiological survey conducted in Shenzhen found that air pollution affected the occurrence of respiratory diseases both on men and women, but there had no significant gender difference [38]. Similar results could be found in another study [39]. However, there were some studies with different results. Duan et al. [12] concluded that men were more susceptible to the effect of PM2.5 and PM10 on pneumonia hospitalization, but their interaction P value was not estimated. Some research detected that the effect of PM2.5 on pneumonia hospitalization was more influential in women than men [40, 41]. In terms of age stratification, infants less than 1-year-old were proved to more sensitive to PM2.5 and PM10 in a study [13]. Cheng et al. [22] found children older than four years were more vulnerable to PM2.5 than those younger than four years. In this study, we found that children younger than four years old seem more susceptible to atmospheric pollutants. The reason why younger children have a higher odds ratio might be due to the vulnerability of airways and alveoli, immature immune systems, and high rates of respiratory infections, which finally leads to their high sensitivity to the atmospheric pollutants. Most studies had inconsistent results in terms of age and gender. The possible reason for this might owing to the inclusion-exclusion criteria, the different composition of air pollutants, and the number of cases in various studies.
On the seasons, the concentration of air pollutants and the number of hospital visits for pneumonia both increased during the cold season. On one hand, this was because of the increased emissions of pollutants caused by winter heating and the transmission pollutants from the northwest Shandong. On the other, Dong et al. [42] found that there a clear seasonal trend in the total microbial level in Qingdao, and the total microbial concentration was significantly increased with the increase of the intensity of air pollutants in winter. This phenomenon might be a crucial reason for the rise in the number of patients who were admitted to the hospital for pneumonia in winter. Consistent with the time series results, the impacts of atmospheric pollutants on the occurrence of pneumonia mainly arose in the cold period. This result was similar to prior studies that found a more potent effect of air pollutions in winter[14, 43]. Low temperature would reduce the ability of the respiratory system to resist infection, which might relate to the decrease of cilia clearance ability of the respiratory system and leukocyte phagocytosis [44]. So, people might be more likely to catch pneumonia in winter, and the effects of atmospheric pollutants were more visible. However, in other people's research, they found that atmospheric pollutants were more noticeable during the warm or transitional season [12, 45]. The reason for this regional difference was possibly owing to seasonal changes in the composition of atmospheric pollutants, people's lifestyles in various regions, and local meteorological conditions. Besides, as the pollution was severe in winter in the Qingdao area, the high incidence of smog in winter might aggravate the harm of air pollution to the human body in Qingdao.
To our knowledge, previous research about pneumonia only added air pollutants into multi-pollutant models without considering the impact of strong correlations between air pollutants. However, the collinearity between atmospheric pollutants might lead to incorrect estimates or unstable models [46]. Besides, the air pollutants in China mainly come from the combustion of fuel, there are obvious homologous characteristics among these pollutants, so it has a slight significance to use the multi-pollutant model directly. Therefore, we used a principal component analysis to investigate the relationship between multiple atmospheric pollutants and hospital visits for pneumonia. In the principal component regression model, the effect of the first principal component which stands for the mixed pollutants was significantly on inducing the hospital visits for pneumonia. This possibly means that it was the mixed pollution (PM2.5, PM10, NO2, SO2) induced hospital visits for pneumonia, and atmospheric particulate matter was the most considerable factor of them. Previous research proved that the mixed effect was mainly driven by PM [47], perhaps this was the reason why PM was more vital in inducing hospital visits for pneumonia. Besides, the former epidemiology research found that PM2.5 was more toxic than PM10 [48]. This possibly the reason why the effect of PM2.5 was higher than PM10.
The strength of this research was that it covers a wide range of people while preceding studies in China have mainly concentrated on special source crowds. Also, the principal component analysis could better explain the effect of various atmospheric pollutants in causing pneumonia. There also have some shortcomings in this study: First, Due to the confidentiality of the data, information such as the patient's home address and behavior habits cannot be obtained, so it was impossible to assess individual exposure accurately; Second, because meteorological monitoring stations in China cannot collect the contents of atmospheric particulates, there still needs a toxicological experiment to verify which air pollutants play a leading role; Last, owing to the heterogeneity of various cities, the result of a single town was difficult to apply to other areas. Therefore, more experiments were needed in the future to verify the mechanism of air pollution on pneumonia occurs.