Introduction
The World Health Organization (WHO) estimates that there are more than 3.8 million annual deaths attributable to indoor air pollution, with a great burden of disease in developing and developed countries (
Lim et al. 2012;
Thomas et al. 2015;
Asikainen et al. 2016). Associations between exposure to various forms of indoor air pollution and lung function outcomes in developed countries have been previously explored (
Mendell et al. 2011;
Lane et al. 2013;
Khreis et al. 2017;
Park et al. 2017). For instance, living in a home with active mold exposure has been associated with asthma development and exacerbation as well as diagnoses of asthma, dyspnea, wheeze, cough, respiratory infection, allergic rhinitis, and other upper respiratory tract symptoms (
Mendell et al. 2011). Home proximity to roadways has been used as an exposure proxy for pollutants emitted through automobile traffic (
Lane et al. 2013;
Khreis et al. 2017;
Park et al. 2017), which has been associated with reduced lung function growth in school-aged children in the United States and Sweden (
Gauderman et al. 2007;
Schultz et al. 2012). Additionally, exposure to environmental tobacco smoke (ETS) (
Chen et al. 2014;
Hu et al. 2017;
Milanzi et al. 2017) allergens is known to play a role in respiratory irritation. Potentially, exposure to endotoxins, pets, pests, and certain microorganisms could also play a similar role (
Bertelsen et al. 2010).
Of course, children possess an increased risk for adverse health effects associated with household air pollution because they have a higher per-minute respiratory rate, underdeveloped immune and respiratory systems, and longer home-stay relative to adults (
Bennett et al. 2007;
Siezen et al. 2009;
McInnes et al. 2011). While numerous epidemiologic studies have assessed associations between exposure to ambient air pollution and lung function decline in children (
Gauderman et al. 2002;
He et al. 2010;
Roy et al. 2012;
Zeng et al. 2016), less is known about the adverse respiratory health outcomes associated with exposure to indoor air pollution in children. This is particularly true of those in developing countries.
Few studies have investigated adverse respiratory health effects associated with exposure to household sources of indoor air pollution in Chinese children. The hypothesis of this investigation was that housing characteristics play a role in child lung function. It examined the associations between lung function measures and exposure to various forms of indoor air pollution in school-aged Chinese children using data from the Seven Northeast Cities (SNEC) Study.
Results
Spirometric means of lung function and standard deviations were calculated for 3382 male and 3358 female participants. A complete list of population characteristics is available in
Table 2, with differences noted by gender.
Calculated odds ratios
Group comparisons of lung function across housing characteristics suggested several statistically significant relationships.
Figure 1 displays odds ratios (ORs) for FVC and FEV1.
Figure 2 shows results for PEF and MMEF. “Ping-fang” housing (the equivalent of a single-family home) had a prevalence rate (PR) of 10% (
Table 1) and was related to increased odds of FEV1 deficit (OR = 1.36, 95% CI: 1.03–1.80) when compared with “dan-yuan-lou-fang” housing (an apartment unit in a multi-story, multi-unit building). However, the “other” housing (PR = 3%) indicated greater FVC volumes (OR = 0.47, 95% CI: 0.26–0.83) compared with dan-yuan-lou-fang housing. Greater housing age was generally associated with better lung function. Residents of homes that were 5–10 years old (PR = 17%) yielded greater FVC volumes (OR = 0.75, 95% CI: 0.57–0.99) and FEV1 volumes (OR = 0.68, 95% CI: 0.51–0.93). Residents of homes that were 10–20 years old (PR = 47%) had greater FVC volumes (OR = 0.75, 95% CI: 0.60–0.94), FEV1 volumes (OR = 0.66, 95% CI: 0.51–0.84), and MMEF rates (OR = 0.78, 95% CI: 0.60–0.99). Finally, residents of homes >20 years old (PR = 23%) had greater FEV1 volumes (OR = 0.74, 95% CI: 0.56–0.97). All of these values reflecting home age were in comparison with houses <5 years old. Participants residing in dwellings within a floor area range of 15–25 m
2 had greater PEF rate (OR = 0.73, 95% CI: 0.57–0.93), whereas those within 35 m
2 or greater had higher MMEF rate (OR = 0.69, 95% CI: 0.49–0.98) when compared with homes with floor areas <25 m
2 (PR = 10%). No group differences were observed for story level or number of rooms.
Environmental conditions proximate to the dwelling contributed to low performance in spirometric measures. Participants residing in homes that were within 20 m of a major roadway (PR = 19%) were more likely to have increased odds of lung function deficit for all measures: FCV deficit (OR = 1.89; 95% CI: 1.55–2.29), FEV1 deficit (OR = 1.71; 95% CI: 1.37–2.13), PEF deficit (OR = 1.37; 95% CI: 1.06–1.76), and MMEF deficit (OR = 1.28; 95% CI: 1.03–1.59). Residents within a range of 20–100 m of a major roadway (PR = 35%) exhibited increased odds of FVC deficit (OR = 1.32; 95% CI: 1.11–1.57) compared with participants living more than 100 m from a major roadway (PR = 46%). Participants residing in homes within 100 m of a pollution source, such as a factory or smokestack (PR = 38%), were more likely to have increased odds of lung function deficit for all measures: FCV deficit (OR = 1.24; 95% CI: 1.06–1.44), FEV1 deficit (OR = 1.22; 95% CI: 1.03–1.46), PEF deficit (OR = 1.22; 95% CI: 1.03–1.46), and MMEF deficit (OR = 1.40; 95% CI: 1.19–1.66).
Other qualities of the home environment were also shown to have bearing on the odds of reduced lung function. Keeping of pets in the home (PR = 22%) was associated with increased odds of lung function deficit in residents for all measures: FCV deficit (OR = 1.29; 95% CI: 1.08–1.54), FEV1 deficit (OR = 1.46; 95% CI: 1.19–1.787), PEF deficit (OR = 1.38; 95% CI: 1.11–1.73), and MMEF deficit (OR = 1.38; 95% CI: 1.14–1.68). Similarly, home renovation within the previous two years (PR = 36%) was associated with increased odds of lung function deficit in residents for all measures: FCV deficit (OR = 1.86; 95% CI: 1.59–2.17), FEV1 deficit (OR = 2.90; 95% CI: 2.43–3.47), PEF deficit (OR = 1.51; 95% CI: 1.24–1.84), and MMEF deficit (OR = 1.94, 95% CI: 1.64–2.30). A pest sighting in the home during the previous 12 months (PR = 36%) was associated only with increased odds of PEF deficit (OR = 1.25; 95% CI: 1.02–1.53). Indoor coal use was notable only in the odds of reduced FEV1 (OR = 1.44; 95% CI: 1.07–1.93). Humidifier use was only associated with a lower MMEF rate (OR = 1.29; 95% CI: 1.06–1.58). The presence of mold or water damage in the home in the previous 12 months and the use of ventilation had no significant impact on the odds of abnormal spirometry measures.
ETS exposure during pregnancy was associated with increased odds of MMEF deficit (OR = 1.29; 95% CI: 1.06–1.58). ETS exposure in the first two years of life was associated with increased odds of FVC deficit (OR = 1.26; 95% CI: 1.05–1.51). Current ETS exposure was associated with increased odds of lung function deficit in residents for several measures: FCV deficit (OR = 1.51; 95% CI: 1.29–1.77), FEV1 deficit (OR = 1.48; 95% CI: 1.24–1.77), and MMEF deficit (OR = 1.33; 95% CI: 1.12–1.57).
Linear modeling of lung function
Several variables related to housing characteristics were significantly associated with altered lung function (
Tables 3 and
4). As shown in
Table 3, ping-fang housing was associated with deficits of FVC (
β = −112; 95% CI: −159 to −64), FEV1 (
β = −85; 95% CI: −127 to −43), and PEF (
β = −190; 95% CI: −282 to −97), relative to dan-yuan-lou-fang housing. Residences with a floor area >15 m
2 were beneficial for lung function across all spirometry outcomes and all floor area categories, though no trend was noted with increasing floor area. Similarly, homes with more than three rooms were associated with better measures of FVC (
β = 40; 95% CI: 11–69), when compared with homes with fewer rooms. The level of the residence being >7 stories was negatively associated with FEV1 (
β = −55; 95% CI: −97 to −13). The age of the home exhibited no significant relationship to any of the four lung function indicators.
Environmental conditions in proximity to the dwelling were also associated with lower relative spirometric outcomes. Homes within <20 m of roadway traffic exhibited lower FEV1 (β = −52; 95% CI: −85 to −20), PEF (β = −91; 95% CI: −163 to −19), and MMEF (β = −61; 95% CI: −115 to −8) values. Participants residing in homes within 100 m of a pollution source had reduced respiratory indicators for all measures.
Other qualities of the home environment of participants also had bearing on lung function measures (
Table 4). The most influential variable was having a home renovation within the previous two years of assessment, which was associated with the greatest estimated deficit in lung function for FEV1 (
β = −241; 95% CI: −265 to −216), PEF (
β = −262; 95% CI: −317 to −207), and MMEF (
β = −240; 95% CI: −279 to −198). The presence of pests in the home, indoor coal use, and ETS exposure were also associated with reduced respiratory indicators for all measures. In contrast, ventilation in the home was associated with slightly improved respiratory function measures for FVC (
β = 55; 95% CI: 3–95), FEV1 (
β = 49; 95% CI: 10–87), and MMEF (
β = 77; 95% CI: 12–141).
ETS exposure during pregnancy was associated with reduced FEV1 volume (β = −31; 95% CI: −61 to −1), and reduced MMEF rate (β = −105; 95% CI: −158 to −55).
Discussion
Few studies have investigated the range of adverse respiratory health effects associated with exposure to household sources of indoor air pollution in Chinese children. The primary strength of this study is the use of objective measures of lung function, building on the work previously performed as part of the SNEC study.
Several residential characteristics and environmental factors that were associated with an increased risk of diminished lung function were identified. Residents of ping-fang housing exhibited FEV1 deficit compared with those living in multi-family structures and dormitories. They also exhibited reduced lung function for FVC, FEV1, and PEF in the linear model. This is contrary to the findings of
Zhang et al. (2013) that children in Wuhan, China, residing in apartment buildings had greater odds of being diagnosed with allergic rhinitis (OR = 1.53; 95% CI: 1.16–2.00) than those residing in single-family housing. In the present study, residents living in ping-fang housing are usually poor and the indoor pollution was much heavier than that in the apartments. For example, 73.1% of subjects living in ping-fang housing reported the use of coal for cooking or heating, while only 2.5% of subjects living in apartments reported this use. In addition, the education level and family annual income were lower in the residents of ping-fang housing than those of apartment housing (parental education < high school: 78% vs. 32%; family annual income <10 000 RMB: 37% vs. 22%). Thus, living in ping-fang housing may represent the social economic characteristics of a poverty-stricken population and may hold higher vulnerability for disease risk when exposed to environmental factors.
Previous work by the authors failed to find any significant relationships between housing type and reported respiratory symptoms consistent with asthma (
Dong et al. 2008).
Langer and Bekö (2013) found higher concentrations of VOCs and formaldehyde in Swedish single-family homes compared with apartment buildings and highlighted the importance of air exchange in residences. This was not considered in this research and could play a role in our observations. Though this study examined the effect of ventilation associated with cooking on respiratory function, it did not address ventilatory air exchange. No significant relationships were observed for ventilation in the logistic model and only small, yet significant, effects were obtained from the linear model.
Taken together, the findings from the age of the residence and recent home renovations bolster previous findings related to respiratory outcomes and new construction in this region of China. In prior work, the authors have speculated that this finding may be relates to self-reported allergic rhinitis in adult women (
Dong et al. 2013) and children (
Dong et al. 2008). However, the consistent findings of the odds of spirometric deficit and the observed, large effects in the linear model of children in homes recently renovated increases the concern of the impact of indoor air contaminants on children. This suggests that builders and renovators may be choosing materials and construction methods that may be hazardous to the respiratory health of children.
Off-gassing of VOCs is a potential source of respiratory irritants and is dependent on factors related to the manufacturing and composition of building materials, paint and coatings, and furniture.
Raw et al. (2004) found increased levels of VOCs and formaldehyde in newer homes and in homes that had undergone recent painting. The work of
Jaakkola et al. (2004,
2006) has provided evidence of both ongoing off-gassing of VOCs from paint following renovation and epidemiological evidence to support respiratory effects in adults following renovation.
Dong et al. (2014) also found that the students living in newly renovated residences had higher rates of rhinitis, cough, and shortness of breath compared with controls. However, these studies did not evaluate the impact of renovations on lung function and found no significant relationships.
A recent review by
Chen et al. (2010) highlighted the inconsistent findings regarding an association between the keeping of dogs and cats and asthma or allergic diseases. The authors concluded that exposure to pets does not have an impact on development of asthma and wheezing (
Chen et al. 2010). They also determined that early exposure to cats may cause sensitization to cat allergen, but early exposure to dogs may cause a protective effect (
Chen et al. 2010). Our findings of respiratory deficits in children living in homes with pets provide quantitative evidence of the effects of pet-keeping on lung function, though not without some curious inconsistency. The odds of respiratory deficit were significant for all measures of spirometric performance, but the linear modeling of effect indicated no significant deficits associated with pets. That being the case, there are still unresolved mechanisms with respect to respiratory effect and the nature and frequency in which pets interact with people as well as how pets are domiciled (i.e., inside or outside the home). This study was not designed to specifically answer these questions, but serves as a potential avenue for future studies. It should also be noted that this study included poultry in the definition of pet-keeping, which may not be typical in other parts of the world.
The presence of mold or water damage in the home within a year had no statistically significant impact on the odds of respiratory deficit and no significant effects on spirometric performance in the linear model. Some estimates, though not significant, indicated a protective effect with mold or water in the home. However, the exposure to mold was based on parental self-reporting and not all mold and water issues are readily visible. It is known that mold proliferation is dependent on the growth medium and the presence of moisture (
Van Loo et al. 2004) and that the presence of mold is linked to respiratory health outcomes (
Mendell and Kumagai 2017). However, many questions remain in our understanding of the interaction between fungal organisms, fungal metabolic products, and environments (
Nevalainen et al. 2015). Related to this potential source of respiratory irritation is the use of a humidifier, which has previously been found to have an impact on respiratory outcomes. In this study, the use of a humidifier was only associated with reduced MMEF rate.
The presence of pests in the home in the previous year was found to be associated with greater odds of PEF deficit and significant deficits to lung function were observed in all four spirometric measures. The role of pests in asthma and allergy has been noted previously and was recently summarized by
Sheehan et al. (2010). Our findings are largely supported by the Sheehan review, but they suggest that future studies should examine the role of pests on objective lung function. They also suggest that intervention studies examining the effect of pest management should consider spirometric performance to assess the effect on pest mitigation in addition to the reporting of asthma symptoms.
Proximity to pollution sources, both from roadway traffic and point sources, was associated with reduced lung function. The Children’s Health Study in Southern California (
Gauderman et al. 2004,
2007), the Oslo Birth Cohort (
Oftedal et al. 2008), and a recent cohort study by
Schultz et al. (2012) represent strong evidence that exposure to pollution from traffic has adverse effects on children’s lung function development, though
Schultz et al. (2012) concluded that the effect was greatest in the first year of life. Our findings are consistent with these studies and add further evidence to this relationship. Children who lived within 20 m of traffic, when compared with children who lived more than 100 m from traffic, experienced greater odds of respiratory deficit for all four measures of lung function and significant effects for FEV1, PEF, and MMEF in the linear model.
Several studies have examined the effects of ETS on child respiratory health outcomes at different stages of growth. Some have associated fetal exposure via maternal smoking with respiratory symptoms (
Lannerö et al. 2006;
Håberg et al. 2007). Others have demonstrated effects due to postnatal exposure to ETS (
Öberg et al. 2011;
Burke et al. 2012). Our findings indicated that current exposure to ETS exhibited the strongest impact on children who were exposed at the time of the study. The odds of respiratory deficits in FVC, FEV1, and MMEF were significant for current ETS, but only a deficit in MMEF was significant for ETS exposure during pregnancy. Similarly, significant odds of respiratory deficit are noted for FVC, FEV1, and MMEF for current ETS exposure, and detrimental effects are indicated for FEV1 and MMEF in the linear model. Exposure to ETS in the first two years of life did not indicate odds of lung function deficit but exhibited effects in the linear model for FEV1 and MMEF.
Although the data were thoroughly detailed and a large sample (
n = 6740) was utilized, our study contained a few limitations. First, the cross-sectional nature of the study cannot establish temporal effects between housing characteristics or environmental exposures and the measured spirometric outcomes. Second, there may have been confounding variables not accounted for in this study that could have played a role in lung function deficit such as body composition (
Scott et al. 2012) and dietary factors (
Veeranki et al. 2015). Third, although new and renovated homes had higher odds of lung function deficit, no measures of indoor air quality were taken that could implicate particular exposure sources in lung function deficit.