Background
Both acute and chronic air pollution exposure have been shown to influence cystic fibrosis (CF) bronchiectatic disease. Ambient concentrations of ozone, PM
10 and NO
2 play a role in triggering an exacerbation in CF and annual average exposures to particulate air pollution was associated with an increased risk of pulmonary exacerbations and a decline in lung function [
1],[
2]. Non-cystic fibrosis bronchiectatic (NCFB) disease is characterized by chronic bronchial inflammation caused by inappropriate clearance of various microorganisms and recurrent or chronic infections [
3]. Mortality in NCFB is known to be influenced by a number of factors such as gender, age, smoking history and
Pseudomonas aeruginosa, but data on the impact of air pollution on NCFB mortality are lacking [
4],[
5]. More recently, De Soyza
and colleagues formulated some research priorities in NCFB disease [
6]. One of the three priorities is more research into epidemiology, to identify certain patient subgroups at risk for worse disease.
As the impact of air pollution in NCFB has not been studied before, our primary aim was to study the association between residential distance to a major road and death in a cohort of patients with NCFB. Secondary aims were to evaluate the distance-weighted traffic density at 100 and 200 meters from the patient’s residence.
Discussion
Our analysis showed that residential proximity to a major road was linked to an increased risk of dying in patients with NCFB. Further analysis suggested that distance-weighted traffic density is an important factor in this risk.
We describe for the first time the impact of traffic related chronic air pollution exposure indicators in a NCFB population. With the exception of cystic fibrosis, there is a striking paucity of research into bronchiectasis [
6]. For cystic fibrosis, chronic and acute air pollution exposure have been linked to exacerbations [
1],[
2], but the impact on mortality has not been established yet. For other respiratory and cardiovascular diseases, research clearly indicates a pronounced negative effect on mortality. Air pollution triggers myocardial infarction [
20], it is associated with cardiopulmonary mortality [
21], and causes an increased risk of infant mortality [
22]. In respiratory disease, it has been linked with lung cancer [
21] and mortality or bronchiolitis obliterans syndrome after lung transplantation [
23]. For COPD, an interquartile range elevation in black carbon concentrations is associated with a 7% increase in COPD mortality [
24]. We now added NCFB to the increasing list of respiratory diseases where mortality is affected by traffic related air pollution.
To further strengthen our data, we also analyzed distance-weighted traffic density, and demonstrated a significant effect in a buffer of 100 and 200 meters from the patient’s residence. This is in line with previous literature showing that cardiopulmonary mortality is associated with living near a major road [
25]-[
27].
It is known that air pollution affects mortality in the general population. Carey
and colleagues found that in their nationwide English cohort of 835,607 people, residential pollution concentration was associated with mortality with a HR of 1.02, 1.03 and 1.04 for PM
2.5, NO
2 and SO
2 respectively. [
28] Similar effect sizes of pollution in the general population have been seen by other groups [
21],[
29], and are much lower than the effects we describe in our NCFB population. A possible reason might be the superimposed effect of air pollution on the vicious cycle of inflammation, infection and mucus plugging with proteolytic enzymatic actions. Whether air pollution also leads to a higher exacerbation rate, remains to be established.
These results again emphasize the ongoing need to sensitize policy makers in reducing traffic and air pollution [
30]. Cesaroni
et al. elegantly showed that a reduction in traffic in the city center of Rome not only effectively reduced air pollution, but also caused an average life gain of 3.4 days per person for the more than a quarter million residents living along the busy roads [
31].
Our study has some limitations. The most important limitation is the low number of events. We aimed at investigating a hard endpoint in NCFB, i.e. death. However, NCFB is not a very lethal condition and our 15 deaths among 183 patients over the course of our study period are in line with the rates from literature. Loebinger
and colleagues showed only 27 deaths among their 91 patients over a 13-year period (29.7%) [
4]. This limited number of events forces us to be cautious when interpreting our results regarding the effect of pollution. We further tried to refine our exposure assessment by adding distance-weighted traffic density analysis. The association between mortality and traffic density within distances of either 100 or 200 meters from the patient’s home address, adds to the suggestion that there is a real effect of air pollution and traffic density on mortality.
On the other hand, a second limitation of our study might be overadjustment for covariates in our multivariate model. Correcting for many factors (gender, age, disease severity, SES, chronic macrolide use, chronic colonization by Pseudomonas aeruginosa and smoking history) when there are only few events might have led to overadjustment and therefore an incorrect estimation of the effect. On the other hand, the unadjusted model and the model where we only correct for age and gender resulted in similar significant effects. The largely similar results of both the adjusted and unadjusted models therefore suggest that this limitation probably had a limited effect on our overall outcome.
Thirdly, as one patient had an unknown cause of death, this death might be unrelated to air pollution. We preferred all-cause mortality as our primary outcome as this is a strong objective end-point which is not biased or influenced by confounders. When excluding this patient from the analysis, residential proximity to a major road was still associated with risk of dying as were the results for distance-weighted traffic density. A second patient had a cardiorespiratory arrest and alcohol intoxication. We included this patient as literature has shown that air pollution is known also to affect cardiovascular mortality [
20].
In this analysis, we did not address seasonality and temporal changes. The traffic counts are based on annual average traffic counts on working days in 2010 and no information on seasonality is available. However, we believe that, as we studied a chronic outcome, even if a seasonal variation in traffic density would exist, the spatial variation is more important than potential temporal variation.
Concerns might be raised on changes in distribution of distance to a major road and traffic density. However, Pearson
et al. previously showed that an increase in traffic over an eleven-year period still shows highly correlated densities [
14].
Finally, for minor roads, the traffic intensity data was missing for some local roads and therefore these missing data were imputed with a default value of 543 vehicles per day. We believe that, as these roads were minor roads, measurement error with regard to defining busy and non-busy roads is likely small. This default setting is based on average counts on local roads and has been used in previous large studies on traffic related air pollution [
32].
In conclusion, we found that living close to a major road is associated with a higher risk of dying in patients with NCFB. We interpret this as reflecting yet another adverse effect of air pollution.
Competing interests
The authors declare that they have no competing interests.