Association transportation noise and MI mortality in adults, independent of air pollution
A systematic analysis of mutual effects of transportation noise and air pollution exposure on myocardial infarction mortality: a nationwide cohort study in SwitzerlandLiterature - Héritier H, Vienneau D, Foraster M et al. - Eur Heart J 2018; doi:10.1093/eurheartj/ehy650
Introduction and methods
Both transportation noise and air pollution have been associated with CV health. An increase in road traffic noise levels is linked to higher risk of coronary heart disease  and road traffic and aircraft noise are positively related to higher risk of myocardial infarction (MI) . Moreover, exposure to NO₂ and PM2.5 increases the risk of MI .
Although the link between co-exposure to air pollution and transportation noise, and CVD mortality has been examined by various studies, the nature of the relationship remains unclear. Some studies reported independent noise effects on CVD risk [4-9], whereas other studies showed that effects of air pollution were attenuated after adjustment for road traffic noise [10,11]. An explanation for these observations might be mutual confounding since transportation noise and air pollution mainly originate from traffic. Other studies did not adjust for exposure to air pollution [12,13].
A limitation of many of these studies was that they assessed noise exposure on the public level rather than on a residence level. Using a different spatial scale for exposure to noise than for subsequent adjustment for exposure to air pollution may introduce potential confounding.
This study (2000-2008) therefore aimed to disentangle the long-term effect of road traffic, railway and aircraft noise levels, as well as the air pollutants NO₂ and PM2.5 on risk of MI mortality in 4.40 million adults (aged >30 years) from the Swiss National Cohort (SNC). The SNC, including 7.28 million adults, probabilistically links national census data with mortality and emigration records.
The authors had previously developed a noise exposure database for the year 2001, that assessed exposure to road traffic, railway noise and aircraft noise emissions for each building in Switzerland, and if possible at different floor levels [14,15]. For each building, transportation noise was estimated at pre-defined façade points. For each façade point, L-den was calculated for each transportation noise source. Noise exposure was assigned on the façade point per dwelling unit with the highest Lden value. NO₂ exposure was estimated using data from cantonal air pollution monitoring authorities, comprising data from 14 days passive measurements collected from 2000-2008 at a total of 1834 locations. Daily PM2.5 at 100m grid cells across Switzerland was predicted for 2003-2008 from satellite, land use, and meteorological data. Using multipollutant models, linear HRs were calculated and adjusted for potential confounders (sex, neighborhood index of socio-economic position, civil status, educational level, nationality, and mother tongue), excluding NO₂ and PM2.5. Outcome was primary cause of death from MI.
- The highest correlation was observed between L-den(road) and NO₂ (Spearman’s correlation coefficient 0.44). Correlations between PM2.5 and road (0.27), railway (0.20), and aircraft noise (0.24) were rather low.
- The correlation coefficient between NO₂ and PM2.5 was 0.62.
- The adjusted HR for MI mortality per 10 dB noise increase was 1.032 (95%CI: 1.014-1.051) for road traffic, 1.020 (95%CI: 1.007-1.033) for railway traffic, and 1.025 (95%CI: 1.006-1.045) for aircraft traffic. These risk estimates were hardly affected by additional adjustment for air pollution.
- The adjusted HR for MI mortality per 10 mg/m3 increase in NO₂ was 1.024 (95%CI: 1.005–1.043) and per 10 mg/m3 increase in PM2.5 it was 1.052 (95%CI: 1.013–1.093). However, these risk estimates lost statistical significance upon adjustment for all noise sources.
- There were no synergistic or antagonistic effects between road traffic noise and PM2.5 or NO₂ in linear-exposure response models including interaction terms. Similarly, no relevant interactions for PM2.5 or NO₂ with road traffic noise exposure were observed by testing interactions in categorical models to evaluate potential thresholds for interaction.
This cohort study suggests that transportation noise is linked to MI mortality in adults, independently of air pollution. This observation is based on a high quality noise exposure model that allowed for assessment of individual exposure at the address and floor level. The risk for MI mortality was higher with increasing air pollution, however, this risk lost significance when adjusting for all noise sources. Studies on air pollution not adequately adjusting for transportation noise may therefore overestimate the CV burden of air pollution.
In their editorial comment , Sørensen and Pershagen mention a strong feature of the study conducted by Héritier et al.: namely the ‘excellent state-of-art exposure assessment models’ assessing exposure to road traffic noise, aircraft noise and railway noise separately. The investigators showed a consistent association between noise and MI mortality across the three modes of transportation. Residential exposure to aircraft and train noise is less correlated with residential air pollution, compared to noise from road traffics and the air pollution near roads. The observed consistent association therefore strongly supports an effect of transportation noise that is independent of air pollution.
The authors also note some limitations of the current study. There was no information available on lifestyle habits such as smoking and alcohol consumption and therefore adjustments relied entirely on socio-economic and demographic variables. This might have led to residual bias, indicating that more research is needed in large cohort studies with more detailed information on potential confounders. Also, the study only focused on mortality, while most MI patients do not die from their disease. It is not certain whether air pollution and transportation noise affect incident and fatal MI in similar ways. Further, air pollution is a highly complex mixture. The investigators did not assess air pollutants other than PM2.5 and NO₂, such as ultrafine particles. These could also be important when disentangling the effects of noise and air pollution. The authors conclude that this study adds to the evidence that transportation noise is an important risk factor for CVD. ‘The study highlights the importance of proper mutual adjustment of transportation noise and air pollution using high-quality exposure assessments.’