Article Text

Download PDFPDF

Original article
Long-term transportation noise exposure and incidence of ischaemic heart disease and stroke: a cohort study
  1. Andrei Pyko1,2,
  2. Niklas Andersson1,
  3. Charlotta Eriksson1,2,
  4. Ulf de Faire1,
  5. Tomas Lind2,
  6. Natalya Mitkovskaya3,
  7. Mikael Ögren4,
  8. Claes-Göran Östenson5,
  9. Nancy L Pedersen6,
  10. Debora Rizzuto7,
  11. Alva Käte Wallas1,
  12. Göran Pershagen1,2
  1. 1 Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden
  2. 2 Centre for Occupational and Environmental Medicine, Stockholm County Council, Stockholm, Sweden
  3. 3 Department of Cardiology and Internal Medicine, Belarusian State Medical University, Minsk, Belarus
  4. 4 Department of Occupational and Environmental Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden
  5. 5 Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden
  6. 6 Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden
  7. 7 Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Stockholm, Sweden
  1. Correspondence to Dr. Andrei Pyko, Institute of Environmental Medicine, Karolinska Institutet, Stockholm 171 77, Sweden; Andrei.pyko{at}ki.se

Abstract

Background There is limited evidence from longitudinal studies on transportation noise from different sources and development of ischaemic heart disease (IHD) and stroke.

Objectives This cohort study assessed associations between exposure to noise from road traffic, railway or aircraft and incidence of IHD and stroke.

Methods In a cohort of 20 012 individuals from Stockholm County, we estimated long-term residential exposure to road traffic, railway and aircraft noise. National Patient and Cause-of-Death Registers were used to identify IHD and stroke events. Information on risk factors was obtained from questionnaires and registers. Adjusted HR for cardiovascular outcomes related to source-specific noise exposure were computed using Cox proportional hazards regression.

Results No clear or consistent associations were observed between transportation noise and incidence of IHD or stroke. However, noise exposure from road traffic and aircraft was related to IHD incidence in women, with HR of 1.11 (95% CI 1.00 to 1.22) and 1.25 (95% CI 1.09 to 1.44) per 10 dB Lden, respectively. For both sexes taken together, we observed a particularly high risk of IHD in those exposed to all three transportation noise sources at≥45 dB Lden, with a HR of 1.57 (95% CI 1.06 to 2.32), and a similar tendency for stroke (HR 1.42; 95% CI 0.87 to 2.32).

Conclusion No overall associations were observed between transportation noise exposure and incidence of IHD or stroke. However, there appeared to be an increased risk of IHD in women exposed to road traffic or aircraft noise as well as in those exposed to multiple sources of transportation noise.

  • epidemiology
  • noise
  • cardiovascular
  • transport

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Key messages

What is already known about this subject?

  • Exposure to transportation noise has been suggested as a risk factor for cardiovascular disease but the evidence from longitudinal studies on ischaemic heart disease (IHD) and stroke is limited, particularly regarding the effect of combined exposure to several noise sources.

What are the new findings?

  • No clear associations were observed between noise exposure from road traffic, railway or aircraft and incidence of IHD or stroke.

  • However, noise exposure from road traffic and from aircraft was related to IHD incidence in women.

  • For both sexes taken together, a particularly high risk of IHD occurred among those exposed to all three transportation noise sources and a similar tendency was seen for stroke.

How might this impact on policy or clinical practice in the foreseeable future?

  • Possible sex-related differences in susceptibility to cardiovascular effects of transportation noise deserve attention.

  • Combined exposure to noise from several transportation noise sources may be particularly harmful.

Introduction

The incidence of cardiovascular disease, primarily ischaemic heart disease (IHD) and stroke, has decreased in recent decades but these diseases are still leading causes of disability and death in Western countries.1 There is growing evidence of adverse cardiovascular effects of transportation noise from different sources, and this evidence was recently evaluated in a systematic review performed within the framework of developing new WHO Environmental Noise Guidelines for the European Region.2 Overall, meta-analyses indicated an association between road traffic noise and incidence of IHD based on seven cohort or case-control studies as well as for aircraft noise based on two ecological studies. In one cohort study, stroke incidence was related to road traffic noise exposure.3 Moreover, several longitudinal studies indicated that exposure to transportation noise was related to obesity,4 5 diabetes6–8 and hypertension,9 which are independent risk factors of cardiovascular disease. The WHO systematic review concluded that not enough studies of good quality are available on cardiovascular and metabolic effects of transportation noise, but that the plausibility of an association calls for further and improved research, particularly studies of longitudinal design. Furthermore, little is known about induction periods for cardiovascular effects of noise and the impact of combined exposure to noise from several sources.

This cohort study aimed to assess individual long-term exposure to transportation noise from different sources, including road traffic, railways and aircraft, in relation to IHD and stroke incidence and mortality. In particular, we investigated the impact of exposure to noise from multiple sources.

Materials and methods

Study population

This longitudinal study is based on pooled analyses of four subcohorts from Stockholm County which have baseline information on risk factors and health parameters related to cardiovascular disease as well as extensive follow-up data. The Stockholm Diabetes Preventive Program (SDPP) is a population-based prospective study of 7949 subjects aged 35–54 years at recruitment in 1992–1998.10 By design none had previously diagnosed diabetes, and approximately half had a family history of diabetes in at least one first-degree relative (mother, father, sister or brother) or in at least two second-degree relatives (grandparent, uncle or aunt). The other half of the subcohort was matched on age and sex but without a family history of diabetes. The SIXTY subcohort included 4232 subjects from a random population sample of one-third of all men and women living in Stockholm County who turned 60 years of age between August 1997 and March 1999.11 The Screening Across the Lifespan Twin Study (SALT) was used as a sampling frame to select 7043 individuals from the Swedish Twin Register born 1958 and earlier, who lived in Stockholm County and were investigated in 1998−2002.12 The Swedish National Study of Aging and Care in Kungsholmen (SNAC-K) included 3363 randomly sampled individuals ≥60 years of age between March 2001 and June 2004 from a central area in Stockholm.13 Taken together the four subcohorts comprise a population of more than 22 314 individuals and constitute the CEANS cohort (Cardiovascular Effects of Air pollution and Noise in Stockholm).

Outcome definition

Data on individual IHD and stroke events were obtained from the National Patient Register and the National Cause of Death Register during the period 5 years before enrolment (February 1987 as the earliest) through December 2011. Each event was defined based on the International Classification of Diseases (ICD) versions 9 and 10: hospitalisation or death with principal diagnosis of IHD (ICD9: 410X–414X; ICD10: I20X–I25X), including the subgroup myocardial infarction (ICD9: 411X; ICD10: I21X–I23X) or stroke (ICD9: 430–436; ICD10: I60–I65) including ischaemic stroke (ICD9: 433X–435X; ICD10: I63X, I65X, I66X). Subjects with IHD or stroke diagnoses before recruitment were excluded from the analyses. All first events during follow-up were classified as incident, whereas National Cause of Death Register records of IHD or stroke, as well as non-traumatic death within 28 days after an IHD or stroke hospitalisation, were classified as fatal cases.

Exposure assessment

We assessed the exposure to noise from road traffic, railways and aircraft at the most exposed façade of each building using a database which incorporated information from several national, regional and local authorities. The database covers Stockholm County during the time-period 1990–2010 and typically contains information on terrain, ground surface, building density, traffic flows on roads (>1000 vehicles/24 hours) and railway lines, speed limits as well as aircraft noise contours around the two major airports in Stockholm (Arlanda and Bromma). Information on noise barriers is also included but not used because of lack of data regarding year of construction.

For road traffic and railway noise, we modelled the 24 hours A-weighted equivalent sound pressure level (LAeq,24h) based on the information in our database and a simplified version of the Nordic prediction method computing noise levels in a range of 30–75 dB. The methodology has been validated against the full Nordic prediction method.14 We converted the LAeq-levels to Lden, assuming a 24 hours traffic flow distribution of 75/20/5% during day, evening and night, respectively, for road traffic and the exact 24 hours distribution for separate segments of the railway lines.

Information on aircraft noise exposure was obtained as noise contours around the Arlanda and Bromma airports for the years 1995, 2000, 2005 and 2010. For the year 1990, we assumed the same noise level as for 1995 since there were no major structural changes at either of the two airports during this time period. The noise contour data ranged from 45 to 70 dB Lden around Arlanda and from 40 to 70 dB Lden around Bromma. By superimposing the noise contour data on a layer of buildings where the study participants had lived, each address could be assigned a corresponding noise level.

In order to estimate the time-weighted-average residential noise exposure for the study participants from each transportation source, we obtained information on their residential address history from the Swedish Taxation Authority. This included information on each address where the participants had lived during the complete follow-up period, starting from 1990 and with precise dates of changes in residency. For each of the residential address coordinates we calculated noise levels separately for each year. If the person moved during a specific year the time-proportional average noise level from the old and new address was used for that year.

A detailed description of covariates is provided in the online supplementary data.

Supplemental material

Statistical analysis

Differences in background characteristics in relation to exposure to road traffic noise at recruitment were tested with Pearson χ² tests for categorical variables.

To analyse associations between exposure to transportation noise (road, railway and aircraft) and incidence of IHD or stroke, we used stratified Cox proportional hazards regression models to compute HR and 95% CI. Person-time at risk was calculated from enrolment into the study until IHD or stroke diagnosis, death from another cause, migration out of Stockholm County (ie, to an address without information noise exposure) or end of follow-up, whichever event occurred first. Age was used as the underlying timescale in all models. To facilitate calculations, we aggregated person-time data into 6-month intervals and computed time-varying noise levels for each of these intervals reflecting exposure in three time windows: during the year of the event (year 0) as well as 1–5 years and 6–10 years prior to the event. For all three noise sources in each of the exposure time-windows, noise levels below 35 dB Lden were recoded as 35 dB to reduce exposure misclassification due to uncertainty in the modelling at low levels. Intervals with more than 20% missing exposure information were excluded from the analyses.

The analyses were performed with a categorical exposure variable (<45 dB, 45–49 dB, 50–54 dB and ≥55 dB Lden) as well as with a continuous exposure variable ranging from 35 to 72 dB Lden, presenting risk estimates per 10 dB Lden increase. We also assessed the effect of combined exposure to multiple noise sources by creating dummy variables, indicating subjects exposed to none, one, two or three transportation noise sources ≥45 dB Lden.

We performed separate analyses for IHD and stroke incidence and mortality assuming a unique baseline hazard for each subcohort. We separately tested crude models adjusted only for sex, enrolment year and year of the event as well as full models additionally adjusted for smoking status, educational level, physical activity during leisure time, alcohol consumption, working status and occupation.

A detailed description of the methodology for effect modification and sensitivity analyses is provided in the online supplementary data.

The study was conducted in accordance with the Helsinki Declaration and approved by the Regional Ethics Review Board at Karolinska Institutet. All participants gave informed consent to the original study which also applied to the present analyses.

Results

From the original sample of 22 314 individuals, 241 (0.1%) were excluded due to missing exposure data and 2061 (9.2%) because of missing data on covariates. The remaining 20 012 subjects had a mean age of 60 years (range 35–104) at study entry and provided 245 000 person-years of observation. During follow-up, we registered 1363 incident IHD and 902 incident stroke events, including 286 and 113 fatal events, respectively. In addition, 527 subjects with IHD and 298 with stroke were diagnosed before recruitment and excluded from the analyses. At recruitment, a total of 7736 (39%) were exposed to one of three transportation noise sources at 45 dB Lden or higher, 4487 (22%) to two sources of transportation noise and 218 (1%) were exposed to all three sources, and 7571 (38%) were exposed to levels below 45 dB Lden from all transportation noise sources. Mean levels of individual exposure to transportation noise at the recruitment address were 46.8, 39.3 and 39.2 dB Lden for road traffic, railways and aircraft, respectively.

Study participants with road traffic noise exposure ≥45 dB Lden were more likely to be women, older, have higher education and socioeconomic status, consume alcohol less frequently and be less physically active (table 1). To some extent, these associations depended on the noise exposure distribution in the separate subcohorts (online supplementary table 1). However, several associations between road traffic noise exposure and covariates were consistent across the subcohorts.

Table 1

Characteristics of participants in the CEANS cohort from Stockholm County in relation to road traffic noise exposure at recruitment

Table 2 and online supplementary table 2 show the associations between transportation noise exposure from different sources 1–5 years preceding the event and incidence of IHD in full and crude models, respectively. In view of the noise exposure differences between men and women sex-specific data are also provided. Overall, no clear or consistent associations were observed between noise exposure from any of the three transportation sources and IHD incidence. However, there were statistically significant associations between exposure to road traffic as well as aircraft noise and incidence of IHD in women, with HRs of 1.11 (95% CI 1.00 to 1.22) and 1.25 (95% CI 1.09 to 1.44) per 10 dB Lden, respectively. On the other hand, inverse associations were seen in men, which was statistically significant for road traffic noise. IHD mortality showed positive tendencies in relation to transportation noise from all three noise sources, which appeared to be stronger in women (online supplementary table 3). Comparing the three sources of noise, for both IHD incidence and mortality, the strongest associations were suggested in relation to aircraft noise exposure.

Table 2

Overall and sex-specific HRs of IHD incidence in relation to transportation noise exposure 1–5 years preceding the event from different sources in fully adjusted model*

Similar to IHD, there were no overall associations between noise exposure from any of the transportation noise sources and stroke incidence (table 3). A borderline statistically significant relation was observed in women for railway noise (HR 1.13; 95% CI 1.00 to 1.27 per 10 dB Lden). Stroke mortality tended to be associated with railway and aircraft noise exposure in both women and men (online supplementary table 3). Similar to IHD, associations with stroke mortality appeared to be strongest in relation to aircraft noise exposure.

Table 3

Overall and sex-specific HRs of stroke incidence in relation to transportation noise exposure 1–5 years preceding the event from different sources in fully adjusted model*

Subanalyses focused on incidence of myocardial infarction in relation to road traffic and aircraft noise generally showed similar patterns as for IHD (online supplementary table 4). Similarly, a stronger association appeared in women between railway noise exposure and ischaemic stroke incidence which reached statistical significance (HR 1.15; 95% CI 1.02 to 1.31 per 10 dB Lden).

We observed higher risks of both IHD and stroke incidence with HR of 1.57 (95% CI 1.06 to 2.32) and 1.42 (95% CI 0.87 to 2.32), respectively, in those exposed to all three noise sources simultaneously, although not statistically significant for stroke (figure 1). These associations were most apparent in women (online supplementary figure 1).

Figure 1

HR of IHD and stroke incidence in relation to number of sources of transportation noise exposures ≥45 dB 1–5 years preceding the event in fully adjusted model (*adjusted for sex, enrolment year and year of the event, smoking status, alcohol consumption, occupation, educational level, physicalactivity during leisure time, marital status and working status with age as underlining time scale. IHD, ischaemic heart disease).

There were no statistically significant interactions in relation to road traffic noise 1–5 years preceding the event and risk of IHD, except for sex (online supplementary figure 2). The association between road traffic noise and IHD incidence also tended to be stronger in the age group younger than 56 years. Effect modification for aircraft noise exposure generally showed the same pattern as for road traffic noise (data not shown).

Sensitivity analyses focused on IHD incidence in relation to road traffic noise and did not indicate consistent differences in the HR estimates between the main model and models with further adjustments for area-based mean income, traffic-related BC or PM2.5, other transportation noise sources (online supplementary figure 3). Similarly, no differences were observed after restricting the study sample to those who did not change address 5 years preceding the event. Moreover, we did not observe substantial differences in the associations between IHD incidence and transportation noise exposure in different time-periods for any of the three noise sources (online supplementary table 5).

In cohort-specific analyses of IHD incidence in relation to road traffic exposure, we observed a positive association in the SDPP cohort (HR 1.19; 95% CI 1.02 to 1.39 per 10 dB Lden), and an inverse association in the SALT cohort (HR 0.87; 95% CI 0.78 to 0.97 per 10 dB Lden) (online supplementary figure 4). Overall, moderate heterogeneity across cohorts was suggested (Higgin’s I2 statistic 68.2%). In the meta-analysis, the estimate for IHD incidence associated with road traffic noise was comparable to the main results with a HR 0.96 (95% CI 0.84 to 1.08) per 10 dB Lden. Low heterogeneity was noted between the cohorts for other transportation noise sources and for stroke (results not shown).

Discussion

This cohort study does not provide clear and consistent evidence of associations between exposure to transportation noise and incidence of IHD or stroke. However, in women, there appeared to be increased risks of IHD (including myocardial infarction) in relation to exposure to road traffic or aircraft noise and for ischaemic stroke in relation to railway noise. Risk estimates seemed particularly high among subjects with combined exposure to noise from road traffic, railways and aircraft.

There is an increasing number of studies on transportation noise and IHD. A recent meta-analysis of three cohorts and four case-control studies calculated a Relative Risk (RR) of 1.08 (95% CI 1.01 to 1.15) per 10 dB Lden for IHD following exposure to noise from road traffic.2 In the present study, the overall result was comparable to a recent Swedish cohort study on myocardial infarction and road traffic noise showing a RR of 0.99 (95% CI 0.86 to 1.14) per 10 dB Lden.15 For IHD mortality, our results were similar to the WHO-review reporting a pooled RR based on one case-control and two cohort studies of 1.05 (95% CI 0.97 to 1.13) per 10 dB Lden. Data are limited for aircraft noise exposure and IHD incidence or mortality. The WHO-review reported a pooled RR of 1.09 (95% CI 1.04 to 1.15) per 10 dB Lden for IHD incidence based on two ecological studies; however, we did not observe an association. One cohort study reported a RR for IHD mortality of 1.04 (95%CI 0.98 to 1.11) per 10 dB Lden related to aircraft noise.16 In the present study, we also observed a tendency towards an increased IHD mortality due to aircraft exposure. For railway noise, the WHO-report did not find any longitudinal studies on IHD incidence or mortality but two subsequent studies indicated positive associations with IHD mortality17 18 with risk estimates consistent with our findings. One possible explanation for the absence of clear associations for different sources of transportation noise in our study may be the comparatively low exposure levels.

The evidence on transportation noise and stroke is limited.2 Our study did not show an association between road traffic exposure and stroke in contrast to a Danish cohort study reporting a RR of 1.14 (95% CI 1.03 to 1.25) per 10 dB Lden and a pooled estimate from the WHO-review with RR of 1.05 (95% CI 0.96 to 1.15) per 10 dB Lden for aircraft noise. Our results showed a tendency of an association between railway noise and stroke incidence, which was statistically significant for ischaemic stroke, which is in line with other studies indicating higher risks for ischaemic than haemorrhagic stroke.18 19 For stroke mortality, our results suggested associations both for railway and aircraft noise in contrast to the WHO-review. Overall, the evidence on transportation noise and stroke appears less consistent than for IHD, which is in line with our data.

Available evidence on gender differences of noise effects on cardiovascular disease is limited and inconsistent. Babisch et al 20 reported that road traffic exposure >70 dB Lday(6-22) was associated with myocardial infarction only in men (OR 1.81; 95% CI 1.02 to 3.21). Similarly, results from a Danish cohort study also suggested stronger effects in men with a RR for myocardial infarction related to road traffic noise of 1.14 (95% CI 1.03 to 1.26) per 10 dB Lden.21 However, Selander et al 22 and Beelen et al 23 did not observe gender differences in cardiovascular incidence and mortality related to road traffic noise. Moreover, Gan et al 24 reported main results with no gender differences but found a 7% non-significant excess risk of coronary mortality in women after adjustment for traffic-related air pollutants. In our results, associations between transportation noise and cardiovascular outcomes were primarily seen in women. This observation is supported by findings that women had particularly elevated levels of salivary cortisol in response to noise exposure, suggesting a higher susceptibility to noise-induced stress responses.25 26

We observed a particularly high IHD incidence in people exposed to all three transportation noise sources simultaneously, which was more pronounced in women. This speaks in favour of the multiple environmental stressors theory which postulates that several stressors may enhance the effects of each other.27 Moreover, these results are in line with our previous findings, which indicate an increased risk of central obesity following exposure to multiple sources of transportation noise.4 It is also supported by a study by Selander et al 28 where an interaction was seen between traffic noise, occupational noise and job strain in relation myocardial infarction.

The risk of IHD due to transportation noise remained stable after adjustment for a number of potential confounders, including individual and contextual socioeconomic characteristics as well as exposure to BC or PM2.5. This confirms earlier evidence suggesting changes of associations between road traffic noise and IHD of less than 10% after adjustment for air pollution.29 The extent of confounding is partly dependent on the quality of the exposure assessment for each source. The lower correlation between noise and air pollution for exposure from aircraft and trains makes disentangling of noise effects less problematic. Overall, we cannot rule out residual or unmeasured confounding. Furthermore, we did not observe any significant interactions between exposure to transportation noise and other factors in relation to IHD incidence, except for sex. The association tended to be stronger in those younger than 56 years, which is in line with some noise studies on hypertension30 31 or obesity.4 On the other hand, some studies showed no age differences in associations of road traffic noise and cardiovascular  outcomes22 23 or stronger associations in higher age groups.21 24 Taken together, it remains unclear if age or other risk factors modify the association between noise and cardiovascular outcomes.

There are several limitations in the present study. Some heterogeneity was indicated between cohorts in risk estimates for IHD related to long-term road traffic noise exposure. This could be due to differences between cohorts in exposure levels, age at enrolment and other characteristics. The heterogeneity across the cohorts was comparable in a study on stroke in relation to long-term air pollution exposure based on the same material.32 It is noteworthy, however, that the risk estimate for IHD incidence in the pooled analysis was the same as the estimate from the meta-analysis. Furthermore, some misclassification of exposure is likely since it was based only on outdoor levels at the façade of each residence and we lacked information on time spent at home. This probably led to attenuation of the exposure-response relationships.

There are also several strengths with the study. The assessments of exposure for all three noise sources were performed objectively and not likely to be dependent on the outcome status. With regard to the outcome assessment, we relied on validated registry information.33 34 Furthermore, the exposure assessment took into account individual residential history and we had extensive information on potential individual and contextual confounders.

Conclusion

Overall, this cohort study provided no clear or consistent evidence of an association between transportation noise from different sources and incidence of IHD or stroke. However, there appeared to be an excess risk of IHD in women related to exposure to noise from road traffic or aircraft. Furthermore, simultaneous exposure to all three sources of transportation noise seemed to be particularly harmful.

References

Footnotes

  • Contributors All coauthors had substantial involvement in drafting and critical revision of the article, handling the reviewer comments as well as have approved the final version of the resubmitted manuscript. Working group of AP, CE, TL and GP was responsible for the study design, data analysis and interpretation of results.

  • Funding This project was funded by the Swedish Research Council for Health, Working Life and Welfare. The SDPP cohort was funded by the Stockholm County Council, the Swedish Research Council, the Diabetes Fund of the Swedish Diabetes Association, Novo Nordisk Scandinavia and GlaxoSmithKline. The SIXTY cohort was funded by the Stockholm County Council and the Swedish Research Council. The SALT cohort was supported by NIH grant AG‐08724. The SNAC‐K cohort was supported by the Ministry of Health and Social Affairs, Sweden and the participating County Councils, Municipalities and University Departments.

  • Competing interests None declared.

  • Patient consent Obtained.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles