Community noise is an important environmental health problem affecting a large number of people. About 50% of the European population live in areas that exceed the World Health Organization guideline value of 55 dB LAeq for outdoor residential areas.1 The hypothesis that exposure to community noise may cause hypertension in humans was first put forward in the 1970s.2 Since then, the question has been explored in both epidemiologic and experimental studies.
Epidemiologic studies of noise exposure and blood pressure (BP) have mainly been performed in occupational settings with high noise levels.3–7 Only a limited number of studies have examined the possible effects of community noise. A recently published German study suggested an association between exposure to road traffic noise and myocardial infarction among men.8 Studies in the Netherlands9 and United Kingdom10 have suggested no association between exposure to road traffic noise and BP, whereas 2 German studies and a recently published Swedish study found positive associations.11–13 Two cross-sectional studies have linked aircraft noise exposure to an increased prevalence of hypertension,2,14 and one study has reported an association between aircraft noise and increased use of medication for cardiovascular diseases.15 There have been no longitudinal studies of aircraft noise and BP.
Experimental studies indicate that noise has the potential to trigger a physiological stress response by activating the sympathetic nervous system and causing an arousal of the neuroendocrine system.16 The release of stress hormones results in various acute hemodynamic and metabolic effects such as elevated BP, aggregation of thrombocytes, and release of free fatty acids into the blood stream.17,18 Further, it has been suggested that long-term noise exposure may lead to chronic dysregulations in the stress mechanism and thus increase the risk for hypertension.19
The aim of this study was to investigate the cumulative incidence of hypertension in relation to residential aircraft noise exposure near the Stockholm international airport, Arlanda.
METHODS
Study Population
The present study is based on 2 epidemiologic surveys conducted within the framework of the Stockholm Diabetes Preventive Program, described in detail elsewhere.20 In 1992–1994, all men age 35–56 years, living in 4 municipalities around Stockholm—Upplands Bro, Sigtuna, Tyresö, and Värmdö—were screened for diabetes and family history of diabetes using a short questionnaire (n = 12,952). Family history of diabetes was defined as at least 1 first-degree relative (mother, father, sister, or brother) or 2 second-degree relatives (grandparents, uncles, or aunts) with diabetes. Of the 10,236 (79%) responding subjects, 2800 provided incomplete information, 1531 had insufficient data on family history of diabetes, 258 had known diabetes, and 212 were of foreign origin. These subjects were excluded from the rest of the study. Since the aim of the original diabetes study was to investigate causes of diabetes in people with and without a family history of the disease, 2 such groups were selected for further investigation. Thus, all 2106 men with a family history of diabetes, along with 2424 men randomly selected within 5-year age groups from among those without family history of diabetes were invited to a baseline survey. Subsequently, 3162 men (70%) agreed to participate. However, a validation process of the family history of diabetes excluded another 33 subjects since the history of diabetes among family members was not confirmed. Altogether, the final sample comprised 3128 men, 1621 with diabetes in the family and 1507 without (1 subject was excluded after not completing the oral glucose tolerance test). All participants answered an extensive questionnaire and underwent a physical examination, including an oral glucose tolerance test and measurements of height, weight, waist and hip circumferences, as well as BP measurements. The questionnaire provided data on various lifestyle factors along with information regarding treatment of hypertension.
After 10 years, ie, between 2002 and 2004, a follow-up survey was conducted. Subjects were not contacted for the follow-up if they (1) were recently diagnosed with diabetes at the baseline examination, (2) were not nationally registered, (3) had moved out of the Stockholm area, or (4) were deceased since the baseline survey (n = 374). Of the contacted subjects (n = 2754), 2392 men (87%) now age 45–66, answered an extended questionnaire and underwent an additional physical examination.
To assess the importance of environmental noise exposure on BP, questions regarding annoyance due to noise, noise sensitivity, and possible diagnosis of hypertension, (including year of diagnosis) were added to the follow-up questionnaire. Incident cases were defined as those who reported that they had been diagnosed with hypertension during the last 10 years in the follow-up questionnaire, or those who had no previous diagnosis of hypertension but attained a BP reading equal of 140/90 mm Hg or higher at the second physical examination. To assess the cumulative incidence of hypertension during follow-up, the analyses were restricted to subjects without treatment for hypertension and a BP reading below 140/90 mm Hg at the baseline survey, and to subjects receiving their diagnosis of hypertension at minimum 1 year after the baseline survey. The cut-off level for high BP (BP ≥140/90 mm Hg) was chosen in accordance with the World Health Organization guideline value for hypertension grade I.21 After these restrictions, the study population comprised 2037 subjects.
Exposure Assessment
For subjects residing near Stockholm-Arlanda airport, residential aircraft noise exposure was assessed by geographical information systems technique. Noise dispersion models from the Swedish Civil Aviation Administration based on air traffic statistics from 1997 were used. The exposure was assessed as time-weighted equal energy (“energy-averaged”) and maximum aircraft noise levels, described in more detail in a previous publication.14 In principle, the energy-averaged aircraft noise levels are based on the 24-hour equivalent sound pressure level (LAeq,24 hours), weighted by time of day, ie, with evening noise events (07.00 pm–10.00 pm) multiplied by a factor of 3 and night-time noise events (10.00 pm–07.00 am) by a factor of 10. The maximum aircraft noise levels are based on the maximum sound pressure level (LAmax) occurring at least 3 times during the average 24-hour period in 1 year.22 Because there had been only minor changes in aircraft types, flight paths, and take-off and landing patterns at the airport during the 1990s, the geographical propagation of aircraft noise was largely unchanged during the follow-up period.23
The energy-averaged aircraft noise levels were provided in 5 dB(A) contours on a digital map; thus, each address was plotted and classified in the exposure categories 50–55, 55–60, 60–65, or >65 dB(A). Some of the exposed subjects had more than one address during the 10-year follow-up period; for them, we estimated a linear time-weighted exposure. Only 3 subjects were classified as having a noise level exceeding 65 dB(A) and these were merged with the exposure category 60-65 dB(A). The exposure of 10 subjects could not be assessed due to incomplete address data, resulting in a final study population of 2027 subjects. In total, 411 subjects (20%) were exposed to energy-averaged levels of 50 dB(A) or above. Subjects exposed to less than 50 dB(A) constituted the reference group (n = 1616).
The exposure assessment according to maximum noise levels was performed in a similar way. The acquired exposure data as maximum noise levels were provided in 1 dB(A) contours ranging from 70 to 85 dB(A). Subjects were classified as unexposed if their address had a maximum aircraft noise level of less than 70 dB(A) or exposed according to 1 of the 3 equally sized exposure categories 70–72, 73–75, and above 75 dB(A). The Spearman's correlation between the categorical energy-averaged and maximum aircraft noise levels in the sample was 0.87.
Noise from other sources was reported in the questionnaire as the degree of annoyance due to road traffic, railway traffic, or occupational noise. Four alternatives were given regarding the frequency of annoyance: “never,” “a few times per month,” “every week,” and “every day.” Subjects were classified as “not annoyed” (“never” or “a few times per month”) by any other noise source or “annoyed” (“every week” or “every day”) by at least one of the other noise sources.
Statistical Analysis
Relative risks (RRs) and 95% confidence intervals (CIs) were estimated by binomial regression with the log link function.24–26 In a few instances the model did not converge and so we used log-Poisson models, which provide consistent, but not fully efficient, estimates of the RR and its CI.27 Categorical exposure variables were used to explore possible exposure-response patterns. We assessed linear increase of hypertension incidence across aircraft noise exposure categories. All RRs were adjusted for age (5-year age groups) and body mass index (BMI) (<25, 25–30, >30 kg/m2). Other models were also evaluated, including family history of diabetes (yes, no), glucose tolerance (normal, impaired/diabetes), smoking (never, former, current), physical activity (sedentary/low, moderate/high), annoyance due to noise from other sources (annoyed, not annoyed), and socioeconomic status (low, medium, high), according to the Swedish Socioeconomic Classification Index.28 However, adjusting for these covariates did not markedly change the estimates.
Subgroup analyses (excluding subjects smoking or using snuff directly preceding the BP measurements) were performed to reduce imprecision in the disease classification. We assessed potential effect modification between aircraft noise exposure and other covariates by stratified analyses, and P values for interaction terms in the regression models are presented. The cut-off value for age was set at the mean of all subjects (57 years). All statistical analyses were performed with Stata 8.0.
RESULTS
Background characteristics of subjects by exposure to aircraft noise are presented in Table 1. The proportion of subjects with a family history of diabetes was slightly higher in the exposed group. Moreover, subjects in the exposed group tended to have a higher BMI, be less physically active, generally be more annoyed by noise from other sources, and have a lower socioeconomic status.
;)
TABLE 1: Selected Characteristics of Men Exposed to Aircraft Noise Around Stockholm Arlanda Airport [Energy-Averaged Levels ≥50 dB(A)] and Unexposed Men [Below 50 dB(A)]
Table 2 displays crude and adjusted RRs for hypertension using continuous and dichotomized exposure variables. After adjustments for age and BMI, the RR for those exposed to energy-averaged levels above 50 dB(A) was 1.19 (95% CI = 1.03–1.37) and 1.20 (1.03–1.40) for those exposed to maximum levels above 70 dB(A). Analyses excluding subjects who smoked or used snuff directly preceding BP measurements resulted in somewhat stronger associations, with an adjusted RR of 1.29 (1.11–1.50) and 1.32 (1.12–1.55), respectively (Table 3). When including noise as a continuous variable in the regression model, the RR increased by 10% per 5 dB(A) of energy-averaged noise levels and per 3 dB(A) of maximum levels. In the analyses with restriction to nonusers of tobacco before the BP measurements the RR increased by 15% per 5 dB(A) of energy-averaged noise levels or per 3 dB(A) of maximum levels.
TABLE 2: Association Between Aircraft Noise Exposure and Cumulative Incidence of Hypertension Among Men in Stockholm
TABLE 3: Association Between Aircraft Noise Exposure and Cumulative Incidence of Hypertension Among Men in Stockholm Following Exclusion of Those Smoking or Using Snuff Directly Preceding Blood Pressure Measurements
Results of the categorical analyses are presented in Figures 1 and 2. Within the categories of energy-averaged noise levels, the percentage of incident cases was 36% (85 of 234), 35% (52 of 147), and 38% (11 of 29). This yielded adjusted RRs of 1.18 (0.99–1.41), 1.20 (0.96–1.48), and 1.22 (0.79–1.87), respectively. An analysis of linear increase over these categories resulted in an adjusted RR of 1.10 (1.01–1.19) per 5 dB(A) increase in energy-averaged noise levels. The corresponding percentage of incident cases in each category of maximum noise levels was 36% (57 of 159), 33% (33 of 100), and 44% (23 of 52), yielding adjusted RRs of 1.18 (0.96–1.44), 1.17 (0.90–1.52), and 1.32 (0.97–1.79), respectively. The adjusted RR for a linear increase over these categories was 1.10 (1.02–1.19) per 3 dB(A) increase in maximum noise levels.
FIGURE 1.:
Relative risk for hypertension among men in Stockholm according to different levels of energy-averaged aircraft noise exposure (bars indicating 95% CI), adjusted for age and BMI.
FIGURE 2.:
Relative risk for hypertension among men in Stockholm according to different levels of maximum aircraft noise exposure (bars indicating 95% CI), adjusted for age and BMI.
The stratified analyses indicated effect modification by age (Fig. 3). The RR for subjects 57 years and older was 1.36 (1.14–1.62) and for younger subjects 1.00 (0.80–1.26). The associations also appeared to be more pronounced among subjects with a normal glucose tolerance (RR = 1.29; 95% CI = 1.10–1.52), nonsmokers (1.33; 1.10–1.62), and subjects not annoyed by noise from other sources (1.27; 1.09–1.48).
FIGURE 3.:
Relative risk for hypertension among men in Stockholm associated with aircraft noise exposure by level of different covariates, adjusted for age and BMI (bars indicating 95% CI). P values for interaction between aircraft noise exposure and each covariate are presented.
DISCUSSION
Our results suggest that aircraft noise exposure is associated with an increased risk of developing hypertension. Cross-sectional studies have reported associations between exposure to aircraft noise and hypertension or medical treatment for hypertension.2,14,15 In a previous Swedish study around Arlanda airport, we investigated the association between aircraft noise exposure and prevalence of hypertension.14 The adjusted odds ratio for subjects exposed to energy-averaged levels above 55 dB(A) was estimated to be 1.6 (1.0–2.5) and there was a linear risk increase of 30% per 5 dB(A). Some overlap of study subjects between our 2 studies might have occurred, although this would apply to a maximum of 25 aircraft-noise-exposed subjects diagnosed with hypertension. Though our findings are generally consistent with previous reports, it is difficult to make direct comparisons of results because of methodologic differences—in particular the absence of earlier longitudinal data.
Our findings suggest an effect primarily among older subjects. This might be due to a prolonged period of exposure (since most elderly subjects in the cohort have lived more than 10 years at their present address), or it may be that older people are more sensitive to noise. The association between aircraft noise and hypertension also appeared primarily among those with normal glucose tolerance, never-smokers, and those not annoyed by noise from other sources. There are several possible reasons for these results. For example, normal glucose tolerance could indicate a lower burden of other cardiovascular risk factors, which were not controlled for. A stronger effect in never-smokers might be due to uncontrolled or residual confounding among smokers. Higher estimates in those not annoyed by noise suggest that concomitant exposure to noise from other sources may blur the picture. It is expected that the estimates for those with normal glucose tolerance, never-smokers, and subjects not annoyed by other noise sources would be less affected by residual confounding. However, the interaction analyses should be interpreted with caution.
The strengths of this study mainly relate to the longitudinal, and largely objective, assessment of both exposure and disease outcome. In addition, extensive covariate data made it possible to evaluate a large number of possible confounding factors. However, uncontrolled or residual confounding, exposure and disease misclassification, and selection bias all need to be considered. Misclassification of noise exposure might be present, especially in the total burden of noise exposure. For example, exposure to aircraft noise may also be present at other locations than at home. However, it is unlikely that the classification of exposure would be dependent on disease status, since these were assessed independently of each other; thus any such bias would instead lead to dilution of the association.
Misclassification of disease may have occurred due to partly subjective assessment by questionnaire and to the large number of subjects who used tobacco prior to the BP measurements. Smoking and nicotine have short-term physiological effects on BP.29,30 Exclusion of those using tobacco before the BP measurements resulted in stronger associations, suggesting that the results may have been affected by nondifferential misclassification of disease.
Our study design over-sampled subjects with a family history of diabetes and thus, the findings may not be directly translated to the general population. However, no association between a family history of diabetes and hypertension was found in the cohort; the cumulative incidence of hypertension was 32% among those with a family history of diabetes and 30% among those without such family history. Moreover, the stratifications indicated an effect of the noise exposure also on subjects without a family history of diabetes, and suggested a stronger association among subjects with a normal glucose tolerance.
In conclusion, our data suggest an association between aircraft noise exposure and hypertension incidence in middle-aged Swedish men.
REFERENCES
1. World Health Organisation (WHO), Berglund B, Lindvall T, et al, eds.
Guidelines for Community Noise. Geneva: WHO; 1999.
2. Knipschild P. V. Medical effects of aircraft noise: community cardiovascular survey.
Int Arch Occup Environ Health. 1977;40:185–190.
3. Powazka EE. A cross-sectional study of occupational noise exposure and blood pressure in steelworkers.
Noise Health. 2003;5:15–22.
4. Talbott EO, Gibson LB, Burks A, et al. Evidence for a dose-response relationship between occupational noise and blood pressure.
Arch Environ Health. 1999;54:71–78.
5. Kristal-Boneh E, Melamed S, Harari G, et al. Acute and chronic effects of noise exposure on blood pressure and heart rate among industrial employees: the Cordis study.
Arch Environ Health. 1995;50:298–304.
6. Fogari R, Zoppi A, Vanasia A, et al. Occupational noise exposure and blood pressure.
J Hypertens. 1994;12:475–479.
7. Hessel PA, Sluis-Cremer GK. Occupational noise exposure and blood pressure: longitudinal and cross-sectional observations in a group of underground miners.
Arch Environ Health. 1994;49:128–134.
8. Babisch W, Beule B, Schust M, et al. Traffic noise and risk of myocardial infarction.
Epidemiology. 2005;16:33–40.
9. Knipschild P, Salle H. Road traffic noise and cardiovascular disease. A population study in The Netherlands.
Int Arch Occup Environ Health. 1979;44:55–59.
10. Babisch W, Ising H, Gallacher JE, et al. Traffic noise and cardiovascular risk: the Speedwell study, first phase. Outdoor noise levels and risk factors.
Arch Environ Health. 1993;48:401–405.
11. Herbold M, Hense HW, Keil U. Effects of road traffic noise on prevalence of hypertension in men: results of the Luebeck blood pressure study.
Soz Praventivmed. 1989;34:19–23.
12. Neus H, Ruddel H, Schulte W, et al. The long-term effect of noise on blood pressure.
J Hypertens. 1983;1(suppl 2):251–253.
13. Bluhm GL, Berglind N, Nordling E, et al. Road traffic noise and hypertension.
Occup Environ Med. 2007;64:122–126.
14. Rosenlund M, Berglind N, Pershagen G, et al. Increased prevalence of hypertension in a population exposed to aircraft noise.
Occup Environ Med. 2001;58:769–773.
15. Franssen EA, van Wiechen CM, Nagelkerke NJ, et al. Aircraft noise around a large international airport and its impact on general health and medication use.
Occup Environ Med. 2004;61:405–413.
16. Ising H, Braun C. Acute and chronic endocrine effects of noise: review of the research conducted at the Institute for water, soil and air hygiene.
Noise Health. 2000;2:7–24.
17. Spreng M. Possible health effects of noise induced cortisol increase.
Noise Health. 2000;2:59–64.
18. Babisch W. Stress hormones in the research on cardiovascular effects of noise.
Noise Health. 2003;5:1–11.
19. WHO.
Report on the Second Meeting on Night Noise Guidelines. Geneva: World Health Organisation, European Centre for Environment and Health; 2004.
20. Carlsson S, Persson PG, Alvarsson M, et al. Weight history, glucose intolerance, and insulin levels in middle-aged Swedish men.
Am J Epidemiol. 1998;148:539–545.
21. WHO. Guidelines for the management of hypertension.
J Hypertens. 1999;17:151–183.
22. Swedish Civil Aviation Administration.
Swedish Aircraft Noise Calculation Model. Revision 4. Stockholm: Swedish Civil Aviation Administration; 2003.
23. Swedish Civil Aviation Administration.
Yearly Environmental Reports, Stockholm-Arlanda Airport (in Swedish), 1992–2002. Stockholm: Swedish Civil Aviation Administration.
24. Wacholder S. Binomial regression in GLIM: estimating risk ratios and risk differences.
Am J Epidemiol. 1986;123:174–184.
25. Skov T, Deddens J, Petersen MR, et al. Prevalence proportion ratios: estimation and hypothesis testing.
Int J Epidemiol. 1998;27:91–95.
26. Greenland S. Model-based estimation of relative risks and other epidemiologic measures in studies of common outcomes and in case-control studies.
Am J Epidemiol. 2004;160:301–305.
27. Zou G. A modified poisson regression approach to prospective studies with binary data.
Am J Epidemiol. 2004;159:702–706.
28. Statistics Sweden.
Swedish Socioeconomic Classification. Stockholm: Statistics Sweden; 1989.
29. Wolk R, Shamsuzzaman AS, Svatikova A, et al. Hemodynamic and autonomic effects of smokeless tobacco in healthy young men.
J Am Coll Cardiol. 2005;45:910–914.
30. Kim JW, Park CG, Hong SJ, et al. Acute and chronic effects of cigarette smoking on arterial stiffness.
Blood Press. 2005;14:80–85.
