Associations between Traffic Noise, Particulate Air Pollution, Hypertension, and Isolated Systolic Hypertension in Adults: The KORA Study

Background: Studies on the association between traffic noise and cardiovascular diseases have rarely considered air pollution as a covariate in the analyses. Isolated systolic hypertension has not yet been in the focus of epidemiological noise research. Methods: The association between traffic noise (road and rail) and the prevalence of hypertension was assessed in two study populations with a total of 4,166 participants 25–74 years of age. Traffic noise (weighted day–night average noise level; LDN) at the facade of the dwellings was derived from noise maps. Annual average PM2.5 mass concentrations at residential addresses were estimated by land-use regression. Hypertension was assessed by blood pressure readings, self-reported doctor-diagnosed hypertension, and antihypertensive drug intake. Results: In the Greater Augsburg, Germany, study population, traffic noise and air pollution were not associated with hypertension. In the City of Augsburg population (n = 1,893), where the exposure assessment was more detailed, the adjusted odds ratio (OR) for a 10-dB(A) increase in noise was 1.16 (95% CI: 1.00, 1.35), and 1.11 (95% CI: 0.94, 1.30) after additional adjustment for PM2.5. The adjusted OR for a 1-μg/m3 increase in PM2.5 was 1.15 (95% CI: 1.02, 1.30), and 1.11 (95% CI: 0.98, 1.27) after additional adjustment for noise. For isolated systolic hypertension, the fully adjusted OR for noise was 1.43 (95% CI: 1.10, 1.86) and for PM2.5 was 1.08 (95% CI: 0.87, 1.34). Conclusions: Traffic noise and PM2.5 were both associated with a higher prevalence of hypertension. Mutually adjusted associations with hypertension were positive but no longer statistically significant. Citation: Babisch W, Wolf K, Petz M, Heinrich J, Cyrys J, Peters A. 2014. Associations between traffic noise, particulate air pollution, hypertension, and isolated systolic hypertension in adults: the KORA Study. Environ Health Perspect 122:492–498; http://dx.doi.org/10.1289/ehp.1306981


METHODS, ADDITIONAL DESCRIPTIONS 2
Noise assessment 2 Disentangling road and railway noise 3 TABLES 5 Table S1: Response statistic 5

Sample selection
For the selection of study subjects a stratified random sample (quota sampling) of 6,640 subjects was drawn from the registry office (appointed date 1 st June 1999). Stratification criteria were age (10 year blocks), gender and region (City of Augsburg, Greater Augsburg), resulting in 10 strata of 664 subjects. In the City of Augsburg stratified random sampling was done directly. In Greater Augsburg a two-step procedure was applied. In a first step 16 of 70 municipalities were randomly selected ("probability proportional to size"), and in a second step the random sampling of subjects was done. Table S1 shows the breakdown statistic of the response. For the agereference the 6 th of June 2000 was chosen, which was in the middle of the data collecting phase.

Noise assessment
Traffic counts including vehicle composition -as obtained from the road construction office of The geo-coded addresses of the study subjects were obtained from the cadastral land register of the City of Augsburg and the surrounding villages of Greater Augsburg. In unclear situations (e. g. missing house numbers), the addresses were visited for visual inspection. Day and night noise

Disentangling road and railway noise
The 2001 noise data did not distinguish explicitly between the two noise sources. For sensitivity analyses (exclusion of subjects), a method was developed to identify participants where railway noise was potentially the dominant noise source at home. This was done using the City of Augsburg noise data from 2009 because separate road and railway noise levels were available from the these noise maps for day (L Aeq16h ) and for the night-time (L Aeq8h ).
Railway noise levels included the so called "railway bonus" of 5 dB(A), meaning that the railway noise levels were 5 dB(A) lower than actually measured or calculated. In a first step the 2009 railway noise levels were shifted by 5 dB(A) to eliminate the '"railway bonus"', and the differences between the road noise levels and the railway noise levels were calculated for the day and the night-time. If either the day noise level or the night noise level of the railway noise was at least 5 dB(A) higher than the respective road noise level, the subjects were classified as predominantly railway noise exposed. This criterion was a pragmatic setting considering that a sound level difference of 3 dB(A) (sound intensity double as high) is just about audible, and that a difference of 10 dB(A) (ten-fold sound intensity) is perceived to be twice as loud. The criterion was fulfilled for 19.0% of participants of the City of Augsburg sample that had not moved on the basis of the 2009 noise data.
In the next step a method was to be found that could be applied to the 2001 noise data and that identified largely the same subjects. It is a common experience that road noise levels in urban streets (no motorways) fall by approximately 7 to 11 dB(A) during the night-time compared with the day-time. This is less the case for railway noise due to increased freight traffic during the night. This was confirmed by the 2009 noise data showing mean differences between day and night noise levels of 9 dB(A) (standard deviation SD = 1.0) dB(A) for road traffic noise, and 3 dB(A) (SD = 2.7) for railway noise, respectively. Using this information and following up the concept that the railway noise level (without railway bonus) has to be at least 5 dB(A) higher than the road noise level in order to be the dominant source, a day-night difference of 6 dB(A) or less of the 2009 total noise levels (road + rail including the railway bonus) was found to be an alternative criterion for the identification of participants that were potentially exposed to dominant railway noise. This was the case for 22.7% of participants of the City of Augsburg.
The statistical sensitivity of this alternative criterion was 0.90, the specificity 0.93 (2009 data).
When this alternative criterion was applied to the 2001 noise data it was estimated that railway noise was the dominant noise source for 25.2 % and 16.3 % of participants in the City of Augsburg and Greater Augsburg, respectively. The "Helmert Index" is based on school education, professional status, family income.
b Households with less than 1.250 € income per 5 x 5 km grid.  1.07, 6.84) a The "Helmert Index" is based on school education, professional status, family income. b Households with less than 1.250 € income per 5 x 5 km grid. Figure S1. Association between traffic noise (noise level categories) and the prevalence of hypertension, adjusted for age, gender, smoking, alcohol intake, body mass index, physical activity, socio-economic status (City of Augsburg). Figure S2. Association between traffic noise (noise level categories) and the prevalence of hypertension, adjusted for age, gender, smoking, alcohol intake, body mass index, physical activity, socio-economic status (Greater Augsburg). Figure S3. Association between traffic noise (noise level categories) and the prevalence of isolated systolic hypertension, adjusted for age, gender, smoking, alcohol intake, body mass index, physical activity, socio-economic status (City of Augsburg).