Road traffic noise and risk for cardiovascular disease and

diabetes

Mette Sørensen, Senior Researcher

Danish Cancer Society Research Centre

Road traffic noise and MI

No adjustment for air pollution

Wolfgang Babisch, Noise and Health, 2008

• Meta-analyses of 7 case-control and cohort studies on road traffic noise and myocardial infarction (MI)

• Dose-response relationship

Babisch W, Noise and Health 2008; 10:27-33

Road traffic noise and MI

Two studies after 2008 included both road traffic noise and air pollution:

1. Selander, Epidemiology, 2009Road traffic noise increase risk for incident MI after adjustment for air pollution (modeled NO2)

2. Beelen, Occup Environ Med, 2009Before air pollution adjustment weak tendency to higher risk for mortality from IHD at road traffic noise > 65 dBAfter adjustment (BS, traffic intensity) no association

Lund, April 2013

Stroke and diabetes

Lund, April 2013

STROKE• Many risk factors in

common with MI• Major risk factor is

hypertension• WHO: 15 million people

suffer stroke each year

DIABETES• Stress and sleep

disturbances has been associated with• Insulin resistance• High morning glucose• Appetite regulation• Type 2 diabetes

• WHO: at least 171 million people worldwide have diabetes

Hypotheses

Exposure to road traffic noise increase the risk for:• Incident myocardial infarction

• Incident stroke

• Diabetes

-> also after adjusting for air pollution

Lund, April 2013

The cohort

Diet, Cancer and Health cohort• Enrolment in 1993-1997

• 160 725 persons aged 50-64 years invited

• Copenhagen and Aarhus

• 57 053 participated

• Questionnaire (diet, smoking, alcohol, education etc.)

• Weight, height and waist circumference measured

Lund, April 2013

Noise exposure

Addresses• Central Population Registry

• Since 1971 all residents addresses (incl. change of address) recorded

• All residential addresses between 1987 and 2006 using the PIN

• Complete address history for ≈ 54 000 participants

• All in all ≈ 62 000 addresses

• Geocoded

Lund, April 2013

Noise exposure

SoundPLAN – the Nordic Prediction Method• Geographical coordinates for each address (≈ 62 000 addresses)

• Height (floor) for each address

• Building polygons

• All road lines with > 1000 vehicles

• Traffic composition (heavy/light)

• Yearly average daily traffic

• Traffic speed

Lund, April 2013

Road noise at enrolment

Lden, dB N Percent

< 50 5541 10

50 – 55 16389 31

55 – 60 12684 24

60 – 65 10011 19

65 – 70 5728 11

70 – 75 2490 5

≥ 75 319 1

Lund, April 2013

Air pollution

Danish AirGIS modeling system• Dispersion model

• At each address from 1987 until 2006

• Calculates the sum of: local, urban background and

regional background contributions

• NOx: indicator of air pollution; correlates with traffic such

as ultrafine particles (r = 0.93) and PM10 (r = 0.70)

Lund, April 2013

Identification of cases

Myocardial infarction (MI) and stroke• Danish National Hospital Registry (1977):

• registers ALL discharges from Danish hospitals• follow-up cohort using PIN (censoring in 2006)

• MI: also validation using medical records• Primary event - exclude participants with stroke/MI before

enrolment• In total 1881 stroke cases and 1600 MI cases

Lund, April 2013

Identification of cases

Diabetes• The National Diabetes Registry (1995)

• Registers all diabetes based on information from other registers:• Hospital admission (National Hospital Registry)• Medication (National Prescription Registry)• Reimbursement for chiropody due to diabetes (National Health

Service Registry) • Blood glucose tests (National Health Service Registry)

• In total 3869 cases• Exclude participants with diabetes before enrolment (questionnaire)

Lund, April 2013

DesignCox proportional hazard model• Follow-up until: stroke/MI/diabetes (event), death, emigration or

censoring (June 2006)

• Age underlying time

• Road traffic noise: time-varying - annual mean of Lden at the address at

time of diagnosis (5-y preceding diagnosis)

• Adjustment:

• sex, calendar-year, socioeconomic factors, railways and airport noise,

air pollution (NOx)

• Lifestyle confounders, e.g. smoking status, BMI and intake of fruit,

vegetables and alcoholLund, April 2013

Results, stroke/MI

Lden, per 10 dB

Stroke Cases

StrokeAdjusted IRR

(95% CI)

MICases

MIAdjusted IRR

(95% CI)

All 1881 1.14 (1.03–1.25) 1600 1.12 (1.02-1.22)

Sex

Men 1109 1.14 (1.02–1.27) 1188 1.14 (1.03-1.26)

Women 772 1.13 (1.00–1.28) 412 1.06 (0.91-1.23)

Age at event

< 65 y 952 1.02 (0.91–1.14)* 871 1.06 (0.95-1.18)

≥ 65 y 929 1.27 (1.13–1.43)* 729 1.19 (1.06-1.34)

* P interaction (< 0.001)

Lund, April 2013

Adjusted for age, sex, BMI, lifestyle habits, socioeconomic factors, railway and airport noise, air pollution

Results, rcs

Stroke Myocardial infarction

Lund, April 2013

Results, stroke age

Lund, April 2013

Results, diabetes

Lund, April 2013

Road traffic noiseLden, per 10 dB

Diabetes Cases

DiabetesAdjusted IRR

(95% CI)

All diabetes

Noise at diagnosis 3 869 1.08 (1.02-1.14)

5-y noise average preceding diagnosis 3 869 1.11 (1.05-1.18)

Confirmed diabetes *

Noise at diagnosis 2 752 1.11 (1.03-1.19)

5-y noise average preceding diagnosis 2 752 1.14 (1.06-1.22)

* Exclusion of cases only included in the National Diabetes Registry based on blood glucose tests.Adjusted for age, sex, BMI, waist circumference, lifestyle habits, socioeconomic factors, railway and airport noise, air pollution

Results, diabetes

Lund, April 2013

Conclusions

• Exposure to residential road traffic noise was associated

with a higher risk for stroke, MI and diabetes

• For stroke the association was strongest among the oldest

• For stroke indications of threshold limit around 60 dB

• For MI no threshold limit

• For diabetes indications of level off around 65 dB

Lund, April 2013

Strenghts

• Large study with good power

• Prospective design

• Follow-up in nationwide registers

• For MI – validation by medical records

• Historical address information

• Adjustment for air pollution

Lund, April 2013

Weaknesses

• No information on noise barriers

• No information on the bedroom location

• Stroke – no information on subtype (hemorrhagic/ischemic)

• Diabetes – unknown whether type I or II (though exclusion of

all diabetes before age 50)

Lund, April 2013

Co-authorsDanish Cancer Society:

Ole Raaschou-Nielsen, Zorana Andersen, Rikke Baastrup, Anne Tjønneland

Danish EPA: Jørgen Jakobsen

Aarhus University, Denmark: Kim Overvad, Martin Hvidberg, Steen Jensen

Rambøll, Denmark: Kenneth Lillelund

Utrecht University, The Netherlands: Rob Beelen

FUNDING:

• Danish EPA

• Research Centre for Environmental Health, Danish Ministry of the Interior and Health

• European Research Council, EU 7th Research Framework Programme (grant 281760)

Lund, April 2013

Thank you for your attention