ETH Zürich, D-MTEC Public and Organizational Health, Zürich, Switzerland, email@example.com
To the Editor:
Huss et al1 present the first large-scale epidemiologic study investigating a link between residential exposure to aircraft noise and mortality from myocardial infarction (MI) in Switzerland. The study was carried out carefully, and the reported results are compatible with the existing literature. However, the paper is somewhat vague as regards the origin and reliability of the air-traffic data underlying exposure assessments, the computational-noise models that were used for the exposure calculations, and, most importantly, the time period for which aircraft noise exposure was assessed and upon which the analysis is based. Because cardiovascular disease is usually the result of accumulated exposure over long time spans, individual integrated-exposure histories dating back as far as possible would be the predictor of choice. The paper does not sufficiently clarify the time period of noise exposure that underlies the analysis of mortality. This could be the average exposure of a reference year (eg, 2005), or an average over several years. Depending on operational parameters, the short-term exposure in a reference year at a particular location is just a proxy for the relevant long-term exposure at that location. The average noise exposure around most airports has steadily decreased in recent decades, owing to the replacement of old and noisy aircraft with newer types. A peculiarity of the Zurich Airport (which accounts for much of the noise exposure in this study) may additionally shift the reported exposure-effect relationship: the years 2001 to 2005 (the time span for which exposure data were apparently available) were characterized by a decrease of aircraft movements at Zurich Airport due to the demise of Swissair in 2001. Because of this, people who have lived 10-15 years or longer at the same place (eg, those from the Model III subpopulation, reported in the original Table 3) may have an increased risk due not only to their longer exposure time, but also because they may enter the model with an underestimated exposure dose. If this is true, hazard ratios would begin to increase at slightly higher exposure values. To reduce such sources of bias, it seems advisable to consider the actual noise exposure history of each person as closely as possible, however, troublesome this may be.
D-MTEC Public and Organizational Health
1.Huss A, Spoerri A, Egger M, Röösli M. Aircraft noise, air pollution, and mortality from myocardial infarction. Epidemiology