Long term effects of residential NO(x) exposure on total and cause-specific mortality and incidence of myocardial infarction in a Swedish cohort

Background and aims: Exposure to air pollution has been linked to total and cardiopulmonary mortality. However, few studies have examined the effects of exposure over decades, or which time windows of long term exposure are most relevant. We investigated the long term effects of residential air pollution on total and cause-specific mortality and incidence of myocardial infarction in a well-characterized cohort of men in Sweden.

Methods: A cohort of 7494 men in Gothenburg was examined in 1970-1973 and followed subsequently to determine predictors of cardiovascular disease. We collected data on residential address and cause-specific mortality for the years 1973-2007. Each individual was assigned yearly nitrogen oxides (NOx) exposure based on dispersion models. Using multivariable Cox regression and generalized additive models with time-dependent exposure, we studied the association between three different time windows of residential NOx exposure, and selected outcomes.

Results: In the years 1973-2007, a total of 5669 deaths, almost half of which were due to cardiovascular diseases, occurred in the cohort. Levels of NOx exposure decreased during the study period, from a median of 38 µg/m(3) in 1973 to 17 µg/m(3) in 2007. Total non-accidental mortality was associated with participants' NOx exposure in the last year (the year of outcome) (HR 1.03, 95% CI 1.01-1.05, per 10 µg/m(3)), with the mean NOx exposure during the last 5 years, and with the mean NOx exposure since enrolment (HR 1.02, 95% CI 1.01-1.04 for both). The associations were similar (HR 1.01-1.03), but generally not statistically significant, for cardiovascular, ischemic heart disease, and acute myocardial infarction mortality, and weaker for cerebrovascular and respiratory mortality. There was no association between NOx exposure and incident myocardial infarction.

Discussion and conclusions: Long term residential exposure to NOx at these relatively low exposure levels in Gothenburg was associated with total non-accidental mortality. The association was as strong for NOx exposure in the last year as for longer exposure windows. The effect was near linear, and only marginally affected by confounders and effect modifiers. The improved air quality in Gothenburg has by these estimates led to a 6% decrease in excess non-accidental mortality during the study period.