Diesel vehicle emission and death rates in Tokyo, Japan: a natural experiment

Evidence linking air pollution with adverse cardiopulmonary outcomes is accumulating. However, few studies have been conducted to evaluate whether vehicle emission control improves public health. We thus evaluated the effect of a diesel emission control law on mortality rates in 23 wards of Tokyo metropolitan area, Japan. We obtained daily counts of mortality and concentrations of nitrogen dioxide (NO(2)) and particulate matter less than 2.5 μm in diameter (PM(2.5)) from April 2003 to December 2008. Time-series and interrupted time-series analysis were employed to analyze the data in two periods: prior to the introduction of tighter restrictions (April 2003 to March 2006) and after the enforcement (April 2006 to December 2008). Concentrations of air pollutants gradually decreased during the study period: from 36.3 ppb (NO(2)) and 22.8 μg/m(3) (PM(2.5)) to 32.1 ppb and 20.3 μg/m(3), respectively. Air pollutants were positively associated with circulatory and pulmonary disease mortality, especially cerebrovascular disease. Each same-day PM(2.5) increase of 10 μg/m(3) was associated with a 1.3% increase in cerebrovascular mortality rate (95% confidence interval: 0.2-2.4). Rate ratios were attenuated after the enforcement in most of the outcomes, probably due to reduced toxicity of the pollutants. In the crude interrupted time-series analysis, reductions of standardized mortality rates after the enforcement were the greatest in high traffic areas. Even after adjustment of longer-time trend, mortality rate from cerebrovascular disease was reduced by 8.50% (p<.001) with dose-response relationship. However, the declines in other cause-specific mortality became equivocal. This natural experiment in Tokyo suggests that emission controls improved air quality. Although suggestive, further data are needed to conclusively demonstrate an impact on mortality rates.