Air pollution and hospital admissions for cardiovascular disease in Tucson

Several recent studies have reported associations between short-term changes in both inhalable particles (PM10) and carbon monoxide and cardiovascular hospital admissions. Here, I seek to replicate those findings in a location where sulfur dioxide concentrations are low and poorly correlated with PM10, and where PM10 concentrations peak in the winter when ozone is lowest. This setting allows the opportunity to separate the effects of different air pollutants. I constructed daily counts of admissions to all hospitals in Tucson, AZ, for cardiovascular disease (International Classification of Diseases, 9th revision, codes 390-429) for persons age 65 years and older. I analyzed these admission counts in a Poisson regression, on temperature, humidity, day of the week indicators, and air pollution. I removed long wavelength patterns using a nonparametric smooth function of day of study. I used regression splines to model possible nonlinearities in the dependence of hospital admissions on weather. I then examined sensitivity analyses to control for weather. Both PM10 and carbon monoxide were associated with increased risk of cardiovascular hospital admissions. Admissions increased by 2.75% [95% confidence limits (CL) = 0.52%, 5.04%] for an interquartile range increase (23 micrograms per m3) in PM10 and by 2.79% (95% CL = 0.51%, 5.41%) for an interquartile range increase (1.66 parts per million) in carbon monoxide. These associations were independent and additive. In contrast, I saw little association with sulfur dioxide [increase of 0.14% (95% CL = -1.3%, 1.6%) for an interquartile range increase in exposure], ozone [increase of 0.54% (95% CL = -2.3%, 3.45%)], or nitrogen dioxide [increase of 0.69% (95% CL = -2.3%, 3.8%)]. The air pollution associations were insensitive to control for a potential interaction between temperature and humidity and to control for temperature and humidity on more than 1 day.