Schwartz J
Department of Environmental Health, Harvard School of Public Health, Boston, MA 02115, USA.
Epidemiology. 1997 Jul;8(4):371-7. doi: 10.1097/00001648-199707000-00004.
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.
最近的几项研究报告了可吸入颗粒物（PM10）和一氧化碳的短期变化与心血管疾病住院之间的关联。在此，我试图在一个二氧化硫浓度较低且与PM10相关性较差、PM10浓度在臭氧浓度最低的冬季达到峰值的地点重现这些发现。这种环境提供了区分不同空气污染物影响的机会。我统计了亚利桑那州图森市所有医院65岁及以上人群因心血管疾病（国际疾病分类第九版，代码390 - 429）的每日住院人数。我在泊松回归模型中分析了这些住院人数，纳入了温度、湿度、星期几指标以及空气污染因素。我使用研究日期的非参数平滑函数去除了长波长模式。我使用回归样条来模拟住院人数对天气的依赖关系中可能存在的非线性。然后我进行了敏感性分析以控制天气因素。PM10和一氧化碳均与心血管疾病住院风险增加相关。PM10每增加一个四分位间距（23微克每立方米），住院人数增加2.75% [95%置信区间（CL）= 0.52%，5.04%]；一氧化碳每增加一个四分位间距（百万分之1.66），住院人数增加2.79%（95% CL = 0.51%，5.41%）。这些关联是独立且相加的。相比之下，我发现二氧化硫 [暴露增加一个四分位间距时增加0.14%（95% CL = -1.3%，1.6%）]、臭氧 [增加0.54%（95% CL = -2.3%，3.45%）] 或二氧化氮 [增加0.69%（95% CL = -2.3%，3.8%）] 与住院人数的关联不大。空气污染与住院人数的关联对控制温度和湿度之间的潜在相互作用以及控制多天的温度和湿度不敏感。
