A case-crossover analysis of the impact of PM(2.5) on cardiovascular disease hospitalizations for selected CDC tracking states

Environ Res. 2014 Oct;134:455-65. doi: 10.1016/j.envres.2014.06.018. Epub 2014 Sep 30.


Background: Information is currently being collected by the CDC Environmental Public Health Tracking (EPHT) network on hospitalizations due to Acute Myocardial Infarction (AMI) and there is interest by CDC in exploring the relationship between fine particulate matter (PM2.5) and other cardiovascular (CVD) outcomes in the context of the EPHT program. The goal of this study was to assess the short term effects of daily PM(2.5) air pollution levels on hospitalizations for CVD for seven states within the CDC EPHT network (Florida, Massachusetts, New Hampshire, New Jersey, New Mexico, New York, and Washington).

Methods: Hospitalization data was obtained for 2001-2008 admissions for circulatory disease (primary discharge diagnosis of ICD-9 codes 390-459) from data stewards in those states and included admission date, age, gender, and zip code of residence. We used CMAQ-derived predicted daily PM2.5 data as estimated by EPA at the centroid of each Census Bureau Zip Code Tabulation Area (ZCTA) and linked to zip code of patient residence. A time-stratified case-crossover study design with conditional logistic regression was used to evaluate the short-term association of PM2.5 on risk of non-elective hospitalizations for CVD. Specifically, we considered all circulatory disease, ischemic heart disease, acute myocardial infarction, heart failure, cardiac arrhythmia, cerebrovascular disease and peripheral vascular disease endpoints.

Results: Data were obtained on over 7,500,000 hospitalizations for this time period. Mean annual PM2.5 exposure levels were lowest for New Mexico and Washington (6.5 μg/m3 PM2.5 and 8.4 μg/m3 PM2.5). New Jersey, New York and Massachusetts exhibited the highest annual averages for PM2.5, (12.8 μg/m3, 11.1 μg/m3 and 10.8 μg/m3), respectively. The Northeast states (Massachusetts, New Jersey, New Hampshire and New York) exhibited significant effects of PM2.5 during the cooler months across most disease categories after adjustment for ozone and maximum apparent temperature. Ischemic heart disease risk per 10 μg/m3 increase in PM2.5 varied from 1.02 to 1.05 for the cooler months. The largest lag effect was noted on lag days 0 and 1. New Mexico and Washington exhibited no cool or warm month significant effects. Although Florida showed no cooler month effects, significant increases were noted in odds ratios for the warm weather months for all outcomes except peripheral vascular disease. This study is one of the first large scale applications of linkage of hospitalization data by state with national US EPA statistically modeled air pollution data. The results demonstrate that state-wide, there are multiple cardiovascular outcomes in addition to AMI which may be impacted by particulate air pollution.

Keywords: Air pollution; Cardiovascular disease; Case-crossover design; Environmental public health tracking; PM(2.5).

Publication types

  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Cardiovascular Diseases / chemically induced*
  • Centers for Disease Control and Prevention, U.S.
  • Cross-Over Studies
  • Hospitalization*
  • Humans
  • Particulate Matter / toxicity*
  • United States


  • Particulate Matter