Inverse probability weighted distributed lag effects of short-term exposure to PM_2.5 and ozone on CVD hospitalizations in New England Medicare participants - Exploring the causal effects

Background: Although many studies have established significant associations between short-term air pollution and the risk of getting cardiovascular diseases, there is a lack of evidence based on causal distributed lag modeling. Methods: Inverse probability weighting (ipw) propensity score models along with conditional logistic outcome regression models based on a case-crossover study design were applied to get the causal unconstrained distributed (lag0-lag5) as well as cumulative lag effect of short-term exposure to PM_2.5/Ozone on hospital admissions of acute myocardial infarction (AMI), congestive heart failure (CHF) and ischemic stroke (IS) among New England Medicare participants during 2000–2012. Effect modification by gender, race, secondary diagnosis of Chronic Obstructive Pulmonary Diseases (COPD) and Diabetes (DM) was explored. Results: Each 10 μg/m³ increase in lag0-lag5 cumulative PM_2.5 exposure was associated with an increase of 4.3% (95% confidence interval: 2.2%, 6.4%, percentage change) in AMI hospital admission rate, an increase of 3.9% (2.4%, 5.5%) in CHF rate and an increase of 2.6% (0.4%, 4.7%) in IS rate. A weakened lagging effect of PM_2.5 from lag0 to lag5 could be observed. No cumulative short-term effect of ozone on CVD was found. People with secondary diagnosis of COPD, diabetes, female gender and black race are sensitive population. Conclusions: Based on our causal distributed lag modeling, we found that short-term exposure to an increased ambient PM_2.5 level had the potential to induce higher risk of CVD hospitalization in a causal way. More attention should be paid to population of COPD, diabetes, female gender and black race.