Several studies have demonstrated positive associations between day-to-day increases in air pollution and stroke. These findings have been inconsistent, and the influence of patient characteristics has been largely ignored. In this study, we investigated the short-term effects of air pollution on stroke using a time-stratified case-crossover design. Data for hospital visits for stroke were extracted from 5927 medical charts of patients who presented to emergency departments between 2003 and 2009 in Edmonton, Canada. Daily concentrations of five air pollutants (NO(2), PM (2.5), CO, O(3), and SO(2)) were obtained from fixed-site monitors. Relative humidity and temperature were obtained from a metrological station operating at the city's airport. Chart data included: disease history, medication use, and smoking status. Conditional logistic regression was used to estimate the odds ratio (OR) of stroke in relation to an increase in the interquartile range for each pollutant. Positive associations were observed between ischemic stroke and air pollution during the 'warm' season (April through September). Specifically, the OR for an increase in 9.4 ppb in the 3-day average of NO(2) was 1.50 (95% CI: 1.12, 2.01). There were no statistically significant associations with any of the other pollutants after adjusting for NO(2) concentrations. Associations with ischemic stroke were stronger for those with a history of stroke (OR=2.31; 95% CI: 1.39, 3.83), heart disease (OR=1.99; 95% CI: 1.20, 3.28), and taking medication for diabetes (OR=2.03; 95% CI: 1.14, 3.59). Temperature was inversely associated with ischemic stroke during the 'warm' season, but no associations were evident with the other stroke subtypes. Air pollution was not associated with hemorrhagic stroke or transient ischemic attacks. The findings suggest that specific patient characteristics modify associations between air pollution and ischemic stroke.