ISEE-0166 Background and Objective: We studied the short-term effect of air pollution on stroke separately for ischemic (IS) and hemorrhagic (HS) strokes. To study the effect on IS of primarily embolic origin, IS were separated into IS with and without atrial fibrillation (AF). Finally, we stratified patients by stroke severity into mild and severe strokes. Methods: We applied a time-stratified case-crossover design to evaluate the association between particulate (PM10, PM2.5, UFP (particles <0.1 μm diameter)) and gaseous (Nitrogen oxides (NOx), Carbon Monoxide (CO), ozone (O3) air pollution to stroke hospital admissions in Copenhagen (2003-2006). We used conditional logistic regression analysis, adjusting for temperature, relative humidity and wind speed, to examine the air pollution effect for a 0-4 day lag. Computed tomography was used to differentiate IS and HS. AF was determined by electrocardiography. Admission stroke severity was measured by the Scandinavian Stroke Scale (SSS). Results: Out of 7485 stroke admissions, 6798 were IS and 687 were HS, 3485 were mild and 2248 were severe strokes (1752 missing SSS), and 1278 were with and 5811 were without AF (396 were missing information on AF). We observed a significant positive association between NOx, CO and UFPs and IS; NOx, CO and UFPs and strokes without AF; and finally UFPs and mild strokes, all at lag 4. No associations were observed between air pollution and HS, IS with AF or with severe strokes. Among 2822 cases with mild IS without AF, the effect of UFPs (per inter-quartile range, 5-day average) was a 21% (95% confidence interval: 4-41%) increase in admissions. Conclusion: We found an air pollution effect on stroke admissions in patients with mild, IS without AF. Traffic related air pollutants, especially UFPs, were most relevant. Our findings indicate that the effect of air pollution is on small IS of thrombotic origin rather than HS and large embolic strokes.