Introduction: Cardioembolic (CE) stroke etiology has been increasing over time. How incidence rates of CE strokes within subgroups of the population have been changing is not known. We sought to determine if strokes of CE etiology have been changing similarly between races and between older versus younger populations over a recent 17-year period.
Methods: Within the Greater Cincinnati/Northern Kentucky (GCNK) catchment area of 1.3 million, all strokes were ascertained in the population from July 1993 to June 1994 and in 1999, 2005 and 2010 using ICD-9 codes 430-436. Stroke subtypes were determined by retrospective physician review and were then examined by age (< 50 or ≥ 50) and race (black or white). The proportions of incident stroke subtypes with 95% confidence intervals were calculated for each study period.
Results: There was an increase in the proportion of ischemic stokes identified as having CE etiology over time (p = 0.0002). The Table shows the percentage of CE and cryptogenic strokes by both race and age for each study period. Whites had a higher percentage of CE strokes compared with blacks, with the largest difference in 2010. The percentage of CE strokes in patients ≥50 years increased in 2010 but remained relatively stable in patients < 50 years old. Examination by race showed just over 10% reduction in cryptogenic strokes over time in both blacks and whites. The proportion of cryptogenic strokes decreased in both those ≥50 years and in those <50 years over time.
Conclusions: In our population, there was an increase in the proportion of CE strokes in whites only, most significantly between the two most recent study periods. Conversely, the percentage of cryptogenic strokes has decreased steadily over the study periods and equally in blacks and whites. Future studies are needed to explore the rates of longer-term cardiac monitoring to evaluate if this fully explains changes seen in CE stroke.