Abstract 3698: Ischemic Stroke Etiology in a Population Based Study: Temporal Trends in Ischemic Stroke Subtyping Over Time
Introduction: Ischemic stroke subtyping is important for genetic research phenotyping, selection into some clinical trials, and for public health purposes regarding optimal stroke prevention. Population-based studies have high rates of unknown etiology due to insufficient testing to allow definitive subtyping. We sought to investigate temporal trends in testing after ischemic stroke and etiologic subtype determination over three periods.
Methods: Our study ascertained all strokes for residents of a 5-county region during three one-year study periods, 7/93- 6/94, 1999 and 2005, by screening ICD-9 codes 430-436. An abstract of hospital records for every potential case was done by a study nurse and then adjudicated by a study physician. Testing performed during ER encounters and hospitalization was recorded, and copies of relevant test results were provided. Our etiologic subtype definitions are based upon the Classification of Cerebrovascular Diseases III and previous work from our center. We calculated the proportion of subjects who had "complete testing" for documenting stroke subtypes, i.e., the minimum required for recently proposed classification systems. For first-ever cases of ischemic stroke ≥20 years of age at onset, we compared the proportions for each possible etiology over time, as well as the proportions of subjects with multiple identified causes.
Results: The proportions of ischemic stroke patients who received relevant testing in each study period and subtype determinations are shown in the table. Notably, significant increases were seen over time in the number of MRIs, MRAs, and echocardiograms. The proportion of patients who received "complete testing" significantly increased over time, but only 57% of patients received complete testing in the most recent study period. .The major difference over time in ischemic stroke subtypes was an increase in “other” causes with a corresponding decline in “unknown” subtype. There was a significant trend towards determining single or multiple etiologies, compared with “unknown” (p = 0.01).
Conclusions: Despite the representative nature and statistical power of a large population based study, many ischemic stroke patients would not have a defined etiologic subtype utilizing recently proposed subtyping methods due to insufficient testing and would be classified as "unknown". Further study of subtyping methods is warranted, as requiring “complete” testing for every stroke patient is not feasible in current community practice and because it is possible to assign a most likely phenotype in some patients who have incomplete testing.
- © 2012 by American Heart Association, Inc.