Yonsei Stroke Registry. Analysis of 1,000 patients with acute cerebral infarctions

Background and purpose: The hospital-based stroke registry is a well-established method useful for understanding diverse clinical characteristics of stroke related to geographical, racial or environmental differences. We analyzed the data from 1,000 patients with acute cerebral infarctions registered with the Yonsei Stroke Registry (YSR) which is the first prospective hospital-based observational study in Korea.

Methods: All patients had cerebral infarctions and presented within 7 days of onset. CT or MRI was performed in all patients and a vascular imaging study (digital subtraction or magnetic resonance angiography) was conducted in 53.9% of the patients. Subtype classification was made through a consensus approach based on the strict application of TOAST criteria.

Results: The mean age of patients was 62 +/- 12 years, and 60.8% were males. Undetermined cause (UD) was the most frequent subtype (40.6%), which was followed by lacunar stroke (LS 21.5%), cardiac embolism (CE 18.3%), large-artery atherosclerosis (LAA 16.5%) and other determined causes (3.1%). Hypertension was found in 64.3%, smoking in 35.2%, diabetes mellitus in 26.9%, hypercholesterolemia in 24.1%, high hematocrit (> or = 50%) in 21.8%, clinically identified potential cardiac sources of embolism in 18.3%, a history of previous stroke in 22.0% and a history of previous transient ischemic attack in 4.7%. Recurrent stroke was associated with a higher number of risk factors (p < 0.001) and a higher incidence of LAA (p = 0.003) than the first stroke. Vertebrobasilar artery territorial infarction was found in 39.8%, which was associated with higher incidences of LAA and LS and a lower incidence of CE than carotid artery territorial lesions (p = 0.001). The 30-day mortality rate was 5.3% and cerebral herniation caused early death in 52%.

Conclusion: The distribution of stroke subtypes in the YSR was largely comparable with that of western registries. The highest incidence of UD might be related to the strict application of TOAST criteria.