Long-term survival of ischemic cerebrovascular disease in the acute inflammatory stroke study, a hospital-based cohort described by TOAST and ASCO

Background: Ischemic cerebrovascular disease (ICVD) comprises multiple etiological phenotypes that share common clinical characteristics. Etiological classification of patients with ICVD is of major clinical interest to achieve optimal medical treatment and predict prognosis. The TOAST classification system has been widely used to describe stroke etiology but provides restricted phenotypic homogeneity within groups. The ASCO classification system has introduced a new approach in phenotypic classification, and aims to describe clinical characteristics without merging concurrent comorbidities. Inflammatory processes have been suggested to mediate stroke etiology and pathology. The Acute Inflammatory Stroke Study (AISS), a hospital-based cohort, is here introduced and described by TOAST and ASCO classification systems. The aim of this first analysis of AISS was to investigate long-term mortality in relation to ischemic stroke subtypes, and clinical and biochemical markers.

Methods: AISS consecutively follows patients on 6 occasions up to 1 year after stroke onset. Complete workup according to ASCO comprised CT or MRI of the head, ECG, duplex of the extracranial arteries or CT/MR angiography and ultrasound of the heart. Level 2 evidence was required in each domain to obtain a comparable system to TOAST (ASCO2). Clinical and biochemical characteristics and mortality rates were documented and compared by the two classification systems.

Results: Of 142 patients consecutively evaluated and recruited in the study, a total of 101 ICVD patients (ischemic stroke, n = 84; transient ischemic attack, n = 17) were included in the final analysis. Agreement between ASCO2 and TOAST was very good. During the mean observation period of 28 months, 26 patients died. The 1- and 4-year mortality rates were 0 and 4% for large artery atherosclerosis (LAA); 23 and 36% for cardioembolism (CE); 0% for small artery occlusion (SAO); 63 and 100% for the subtype with unknown etiology due to incomplete workup (Unknown), and 12 and 29% for the cryptogenic subtype. As for the ASCO2 groups, the 1- and 4-year mortality rates were 0 and 6% in LAA, 25 and 36% in CE, 0% in SAO, 0 and 14% in LAA + CE, 0% in SAO + CE, 16 and 36% in the subgroup with undetermined etiology despite complete workup, and 56 and 100% in Unknown. Regression analysis showed that age, white blood cell count, fibrinogen and bilirubin, but not etiological subgroup, were independent predictors of mortality.

Conclusion: Our findings indicate that clinical and biochemical markers may differentiate phenotypically homogeneous etiological subtypes and predict long-term mortality. Further studies with larger patient numbers are needed to investigate possible causative mechanisms.