SMASH-U: a proposal for etiologic classification of intracerebral hemorrhage.

BACKGROUND AND PURPOSE The purpose of this study was to provide a simple and practical clinical classification for the etiology of intracerebral hemorrhage (ICH). METHODS We performed a retrospective chart review of consecutive patients with ICH treated at the Helsinki University Central Hospital, January 2005 to March 2010 (n=1013). We classified ICH etiology by predefined criteria as structural vascular lesions (S), medication (M), amyloid angiopathy (A), systemic disease (S), hypertension (H), or undetermined (U). Clinical and radiological features and mortality by SMASH-U (Structural lesion, Medication, Amyloid angiopathy, Systemic/other disease, Hypertension, Undetermined) etiology were analyzed. RESULTS Structural lesions, namely cavernomas and arteriovenous malformations, caused 5% of the ICH, anticoagulation 14%, and systemic disease 5% (23 liver cirrhosis, 8 thrombocytopenia, and 17 various rare conditions). Amyloid angiopathy (20%) and hypertensive angiopathy (35%) were common, but etiology remained undetermined in 21%. Interrater agreement in classifying cases was high (κ, 0.89; 95% CI, 0.82-0.96). Patients with structural lesions had the smallest hemorrhages (median volume, 2.8 mL) and best prognosis (3-month mortality 4%), whereas anticoagulation-related ICHs were largest (13.4 mL) and most often fatal (54%). Overall, median ICH survival was 5&frac12; years, varying strongly by etiology (P<0.001). After adjustment for baseline characteristics, patients with structural lesions had the lowest 3-month mortality rates (OR, 0.06; 95% CI, 0.01-0.37) and those with anticoagulation (OR, 1.9; 1.0-3.6) or other systemic cause (OR, 4.0; 1.6-10.1) the highest. CONCLUSIONS In our patients, performing the SMASH-U classification was feasible and interrater agreement excellent. A plausible etiology was determined in most patients but remained elusive in one in 5. In this series, SMASH-U based etiology was strongly associated with survival.