Publication
The AVICH Score: A Novel Grading System to Predict Clinical Outcome in Arteriovenous Malformation-Related Intracerebral Hemorrhage
Journal Paper/Review - May 2, 2016
Neidert Marian Christoph, Lawton Michael T, Mader Marius, Seifert Burkhardt, Valavanis Antonios, Regli Luca, Bozinov Oliver, Burkhardt Jan-Karl
Units
PubMed
Doi
Citation
Type
Journal
Publication Date
Issn Electronic
Pages
Brief description/objective
OBJECTIVE
To establish an arteriovenous malformation (AVM) grading score for patients with ruptured AVM and associated intracerebral hemorrhage (ICH) to predict clinical outcome.
METHODS
Patient data from January 2006 to December 2013 with newly diagnosed ICH caused by ruptured AVM and a modified Rankin Scale (mRS) score <2 before ICH were included for this analysis. Clinical outcome was dichotomized in favorable (mRS score, 0-2) and unfavorable (mRS score, 3-6) to find predictors for outcome and to establish a new score based on the areas under the receiver-operating characteristic curves (AUROC) at 3 months, 1 year, and at last follow-up (mean, 31 months).
RESULTS
A total of 67 patients (mean age, 41 years; 66% male) were analyzed including 39 patients with favorable and 28 with unfavorable outcome. Intraventricular hemorrhage (P = 0.048), ICH score (P = 0.003), AVM size (P < 0.001), Spetzler-Martin grade (P < 0.001), nidus structure (P = 0.005), Lawton-Young grade (P = 0.015), and supplemented Spetzler-Martin score (P < 0.001) were significant predictors for clinical outcome in ruptured AVMs. Based on these results, we created a new score named the AVICH (AVM-related ICH) score, which showed an AUROC of 0.842 compared to 0.789 for the supplemented Spetzler-Martin grading system and 0.703 for the ICH score regarding clinical outcome at last follow-up.
CONCLUSIONS
Based on the AUROC analysis, the AVICH score predicts outcome of patients with ruptured AVM and associated ICH better than the ICH score, the Spetzler-Martin, or the supplemented Spetzler-Martin grading system. An external validation is needed before the AVICH score is tested in a prospective multicenter cohort.