Predicting outcome after arteriovenous malformation-associated intracerebral hemorrhage with the original ICH score

Abstract

Objective: To evaluate the predictive ability of the original ICH Score (oICH) in a large independent cohort of patients with arteriovenous malformation-associated intracerebral hemorrhage (AVM-ICH), an important cause of intracerebral hemorrhage (ICH) that is associated with significantly different epidemiology, clinical course, and outcome compared with primary ICH.

Methods: During the period 1997-2009, 91 patients were admitted to Columbia Medical Center with acute AVM-ICH. Demographic and admission clinical and radiographic variables were obtained for 84 patients through retrospective chart review. Admission oICH and Spetzler-Martin grading scale (SMGS) were calculated. Outcome was assessed at 3 months using the modified Rankin Scale (mRS). Maximum Youden Indices were used to identify cutoffs for age and ICH volume that are associated with optimal predictive accuracy for an unfavorable outcome (mRS ≥ 3). Receiver operating characteristic (ROC) analysis was used to evaluate the predictive performance of oICH, and oICH with new age and ICH cutoff points (new AVM-ICH score based on original ICH Score [AVM-oICH]).

Results: The mean age was 35 years ± 14, and mean ICH volume was 22 mL ± 20. At 3-month follow-up, 3 (4%) patients were dead, and 15 (18%) had an unfavorable outcome. Two of the patients who died had oICH of 3, and one had oICH of 5. ICH volume of 37 mL and age of 41 years were identified as optimal cutoffs for predicting an unfavorable outcome. oICH and AVM-oICH showed good predictive accuracies with area under the curve of 0.914 and 0.891 (P = 0.422). AVM-oICH and oICH had similarly high sensitivities (0.889 and 0.944; P = 1.00), but the former had significantly greater specificity (0.879 vs. 0.682; P < 0.001).

Conclusions: oICH is a valid clinical grading scale with high predictive accuracy for functional outcome after AVM-ICH. It is unclear whether the score is appropriate for risk stratification with regard to mortality because of the low risk of death associated with AVM-ICH. Simple adjustments of the age and ICH volume cutoff points improve performance of the score and reduce the probability of overestimating a patient's risk of an unfavorable outcome after AVM-ICH.