Redefining Hematoma Expansion With the Inclusion of Intraventricular Hemorrhage Growth

Abstract

Background and Purpose— Definitions of significant hematoma expansion traditionally focus on changes in intraparenchymal volume. The presence of intraventricular hemorrhage (IVH) is a predictor of poor outcome, but current definitions of hematoma expansion do not include IVH expansion. We evaluated whether including IVH expansion to current definitions of hematoma expansion improves the ability to predict 90-day outcome. Methods— Using data from the PREDICT-ICH study (Predicting Hematoma Growth and Outcome in Intracerebral Hemorrhage Using Contrast Bolus CT), we compared a standard definition of hematoma expansion (≥6 mL or ≥33%) to revised definitions that includes new IVH development or expansion (≥6 mL or ≥33% or any IVH; ≥6 mL or ≥33% or IVH expansion ≥1 mL). The primary outcome was poor clinical outcome (modified Rankin Scale score, 4–6) at 90 days. Diagnostic accuracy measures were calculated for each definition, and C statistics for each definition were compared using nonparametric methods. Results— Of the 256 patients eligible for primary analysis, 127 (49.6%) had a modified Rankin Scale score of 4 to 6. Sensitivity and specificity for the standard definition (n=80) were 45.7% (95% CI, 36.8–54.7) and 82.9% (95% CI, 75.3–88.9), respectively. The revised definition, ≥6 mL or ≥33% or any IVH (n=113), possessed a sensitivity of 63.8% (95% CI, 54.8–72.1) and specificity of 75.2% (95% CI, 66.8–82.4). Overall accuracy was significantly improved with the revised definition ( P =0.013) and after adjusting for relevant covariates, was associated with a 2.55-fold increased odds (95% CI, 1.31–4.94) of poor outcome at 90 days. A second revised definition, ≥6 mL or ≥33% or IVH expansion ≥1 mL, performed similarly (sensitivity, 56.7% [95% CI, 47.6–65.5]; specificity, 78.3% [95% CI, 40.2–85.1]; aOR, 2.40 [95% CI, 1.23–4.69]). Conclusions— In patients with mild-to-moderate ICH, including IVH expansion to the definition of hematoma expansion improves sensitivity with only minimal decreases to specificity and improves overall prediction of 90-day outcome.