Abstract Body: INTRODUCTION:
Hematoma expansion (HE) is strongly associated with mortality and disability in intracerebral hemorrhage (ICH). While it is accepted that the majority of this expansion occurs early after onset, there is limited data on HE and associations with outcome in the ultra-early time window (within 3 hours of symptom onset). We aimed to (a) estimate the incidence of HE within the ultra-early period of ICH, (b) describe hematoma growth dynamics over time, and (c) investigate the associations between ultra-early HE and clinical outcomes after ICH.

METHODS:
We performed a planned secondary analysis of the STOP-MSU international multicenter randomized controlled trial. The trial compared tranexamic acid (TXA) with placebo in 201 patients with primary ICH presenting within 2 hours of symptom onset. Repeat CT imaging 1 hour after treatment commencement was optional, and patients who underwent this ultra-early re-imaging were included in this descriptive study. Hematoma expansion was defined as growth by either >33% or >6ml from the baseline hematoma volume.

RESULTS:
We included 105 patients with ultra-early re-imaging (median age 66 years, 40% female, 53% TXA). Median time from onset to baseline imaging was 74min (IQR 56-87min), and between baseline and ultra-early re-imaging was 95min (IQR 74-132min). Forty-one patients (39%) had ultra-early HE. These patients had larger baseline hematoma volumes (15.9ml vs 9.1ml, p=0.03) and similar rates of intraventricular hemorrhage (22% vs 29.7%, p=0.38), compared to those without ultra-early HE. No effect of TXA on ultra-early HE was observed (TXA 41.1% vs placebo 36.5%, p=0.61).
Of 92 patients with both ultra-early and 24 hour re-imaging available, 31 patients had ultra-early HE. Of these, 9 (29%) had further growth at 24 hours, compared to 4/61 patients (6.6%) with no ultra-early growth (p<0.01). Median hematoma growth rate significantly reduced over time compared to the onset-to-baseline imaging period (clustered median regression p<0.01) (Figure 1).
Ultra-early hematoma expansion was associated with poor functional outcomes (mRS 3-6; aOR 3.87 [1.21-12.40], p=0.02) and mortality (aOR 6.16 [95% CI 2.15-17.68], p<0.01), adjusted for treatment group.

CONCLUSIONS:
Most hematoma growth occurs in the ultra-early period. The presence of ultra-early HE is associated with ongoing hematoma growth, poor functional outcomes and mortality, and could be a therapeutic target for clinical trials.