Background: As clinical and research interventions for intracerebral hemorrhage (ICH) migrate to the prehospital setting, patients are being treated earlier than ever before. The frequency, determinants, and outcomes of hematoma expansion among hyperacute, EMS-transported patients have not been previously characterized. Methods: We analyzed all ICH patients enrolled in the NIH Phase 3 Field Administration of Stroke Therapy - Magnesium (FAST-MAG) prehospital trial who had both ED-arrival and 6 to 48-hour follow-up brain CT or GRE-MRI. Hematoma expansion (HE) was defined as hematoma volume growth by ≥ 6 ml. Results: Among 262 ICH patients meeting entry criteria, age was 64.8±12.9, 32.4% were female, and last known well to ED arrival was 52 mins (IQR 44-70). Hematoma expansion occurred in 80 (30.5%). Clinical predictors of HE included: anticoagulant use at onset, 11.3% vs 3.8%, p=0.03; lower level of consciousness at ED arrival, median GCS 14 vs 15, p<0.001; greater neurologic deficit at ED arrival, NIHSS 21 vs 13, p<0.001, basal ganglia ICH location, 63.7% vs 46.2%, p =0.009; and shorter interval from onset to first imaging, 80.2 vs 84.5 mins, p=0.04. Features not predictive of HE in this hyperacute cohort included: age, blood pressure, and intraventricular hemorrhage extension on initial imaging. In multivariable analysis, independent predictors of HE were: anticoagulants at onset (aOR 4.2, 95%CI 1.5-12.9), GCS at arrival (aOR 1.2, 95%CI 1.0 – 1.4), and NIHSS at arrival (aOR 1.1, 95%CI 1.1 – 1.2), with c statistic 0.73. HE was associated with increased early (48h) neurologic deterioration (45.6% vs 15.5%, aOR 5.3, 95%CI 2.8-10.1), reduced 90d functional independence (mRS 0-2, 15.0% vs 36.5%, aOR 0.27, 95%CI 0.13-0.56), and increased 90d mortality (46.3% vs 11.0%, aOR 8.1, 95%CI 4.0-16.6). Conclusions: Hematoma expansion is frequent among hyperacute, EMS-transported ICH patients, occurring in more than 3 in 10 patients. Anticoagulant use, reduced level of consciousness, and greater neurologic deficit are independent determinants of HE, and collectively predict HE moderately well. Hematoma expansion is clinically consequential, associated with a 5-fold increase in odds of early neurology deterioration and 8-fold increase in odds of death by 90 days.