Coagulopathy in patients with intracranial haemorrhage or traumatic brain injury (TBI) is associated with clinical deterioration and worse outcome. Whole blood viscoelastic haemostatic assays, like thrombelastography (TEG), might aid conventional coagulation assays in identification of patients with worse prognosis. We performed a review of patients (totalling 78 patients) with primary acute intracranial haemorrhage or isolated TBI admitted to a neurointensive care unit (NICU) for more than 24 h during a period of 9 months, who had TEG analysis performed at admission. Primary outcome was all-cause 30-day mortality, whereas decline in Glasgow Coma Scale (GCS) score at 24 h after admission or death due to cerebral incarceration were secondary outcomes. Patients were defined as hypocoaguable if TEG reaction time was more than 8 min, angle less than 55° and/or maximal amplitude less than 51 mm. Patients were defined hypocoaguable according to conventional coagulation assays if international normalized ratio was more than 1.3, platelet counts less than 100 × 109/l and/or activated partial thromboplastine time more than 35 s. Eight patients were hypocoaguable by TEG on admission to NICU and had higher 30-day mortality (63% vs. 16%, P = 0.008), more often declined in GCS (57% vs. 16%, P = 0.02) and expired due to cerebral incarceration (50% vs. 6%, P = 0.02). Hypocoagulability by TEG, lower admission GCS and subarachnoid haemorrhage were independently associated with higher 30-day mortality [TEG: odds ratio (OR) 14.8 (2.2–100.1), P = 0.006; GCS: OR 1.3 (1.1–1.5), P = 0.006; subarachnoid haemorrhage: OR: 5.3 (1.3–22.3), P = 0.02]. Only two patients were hypocoaguable by both conventional coagulation assays and TEG. The current data indicate that hypocoagulability by TEG at admission to NICU predicts worse prognosis. Low concordance with conventional coagulation assays indicates that TEG might be valuable in identifying patients with clinically relevant coagulopathy.