The aim of this study was to determine if neurologic findings at the time of initial resuscitation can predict coagulation abnormalities resulting from head injury. Fifty-three children with head injury were reviewed for Glasgow Coma Scale (GCS), prothrombin time (PT), international normalized ratio (INR), partial thromboplastin time (PTT), use of fresh frozen plasma (FFP) and outcome. Twenty-six of the 53 children (49%) presented with a GCS of 15 and 27 (51%) had a GCS less than 14. The incidence of computed tomography (CT)-documented intracranial injury was 12% in those children with a GCS of 15 versus 78% when GCS < or = 14 (P < .05). Abnormal coagulation (PT > 14.5, INR > 1.2, PTT > 38) in children with a GCS = 15 was 7% v 67% when GCS was < or = 14 (P < .05). A mean of 1 unit of FFP per patient was required in children with a GCS of < or = 14. No child with GCS of 15 and CT evidence for intracranial injury had a coagulopathy, and no child with GCS of 15 required FFP. In head injured children, significant coagulation abnormalities requiring treatment are excluded by the presence of a normal GCS at presentation. Children with GCS less than 14 are at risk for intracranial injury and coagulopathy, this risk increases inversely with the GCS. Children who present with a GCS less than 8 should have FFP prepared at the time of admission. These data may guide the use of laboratory tests and blood bank resources during trauma resuscitation.