Analyzing risk factors for late posttraumatic seizures: a prospective, multicenter investigation.

Objectives: To ascertain the natural history and to stratify risks for the development of late posttraumatic seizures in individuals with moderate to severe traumatic brain injury (TBI).

Design: Prospective, observational study of individuals with TBI admitted to 4 trauma centers within 24 hours of injury.

Setting: Four tertiary care trauma centers in urban areas.

Participants: A total of 647 individuals (>/=16 y) with any of the following abnormal computed tomography (CT) scan findings: extent of midline shift and/or cisternal compression or presence of any focal pathology (eg, punctate, subarachnoid, or intraventricular hemorrhage; cortical or subcortical contusion; extra-axial lesions) during the first 7 days postinjury or best Glasgow Coma Scale (GCS) score of
Interventions: Not applicable.

Main Outcome Measures: Cumulative probability, relative risk, and survival analyses were used to stratify risks for development of late postttraumatic seizures on the basis of demographic factors, etiology of injury, initial GCS, early posttraumatic seizures, time post-TBI, types of intracerebral lesion by CT scan, and number and types of intracranial procedures.

Results: Sixty-six individuals had a late posttraumatic seizures; 337 had no late posttraumatic seizures during full 24-month follow-up; 167 had no late posttraumatic seizures during time followed (<24 mo); and 54 were placed on anticonvulsants without a late posttraumatic seizures, whereas 23 died before their first late posttraumatic seizures. The highest cumulative probability for late posttraumatic seizures included biparietal contusions (66%), dural penetration with bone and metal fragments (62.5%), multiple intracranial operations (36.5%), multiple subcortical contusions (33.4%), subdural hematoma with evacuation (27.8%), midline shift greater than 5mm (25.8%), or multiple or bilateral cortical contusions (25%). Initial GCS score was associated with the following cumulative probabilities for development of late posttraumatic seizures at 24 months: GCS score of 3 to 8, 16.8%; GCS score of 9 to 12, 24.3%; and GCS score of 13 to 15, 8.0%.

Conclusions: Stratification by CT scan findings and neurosurgical procedures performed were the most useful findings in defining individuals at highest risk for late posttraumatic seizures.