TBI risk stratification at presentation: a prospective study of the incidence and timing of radiographic worsening in the Parkland Protocol.

Background: We have created a theoretical algorithm for venous thromboembolism prophylaxis after traumatic brain injury (TBI) known as the Parkland Protocol, which stratifies patients into low-, medium-, and high-risk categories for spontaneous progression of hemorrhage. This prospective study characterizes the incidence and timing of radiographic progression of the TBI patterns in these categories.

Methods: Inclusion criterion was presentation with intracranial blood between February 2010 and March 2011; exclusion was receipt of only one computed tomographic scan of the head during the inpatient stay or preinjury warfarin. At admission, all patients were preliminarily categorized per the Parkland Protocol as follows: low risk (LR), patients meeting the modified Berne-Norwood criteria; moderate risk (MR), injuries larger than the modified Berne-Norwood criteria without requiring a neurosurgical procedure; high risk (HR), any patient with a craniotomy/monitor.

Results: A total of 245 patients with intracranial hemorrhage were enrolled during the 13-month study period. Of patients preliminarily classified as LR at admission (n = 136), progression was seen in 25.0%. Spontaneous worsening was seen in 7.4% of LR patients at 24 hours after injury, and no LR patients progressed at 72 hours after injury. In patients initially classified as MR at admission (n = 42), progression was seen in 42.9%, with 91.5% of patients demonstrating stable computed tomographic head scans at 72 hours after injury. In patients initially classified as HR (n = 67), 64.2% demonstrated spontaneous progression of their TBI patterns, with 10.5% continuing to progress at 72 hours after injury. Most repeat scans were performed as routinely scheduled studies (81-91%).

Conclusion: Increases in the incidence of spontaneous worsening were seen as severities of injury progressed from the Parkland Protocol's LR to MR to HR arms. The time frames for these spontaneous worsenings seem to be such that the protocol's theoretical recommendations for venous thromboembolism prophylaxis are worth pursuing as future points of investigation.