The prophylactic hemicraniectomy score: Predicting need for decompressive craniectomy after traumatic brain injury due to refractory intracranial hypertension

Decompressive craniectomy (DC) is performed in severe traumatic brain injury (sTBI) to treat medically refractory intracranial hypertension. We sought to identify factors present early in admission to predict at risk for treatment-refractory intracranial hypertension requiring DC.

Adults with sTBI at our facility from 2016 to 2023 were identified from the National Trauma Database local registry. Patients with DC within 24 h of admission were excluded. Measured primary outcome was DC. After univariable analysis, multivariable logistic regression was performed to identify predictive variables to develop a predictive score.

205 patients met inclusion criteria; 30 patients (14.6 %) underwent DC. Patients in the DC cohort were significantly younger (median = 29 years-old, interquartile range [IQR] = 22.0-45.3 versus39 years, IQR = 25.0-59.0; p = 0.021). Initial ICP was higher in the DC cohort (median = 20.3 mmHg, IQR = 13-31.6 versus 10.7 mmHg, IQR = 7.1-13.9; p < 0.001). On admission, the DC cohort more frequently had midline shift (p = 0.001) and epidural hematomas (p = 0.007) but similar proportions of contusions and subdural hematomas. CT scans were less likely to be initially stable in the DC cohort (24.1 % versus 68.8 %, p < 0.001). In multivariable analysis, stability between first and second CT scans was protective of DC (odds ratio[OR] = 0.084, 95 % confidence interval [95CI] = 0.019-0.373, p = 0.001), presence of midline shift, and higher initial ICP were associated with higher odds of secondary DC (midline shift: OR = 6.41, 95CI = 1.72-23.97, p = 0.006; ICP: OR = 1.08, 95CI = 1.03-1.14, p = 0.003). The prophylactic hemicraniectomy (PH) score was created where each of the following was graded on a binary scale: scan instability, initial midline shift, initial ICP > 17.8 mmHg. Higher score (0-3) was associated with increased odds of DC in a stepwise fashion (OR = 9.11, 95CI = 3.5-23.7, p < 0.001).

The PH score can predict secondary DC using three variables: first-to-second CT head instability, presence of midline shift, and initial ICP over 17.8 mmHg. This may serve as a risk-stratification tool to identify patients at high risk for DC and with prospective external validation may inform on patients who would benefit from earlier DC.