Incremental Prognostic Value of Coagulopathy in Addition to the Crash Score in Traumatic Brain Injury Patients

BACKGROUND/OBJECTIVE: Multivariable prognostic scores play an important role for clinical decision-making, information giving to patients/relatives, benchmarking and guiding clinical trial design. Coagulopathy has been implicated on trauma and critical care outcomes, but few studies have evaluated its role on traumatic brain injury (TBI) outcomes. Our objective was to verify the incremental prognostic value of routine coagulopathy parameters in addition to the CRASH-CT score to predict 14-day mortality in TBI patients. METHODS: This is a prospective cohort of consecutive TBI patients admitted to a tertiary university hospital Trauma intensive care unit (ICU) from March/2012 to January/2015. The prognostic performance of the coagulation parameters platelet count, prothrombin time (international normalized ratio, INR) and activated partial thromboplastin time (aPTT) ratio was assessed through logistic regression adjusted for the original CRASH-CT score. A new model, CRASH-CT-Coag, was created and its calibration (Brier scores and Hosmer–Lemeshow (H–L) test), discrimination [area under the receiver operating characteristic curve (AUC-ROC) and the integrated discrimination improvement (IDI)] and clinical utility (net reclassification index) were compared to the original CRASH-CT score. RESULTS: A total 517 patients were included (median age 39 years, 85.1% male, median admission glasgow coma scale 8, neurosurgery on 44.9%). The 14-day mortality observed and predicted by the original CRASH-CT was 22.8% and 26.2%, respectively. Platelet count < 100,000/mm(3), INR > 1.2 and aPTT ratio > 1.2 were present on 11.3%, 65.0% and 27.2%, respectively, (at least one of these was altered on 70.6%). All three variables maintained statistical significance after adjustment for the CRASH-CT score. The CRASH-CT-Coag score outperformed the original score on calibration (brier scores 0.122 ± 0.216 vs 0.132 ± 0.202, mean difference 0.010, 95% CI 0.005–0.019, p = 0.036, respectively) and discrimination (AUC-ROC 0.854 ± 0.020 vs 0.813 ± 0.024, p = 0.014; IDI 5.0%, 95% CI 1.3–11.0%). Both scores showed the satisfactory H–L test results. The net reclassification index favored the new model. Considering the strata of low (< 10%), moderate (10–30%) and high (> 30%) risk of death, the CRASH-CT-Coag model yielded a global net correct reclassification of 22.9% (95% CI 3.8–43.4%). CONCLUSIONS: The addition of early markers of coagulopathy—platelet count, INR and aPTT ratio—to the CRASH-CT score increased its accuracy. Additional studies are required to externally validate this finding and further investigate the coagulopathy role on TBI outcomes.