Citation

Lingsma H, Andriessen TM, Haitsema I, Horn J, van der Naalt J, Franschman G, Maas AI, Vos PE, Steyerberg EW. J. Trauma Acute Care Surg. 2013; 74(2): 639-646.

Affiliation

From the Departments of Public Health (H.L., E.W.S.) and Neurosurgery (I.H.), Center for Medical Decision Making, Erasmus Medical Center, Rotterdam, The Netherlands; Department of Neurology (T.M.J.C.A., P.E.V.), Radboud University Nijmegen Medical Center, Nijmegen, the Netherlands; Department of Intensive Care Medicine (J.H.), Academic Medical Center, Amsterdam, the Netherlands; Department of Neurology (J.v.d.N.), University Medical Center Groningen, Groningen, the Netherlands; Department of Anesthesiology (G.F.), VU University Medical Center, Amsterdam, Amsterdam, the Netherlands; and Department of Neurosurgery (A.I.R.M.), University Hospital Antwerp, Antwerp, Belgium.

Copyright

(Copyright © 2013, Lippincott Williams and Wilkins)

DOI

10.1097/TA.0b013e31827d602e

PMID

23354263

Abstract

BACKGROUND: Several prognostic models to predict outcome in traumatic brain injury (TBI) have been developed, but few are externally validated. We aimed to validate the International Mission on Prognosis and Analysis of Clinical Trials in TBI (IMPACT) prognostic models in a recent unselected patient cohort and to assess the additional prognostic value of extracranial injury. METHODS: The Prospective Observational COhort Neurotrauma (POCON) registry contains 508 patients with moderate or severe TBI, who were admitted in 2008 and 2009 to five trauma centers in the Netherlands. We predicted the probability of mortality and unfavorable outcome at 6 months after injury with the IMPACT prognostic models. We studied discrimination (area under the curve [AUC]) and calibration. We added the extracranial component of the Injury Severity Score (ISS) to the models and calculated the increase in AUC. RESULTS: The IMPACT models had an adequate discrimination in the POCON registry, with AUCs in the external validation between 0.85 and 0.90 for mortality and between 0.82 and 0.87 for unfavorable outcome. Observed outcomes agreed well with predicted outcomes. Adding extracranial injury slightly improved predictions in the overall population (unfavorable outcome: AUC increase of 0.002, p = 0.02; mortality: AUC increase of 0.000, p = 0.37) but more clearly in patients with moderate TBI (unfavorable outcome: AUC increase of 0.008, p < 0.01, mortality: AUC increase of 0.012, p = 0.02) and patients with minor computed tomographic result abnormalities (unfavorable outcome: AUC increase of 0.013, p < 0.01; mortality: AUC increase of 0.001, p = 0.08). CONCLUSION: The IMPACT models performed well in a recent series of TBI patients. We found some additional impact of extracranial injury on outcome, specifically in patients with less severe TBI or minor computed tomographic result abnormalities. LEVEL OF EVIDENCE: Epidemiologic/prognostic study.

Language: en