• Rachel D Appelbaum, Katherine M Riera, Madeline R Fram, Gregory B Russell, Samuel P Carmichael Ii, R Shayn Martin, J Jason Hoth, Nathan T Mowery, and Andrew M Nunn.
• Vanderbilt University Medical Center, Vanderbilt University School of Medicine, USA. Electronic address: rachel.appelbaum@vumc.org.
• Injury. 2023 May 1; 54 (5): 137413781374-1378.

BackgroundCirrhosis in trauma patients is an indicator of poor prognosis, but current trauma injury grading systems do not take into account liver dysfunction as a risk factor. Our objective was to construct a simple clinical mortality prediction model in cirrhotic trauma patients: Cirrhosis Outcomes Score in Trauma (COST).MethodsTrauma patients with pre-existing cirrhosis or liver dysfunction who were admitted to our ACS Level I trauma center between 2013 and 2021 were reviewed. Patients with significant acute liver trauma (AAST Grade ≥ 3) or those that developed acute liver dysfunction while admitted were excluded. Demographics as well as ISS, MELD, complications, and mortality were evaluated. COST was defined as the sum of age, ISS, and MELD. Univariate and multivariable analysis was used to determine independent predictors of mortality. The area under the receiver operating curve (AUROC) was calculated to assess the ability of COST to predict mortality.ResultsA total of 318 patients were analyzed of which the majority were males 214 (67.3%) who suffered blunt trauma 305 (95.9%). Mortality at 30-days, 60-days, and 90-days was 20.4%, 23.6%, and 25.5%, respectively. COST was associated with inpatient, 30-day, and 90-day mortality on regression analyses and the AUROC for COST predicting mortality at these respective time points was 0.810, 0.801, and 0.813.ConclusionCurrent trauma injury grading systems do not take into account liver dysfunction as a risk factor. COST is highly predictive of mortality in cirrhotic trauma patients. The simplicity of the score makes it useful in guiding clinical care and in optimizing goals of care discussions. Future studies to validate this prediction model are required prior to clinical use.Copyright © 2023. Published by Elsevier Ltd.

13 keywords...

hide…