The Denver and Sequential Organ Failure Assessment (SOFA) scores have been used widely to describe the epidemiology of postinjury multiple-organ failure; however, differences in these scores make it difficult to compare incidence, duration, and mortality of multiple-organ failure. The study aim was to compare the performance of the Denver and SOFA scores with respect to the outcomes of mortality, intensive care unit length of stay (ICU LOS), and ventilator days.
A 60-month prospective epidemiologic study was undertaken at an Australian Level I trauma center. Data were collected on trauma patients that met inclusion criteria (ICU admission, Injury Severity Score [ISS] > 15, age > 18 years, head Abbreviated Injury Scale [AIS] score < 3, survival for >48 hours). Demographics, ISS, physiologic parameters, SOFA and Denver scores, and outcome data were prospectively collected. Sensitivity/specificity and receiver operating characteristic curve were calculated for both scores. Analysis was also completed for a Day 3 postinjury SOFA and Denver score.
A total of 140 patients met the inclusion criteria (mean [SD] age, 47  years; ISS, 30; male, 69%; mortality rate, 6%; mean [SD] ICU LOS, 9  days; mean [SD] ventilation period, 6  days). There was no difference in the score performance predicting mortality. Day 3 SOFA score of 4 or greater outperformed the Denver score of greater than 3 when predicting ICU LOS and ventilator days (area under the curve, 0.83 vs. 0.69, 0.86 vs. 0.73, respectively). The SOFA score was more sensitive and the Denver score was more specific when predicting mortality, ICU LOS, and ventilator days.
Both scores had similar performance predicting mortality; however, the Day 3 SOFA score outperforms the Denver score when predicting ICU LOS and ventilator days. Either score could be superior based on whether one is seeking to optimize specificity or sensitivity. It is important to note that these findings are in a non–head-injured population and that there are practical difficulties using the SOFA in head-injured patients.
Diagnostic study, level II.
From the Department of Traumatology (D.C.D., A.W., S.M.T., K.L.K., Z.J.B.), and Division of Medicine (J.A.), John Hunter Hospital, and Centre for Clinical Epidemiology and Biostatistics (J.A.), and Hunter Medical Research Institute, University of Newcastle, Newcastle, New South Wales, Australia.
Submitted: May 4, 2014, Revised: May 27, 2014, Accepted: May 30, 2014.
Address for reprints: Zsolt Balogh, MD, PhD, Department of Traumatology, Division of Surgery, John Hunter Hospital and University of Newcastle, Locked Bag 1, Hunter Region Mail Centre, Newcastle, NSW 2310, Australia; email: firstname.lastname@example.org.