|
Citation
|
Keskey RC, Slidell MB, Bohr NL, Biermann H, Cirone J, Zakrison T, Cone J, Wilson K, Hampton D. J. Trauma Acute Care Surg. 2021; ePub(ePub): ePub. |
|
Abstract
|
Injury severity score (ISS) is a widely used metric for trauma research and center verification; however, it does not account for age-related physiologic parameters. We hypothesized that a novel age-based injury severity metric would better predict mortality.
METHODS: Adult patients(≥18y) sustaining blunt (BT) or penetrating(PT) trauma were abstracted from the 2010-2016 National Trauma Data Bank. Admission vitals, GCS, ISS, mechanism, and outcomes were analyzed. Patients with incomplete/non-physiologic vital signs were excluded. For each age: 1) a cut point analysis was used to determine the ISS with the highest specificity and sensitivity for predicting mortality and 2) a linear discriminant analysis was performed using ISS, ISS > 16, TRISS, and RTS to compare each scoring system's mortality prediction. A novel injury severity metric, the trauma component score (TCS), was developed for each age using significant (p < 0.05) variables selected from AIS scores, GCS, vital signs, and gender. Receiver operator curves were developed and the areas under the curve (AUC) were compared between the TCS and other systems.
RESULTS: 777,794 patients were studied (BT: 91.1%; PT: 8.9%). BT patients were older (53.6y ± 21.3 vs 34.4y ± 13.8), had higher ISS scores (11.1 ± 8.5 vs 8.5 ± 8.9), and lower mortality (2.9% vs 3.4%) than PT patients (p < 0.05). When assessing the entire PT and BT cohort the optimal ISS cut point was 16. The optimal ISS was between 20 and 25 for BT under 70y. Over 70y, the optimal BT ISS steadily declined as age increased PT's cut point was ≤16 for all ages assessed. When the injury metrics were compared by AUC, our novel TCS more accurately predicted mortality across all ages in both BT and PT (p < 0.001).
CONCLUSIONS: ISS is a poor mortality predictor in older patients and those sustaining penetrating trauma. The age based TCS is a superior metric for mortality prediction across all ages. LEVEL OF EVIDENCE: IV, clinical outcomes.
Language: en |