(1587) Defining and Predicting Prolonged Mechanical Ventilation in a Civilian Trauma Population

Introduction: Acute lung injury is a common complication in patients with traumatic injuries. However, only a minority of these patients will develop respiratory failure requiring prolonged mechanical ventilation (PMV). In our previous study in a combat population we defined PMV as ≥9 days using a receiver operating characteristic (ROC) curve which plotted ventilator days against mortality. In the current study among a civilian trauma population, we sought to (1) validate our previous findings, and (2) explore factors associated with PMV.

Methods: We obtained data from the Trauma Quality Improvement Program (TQIP) database for analysis. We isolated subjects that had a documented age of ≥15 years and at least one day of mechanical ventilation. We excluded subjects that had a major TBI defined as an abbreviated injury scale of ≥3 for the head region or if they died in < 24 hours. An ROC curve was plotted for ventilator days against mortality with Youden’s index used to determine the threshold. Multivariable logistic regression (MVLR) models using a forward stepwise approach were used to test for associations and adjust for confounders.

Results: There were 1,209,097 in the 2021-year dataset, of which, 48,972 met inclusion for this analysis. Based on our ROC, we found a Youden’s index of 9 days (AUROC 0.579, sensitivity 0.324, specificity 0.807). Within our dataset, there were 13,856 (28%) that met our definition of PMV. The PMV group had a median of 15 (IQR 11-21) ventilator days compared to 2 (IQR 2-4) in the non-PMV group and a total hospital days of 27 (IQR 18-39) versus 7 (IQR 3-14). On MVLR testing, the injury severity score, mechanism of injury, arrival Glasgow coma scale, 4-hour plasma administration volume, sex, age, emergency surgery to the liver and/or pelvis, cocaine use, and body mass index were all associated with PMV.

Conclusions: PMV was predicted by patient factors including advanced age, female sex, obesity and cocaine use. PMV was also associated with injury severity, low GCS, and with patients requiring emergent surgical intervention of the liver and/or pelvis. Initial plasma transfusion volumes were significantly higher in patients who later required PMV. Providers should consider these risk factors for PMV to help inform decisions regarding the potential clinical course of trauma patients.