Prehospital serum lactate as a predictor of outcomes in trauma patients: a retrospective observational study.

BACKGROUND Lactate is associated with morbidity and mortality; however, the value of prehospital lactate (pLA) is unknown. Our objective was to determine whether pLA improves identification of mortality and morbidity independent of vital signs. METHODS We measured pLA in 1,168 patients transported by rotorcraft to a Level I trauma center over 18 months. The primary outcome was in-hospital mortality; secondary outcomes were emergent surgery and multiple organ dysfunction syndrome (MODS). Covariates include age, sex, prehospital vital signs, and mental status. We created multivariable logistic regression models and tested them for interaction terms and goodness of fit. Cutoff values were established for reporting operating characteristics using shock (defined as shock index >0.8, heart rate >110, and systolic blood pressure <100), tachypnea (RR ≥30), and altered sensorium (Glasgow Coma Scale score <15). RESULTS In-hospital mortality was 5.6%, 7.4% required surgery and 5.7% developed MODS. Median lactate was 2.4 mmol/L. Lactate was associated with mortality (odds ratio [OR], 1.23; p < 0.0001), surgery (OR, 1.13; p < 0.001), and MODS (OR, 1.14; p < 0.0001). Inclusion of pLA into a logistic model significantly improved the area under the receiver operator curves from 0.85 to 0.89 for death (p < 0.001), 0.68 to 0.71 for surgery (p = 0.02), and 0.78 to 0.81 for MODS (p = 0.002). When a threshold lactate value of >2 mmol/L was added to a predictive model of shock, respiratory distress, or altered sensorium, it improved sensitivity from 88% to 97% for death, 64% to 86% for surgery, and 94% to 99% for MODS. CONCLUSION The pLA measurements improve prediction of mortality, surgery, and MODS. Lactate may improve the identification of patients who require monitoring, resources, and resuscitation.