INTRODUCTION:: Early prediction of massive transfusion (MT) is critical in the management of severely injured trauma patients. Variables available early after injury including physiologic, laboratory, and rotation thromboelastometric (ROTEM) parameters were evaluated as predictors for the need of MT. METHODS:: After Institutional Review Board approval, we retrospectively reviewed a cohort of severely injured trauma patients (Injury Severity Score ≥16) admitted to a Level I trauma center with available ROTEM measurements on hospital admission during a 1-year study period. Patients with isolated head injury (Abbreviated Injury Scale head ≥3 and Abbreviated Injury Scale chest, abdomen, and extremity <3) and patients with a penetrating mechanism of injury were excluded. Patients who received a MT (≥10 units packed red blood cell within 24 hours of admission) were compared with patients who did not. Variables independently associated with MT were identified using stepwise logistic regression. RESULTS:: A total of 53 patients met inclusion criteria. Of these, 18 patients (34.0%) received a MT and 35 patients (66.0%) did not. Massively transfused patients had significantly lower baseline hemoglobin values (7.9 g/dL ± 0.4 g/dL vs. 11.4 g/dL ± 0.4 g/dL; p < 0.001) and a trend toward higher lactate (4.8 mmol/L ± 0.8 mmol/L vs. 3.0 mmol/L ± 0.3 mmol/L; p = 0.056) and base deficit values (5.9 mmol/L ± 1.1 mmol/L vs. 3.6 mmol/L ± 0.6 mmol/L; p = 0.052). Mean international normalized ratio (1.46 ± 0.07 vs. 1.22 ± 0.05; p = 0.001) and partial thromboplastin times (42.4 seconds ± 5.0 seconds vs. 29.7 seconds ± 1.8 seconds; p < 0.001) were significantly higher in MT patients. Patients receiving a MT had significantly altered ROTEM values on admission compared with non-MT patients. An increase in the clot formation time (471.3 seconds ± 169.9 seconds vs. 178.1 seconds ± 19.9 seconds; p = 0.001), a shortening of the maximum clot firmness (37.5 mm ± 2.9 mm vs. 50.7 mm ± 1.4 mm; p < 0.001), and a shortening of the clot amplitude at all time points (10/20/30 minutes) were observed in massively transfused trauma patients. Variables independently associated with MT included a hemoglobin level ≤10 g/dL and an abnormal maximum clot firmness value (area under the receiver operator characteristic curve: 0.831 [95% confidence interval: 0.719-0.942; p < 0.001]). CONCLUSION:: Hemoglobin ≤10 g/dL and an abnormal maximum cloth firmness measured by rotation thromboelastometry on admission reliably predict the need for MT. Prospective validation of the effectiveness of thromboelastometry to guide the transfusion practice after trauma is warranted.