Send to

Choose Destination
J Trauma Acute Care Surg. 2013 Sep;75(3):359-64. doi: 10.1097/TA.0b013e31829e2248.

Identifying the bleeding trauma patient: predictive factors for massive transfusion in an Australasian trauma population.

Author information

From the Trauma Service (J.M.H.), and Department of Surgery (K.H., J.P.F.), Westmead Hospital; Discipline of Surgery (J.M.H., K.H., J.P.F.), University of Sydney, Sydney, New South Wales, Australia.



Military and civilian data would suggest that hemostatic resuscitation results in improved outcomes for exsanguinating patients. However, identification of those patients who are at risk of significant hemorrhage is not clearly defined. We attempted to identify factors that would predict the need for massive transfusion (MT) in an Australasian trauma population, by comparing those trauma patients who did receive massive transfusion with those who did not.


Between 1985 and 2010, 1,686 trauma patients receiving at least 1 U of packed red blood cells were identified from our prospectively maintained trauma registry. Demographic, physiologic, laboratory, injury, and outcome variables were reviewed. Univariate analysis determined significant factors between those who received MT and those who did not. A predictive multivariate logistic regression model with backward conditional stepwise elimination was used for MT risk. Statistical analysis was performed using SPSS PASW.


MT patients had a higher pulse rate, lower Glasgow Coma Scale (GCS) score, lower systolic blood pressure, lower hemoglobin level, higher Injury Severity Score (ISS), higher international normalized ratio (INR), and longer stay. Initial logistic regression identified base deficit (BD), INR, and hemoperitoneum at laparotomy as independent predictive variables. After assigning cutoff points of BD being greater than 5 and an INR of 1.5 or greater, a further model was created. A BD greater than 5 and either INR of 1.5 or greater or hemoperitoneum was associated with 51 times increase in MT risk (odds ratio, 51.6; 95% confidence interval, 24.9-95.8). The area under the receiver operating characteristic curve for the model was 0.859.


From this study, a combination of BD, INR, and hemoperitoneum has demonstrated good predictability for MT. This tool may assist in the determination of those patients who might benefit from hemostatic resuscitation.


Prognostic study, level III.

[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Wolters Kluwer
Loading ...
Support Center