ABC Score for massive transfusion

Clinical background

In severe trauma, haemorrhage is the leading cause of preventable death, and early activation of a massive transfusion protocol improves survival. The difficulty lies in identifying which patients will actually need massive transfusion before laboratory results and imaging are available. The decision must be made within minutes of arrival, and both too much and too little have consequences: unnecessary activation wastes blood products and resources, while delayed activation costs lives.

The ABC score was constructed for precisely this decision: to screen rapidly for the need for massive transfusion using four variables available at the bedside on arrival. The score is not intended to replace clinical judgement, but to give a structured basis for when a massive transfusion protocol should be activated.

Calculating the ABC score

The ABC score consists of four binary variables, each worth 0 or 1 point:

$$\text{ABC} = \text{penetrating} + \mathbb{1}(\text{SBP} \leq 90) + \mathbb{1}(\text{HR} \geq 120) + \text{positive FAST}$$

where $\text{penetrating}$ refers to a penetrating mechanism of injury, $\text{SBP}$ is the systolic blood pressure in the emergency department, $\text{HR}$ is the heart rate in the emergency department, and $\text{positive FAST}$ refers to a positive focused assessment with sonography for trauma. The total ranges from 0 to 4. A value of 2 or more constitutes a positive screen.

The derivation cohort consisted of 596 adult trauma patients transported directly from the scene to a US level I trauma centre in Nashville between July 2005 and June 2006 [1]. Massive transfusion was defined as 10 or more units of red cells within 24 hours, and its frequency was 12.4%. In this cohort the ABC score achieved an AUROC of 0.842, comparable to more complex scores such as TASH (0.842) and McLaughlin (0.846) [1]. At a threshold of 2 points, sensitivity was 75% and specificity 86%, with 85% correctly classified [1].

Interpretation in practice

The threshold of 2 points is the calculator's positive screening limit. It is important to note that a score below 2 does not exclude massive transfusion. Sensitivity at this threshold was 75% in the derivation cohort, meaning that one in four patients who actually need massive transfusion is missed by the score [1]. Clinical judgement, mechanism, history and the patient's course must always be weighed in.

Validation and performance

In a large external validation in the German TraumaRegister DGU (5,147 patients, of whom 95% had blunt trauma, with massive transfusion in 5.6%), the ABC score performed less well than in the derivation [2]. The AUC for ABC was lower than for TASH (0.889), PWH (0.860) and Vandromme (0.840), and the authors concluded that weighted and more complex systems outperform simple unweighted models such as ABC [2]. The study was dominated by blunt trauma, which differs from the derivation cohort, where the proportion of penetrating trauma was probably higher, as is typical of US trauma centres.

In the Swiss trauma registry (13,222 patients with severe trauma, ISS 16 or higher), the discrimination of the ABC score for massive transfusion was even lower: a c-statistic of 0.66 (95% CI 0.64 to 0.69) in prehospital use [3]. At a threshold of 2 points, sensitivity for early death (within 24 hours) was only 10%, that is, 90% of patients who died early had an ABC score below 2 [3]. Specificity was 90% or higher, meaning that the score is relatively good at confirming high risk but poor at ruling out serious haemorrhage. The authors concluded that ABC, TASH and the shock index all have too low a sensitivity to serve as tools for excluding life-threatening bleeding [3].

Calibration could not be evaluated for ABC in the Swiss study, since the score does not produce a probability estimate but only a binary screen [3].

Limitations

The ABC score rests on only four variables and ignores several factors of importance for bleeding risk: age, coagulopathy, lactate, base excess, haemoglobin and the detailed pattern of injury. The score is designed for adult trauma patients and should not be applied to children, pregnant women or patients with non-traumatic bleeding.

Perhaps the most important limitation is its sensitivity. In external cohorts, particularly those dominated by blunt trauma, the ABC score at a threshold of 2 misses a substantial proportion of patients who do require massive transfusion [2, 3]. This means the score cannot be used as an exclusion tool. A patient with an ABC score of 0 or 1 may still have serious haemorrhage, and clinical suspicion should govern activation regardless of the score.

A further limitation is that FAST is not always available or interpretable in prehospital use or on arrival at smaller units, which may reduce the score to three variables and worsen performance further.

References

1. Nunez TC, Voskresensky IV, Dossett LA, et al. Early prediction of massive transfusion in trauma: simple as ABC (assessment of blood consumption)? J Trauma 2009. PMID: 19204506
2. Brockamp T, Nienaber U, Mutschler M, et al. Predicting on-going hemorrhage and transfusion requirement after severe trauma: a validation of six scoring systems and algorithms on the TraumaRegister DGU. Crit Care 2012. PMID: 22818020
3. Costa A, Carron PN, Zingg T, et al. Early identification of bleeding in trauma patients: external validation of traumatic bleeding scores in the Swiss Trauma Registry. Crit Care 2022. PMID: 36171598