If a patient is hemorrhaging and anticipated to require ongoing resuscitation, replace blood lost with blood products and do it quickly. For something that is so axiomatic in critical care, the decision to pull the trigger and initiate a massive transfusion protocol (MTP) is a decidedly complex and occasionally “weighty” one. Especially if you’re using some of the proposed logistic regression algorithms.
Hemorrhage is the most common cause of mortality in the first hour of arrival to a trauma center, and exsanguination from hemorrhage and coagulopathy combined account for nearly 50% of the deaths in the first 24 hours following an injury. (1-3) Coagulopathy develops early in these patients with rates on arrival between 25-40%. (4,5) Damage control resuscitation using an MTP mitigates trauma-induced hemorrhage and coagulopathy by delivering blood products in organized fashion with predefined blood component ratios. (6-8) Ultimately, just 3-5% of civilian trauma patients receive a massive transfusion (MT) - traditionally defined as ≥ 10 Units of packed red blood cells (pRBCs) in 24 hours - but they still consume an overwhelming 70% of all blood transfused at trauma centers. (9)
Massive transfusion is a necessarily highly coordinated operation, which mobilizes processing and delivery of large amounts of blood products for rapid transfusion. Massive transfusion protocols are thus designed by institution-specific, interdisciplinary committees to ensure appropriate activation of these valuable resources and ensure that patients are not unnecessarily exposed to blood products. Several studies have demonstrated that MTP is associated with reductions in overall mortality among severely injured trauma patients, inter-provider variability in practice, and overall blood product use. (9-12)
So you have a critically ill, hemorrhaging patient who just arrived in your trauma bay. Your spidey sense is tingling, and you want to start balanced transfusion, but should you initiate MTP? How can you predict if this patient will require continued transfusion of blood products over the ensuing 24 hours? In this post, we’ll look into the crystal ball of MTP trigger prediction tools with particular attention paid to the Assessment of Blood Consumption (ABC) Score, its attributes and limitations, and how it fares in comparison to other scores.
The ABC Score was developed at Vanderbilt University Medical Center (VUMC) based on the experiences of faculty who were noted to activate MT early. It consists of four, dichotomous, equally weighted variables:
Each variable is assigned one point if present, and, based on Nunez and colleague’s initial validation study, a score of 2 or more warrants MTP activation. The ABC Score was first assessed in a single center retrospective review of all patients admitted to VUMC’s trauma service in a one-year period. Their study population included all patients who were level I major trauma activations, were transferred directly from the scene, and received any blood transfusion during their hospitalization. The ABC Score along with the Trauma-Associated Severe Hemorrhage (TASH) and McLaughlin Scores were then applied to each patient and the ability of each to predict MT compared using area under the receiver operating characteristics curve (AUROC).
As a brief introduction to the other scores mentioned, TASH uses seven independent, weighted variables to identify patients requiring MT. These include blood pressure, gender, hemoglobin, FAST, pulse, base excess, and extremity or pelvic fractures. The McLaughlin Score is slightly less complicated as it uses four non-weighted dichotomous components – pulse > 105 bpm, SBP < 110 mm Hg, pH < 7.25, and hematocrit < 32%. Still, the McLaughlin Score is limited by its reliance on laboratory testing and in its generalizability as it was developed based on military casualties.
In the initial ABC study, 596 patients were included in the cohort of whom 12.7% (n=76) received MT. Fifty-two percent (n = 309) of these patients did not meet any ABC criteria. Of these, there was an MT rate of 2% (n=8). Nunez and colleagues chose a cutoff score of 2 based on a sensitivity of 75% and specificity of 86%. The likelihood of requiring an MT increased from 10% to 40% with a score of 2. While the ABC Score’s positive predictive value (PPV) was poor at 50-55%, it identified more than 95% of patients requiring an MT with a negative predictive value (NPV) of < 5%. It also performed well in its estimation compared to the TASH and McLaughlin Scores. ABC had the highest overall accuracy with an AUROC of 0.859 followed by TASH (0.842) and McLaughlin (0.767). Importantly, the difference between the ABC and TASH scores was not significantly different.
The authors concluded that the ABC Score was just as accurate and simpler to use than the TASH and McLaughlin scores as it does not require any laboratory testing and, subsequently, does not delay transfusion. The retrospective nature of the study and its evaluation with data from a single center were major limitations, and critics often point towards to the operator dependent nature of the FAST exam, which could other influence the score’s accuracy. (13)
Fortunately, Cotton and colleagues addressed one of these limitations in their subsequent validation study of the ABC Score, which applied the same inclusion and exclusion criteria to patients at three Level I trauma centers – VUMC, Parkland Memorial Hospital (PMH), and Johns Hopkins University (JHU). Across all three trauma centers, the sensitivity and specificity of ABC Score ranged from 75% to 90% and 67 to 88%, respectively. Correctly classified patients were similar (84-87%) and AUROC ranged between 0.83-0.90 though it was not statistically different by institution. PPV remained low (55%) and NPV high (97%). (14)
Comparisons and Critiques:
Despite the ABC Score’s ease of use and comparable accuracy in these early studies, it has undergone further scrutiny regarding its accuracy in head-to-head tests against other scores and even clinician gestalt.
ABC versus Shock Index (SI):(15)
Rationale: The Shock Index is defined as heart rate/systolic blood pressure, and greater values reflect increased need for transfusion. SI is thought to reflect degree of shock, decreased tissue oxygenation, and left ventricular performance.
Method: Schroll and colleagues performed a retrospective review of all trauma activations between January 1, 2009 and December 31, 2013 at an urban level 1 trauma center, calculating an ABC and SI score for each patient.
Results: The study ultimately analyzed 644 patients. The Shock Index ≥ 1 had a sensitivity of 67.7% and a specificity of 81.3% for predicting MTP compared to the ABC Score ≥ 2, which had a sensitivity of 47% and a specificity of 89.8%. SI also had the greater AUROC with value of 0.83 whereas the ABC Score had a value of 0.74.
Conclusions: Ultimately, they found that both scores can be used to predict need for MT, but SI is more sensitive and requires less technical skill than the ABC Score as it removes the necessity for ultrasound training and equipment. They are argued that while the ABC Score was more specific, the decrease in false positives is overshadowed the number of false negatives. Furthermore, it is easy for all providers to use and even quicker to calculate.
ABC Score versus all comers: (16)
Method: Brockcamp and colleagues performed a retrospective validation of six scoring systems and algorithms –TASH, Prince of Wales Hospital/Rainer Score (PWH), Vandromme score, ABC Score, Schreiber score, and Larson score - and stratified patients at risk of MT using a single dataset of severely injury patients derived from the TraumaRegister DGU of the German Trauma Society database. They applied the scores to each patient and then compared the AUC for each scoring system.
Prince of Wales Hospital/Rainer Score
This score was developed based on a retrospective analysis of 1,891 civilian trauma patients from a single center analysis in Hong Kong between 2001 and 2009.
It includes seven variables that predict need for MT: pulse ≥ 120 bpm, SBP ≤ 120 mmHg, Glasgow coma scale ≤ 8, displaced pelvic fracture, CT scan or FAST-positive for fluid, base deficit > 5 mmol/l, hemoglobin ≤ 7 g/dl, and hemoglobin 7.1-10.0 g/dl
This score evolved from a retrospective analysis of civilian trauma patients at a single level I trauma center in the U.S.
Its variables include blood lactate (BL) ≥ 5 mmol/l, pulse > 105 bpm, INR > 1.5, hemoglobin ≤ 11 g/dl, and SBP < 110 mmHg.
This was one of the first scores to predict MT in a military setting. The score was developed based on a retrospective cohort analysis of 558 victims who presented to two combat support hospitals in Iraq, of whom 247 (44.3%) required MT.
Based on their analysis, the independent predictors of MT included hemoglobin, INR, and penetrating mechanism of injury.
This score was developed from a retrospective review of the Joint Theater Trauma Registry transfusions database for all US personnel injured in combat during overseas contingency operations. Data from 1,124 patients between 2003-2008 was reviewed with 420 patients (37%) receiving MT.
The Larson score include the following variables in its assessment: pulse, SBP, hemoglobin, and base deficit.
Results: Brockcamp and colleagues analyzed a total of 5,147 patients in their review. The overall MT rate was 5.6% (n =289). The TASH score had the highest overall accuracy with an AUC of 0.889 followed by the PWH score with an AUC of 0.860 and the score developed by Vandromme and colleagues with an AUC of 0.840. The ABC Score performed less accurately than all other scores as reflected by an AUC of 0.763. At defined cut-off values for each score, the Schreiber score had the highest sensitivity at 85.8%, but the lowest specificity at 61.7%. The TASH score at a pre-defined cut-off of ≥ 8.5 had a sensitivity of 84.4% as well as a high specificity of 78.4%.
Conclusion: Weighted and more sophisticated systems like TASH and PWH scores, which necessarily include more and often very similar variables performed better than those which use simple, non-weighted models.
How do these scores compare against arguably the easiest and quickest assessment of patient need for massive transfusion – clinical gestalt? In short, better, but still not great.
In one study, Pommerening and colleagues asked trauma surgeons if patients were likely to be transfused 10-minutes after a patient’s arrival and then compared their answers against the TASH, McLaughlin, and ABC Scores using ROC. Their study enrolled 1245 patients of whom 966 met the inclusion criteria and 221 (23%) received MT. They found that gestalt sensitivity was 65.6% and specificity was 63.8%. PPV and NPV were 34.9% and 86.2%, respectively. Among the 486 patients analyzed for comparison, the TASH score correctly classified more patients than gestalt, but no statistically significant difference in predictive ability was discerned between gestalt and the ABC or McLaughlin scores.
Motameni and colleagues revisited this question in a study of 3421 patients seen at a single center in 2016. They found that only 33% of patients who would have had activation based on the ABC criteria used more than 5 units of blood products in 24 hours compared with 65% of patients in whom clinical judgement was used. 67% of all MTP activations based on clinical judgement would have been activated by ABC criteria in the ED. This discrepancy was similarly observed in the Operating Room.
These studies suggest that while our current algorithms like the ABC Score are limited and overestimate the need for massive transfusion, they still perform better than clinical gestalt alone.
What about specific population-based scores? (19)
Ohmori and colleagues evaluated the ABC, TASH, and PWH scores and found that all 3 were less accurate in predicting MT in trauma patients aged 65 years or greater.
The most important risk factors predicting the need for MT were FAST, unstable pelvic fractures, and long bone open fractures of the lower limbs as well as pre-injury anticoagulant use, antiplatelet agent use, lactate levels, and shock index.
In the course of an 8-year, 3-month study, they analyzed 714 patients and found that the AUC for the TASH and PWH scores were accurate in younger patients, but scores for all three systems were less accurate in older patients.
We are now approaching 10 years since Nunez and colleagues first shared the ABC Score. In the seeming flood of comparative studies that followed, the ABC Score has been variable in its predictive value, never performing quite as well as it did in Nunez and Cotton’s studies. It is only right then that it continues to be reassessed as new scores are validated and point-of-care testing evolves. Ultimately though, there the American College of Surgeons continues to recommend its adoption above other tools to trigger MTP. (20) Why?
The ABC Score is easy to remember and to use – four simple variables without weights attached.
The ABC Score does not rely on time-consuming tests. In general, scores that incorporate more sophisticated variables and weighting unsurprisingly demonstrated superior predictive value compared to the ABC Score. However, they relied on laboratory or radiographic analysis, which could critically delay transfusion. (16).
Finally, while the ABC Score overestimates the need for massive transfusion, it also consistently identifies those patients who will require a massive transfusion. It is important to remember that these scores are not predicting who will need transfusion, only who will need the additional resources of Massive Transfusion. Ultimately over-triaging is an unfortunate, but acceptable consequence in exchange for a patient who can be appropriately resuscitated and survive.
Post by Colleen Laurence, MD
Dr. Laurence is a PGY-1 resident in Emergency Medicine at the University of Cincinnati
Peer Editing by Ryan LaFollette, MD
Dr. LaFollette is an Assistant Residency Directory at the University of Cincinnati
Sauaia A, Moore FA, More EE, et al. Epidemiology of trauma deaths: a reassessment. Journal of Trauma. 1995; 38; 185-193.
Kauvar DS, Lefering R, Wade CE. Impact of hemorrhage on trauma outcome: an overview of epidemiology, clinical presentations, and therapeutic considerations. Journal of Trauma. 2006; 60: S3-S11.
Acosta JA, Yang JC, Winchell RJ, et al. Lethal injuries and time to death in a level I trauma center. Journal of American College of Surgeons. 1998; 186: 528-533.
Brohi K, Singh J, Heron M, Coats T. Acute traumatic coagulopathy. Journal of Trauma. 2003; 54: 1127-1130.
MacLeod JB, Lynn M, McKenney MG, Cohn SM, Murtha M. Early coagulopathy predicts mortality in trauma. Journal of Trauma. 2003; 55: 39-44.
Holcomb JB. Damage control resuscitation. Journal of Trauma. 2007; 62:S36-S37.
Cotton BA, Gunter OLD, Ishell J, et al. Damage control hematology: the impact of a trauma exsanguination protocol on survival and blood product utilization. Journal of Trauma. 2008; 64: 1177-1182; discussion 1182-1183.
Tieu BH, Holcomb JD, Schreiber MA. Coagulopathy: its pathophysiology and treatment in the injured patient. World Journal of Surgery. 2007; 31: 1055-1064.
Como JJ, Dutton RP, Scalea TM, et al. Blood transfusion rates in the care of acute trauma. Transfusion. 2004; 44: 809-813.
Malone DL, Jess JF, Fingerhut A. Massive transfusion practices around the globe and a suggestion for a common massive transfusion protocol. Journal of Trauma. 2006; 60: S91-S96.
Borgman MA, Spinella PC, Perkins JG, et al. The ratio of blood products transfused affects mortality in patients receiving massive transfusions at a combat support hospital. Journal of Trauma. 2007; 63:805-813.
O’Keeffe T, Refaai M, Tschorz K, Forestner JE, Sarode R. A Massive transfusion protocol to decrease blood component use and costs. Archives of Surgery. 2008; 143 (7): 686-690.
Nunez TC, Voskrensensky IV, Dossett LA, Shinall R, Dutton WD, Cotton BA. Early Prediction of Massive Transfusion in Trauma: Simple as ABC (Assessment of Blood Consumption)? The Journal of Trauma: Injury, Infection, and Critical Care. 2009; 66: 366-352.
Cotton BA, Dossett LA, Haut ER, Shafi S, Nunez TC, Au BK, Zaydfudi V, Johnston M, Arbogast P, Young PP. Multicenter Validation of a Simplified Score to Predict Massive Transfusion in Trauma. Journal of Trauma: Injury, Infection, and Critical Care. 2010; 69: S33-S39.
Schroll R, Swift D, Tatum D, Couch D, Heaney JB, Llado-Farrulla M, Zucker S, Gill F, Brown G, Buffin N. Shock index versus ABC score to predict need for massive transfusion in trauma patients. Injury. 2017; 49 (1): 15-19.
Brockcamp T, Nienaber U, Mutschler M, Wafaisade A, Peiniger S, Leferine R, Bouillon B, Maegele M. Predicting on-going hemorrhage and transfusion requirement after severe trauma: a validation of six scoring systems and algorithms on the TraumaRegister DGU ®. Critical Care. 2012; 16: R129-R138.
Pommerening MJ, Goodman MD, Holcomb JB, Wade CE, Fox EE, del Junco DJ, Brasel KJ, Bulger EM, Cohen MJ, Alarcon LH, Schreiber MA, Myers JG, Phelan HA, Muskat P, Rahbar M, Cotton BA. Clinical gestalt and the prediction of massive transfusion after trauma. Injury. 2015; 46: 807-813.
Motameni AT, Hodge RA, McKinley WI, Georgel JM, Strollo B, Benns MV, Miller KR, Harbrecht BG. The use of ABC Score in activation of massive transfusion: The yin and the yang. Journal of Traumatic Acute Care Surgery. 2018; 85 92): 298-302.
Ohmori T, Kitamura T, Ishihara J, Onishi H, Nojima T, Yamamoto K, Tamura R, Muranishi K, Matsumoto T, Tokioka T. Early predictors for massive transfusion in older adult severe trauma patients. Injury. 2017; 48: 1006-1012.
ACS Trauma Quality Improvement Program. Massive Transfusion in Trauma Guidelines. https://www.facs.org/~/media/files/quality%20programs/trauma/tqip/massive%20transfusion%20in%20trauma%20guildelines.ashx. Accessed on December 21, 2018.