Exsanguinating hemorrhage continues to be one of the

Transcription

1 ORIGINAL ARTICLE All Massive Transfusion Criteria Are Not Created Equal: Defining the Predictive Value of Individual Transfusion Triggers to Better Determine Who Benefits From Blood Rachael A. Callcut, MD, MSPH, Jay A. Johannigman, MD, Kurt S. Kadon, BA, Dennis J. Hanseman, PhD, and Bryce R. H. Robinson, MD Background: As familiarity with military massive transfusion (MT) triggers has increased, there is a growing interest in applying these in the civilian population to initiate MT protocols (MTP) earlier. We hypothesize that these triggers do not have equal predictability for MT and understanding the contribution of each would improve our ability to initiate the MTP earlier. Methods: All patients presenting to a Level I trauma center from October 2007 to September 2008 requiring immediate operation were included in this study. Emergency department records, operative logs, and blood transfusion data from arrival to procedure end were analyzed using multivariate regression techniques. Triggers included systolic blood pressure (SBP) 90 mm Hg, hemoglobin 11 g/dl, temperature 35.5 C, International normalized ratio (INR) 1.5, and base deficit 6. Results: One hundred seventy patients required immediate operation with an overall survival of 91%. Transfusion of packed red blood cells was noted in 45% (77 of 170) with the mean number of transfused units highest in those meeting SBP (12.9 Units) or INR (12.3 Units) triggers. The triggers do not contribute equal predictive value for the need for transfusion with INR being the most predictive (odds ratio, 16.7; 95% confidence interval, 2 137) for any transfusion and highly predictive for the need for MT (odds ratio, 11.3; 95% confidence interval, 3 47). In fact, if patients met either INR or SBP triggers alone, they were likely to receive MT (p and 0.003, respectively). Conclusion: Triggers have differential predictive values for need for transfusion. Defining the individual utility of each criterion will help to identify those most likely to benefit from an early initiation of the MTP. Key Words: Massive transfusion, Transfusion triggers, International normalized ratio. (J Trauma. 2011;70: ) Exsanguinating hemorrhage continues to be one of the leading causes of death for trauma patients and the majority of these deaths occur in the first 6 hours after Submitted for publication September 24, Accepted for publication January 26, Copyright 2011 by Lippincott Williams & Wilkins From the Department of Surgery (R.A.C.), Section of Trauma and Critical Care, Stanford University, Stanford, California; Department of Surgery (J.A.J., K.S.K., D.J.H., B.R.H.R.), Division of Trauma and Critical Care, University of Cincinnati, Cincinnati, Ohio; and Center for Sustainment of Trauma and Readiness Skills (D.J.H.), United States Air Force, Cincinnati, Ohio. Presented at the 69th Annual Meeting of the American Association for the Surgery of Trauma, September 22 25, 2010, Boston, Massachusetts. Address for reprints: Rachael A. Callcut, MD, Department of Surgery, Stanford University, 300 Pasteur Drive, H3680, Stanford, CA ; rcallcut@stanford.edu. DOI: /TA.0b013e e40 injury. 1 3 Recently, improvements in mortality have been found in patients who are ultimately resuscitated in a more equal fresh frozen plasma (FFP) to packed red blood cells (PRBC) transfusion ratio (or plasma rich resuscitation). 4 8 As a result, many trauma centers have altered their massive transfusion protocols (MTP) to reflect this new resuscitation strategy. Despite improvements in knowing how to resuscitate exsanguinating patients, one of the fundamental remaining keys to improve patient outcome is to expeditiously and reproducibly identify the patients most likely to require massive transfusion (MT). 4,7 11 Although the criteria for initiation of the MTP remain highly center and provider dependent, the benefits of early initiation if we could identify the appropriate patient population have been described. 6,12,13 Also importantly, there remains concern regarding inappropriate initiation of the MTP as it may be potentially harmful to patients not ultimately requiring MT volumes. Some of these well-described risks include the immunosuppressive and infectious risks from unnecessary exposure to blood component therapy. The military recognized the importance of early identification of those needing MT more than 5 years ago and has proposed the use of a standard set of physiologic and hemodynamic transfusion triggers identified on the initial evaluation of the soldier. 4,8,11,14 17 Commonly proposed triggers that were correlated with the need for MT include systolic blood pressure (SBP) 90 mm Hg, hemoglobin (hgb) 11 g/dl, temperature 35.5 C, international normalized ratio (INR) 1.5, and base deficit 6. When three or more of these triggers were present, the likelihood of MT was increased. To date, the applicability of these same proposed criteria for predicting MT in the civilian trauma patient population has not been well described. In clinical practice, patients presenting with significant hemorrhage are likely to display a variable number of the triggers, and defining the predictive utility of each individual criterion for determining whether patients will require transfusion would improve our ability to have reliable clinical indicators to initiate earlier transfusion. This study attempts to define the transfusion predictive value of each proposed trigger to better define the need for and quantity of early blood usage in patients requiring immediate operative intervention. 794 The Journal of TRAUMA Injury, Infection, and Critical Care Volume 70, Number 4, April 2011

2 The Journal of TRAUMA Injury, Infection, and Critical Care Volume 70, Number 4, April 2011 Massive Transfusion Triggers Are Not Created Equal PATIENTS AND METHODS All consecutive trauma patients presenting to an urban university-based Level I trauma center from October 1, 2007, to September 30, 2008, who required immediate major operative intervention, were selected for analysis from the institutional trauma database. Patients undergoing isolated orthopedic or neurosurgical procedures or undergoing minor procedures with a trauma surgeon were excluded from the analysis. Major operative interventions included thoracotomy, sternotomy, neck exploration, laparotomy for major injury, and major extremity trauma for control of hemorrhage. Patients undergoing diagnostic laparoscopy or nontherapeutic laparotomy were excluded from the analysis. This study was approved by the institutional review board at the University of Cincinnati. Patient records were reviewed from time of arrival to immediately postprocedure as a surrogate for estimation of the first 6 hours of care. Emergency department (ED) records, operative case logs, anesthesia records, laboratory data, and blood transfusion data were queried. ED vital signs, ED arrival time, demographics (age, race, and sex), and injury characteristics (injury severity score [ISS], injury type [penetrating vs. blunt]) were collected. Operative case logs were used to extrapolate operative procedure total time. Postanesthesia care unit or surgical intensive care unit flow sheets were used to extract immediate postoperative vital sign data. Laboratory data were extracted from the laboratory results and correlated to the ED records, operative times, and operative logs. The first laboratory data available on arrival was counted as ED laboratory results if they were drawn in the ED or performed as a point of care testing in the ED. Operative laboratories were obtained by correlating the draw times with the operative procedure times. Because multiple laboratories were drawn during operative procedures, the worst result was selected to represent the operative data. Immediate postoperative laboratory values were selected as the first laboratory value drawn after the procedure end time and were only counted as an immediate value if they were drawn within 2 hours of the procedure end time. Blood product transfusion data (units of PRBC, FFP, platelets, and cryoprecipitate) and crystalloid infusion volumes were determined by query of the transfusion database, operative reports, and anesthesia records. Units were counted if they were initiated from the time of arrival to the conclusion time of the operative procedure. Before the study period, the institution had adopted an MTP that included 10 Units of PRBCs, 8 Units of FFP, and 1 platelet dose per MTP round, and there were no preset criteria for initiation of the MTP. Intraoperative estimated blood loss was determined from anesthesia records and surgeon operative reports. Operating surgeon and extubation in the operative case was also recorded. Intensive care unit length of stay, hospital length of stay, and patient outcome (alive vs. dead) at the time of discharge were determined. Data were compared between survivors and nonsurvivors. Outcome was determined at the time of discharge or death. Categorical variables were analyzed with 2 and continuous variables were analyzed using Student s t tests with significance of p FFP:PRBC ratios were calculated and the odds ratio (OR) of death was determined using multivariate logistic regression for those transfused at a ratio of more FFP than PRBCs or plasma rich ( 1:1.5) compared with those transfused at a ratio of 1:1.5 (more PRBCs compared with FFP or plasma poor). Transfusion volumes received were determined and comparisons were made for those who met and did not met each transfusion trigger. The triggers and thresholds analyzed include SBP 90 mm Hg, hgb 11 g/dl, temperature 35.5 C, INR 1.5, and base deficit 6. To determine the individual predictive value of a criterion for transfusion, OR for receiving any PRBC transfusion was determined at the 95% confidence interval (CI) using logistic regression analysis. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for each individual trigger for predicting MT ( 10 Units PRBCs in 6 hours). To compare the proposed original application of the criterion with equal weighting (thus, initiating a plasma-rich resuscitation when 3 of the 5 criteria are met) versus the utility of using the criterion individually to predict need for transfusion, additional analysis was performed in only the patients who had data available for all five transfusion triggers. The average amount of PRBCs transfused for meeting 0 to 5 triggers and the likelihood of needing an MT were determined. The goodness of fit (c-statistic) of the model equally weighting each of the triggers for predicting the probability of transfusion was calculated. All analyses were performed using SAS v9.2 (SAS Institute Inc., Cary, NC). RESULTS During the 1-year study period, 170 patients required immediate operative intervention. The mean ISS was (range, 9 66) with an average age of 33 years 13.6 years (range, years). The majority of patients were men (147 of 170, 86%) and suffered from a penetrating mechanism (118 of 170, 69%). Overall cohort survival was 91% (155 of 170) and characteristics of survivors and nonsurvivors are shown in Table 1. Penetrating trauma patients were less likely to die compared with blunt trauma patients (5.1% vs. 17.3%, p 0.016). In the first 6 hours of care, 55% of the patients did not require any transfusion and there were no deaths in this group. PRBC transfusion was required in the other 45% (77 of 170). The overall mortality in the group requiring 1 Units to 9 Units PRBCs was 8% (15 of 77) and was 39% (11 of 28) in the group that required MT. For those requiring transfusion, mean units of PRBC and FFP were 9.8 Units 11.7 Units and 8 Units 7.6 Units, respectfully, with a range of 1 Units to 80 Units PRBCs and 0 Units to 37 Units FFP. In the first 6 hours of hospitalization, 36% (28 of 77) received an MT as defined as 10 Units PRBCs in 6 hours, and 57% (44 of 77) required 10 Units of total blood products (FFP PRBCs). The mean FFP:PRBC transfusion ratio was 1 Unit FFP:1.26 Units PRBCs, and 57% received transfusion in a ratio of 1 Unit FFP:1 Unit to 1.5 Units of PRBCs. Regardless of the total required transfusion volume, there was a survival advantage 2011 Lippincott Williams & Wilkins 795

3 Callcut et al. The Journal of TRAUMA Injury, Infection, and Critical Care Volume 70, Number 4, April 2011 TABLE 1. Comparison of Survivors vs. Nonsurvivors for Entire Cohort of 170 Consecutive Patients Undergoing Operative Intervention Variables Survived Died p Sex Female 21 2 Male Trauma type Blunt 43 (82.7%) 9 (17.3%) Penetrating 112 (94.9%) 6 (5.1%) Blood products No 1 (100.0%) 0 (0.0%) Yes 63 (80.8%) 15 (19.2%) ISS Age (yr) ED SBP (mm Hg) ED pulse (bpm) ED GCS ED hgb (g/dl) ED INR ED temperature ( C) ED base deficit ED ph Operative time (min) EBL (ml) PRBC (Units) FFP (Units) PLT (Units) Cryo (Units) IVF (ml) UOP (ml) bpm, beats per minute; GCS, Glasgow coma scale; EBL, estimated blood loss; PLT, platelets; cryo, cryoprecipitate; IVF, intravenous fluid; UOP, urine output. for patients receiving a plasma-rich resuscitation (more equal FFP:PRBC ratio) compared with those receiving a more plasma-poor resuscitation (84% vs. 60%, respectfully, p 0.05, Table 2). In fact, the odds of death were threefold higher if patients received a plasma-poor resuscitation (OR death, 3.57; 95% CI, ; p 0.05). The only preoperative statistically significant differences in the group receiving blood between survivors and nonsurvivors were ISS and Glasgow coma scale score (Table 2). The cohort was examined to determine the amount of blood transfused based solely on the individual proposed transfusion triggers (SBP 90 mm Hg, hgb 11 g/dl, temperature 35.5 C, INR 1.5, and base deficit 6), and the mean PRBC units transfused are shown in Table 3. Meeting either the ED SBP (12.9 Units) or ED INR (12.3 Units) triggers alone resulted in the highest mean PRBC units transfused and also exceeded the threshold for MT in the first 6 hours of care (Table 3). Importantly, in comparing between those who met and did not meet each trigger, there was a statistically significant difference in total PRBCs received, or trend toward significance, in each criteria (except temperature) even when the total volumes of PRBCs fell below the MT thresholds (Table 3). 796 TABLE 2. Differences Between Survivors and Nonsurvivors in Those Requiring Any Blood Transfusion Variables Survived Died p Trauma type 0.14 Blunt 22 (71%) 9 (29%) Penetrating 40 (87%) 6 (13%) ISS Age (yr) ED SBP (mm Hg) ED pulse (bpm) ED GCS ED hgb (g/dl) ED INR ED temperature ( C) ED base deficit ED ph Operative time (min) EBL (ml) PRBC (Units) FFP (Units) PLT (Units) Cryo (Units) IVF (ml) FFP:PRBC ratio 0.05 Plasma rich 48 (84%) 9 (16%) Plasma poor 9 (60%) 6 (40%) bpm, beats per minute; GCS, Glasgow coma scale; EBL, estimated blood loss; PLT, platelets; cryo, cryoprecipitate; IVF, intravenous fluid; UOP, urine output; plasma rich, FFP:PRBCs of 1:1.5; plasma poor, FFP:PRBCs 1:1.5. Overall, the individual trigger criterion had differential utility in predicting need for MT (Table 4). The individual triggers each had a high specificity (except base deficit) and NPV for the need for early MT with INR exceeding a 95% specificity and 92% NPV (Table 4). The base deficit was the least frequently (44% of cohort) obtained parameter of the five triggers, and thus, specificity was relatively poor (Table 4). The accuracy for predicting need for MT for the individual triggers was generally high with INR predicting correctly 88% of the time (Table 4). The individual triggers had differential utilities for determining who received MT with INR (OR, 11.3; 95% CI, ) and SBP (OR, 8.5; 95% CI, ) being highly predictive (Table 4). In addition, the individual triggers did not contribute equal predictive value for determining the need for any transfusion. The likelihood of receiving a transfusion was greatest when the INR trigger was exceeded (OR, 16.7; 95% CI, ; Table 3). The least predictive trigger criterion was temperature (OR, 3.4; 95% CI, ; Table 3). To account for the colinear nature of INR, the likelihood of transfusion was also determined for INR and each other variable. The highest likelihood of needing any transfusion was in those patients meeting both INR and SBP criteria (OR, 10.4; 95% CI, ), followed by INR hgb (OR, 5.2; 95% CI, ) and INR temperature (OR, 4.6; 95% CI, ; Table 5) Lippincott Williams & Wilkins

4 The Journal of TRAUMA Injury, Infection, and Critical Care Volume 70, Number 4, April 2011 Massive Transfusion Triggers Are Not Created Equal TABLE 3. Mean Units of PRBCs Transfused and Likelihood of Needing Any Transfusion for Each Individual Transfusion Trigger Mean Units PRBCs Transfused Likelihood of Transfusion OR Trigger Sample Size (n) Trigger, Yes Trigger, No p (95% CI) INR ( ) SBP 90 mm Hg ( ) Hgb 11 g/dl ( ) Base deficit ( ) Temperature 35.5 C ( ) TABLE 4. Trigger Exceeded Utility of Each Individual Trigger for Prediction of Need for Massive Transfusion SENS SPEC PPV NPV Correctly Classified p MT Received Likelihood of MT OR (95% CI) INR ( )* SBP 90 mm Hg ( )* Hgb 11 g/dl ( )* Base Deficit ( )* Temperature 35.5 C ( )* PPV, positive predictive value. *p TABLE 5. Likelihood of Any Transfusion With Two or More Criteria Weighted Individually Triggers N OR c-stat INR hgb ( ) INR BD ( ) INR SBP ( ) INR temp ( ) Hgb BD temp ( ) Temp, temperature; BD, base deficit; c-stat, measure of goodness of fit. To demonstrate the impact of using the triggers to predict MT whether they were assumed to have equal utility and be interchangeable as was originally intended in the military population, patients in whom data were available for all the five triggers were analyzed. The cohort contained 53 patients (31%) in whom all the five triggers were known and the average number of units of PRBCs and FFP transfused is shown in Table 6. Using this equal weighting, when any three or more triggers were exceeded, there was a marked increase in PRBC units transfused (Table 6). Similarly, with this method, any three or more triggers needed to be reached before the probability of receiving MT surpassed a one of three chances (Table 6). DISCUSSION Preventing early death from exsanguinating hemorrhage continues to be a focus of ongoing research in resuscitation. 14,18 Military experience garnered during the recent conflicts in Iraq and Afghanistan identified a set of transfusion trigger criteria that increased the likelihood of need for MT when three or more were present. 4,8,11,19 Commonly cited triggers include SBP 90 mm Hg, hgb TABLE 6. For Patients Having All Five Transfusion Trigger Data Available, the Average Blood Usage and Probability of Massive Transfusion With Equal Weighting of Triggers Number of Triggers Met Average PRBC Units Average PRBC FFP Units Probability of Blood >10 Units c-stat, measure of goodness of fit. c-stat g/dl, temperature 35.5 C, INR 1.5, and base deficit 6. As the civilian trauma community has become more familiar with the military findings, there has been a growing interest in applying similar triggers in the civilian population but precise criterion for initiating an MTP are still lacking 11,14. It remains challenging to expeditiously and reproducibly identify patients most likely to need significant transfusion early in their care. Traditionally, MT has been defined as 10 Units of PRBCs in 24 hours; however, recent literature has advocated redefining this criteria as 10 Units of PRBCs in 6 hours, 2,11,13,20 given that most of the hemorrhage that occurs does so early. 1,3 This early phase of care is the time interval in which patients are most likely to benefit from earlier initiation of an MTP that includes more equal plasma to PRBC transfusion ratios. 6,14 We hypothesized that understanding the contribution of each proposed trigger would improve our ability to initiate earlier MTP even in situations where only a few of the triggers are known. This study 2011 Lippincott Williams & Wilkins 797

5 Callcut et al. The Journal of TRAUMA Injury, Infection, and Critical Care Volume 70, Number 4, April 2011 examined the utility of proposed individual transfusion triggers for determining the likelihood of needing early MT (defined as 10 Units in 6 hours) in patients requiring immediate operative intervention. In this study of civilian trauma patients requiring immediate operative intervention, these proposed transfusion triggers were useful (high specificity) and accurate in differentiating between those receiving and not receiving both early MT and any transfusion of PRBCs when individual criteria were met (Tables 3 and 4). Conversely, if the triggers in this study were weighted equally and interchangeably as originally proposed in military uses, we demonstrated a similar finding to the military that the utility for predicting need for MT was less helpful until three or more triggers were exceeded (Table 6). This finding can be explained by the fact that the individual criterion had differential predictive utility and therefore should not be equally interchanged. In fact, INR was the most helpful for determining the need for both MT (highest specificity, sensitivity, and NPV of any individual trigger; Table 4) and any transfusion (OR, 16.7; Table 3). Because of the potential collinear nature of these triggers points, the predictive utility for the volume of transfusion is variable when one examines pairings of these triggers. Analysis of our data set reveals that the combination of INR and SBP was the most predictive of a patient who will receive an MT at our institution (Table 5). This is consistent with findings from other studies supporting INR as a major marker for injury in trauma patients. 21 An ideal screening test would be sensitive, specific, precise (positive predictive value), and accurate (correctly classified) to avoid both under triage and over triage of patients needing significant blood volumes. When applied individually for the prediction of MT, these triggers have relatively low sensitivities but high NPVs (Table 4). This indicates that although not all patients will be captured by an individual criterion, if a patient did not meet a trigger, they were unlikely to require an MT. This is not surprising in that it is impractical to think that application of a single trigger alone would be sensitive enough to identify all patients needing MT or even any transfusion. This is especially noteworthy in considering that unlike the military population, the civilian population contains many older individuals with multiple comorbidities that may impact whether the individual trigger may be sensitive in an individual patient. Despite the low sensitivity, when a trigger was exceeded, it was likely that a patient would require a significant early transfusion volume as there was a high specificity and accuracy for each criterion (except base deficit). Therefore, these criteria should be determined early in patient evaluation. The base deficit has a significantly lower specificity (59%) among the triggers; however, it was the least-oftenknown parameter in our cohort and the specificity may be impacted by bias in whom had a base deficit determined in the trauma bay. In the group who did have a base deficit obtained, it was highly sensitive and this is consistent with previous literature showing it to be an important predictor of coagulopathy and death in trauma patients. 17, Although other scoring algorithms have been investigated for their utility in predicting need for MT early in patient evaluations (i.e., assessment of blood consumption [ABC] score, trauma-associated severe hemorrhage, McLaughlin score), most have had limited real-time applicability due to cumbersome mathematical calculations or complex scoring algorithms that are required to determine the patients who will need MT. The ABC score (which assigns equal weight to 4 criterion: penetrating mechanism, positive FAST, heart rate 90 beats/min, or SBP 90 mm Hg) is the simplest of the other scoring algorithms as it does not rely on laboratory evaluation. 7,12 Although helpful in predicting MT if patients have all four criteria, the accuracy and specificity are variable. In fact, when no criteria were present, 13% of the patients were misclassified. 12 In addition, two or more factors had to be present with the ABC score before the score would accurately (85% of the time) predict who received MT. 12 In contrast, a single individual trigger in our study (with the exception of base deficit) would accurately predict between 77% and 88% of the time who received MT (Table 4). Thus, reliance on these other strategies is unlikely to provide additional aid in determining early need for MT and even less helpful in determining who needs blood at quantities that will not meet MT volumes. Admittedly, this cohort of patients was selected because they needed immediate operative intervention and would be the group of patients most likely to suffer from exsanguinating hemorrhage. However, it is in patients similar to this cohort for whom often there exists little information beyond these triggers during the critical first minutes of evaluation. Early, reliable prediction of the need for MT is necessary to maximize mortality benefits from resuscitation. 4,5,7,13,14,23 Given that most major trauma centers now have availability of point of care testing in the trauma bay, 4,5 reliance on the original proposed MT criteria may help to identify those most likely to benefit from early MTP activation in the civilian population. Although not all civilian studies have definitively demonstrated a mortality benefit for patients resuscitated in a plasma-rich manner, advocates have favored civilian center MTP development with an emphasis on achieving transfusion in a 1:1 1:1.5 ratio of FFP:PRBCs (or a plasma-rich strategy). 7,8,10,11,14,24 Proponents of this strategy assert that total volumes of transfusion required with this technique prevent the development (or worsening) of the lethal triad of hemorrhage. Although not the primary focus of this study, patients still seemed to benefit from a plasma-rich resuscitation irrespective of the total transfused volume they received (Table 2). This morality reduction has also been demonstrated in two other recent studies in which patients required 10 Units of PRBCs. 17,23 This finding lends further support that patients who, based on these trigger cutoffs, would have the MTP activated, seem to not suffer undo harm if they receive transfusion that falls short of the MT volumes. It is reasonable to postulate that the new alterations to the MTP strategies do indeed help us stay out of trouble rather than get out of trouble and thus, some patients who previously may have needed MT may never require an MT Lippincott Williams & Wilkins

6 The Journal of TRAUMA Injury, Infection, and Critical Care Volume 70, Number 4, April 2011 Massive Transfusion Triggers Are Not Created Equal Determining which patients are likely to require MT remains challenging in the trauma bay, and this study supports the use of the five proposed triggers for early identification of who may benefit from early MTP activation. Future efforts of this group will focus on the validation that each of these factors has variable, individual utility for predicting who benefits from early MT, who benefits from any blood, and what quantity of blood is anticipated for patients requiring immediate operative intervention. In doing so, this will help to establish a reproducible, easily applied, and accurate set of criterion to aid in the initiation of an early MTP while minimizing the overutilization of the limited resources of plasma and blood. CONCLUSIONS A plasma-rich resuscitation strategy was associated with an improved survival in trauma patients requiring immediate operative intervention. Determining who is likely to need any transfusion and especially MT will allow for early initiation of MTPs to maximize the benefit for hemorrhaging trauma patients. An INR 1.5 in a patient requiring operative intervention portends the transfusion of 10 Units of blood within 6 hours at our institution. The five proposed transfusion triggers (hgb, INR, base deficit, temperature, and SBP) have more utility as individually weighted criterion and were early reliable indicators of the need for and quantity of transfusion volumes. The individual predictive nature of these triggers should be validated in a prospective multicenter study. REFERENCES 1. Moore FA, Nelson T, McKinley BA, et al. Is there a role for aggressive use of fresh frozen plasma in massive transfusion of civilian trauma patients? Am J Surg. 2008;196: Kashuk JL, Moore EE, Johnson JL, et al. Postinjury life threatening coagulopathy: is 1:1 fresh frozen plasma:packed red blood cells the answer? J Trauma. 2008;65: Synder CW, Weinberg JA, McGwin G Jr, et al. The relationship of blood product ratio to mortality: survival benefit or survival bias? J Trauma. 2009;66: McLaughlin DF, Niles SE, Salinas J, et al. A predictive model for massive transfusion in combat casualty patients. J Trauma. 2008;64: S57 S Holcomb JB, Wade CE, Michalek JE, et al. Increased plasma and platelet to red blood cell ratios improves outcome in 466 massively transfused civilian trauma patients. Ann Surg. 2008;248: Stansbury LG, Dutton RP, Stein DM, Bochicchio GV, Scalea TM, Hess JR. Controversy in trauma resuscitation: do ratios of plasma to red blood cells matter? Transfus Med Rev. 2009;23: Nunez TC, Cotton BA. Transfusion therapy in hemorrhagic shock. Curr Opin Crit Care. 2009;15: Schreiber MA, Perkins J, Kiraly L, et al. Early predictors of massive transfusion in combat casualties. J Am Coll Surg. 2007;205: Cotton BA, Dossett LA, Au BK, Nunez TC, Robertson AM, Young PP. Room for (performance) improvement: provider-related factors associated with poor outcomes in massive transfusion. J Trauma. 2009;67: Riskin DJ, Tsai TC, Riskin L, et al. Massive transfusion protocols: the role of aggressive resuscitation versus product ratio in mortality reduction. J Am Coll Surg. 2009;209: Dente CJ, Shaz BH, Nicholas JM, et al. Early predictors of massive transfusion in patients sustaining torso gunshot wounds in a civilian level I trauma center. J Trauma. 2010;68: Nunez TC, Voskrensensky IV, Dossett LA, Shinall R, Dutton WD, Cotton BA. Early predictors of massive transfusion in trauma: simple as ABC (assessment of blood consumption)? J Trauma. 2009;66: Zink KA, Sambasivan CN, Holcomb JB, Chisholm G, Schreiber MA. A high ratio of plasma and platelets to packed red blood cells in the first 6 hours of massive transfusion improves outcomes in a large multicenter study. Am J Surg. 2009;197: Spinella PC, Holcomb JB. Resuscitation and transfusion principles for traumatic hemorrhagic shock. Blood Rev. 2009;23: Niles SE, McLaughlin DF, Perkins JG, et al. Increased mortality associated with the early coagulopathy of trauma in combat casualties. J Trauma. 2008;64: Maani CV, DeSocio PA, Holcomb JB. Coagulopathy in trauma patients: what are the main influence factors? Curr Opin Anaesthesiol. 2009;22: Spinella PC, Perkins JB, Grathwohl KW, et al. Effect of plasma and red blood cell transfusions on survival in patients with combat related traumatic injuries. J Trauma. 2008;64:S69 S 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. J Trauma. 2007;63: Arthurs Z, Cuadrado D, Beekley A, et al. The impact of hypothermia on trauma care at the 31st combat support hospital. Am J Surg. 2006;191: Duchesne JC, Islam TM, Stuke L, et al. Hemostatic resuscitation during surgery improves survival in patients with traumatic-induced coagulopathy. J Trauma. 2009;67: Brohi K, Cohen MJ, Davenport RA. Acute coagulopathy of trauma: mechanism, identification and effect. Curr Opin Crit Care. 2007;13: Cosgriff N, Moore EE, Sauaia A, Kenny-Moynihan M, Burch JM, Galloway B. Predicting life-threatening coagulopathy in the massively transfused trauma patient: hypothermia and acidoses revisited. J Trauma. 1997;42: Duchesne JC, Hunt JP, Wahl G, et al. Review of current blood transfusions strategies in a mature level I trauma center: were we wrong for the last 60 years? J Trauma. 2008;65: Gunter OL Jr, Au BK, Isbell JM, Mowery NT, Young PP, Cotton BA. Optimizing outcomes in damage control resuscitation: identifying blood product ratios associated with improved survival. J Trauma. 2008;65: DISCUSSION Dr. M. Margaret Knudson (San Francisco, California): It all started with the Germans and the TASH score or the Trauma Associated Hemorrhage Score, a list of physiologic and anatomic variables applied retrospectively and correlated with receiving more than 10 units of blood from the time the patient presented to the emergency department until they reached the ICU. Interestingly, the German scoring system gives you one point for being male and zero points for being female. But I m not going to go there today. This was followed by the McLaughlin and the Schreiber Scores, and a number of other scores including Nunez and Cotton s the ABC Score which requires a little less math than some of the others. All of these scores, however, have a positive predictive value of no higher than 55 percent. And although we trauma surgeons like to think that we have a super gestalt about who is going to need a massive transfusion, the PROMMTT Study (Prospective Observational Multicenter Massive Transfusion Study)has demonstrated to us that our gestalt is no better than 50:50 or as accurate as flipping a coin. So what does the Cincinnati Score, so eloquently presented by Dr. Callcut, add to our knowledge in this area? First, the authors have chosen to concentrate on patients who were 2011 Lippincott Williams & Wilkins 799

7 Callcut et al. The Journal of TRAUMA Injury, Infection, and Critical Care Volume 70, Number 4, April 2011 massively transfused within the first six hours of injury, that is the patients who are most likely to die from hemorrhage. Second, they focused on the criteria that were most accurate in predicting the need for activating the Cincinnati Massive Transfusion Protocol which is both plasma and platelet rich. The wisdom of this approach is obvious from their remarkably low mortality of only 16 percent in those patients who received these plasma rich transfusions. And, third, while four variables in the Cincinnati Score were correlated with the likelihood of receiving any transfusion if you add four and five of these variables together they correctively predicted the the patients who need a massive transfusion with the probability of between 55 and 78 percent. I have three questions for the authors. First, you have chosen to focus on the physiologic and point of care laboratory results already available when the patient arrives in the emergency department but when I m in the trauma room and see an open pelvic fracture or a massive hemothorax using ultrasound or x-ray I am very likely to activate my massive transfusion protocol, even if the other variables have not yet become abnormal. Do you agree with this approach? Question 2, two of the prior military studies have found that penetrating mechanism was useful in predicting the need for a massive transfusion. Have you analyzed your data comparing blunt mechanism versus penetrating? And, Number 3, equally important in activating a massive transfusion protocol is recognizing when to terminate it. From your experience in this area can you give us some insight in this process? In conclusion, although the study is provocative and gives us some insight as to the relative contributions of the selected variables in the predictive model, we must be cautious in our interpretation. Just because a patient with a low systolic blood pressure receives blood or a patient with an elevated INR on arrival was given plasma does not prove that the patient needed it. Only data collected prospectively, including serial point of care laboratory evaluation and continuous physiologic monitoring, will assure us that the right amount of blood and blood product is given to the right patient at the right time. Dr. John B. Holcomb (Houston, Texas): Dr. Callcut, that was a great presentation. You mentioned it wasn t the focus of your paper but you mentioned plasma resuscitation and you talked about platelets a little bit. As we have looked at platelets more and more I think they re becoming more important in our transfusion protocols. I think there was a publication by Dr. Inaba this week in the Journal of Trauma that suggested this. Can you comment on the effect of platelets in your retrospective study? Dr. Jameel Ali (Toronto, Ontario): You seem to equate blood transfusion requirement with blood administered. I don t think we can presume that because patients received transfusions that those were required. Could you offer some explanation for that approach? If you had some endpoints of resuscitation and you aim for those, when they are reached, you can say these are the transfusions requirements to meet those endpoints. 800 Did you do that? If you didn t, do you have an explanation for why it s not needed. Thanks. Dr. Lena M. Napolitano (Ann Arbor, Michigan): One other comment. Rachael can you address for us why you chose just to focus on the operative management patients? There are many non-operative management patients that require massive transfusion like the pelvic fracture patients. Dr. Rachael Callcut (Stanford, California): Obviously, we don t have the whole story figured out and people are working very hard in other studies like PROMMTT to prospectively validate other transfusion triggers. With respect to distinguishing between those who need and those who receive massive transfusion, I concur with the insight of both of the discussants who ask questions about this. Assuming that someone needed what they received in a retrospective study is a stretch. It s what has plagued all the retrospective studies that have looked at transfusion triggers. What reassures me, to some degree, in our study is if we didn t give you any blood you, there were no mortalities in that group. In addition to that, when we present our data and talk about if you had a trigger you were likely to require a massive transfusion, we show a very high number of patients who would for a single individual trigger, for example, the INR in 50 percent of the patients need a massive transfusion. There has been a lot of questions about whether markers such as INR and some of these transfusion triggers are a marker of illness or they re actually a marker of bleeding. And when we put the triggers together, which we did not present for the sake of time, we show that the predictive utility of a massive transfusion increases as the number of triggers are known. Even knowing a couple of these specific triggers is quite predictive, and as an example, if only 2 triggers were known, INR and systolic blood pressure are the combination that most closely correlated with need for massive transfusion. This was followed very closely with the combination of INR and hemoglobin, which suggests to me that having an elevation of INR or a low hemoglobin or a low systolic blood pressure, that those patients actually were bleeding and we re not just ill, but are actually hemorrhaging. But it, again, is something that hopefully will be answered in a number of the prospective studies that can speak more to what people actually required, rather than received. In regards to whether or not this should be activated, whether or not you should active it based upon your clinical judgment if patients don t have one of these transfusion triggers, it s important to note that if you have none of these there is still a small but measurable risk that you need massive transfusion and this probably speaks to clinical judgment. No scoring system that will ever be able to fully capture what we see in the trauma bay in terms of clinical judgment. And I think if you have a patient who has a significant mechanism that you think is likely to bleed, we still don t know for sure the answer to this question and given the fact that this doesn t completely answer the question and the fact that if you had none of these there was still a 5 percent risk 2011 Lippincott Williams & Wilkins

8 The Journal of TRAUMA Injury, Infection, and Critical Care Volume 70, Number 4, April 2011 Massive Transfusion Triggers Are Not Created Equal of massive transfusion you should default to your clinical judgment and expertise. The goal of this study and subsequent studies from our group to try to validate these in a prospective fashion is going to try to allow us to be able to apply these across the board in centers that maybe have less expertise than some of the larger urban centers that have focused on this and give us a set of criteria that assists us in terms of combining it with clinical judgment. With regard to the subgroup analysis between penetrating and blunt patients, it is a very important concept. Because we have such a large number of penetrating patients in our study group we are able to look at the penetrating patients and validate the I should say find the same results as we found here in terms of the trigger criteria. The blunt patients, also, that we look at are a relatively small group and are fairly underpowered to arrive at the same conclusions but they all do trend towards significance with INR actually being a significant predictor and the others very close. And it s probably representative of underpowered so I think it probably does hold true for the blunt patients as well. And we hope to continue the work in a larger group of patients that also included a variety of blunt patients which also speaks to why we focused on operative versus nonoperative patients. We wanted to select out a group of patients to decide as a preliminary study if these factors were even correlated and if we could then use them in subsequent studies to apply them to a larger group. So we ve selected the group of patients that we felt would be most likely to require the massive transfusion and based upon the prior studies that have looked at penetrating mechanisms or patients who needed to go to the operating room for an exsanguinating hemorrhage in that short period of time that this was the best patient group in which to perform our preliminary study. And, finally, the big question everyone has is now we re in search of triggers to try to activate massive transfusion but we also need to be in a quest to find the triggers to stop massive transfusion and our endpoints of resuscitation. There is a lot of theoretic things that we could look at the normalization of TEG, the normalization of lactate, the normalization of base deficit, looking at oxygen delivery markers, but the jury is really still out on this. This should be the subject of subsequent prospective evaluations in order to help us define when we can stop the massive transfusion protocol Lippincott Williams & Wilkins 801