Secondary Logo

Journal Logo

Massive Transfusion in Cardiac Surgery: The Impact of Blood Component Ratios on Clinical Outcomes and Survival

Delaney, Meghan DO, MPH*; Stark, Paul C. MS, ScD; Suh, Minhyung MPH; Triulzi, Darrell J. MD; Hess, John R. MD, MPH§; Steiner, Marie E. MD, MS; Stowell, Christopher P. MD, PhD; Sloan, Steven R. MD, PhD#

doi: 10.1213/ANE.0000000000001926
Hemostasis: Original Clinical Research Report

BACKGROUND: Cardiac surgery is the most common setting for massive transfusion in medically advanced countries. Studies of massive transfusion after injury suggest that the ratios of administered plasma and platelets (PLT) to red blood cells (RBCs) affect mortality. Data from the Red Cell Storage Duration Study (RECESS), a large randomized trial of the effect of RBC storage duration in patients undergoing complex cardiac surgery, were analyzed retrospectively to investigate the association between blood component ratios used in massively transfused patients and subsequent clinical outcomes.

METHODS: Massive transfusion was defined as those who had ≥6 RBC units or ≥8 total blood components. For plasma, high ratio was defined as ≥1 plasma unit:1 RBC unit. For PLT transfusion, high ratio was defined as ≥0.2 PLT doses:1 RBC unit; PLT dose was defined as 1 apheresis PLT or 5 whole blood PLT equivalents. The clinical outcomes analyzed were mortality and the change in the Multiple Organ Dysfunction Score (ΔMODS) comparing the preoperative score with the highest composite score through the earliest of death, discharge, or day 7. Outcomes were compared between patients transfused with high and low ratios. Linear and Cox regression were used to explore relationships between predictors and continuous outcomes and time to event outcomes.

RESULTS: A total of 324 subjects met the definition of massive transfusion. In those receiving high plasma:RBC ratio, the mean (SE) 7- and 28-day ΔMODS was 1.24 (0.45) and 1.26 (0.56) points lower, (P = .007 and P = .024), respectively, than in patients receiving lower ratios. In patients receiving high PLT:RBC ratio, the mean (SE) 7- and 28-day ΔMODS were 1.55 (0.53) and 1.49 (0.65) points lower (P = .004 and P = .022), respectively. Subjects who received low-ratio plasma:RBC transfusion had excess 7-day mortality compared with those who received high ratio (7.2% vs 1.7%, respectively, P = .0318), which remained significant at 28 days (P = .035). The ratio of PLT:RBCs was not associated with differences in mortality.

CONCLUSIONS: This analysis found that in complex cardiac surgery patients who received massive transfusion, there was an association between the composition of blood products used and clinical outcomes. Specifically, there was less organ dysfunction in those who received high-ratio transfusions (plasma:RBCs and PLT:RBCs), and lower mortality in those who received high-ratio plasma:RBC transfusions.

Published ahead of print March 22, 2017.

From the *Medical Division and Department of Laboratory Medicine, University of Washington, Seattle, Washington; Center for Epidemiological and Statistical Research, New England Research Institutes (Data Coordinating Center), Watertown, Massachusetts; Division of Transfusion Medicine, Department of Pathology, University of Pittsburgh, Pittsburgh, Pennsylvania; §Harborview Medical Center, Department of Laboratory Medicine and Division of Hematology, University of Washington, Seattle, Washington; Department of Pediatrics, University of Minnesota, Minneapolis, Minnesota; Department of Pathology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts; and #Department of Laboratory Medicine, Boston Children’s Hospital and Department of Pathology, Harvard Medical School, Boston, Massachusetts.

Published ahead of print March 22, 2017.

Accepted for publication December 27, 2016.

Funding: Transfusion Medicine and Hemostasis Clinical Trials Network (TMHCTN) was supported by the National Heart, Lung, and Blood Institute ( # NCT00991341).

The authors declare no conflicts of interest.

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website.

Each of the participating institutions in the RECESS study obtained human subject approval at each center.

Reprints will not be available from the authors.

Address correspondence to Meghan Delaney, DO, MPH, Bloodworks and University of Washington, 921 Terry Ave, Seattle, WA 98104. Address e-mail to

Cardiac surgery is the most common setting for massive transfusion in developed countries.1 Each year, half a million people in the United States undergo surgery for coronary artery bypass grafting or valve replacement.2 Ninety percent of these patients receive <5 units of red blood cells (RBCs), and their mortality is generally less than 1%.3 However, for patients who receive >4 units of RBCs, mortality increases progressively with each additional unit of RBCs given. In a large single-center series of 15,000 consecutive patients, mortality was 1% among those receiving 5 units of RBCs, 3% with 6 units, 7% with 7 units, 18% with 8 or 9 units, and 30% with 10 units or more.3 Although there is not a universally accepted definition for massive transfusion, the transfusion of 5 or more units of RBCs in 6 hours is one of the definitions that is becoming widely used.4,5

Bleeding in cardiac surgery is associated with platelet dysfunction caused by contact with the nonbiological cardiopulmonary bypass (CPB) machine surfaces.6 Moreover, platelets tolerate the shear of blood pumping poorly, becoming partially activated and shedding the contents of their granules.7 During the course of the surgical procedure, the platelet count decreases as they are consumed in the circuit, and the function of the remaining platelets is impaired. As a result, platelet transfusions are commonly used to treat the platelet function defect and thrombocytopenia in patients undergoing CPB.8 A landmark randomized clinical trial studied different approaches for the administration of RBCs, platelets, and plasma during massive hemorrhage as part of the resuscitation of trauma patients.9 The authors found that transfusing plasma, platelets, and RBCs in a 1:1:1 ratio (plasma:platelets:RBCs) compared with transfusing at a 1:1:2 ratio led to more patients achieving hemostasis (86% vs 78%, P = .006). Moreover, fewer patients receiving 1:1:1 ratio died because of exsanguination at 24 hours (9.2% vs 14.6% in 1:1:2 group; difference, −5.4% [95% confidence interval (CI), −10.4% to −0.5%]; P = .03). The authors did not find significant differences in overall mortality at 24 hours or at 30 days. We hypothesized that massively transfused cardiac surgery patients who received plasma and/or platelets at a high ratio to RBCs might have better clinical outcomes than those who did not.10

The Red Cell Storage Duration Study (RECESS) was a multicenter, prospective clinical trial in patients undergoing complex cardiac surgery via midline sternotomy who were selected to be at high risk of transfusion using a “Transfusion Risk Understanding Scoring Tool” (TRUST), which was developed using multivariable logistic regression modeling techniques to determine independent variables that correlate with exposure to allogeneic blood transfusion.11 Patients with a TRUST score ≥3, which predicts a 60% chance of transfusion, were randomized to receive RBCs that had been stored either ≤10 days or ≥21 days, to evaluate the effect of the duration of RBC storage on postoperative changes in the Multiple Organ Dysfunction Score (ΔMODS) and mortality ( number, NCT0099134).12 The study found that RBC storage duration was not associated with significant differences in ΔMODS at either 7 or 28 days, nor did it affect 7- or 28-day mortality. Adverse events also did not differ between the 2 arms, except for mild hyperbilirubinemia, which was more common in the subjects randomized to longer stored red cells.

To investigate whether massively transfused complex cardiac surgery patients who received blood product transfusions in a high ratio of plasma:RBCs or platelet:RBCs have improved organ function and decreased mortality, a subset of the RECESS study population was studied. The goal of this analysis of the RECESS study data was to determine whether outcomes were different among subjects who received either a low ratio or a high ratio of platelet (PLT) or plasma in relation to RBC units during massive transfusion after cardiac surgery, focusing separately on plasma units in ratio to RBC units (plasma:RBC ratio) and platelet doses in ratio to RBC units (platelet dose:RBCs).13

Back to Top | Article Outline


Subjects were enrolled in the RECESS study if they met the inclusion criteria ≥12 years of age, weighed ≥40 kg, and were scheduled for complex cardiac surgery via median sternotomy. Patients ≥18 years of age were also required to have a TRUST score of ≥3, indicating ≥60% probability of RBC transfusion.11 The study was approved by the institutional review board of each participating center, and written informed consent was obtained from all subjects. Participants were randomized to receive RBC units that had been stored ≤10 days, or RBC units stored ≥21 days, for all transfusions from randomization through postoperative day 28, hospital discharge, or death, whichever occurred first. Because most of the outcome measures in RECESS, including ΔMODS and mortality, were not different between the 2 study arms, we pooled the subjects from both arms for this study, considering them to be at equivalent risk for any consequences of differences in the volumes of blood components they received. The MODS score is a validated scoring system for changes in organ function that incorporates mortality.14 It has been used and validated in studies of patients with cardiovascular disease in the critical care setting.15,16 If subjects died, their MODS was assigned the maximum value of 24. Patients who died within 24 hours were excluded. Within the study, the date and time of the start of surgery and the issue of all blood products were recorded; the use of an intraoperative cell washer was not. Further details of the RECESS study approach, enrollment, and findings have been described.12,17 The RECESS trial was registered at number NCT00991341 on October 7, 2009.

Subjects who received ≥6 RBC units or ≥8 total blood components were identified and categorized into groups based on ratio of units of plasma and RBCs (high plasma:RBC ratio, ≥1:1; or low plasma:RBC ratio, <1:1) and the ratio of platelet doses and RBC units (high platelet:RBC ratio, ≥0.2:1; or low platelet:RBC ratio <0.2:1) that were transfused. A single unit of apheresis platelets was considered to be the equivalent of 5 units of whole blood–derived platelets. Blood products were included if they were administered from the start time of the surgery until 24 hours later. Because there is no one universally accepted definition for massive transfusion, we included patients who had 6 or more RBC units or had 8 or more total blood components. The survival outcomes (all-cause mortality at postoperatives day 2, 7, and 28) were compared between the 2 groups (defined by transfusion amounts) using Cox proportional hazards regression and Kaplan−Meier product-limit survival analysis with the log rank test. The continuous outcomes (intensive care unit [ICU] length of stay, 7-day ΔMODS and 28-day ΔMODS) were analyzed using linear regression. For each outcome, we first performed bivariate analyses. Any variable with a P < .2 was considered as a candidate for the final model; multivariable models were then constructed stepwise. Factors (key covariates) that might confound the relationship between plasma and platelet ratios and outcomes were added to the models to control for bias. These were baseline MODS, randomized treatment arm, cross-clamp time, CPB time, and the total number of blood products received in the 24-hour window; gender and baseline hemoglobin were not. In addition, we analyzed postoperative days 0, 1, and 2 laboratory values: hemoglobin; creatinine (as a surrogate for renal impairment) and bilirubin using linear regression. Interactions between variables represented in the final regression models were also explored. The validity of the regression assumptions (proportional hazards for the survival models and linearity and homoscedasticity for the linear models) was assessed for each final model. None of the models presented violated any regression assumptions. Any P < .05 was considered statistically significant. Analyses were performed using commercial statistical software (SAS version 9.3, Cary, NC).

Back to Top | Article Outline


Of the 1098 RECESS subjects, 324 met the massive transfusion criteria by having received ≥6 units of RBCs and/or a total of at least 8 blood components (≥8 blood products). Subject demographics were not different between groups defined by transfusion ratios. However, the massively transfused subjects were more likely to be male, taller, have a lower baseline hemoglobin, and higher baseline bilirubin, creatinine and MODS scores than those who were not massively transfused (Supplemental Digital Content, Table 1, They also tended to have longer CPB time and aortic cross-clamp time. Of the 324 subjects, 117 (36.1%) had a high ratio of plasma to RBC units, and 255 (78.7%) had a high ratio of platelet to RBC units (Tables 1 and 2).

Table 1.

Table 1.

Table 2.

Table 2.

All 324 subjects survived at least 24 hours. The 7-day ΔMODS was on average lower (more favorable) by 1.55 points (SE = 0.53, P = .004) for high platelet:RBC transfusion ratio, compared with low platelet:RBC ratio, after adjusting for plasma:RBC transfusion ratio, baseline MODS, the total number of blood products received in the 24-hour window, and randomized treatment arm (Table 3). The subjects receiving a high plasma:RBC transfusion ratio also had a 1.24 point lower 7-day ΔMODS compared with the low plasma:RBC ratio (SE = 0.45, P = .007), independent of the effect of platelet:RBC transfusion ratio and baseline MODS, the total number of blood products received in the 24-hour window, and randomized treatment arm. The subjects who received a high platelet:RBC transfusion ratio also had a 28-day ΔMODS score that was 1.49 points lower compared with low platelet:RBC ratio (SE = 0.65, P = .022). The subjects who were transfused with a high plasma:RBC ratio had a 1.26 lower 28-day ΔMODS score compared with low platelet:RBC ratio (SE = 0.56, P = .024). The interactions between patients who received both plasma:RBC and PLT:RBC in high or low ratios were explored and not found to be significant.

Table 3.

Table 3.

Cox proportional hazards regression models, controlling for baseline MODS, aortic cross-clamp, CPB, and total number of blood products received, showed that high plasma:RBC ratio transfusion was associated with a lower risk of death; hazard ratio of 0.23 (95% CI, 0.05, 1.01, P = .052) through 7 days and a hazard ratio of 0.36 (95% CI, 0.14, 0.97, P = .042) through 28 days. This association was not seen with the high platelet:RBC ratio at either time point. Kaplan−Meier survival analysis showed consistently better survival in the high plasma:RBC ratio subjects (Figure). Deaths began to accrue on day 1 and continued to accrue throughout the postoperative period, but the majority of all excess deaths had occurred by day 10. By the end of 28 days of follow-up, there was a 3-fold increased survival benefit in the high plasma:RBC ratio (Log rank, P = .0135) (Table 4). Analyses of hospital and ICU discharge by day 31 did not find an association with either platelet:RBC or plasma:RBC ratios (data not shown).

Table 4.

Table 4.



Table 5.

Table 5.

The only laboratory value to be statistically significantly associated with the ratios was hemoglobin. Patients with high plasma:RBC tended to have values of hemoglobin on day 2 that were 0.36 g/L lower than those with low plasma:RBC ratios (P = .016, Table 5). Patients with high platelet: RBC in the longer blood storage arm tended to have values of hemoglobin on day 2 that were 0.62 g/L lower than those with low platelet:RBC ratios (P = .009). This hemoglobin difference is not likely to be of clinical significance and was not associated with any excess mortality either in the original trial or in our reanalysis.

Back to Top | Article Outline


In this retrospective analysis, high transfusion-risk cardiac surgery patients who were part of a randomized prospective trial and received ≥6 units of RBCs or ≥8 total blood components appeared to have a survival benefit when receiving plasma in a high ratio to RBCs. This observation is consistent with retrospective and prospective studies in trauma patients, which suggest that the transfusion of high ratio plasma and PLTs during resuscitation decreases the risk of hemorrhagic death.9,18,19 Because these associations between ratios and outcomes in cardiac surgery patients were derived from a retrospective analysis, they must be considered to be hypothesis generating, and they will require further prospective studies to confirm them and before considering any practice changes.

Recent analyses in trauma and nontrauma settings have focused on survival after massive transfusion. In RECESS, there were 52 deaths; 31(9.5%) in our massively transfused group. This is consistent with survival rates reported by Dzik et al1 among nontrauma patients undergoing massive transfusion in different clinical settings. In this study, 22% of patients who received ≥20 RBC units in 48 hours were cardiac and vascular surgery patients. The authors found an overall survival of 71% (5-day) and 60% (30-day). The ratio of plasma and platelets to RBCs was noted in this study, but the association with patient survival was not reported.

In our study, we found that a high PLT transfusion ratio was associated with a small reduction in MODS. It is not clear whether the MODS is a reliable predictor of mortality in the massively transfused patients, but if it is, then the high frequency of platelet transfusions in this population may contribute to unexpected increased survival in that analysis.

There are several limitations with this analysis: it is retrospective, the examined population is small, and the statistical power of the conclusion is weak; additional subgroup analyses were not undertaken given the relatively small sample size. The subjects in this analysis were a subset of the RECESS study population who were entered into RECESS based on a predicted higher risk of RBC transfusion, that is, by having a TRUST score that predicted a 60% chance of transfusion. The TRUST score was developed using robust statistical modeling of more than 15,000 subjects and found that low hemoglobin, female gender, low body weight, greater age, elevated creatinine, and type of surgery (repeat sternotomy, nonelective, or nonisolated procedure), were risk factors for RBC transfusion (ie, favors a higher score). These selection criteria weight these characteristics in this analysis. In addition, the RECESS data set did not include center-specific transfusion policies, and so this could not be compared in this exploratory analysis. The center-to-center variation was presumably minimized by the fact that randomization was balanced by center. Surgeon skill (or a surrogate such as years of experience) was not included in the original RECESS dataset, so it was not able to be incorporated into this analysis. Despite these limitations, it is the largest population of massively transfused cardiac surgery patients in which the association of blood component ratios with clinical outcomes has been investigated.

These data suggest that transfusion of higher ratios of plasma to RBCs in massively transfused cardiac surgery patients was associated with improved outcomes. Further studies are needed to gain a deeper understanding about the impact of transfusion ratios in other cardiac surgery patient populations that are at high risk for transfusion. In the PROPPR trial, giving plasma at a 1:1 ratio with RBCs was not associated with any excess in a number of complications, more patients achieved hemostasis, and hemorrhage stopped sooner for patients at risk of hemorrhagic death. Further study on the best approach to massive transfusions in cardiac surgery is warranted to deepen our knowledge and improves patient care.

Back to Top | Article Outline


The authors thank Joe Gu for his assistance with the data analysis.

Back to Top | Article Outline


Name: Meghan Delaney, DO, MPH.

Contribution: This author helped conceive of the project idea, design the study, interpret the results, and write and revise the manuscript.

Name: Paul C. Stark, MS, ScD.

Contribution: This author helped perform analyses, interpret the results, and write and revise the manuscript.

Name: Minhyung Suh, MPH.

Contribution: This author helped formulate and perform analyses, and revise the manuscript.

Name: Darrell J.Triulzi, MD.

Contribution: This author helped design the project, interpret the results, and revise the manuscript.

Name: John R. Hess, MD, MPH.

Contribution: This author helped design the project, interpret the results, and revise the manuscript.

Name: Marie Steiner, MD, MS.

Contribution: This author helped lead the primary RECESS study, design the project, interpret the results, and revise the manuscript.

Name: Christopher P. Stowell, MD, PhD.

Contribution: This author helped design the study, interpret the results, and revise the manuscript.

Name: Steven R. Sloan, MD, PhD.

Contribution: This author helped design the study, interpret the and revise the results, and revise the manuscript.

This manuscript was handled by: Roman M. Sniecinski, MD.

Back to Top | Article Outline


1. Dzik WS, Ziman A, Cohen C, et al.; Biomedical Excellence for Safer Transfusion CollaborativeSurvival after ultramassive transfusion: a review of 1360 cases. Transfusion. 2016;56:558–563.
2. Jacobs JP, Shahian DM, Prager RL, et al. Introduction to the STS national database series: outcomes analysis, quality improvement, and patient safety. Ann Thorac Surg. 2015;100:1992–2000.
3. Koch CG, Li L, Duncan AI, et al. Morbidity and mortality risk associated with red blood cell and blood-component transfusion in isolated coronary artery bypass grafting. Crit Care Med. 2006;34:1608–1616.
4. Stanworth SJ, Morris TP, Gaarder C, et al. Reappraising the concept of massive transfusion in trauma. Crit Care. 2010;14:R239.
5. Zatta AJ, McQuilten ZK, Mitra B, et al.; Massive Transfusion Registry Steering Committee. Elucidating the clinical characteristics of patients captured using different definitions of massive transfusion. Vox Sang. 2014;107:60–70.
6. Kehara H, Takano T, Ohashi N, Terasaki T, Amano J. Platelet function during cardiopulmonary bypass using multiple electrode aggregometry: comparison of centrifugal and roller pumps. Artif Organs. 2014;38:924–930.
7. Harker LA, Malpass TW, Branson HE, Hessel EA II, Slichter SJ. Mechanism of abnormal bleeding in patients undergoing cardiopulmonary bypass: acquired transient platelet dysfunction associated with selective alpha-granule release. Blood. 1980;56:824–834.
8. Perek B, Stefaniak S, Komosa A, Perek A, Katynska I, Jemielity M. Routine transfusion of platelet concentrates effectively reduces reoperation rate for bleeding and pericardial effusion after elective operations for ascending aortic aneurysm. Platelets. 2016:1–7.
9. Holcomb JB, Tilley BC, Baraniuk S, et al.; PROPPR Study Group. Transfusion of plasma, platelets, and red blood cells in a 1:1:1 vs a 1:1:2 ratio and mortality in patients with severe trauma: the PROPPR randomized clinical trial. JAMA. 2015;313:471–482.
10. Sniecinski RM, Levy JH. Bleeding and management of coagulopathy. J Thorac Cardiovasc Surg. 2011;142:662–667.
11. Alghamdi AA, Davis A, Brister S, Corey P, Logan A. Development and validation of Transfusion Risk Understanding Scoring Tool (TRUST) to stratify cardiac surgery patients according to their blood transfusion needs. Transfusion. 2006;46:1120–1129.
12. Steiner ME, Ness PM, Assmann SF, et al. Effects of red-cell storage duration on patients undergoing cardiac surgery. N Engl J Med. 2015;372:1419–1429.
13. Holcomb JB, Jenkins D, Rhee P, et al. Damage control resuscitation: directly addressing the early coagulopathy of trauma. J Trauma. 2007;62:307–310.
14. Marshall JC, Cook DJ, Christou NV, Bernard GR, Sprung CL, Sibbald WJ. Multiple organ dysfunction score: a reliable descriptor of a complex clinical outcome. Crit Care Med. 1995;23:1638–1652
15. Hébert PC, Yetisir E, Martin C, et al.; Transfusion Requirements in Critical Care Investigators for the Canadian Critical Care Trials Group. Is a low transfusion threshold safe in critically ill patients with cardiovascular diseases? Crit Care Med. 2001;29:227–234.
16. Buckley TA, Gomersall CD, Ramsay SJ. Validation of the multiple organ dysfunction (MOD) score in critically ill medical and surgical patients. Intensive Care Med. 2003;29:2216–2222.
17. Steiner ME, Assmann SF, Levy JH, et al. Addressing the question of the effect of RBC storage on clinical outcomes: the Red Cell Storage Duration Study (RECESS) (Section 7). Transfus Apher Sci. 2010;43:107–116.
18. Cotton BA, Reddy N, Hatch QM, et al. Damage control resuscitation is associated with a reduction in resuscitation volumes and improvement in survival in 390 damage control laparotomy patients. Ann Surg. 2011;254:598–605.
19. Holcomb JB, del Junco DJ, Fox EE, et al.; PROMMTT Study Group. The prospective, observational, multicenter, major trauma transfusion (PROMMTT) study: comparative effectiveness of a time-varying treatment with competing risks. JAMA Surg. 2013;148:127–136.

Supplemental Digital Content

Back to Top | Article Outline
Copyright © 2017 International Anesthesia Research Society