Free
Perioperative Medicine  |   February 2016
Morbidity and Mortality after High-dose Transfusion
Author Notes
  • From the Department of Anesthesiology and Critical Care Medicine (D.J.J., A.V.S., V.M.B., S.P., J.O.W., S.M.F.) and Department of Pathology (P.M.N., T.G.), The Johns Hopkins Medical Institutions, Baltimore, Maryland.
  • Submitted for publication April 2, 2015. Accepted for publication October 5, 2015.
    Submitted for publication April 2, 2015. Accepted for publication October 5, 2015.×
  • Address correspondence to Dr. Frank: Department of Anesthesiology, Johns Hopkins Hospital, Sheik Zayed Tower 6208, 1800 Orleans Street, Baltimore, Maryland 21287. sfrank3@jhmi.edu. Information on purchasing reprints may be found at www.anesthesiology.org or on the masthead page at the beginning of this issue. Anesthesiology’s articles are made freely accessible to all readers, for personal use only, 6 months from the cover date of the issue.
Article Information
Perioperative Medicine / Clinical Science / Coagulation and Transfusion / Infectious Disease
Perioperative Medicine   |   February 2016
Morbidity and Mortality after High-dose Transfusion
Anesthesiology 2 2016, Vol.124, 387-395. doi:10.1097/ALN.0000000000000945
Anesthesiology 2 2016, Vol.124, 387-395. doi:10.1097/ALN.0000000000000945
Abstract

Background: It is well recognized that increased transfusion volumes are associated with increased morbidity and mortality, but dose–response relations between high- and very-high-dose transfusion and clinical outcomes have not been described previously. In this study, the authors assessed (1) the dose–response relation over a wide range of transfusion volumes for morbidity and mortality and (2) other clinical predictors of adverse outcomes.

Methods: The authors retrospectively analyzed electronic medical records for 272,592 medical and surgical patients (excluding those with hematologic malignancies), 3,523 of whom received transfusion (10 or greater erythrocyte units throughout the hospital stay), to create dose–response curves for transfusion volumes and in-hospital morbidity and mortality. Prehospital comorbidities were assessed in a risk-adjusted manner to identify the correlation with clinical outcomes.

Results: For patients receiving high- or very-high-dose transfusion, infections and thrombotic events were four to five times more prevalent than renal, respiratory, and ischemic events. Mortality increased linearly over the entire dose range, with a 10% increase for each 10 units of erythrocytes transfused and 50% mortality after 50 erythrocyte units. Independent predictors of mortality were transfusion dose (odds ratio [OR], 1.037; 95% CI, 1.029 to 1.044), the Charlson comorbidity index (OR, 1.209; 95% CI, 1.141 to 1.276), and a history of congestive heart failure (OR, 1.482; 95% CI, 1.062 to 2.063).

Conclusions: Patients receiving high- or very-high-dose transfusion are at especially high risk for hospital-acquired infections and thrombotic events. Mortality increased linearly over the entire dose range and exceeded 50% after 50 erythrocyte units.

Abstract

Mortality increased linearly with erythrocyte dose, reaching 50% in patients given more than 50 units of blood. Infection and thrombotic events were the most common complications.

What We Already Know about This Topic
  • Massive transfusion is associated with serious complications and mortality; however, the dose dependence remains unclear

  • The authors evaluated the records of more than 272,000 patients, of whom 3,523 were given at least 10 units of erythrocytes

What This Article Tells Us That Is New
  • Mortality increased linearly with erythrocyte dose, reaching 50% in patients given more than 50 units of blood

  • Infection and thrombotic events were the most common complications

BLOOD transfusion, the most commonly performed procedure in U.S. hospitals,1  can be a life-saving measure in hemorrhaging patients or in those with moderate-to-severe anemia. In the past decade, massive transfusion protocols have evolved, based on the studies attempting to identify the ideal ratio of blood components administered,2–4  which continues to be somewhat controversial. What remains to be determined, however, are the clinical factors that predict outcomes after high- or very-high-dose transfusion, and the relation between transfusion dose and rates of morbidity and mortality.
It is recognized that large transfusion volumes are associated with increased morbidity and mortality.5–11  However, as more aggressive approaches to transfusion are being adopted, more patients are receiving high- or very-high-volume transfusions. In fact, a massive transfusion is generally regarded as a patient receiving more than 10 units within 24 h,12  but, to the best of our knowledge, no studies have characterized the effect of incrementally increasing doses greater than 10 units on important clinical outcomes. The purpose of this investigation was to determine the dose–response relation for transfusion volume and mortality as a primary outcome. The secondary outcomes were to determine (1) the dose–response relation for transfusion volume and morbid events, (2) the morbid events that were most common after high-dose transfusion, and (3) the clinical variables associated with adverse outcomes in high-dose transfused patients.
Materials and Methods
After receiving approval from the institutional review board at the Johns Hopkins Medical Institutions (Baltimore, Maryland), we acquired electronic medical record data from a Web-based intelligence portal (IMPACT Online; Haemonetics Corp., USA) and from our hospital billing database for 283,025 inpatients discharged from the Johns Hopkins Hospital between January 2009 and October 2014. Our use of these databases and our quality control methods have been described previously.13,14  Collected data included up to 29 prehospital comorbidities for each patient, the Charlson comorbidity index, and the number of erythrocyte units each patient received throughout his/her hospitalization. In addition, in-hospital morbid events were determined by International Classification of Diseases, 9th edition (ICD-9) codes upon discharge, as we have described previously.14 
We excluded 10,433 patients whose final diagnosis-related group description contained the following words: “bone marrow transplant,” “leukemia,” or “lymphoma,” because patients with hematologic malignancies have unique blood transfusion requirements (due to bone marrow aplasia and chemotherapy). The remaining 272,592 patients were further analyzed. Of these patients, 3,523 received high-dose transfusion (10 or greater erythrocyte units during their hospitalization). We analyzed this subset of patients to determine the predictors of clinical outcomes.
On the basis of the total number of allogeneic erythrocyte units transfused over the course of their hospitalization, we stratified patients into the following eight groups: 0, 1 to 9, 10 to 19, 20 to 29, 30 to 39, 40 to 49, 50 to 75, and greater than 75 units. We also compared the baseline characteristics of these groups. The primary outcomes were composite morbidity (occurrence of any of the five morbid events) and mortality during the hospital stay. Morbid events were defined as (1) infection (Clostridium difficile, sepsis, surgical-site infection, or drug-resistant infection), (2) thrombotic event (deep venous thrombosis, pulmonary embolus, or disseminated intravascular coagulation), (3) kidney injury, (4) respiratory event, and (5) ischemic event (myocardial infarction, transient ischemic attack, or cerebrovascular accident). Both individual and composite morbidities are reported. The composite morbidity was defined as having any of the above listed five morbid events during their hospitalization. Data to determine morbid events were obtained through the hospital’s billing database by using ICD-9 codes as outlined in detail in the appendix. Conditions that were flagged as present on admission were not considered to be hospital-acquired morbid events. In our institution, as required by the Centers for Medicare and Medicaid Services in the state of Maryland since 2007, these conditions are routinely identified as present on admission in the database, which allowed us to exclude them as hospital-acquired events.
Outcome Assessment and Statistical Analysis
Using the eight groups of total number of allogeneic erythrocyte units transfused (which included all 272,592 patients), dose–response curves were constructed for both morbidity and mortality. Once plotted, a linear and logarithmic model was performed to fit each curve respectively. The model that best predicted the dose–response curve was determined by using the R2 value. These dose–response curves were repeated for the nonsurgical and surgical patient subsets.
For the subset of patients receiving high-dose transfusion (10 or greater erythrocyte units during their hospitalization), we analyzed the relation between the number of erythrocyte units transfused and clinical outcomes (both morbidity and mortality) in a risk-unadjusted and risk-adjusted manner by using univariable and multivariable logistic regressions, respectively. All 17 of the baseline patient characteristics and comorbidities were included in the multivariable logistic regression model, with the primary aim of determining independent predictors of adverse outcomes in high-dose transfused patients.
Continuous variables that were normally distributed are reported as mean ± SD; those that were not normally distributed as well as ordinal variables are reported as median and interquartile range. Normality of distribution was tested by the Shapiro–Wilk test. P value less than 0.05 was used to define significance; however, we applied a Bonferroni correction to adjust for multiple comparisons using an α of 0.0028 (0.05/18), 18 being the number of independent variables tested as potential predictors of outcomes. Odds ratios (ORs) and 95% CIs are reported. Analyses were generated with JMP version 9.0.2 (SAS Institute, Inc., USA).
Results
For each of the groups stratified by erythrocyte dose, we compared patient characteristics and prehospital admission comorbidities (table 1). As expected, patients who were transfused had more comorbidities, as did those receiving more erythrocyte units. A dose–response curve was then created for morbidity according to the total number of erythrocyte units transfused during the hospital stay (fig. 1). The incidence of composite morbidity (having one or more of the five morbid events) increased in a curvilinear manner over the entire range of erythrocyte units administered. Patients who received no transfusion had a morbidity rate of 9.8% (22,919 of 233,460); those who received 1 to 9 units, 28.5% (10,149 of 35,609); 10 to 19 units, 55.7% (1,348 of 2,418); 20 to 29 units, 66.8% (385 of 576); 30 to 39 units, 73.9% (153 of 207); 40 to 49 units, 76.7% (79 of 103); 50 to 75 units, 78.9% (101 of 128); and greater than 75 units, 80.2% (73 of 91). These values resulted in a best-fit curve of percent morbidity = 36.53 ln(erythrocyte units) + 10.4, with an R2 value of 0.96. A 50% morbidity rate occurred in patients who received 10 or greater erythrocyte units.
Table 1.
Patient Characteristics for Groups Defined by Units of Erythrocytes Transfused
Patient Characteristics for Groups Defined by Units of Erythrocytes Transfused×
Patient Characteristics for Groups Defined by Units of Erythrocytes Transfused
Table 1.
Patient Characteristics for Groups Defined by Units of Erythrocytes Transfused
Patient Characteristics for Groups Defined by Units of Erythrocytes Transfused×
×
Fig. 1.
In-hospital morbidity and mortality rates according to the number of erythrocyte units transfused. In-hospital morbidity (a composite of all five morbid events shown in fig. 2) increased with erythrocyte dose in a curvilinear manner, reaching a 50% rate of morbidity at 10 or greater erythrocyte units. The slope was steepest up to 30 erythrocyte units, with an inflection point and plateau at higher doses. The formula defining the curve is y = 36.5 ln(x) + 10.4 (R2 = 0.962). Mortality increased in a linear manner with a slope close to 10, indicating that for each 10-erythrocyte unit increment, mortality increased approximately 10%. After transfusion of 50 units, mortality exceeded 50%. The formula defining the curve is y = 9.47 (x) − 10.56 (R2 = 0.99). RBC = erythrocyte.
In-hospital morbidity and mortality rates according to the number of erythrocyte units transfused. In-hospital morbidity (a composite of all five morbid events shown in fig. 2) increased with erythrocyte dose in a curvilinear manner, reaching a 50% rate of morbidity at 10 or greater erythrocyte units. The slope was steepest up to 30 erythrocyte units, with an inflection point and plateau at higher doses. The formula defining the curve is y = 36.5 ln(x) + 10.4 (R2 = 0.962). Mortality increased in a linear manner with a slope close to 10, indicating that for each 10-erythrocyte unit increment, mortality increased approximately 10%. After transfusion of 50 units, mortality exceeded 50%. The formula defining the curve is y = 9.47 (x) − 10.56 (R2 = 0.99). RBC = erythrocyte.
Fig. 1.
In-hospital morbidity and mortality rates according to the number of erythrocyte units transfused. In-hospital morbidity (a composite of all five morbid events shown in fig. 2) increased with erythrocyte dose in a curvilinear manner, reaching a 50% rate of morbidity at 10 or greater erythrocyte units. The slope was steepest up to 30 erythrocyte units, with an inflection point and plateau at higher doses. The formula defining the curve is y = 36.5 ln(x) + 10.4 (R2 = 0.962). Mortality increased in a linear manner with a slope close to 10, indicating that for each 10-erythrocyte unit increment, mortality increased approximately 10%. After transfusion of 50 units, mortality exceeded 50%. The formula defining the curve is y = 9.47 (x) − 10.56 (R2 = 0.99). RBC = erythrocyte.
×
The dose–response curve for percent mortality according to the number of erythrocyte units transfused during the hospital stay is shown in figure 1. Mortality increased in a linear manner with increasing transfusion dose. Patients who received no transfusion had a mortality of 0.74% (1,728 of 233,460); those who received 1 to 9 units, 4.7% (1,674 of 35,609); 10 to 19 units, 18.1% (437 of 2,418); 20 to 29 units, 30.2% (174 of 576); 30 to 39 units, 36.7% (76 of 207); 40 to 49 units, 42.7% (44 of 103); 50 to 75 units, 58.6% (75 of 128); and greater than 75 units, 61.5% (56 of 91). These values gave a best-fit line of percent mortality = 9.47(erythrocyte units) − 10.56, with an R2 value of 0.99. The slope of the mortality line of fit (9.47) indicates that mortality increased by approximately 10% for every 10 erythrocyte units. A 50% mortality rate occurred at a dose of 50 erythrocyte units.
The dose–response curves for morbidity and mortality were plotted for the surgical (n = 79,865) and nonsurgical (n = 192,727) patient subsets. To test for interaction between erythrocyte dose and surgical status, two independent variables (erythrocyte dose as a continuous variable and surgical status) were entered into a multiple logistic regression with either morbidity or mortality as dependent variables. Being a surgical patient was associated with decreased morbidity when adjusted for erythrocyte dose (OR, 0.892; 95% CI, 0.853 to 0.932). Likewise, surgical status was associated with decreased mortality when adjusted for erythrocyte dose (OR, 0.493; 95% CI, 0.449 to 0.541).
For the secondary analysis, we included only patients who were transfused with 10 or greater erythrocyte units (n = 3,523). We performed univariable and multivariable analyses to identify the predictors of morbidity in the high-dose transfused patients (table 2). The comorbidities associated with increased in-hospital morbidity included congestive heart failure, peripheral vascular disease, renal disease, and liver disease in the univariate model. The Charlson comorbidity index and the number of erythrocyte units were associated with increased morbidity in both the risk-adjusted and risk-unadjusted models.
Table 2.
Predictors of Morbidity* in High-dose Transfused Patients†
Predictors of Morbidity* in High-dose Transfused Patients†×
Predictors of Morbidity* in High-dose Transfused Patients†
Table 2.
Predictors of Morbidity* in High-dose Transfused Patients†
Predictors of Morbidity* in High-dose Transfused Patients†×
×
Next, we performed univariable and multivariable analyses to identify the predictors of mortality in the high-dose transfused patients (table 3). The comorbidities associated with increased in-hospital mortality included congestive heart failure and liver disease in the univariate model. The Charlson comorbidity index, number of erythrocyte units, being a nonsurgical patient, and congestive heart failure were associated with increased mortality in both the risk-unadjusted and risk-adjusted models.
Table 3.
Predictors of Mortality* in High-dose Transfused Patients†
Predictors of Mortality* in High-dose Transfused Patients†×
Predictors of Mortality* in High-dose Transfused Patients†
Table 3.
Predictors of Mortality* in High-dose Transfused Patients†
Predictors of Mortality* in High-dose Transfused Patients†×
×
Finally, we analyzed the entire cohort of 272,592 patients to identify the relation between erythrocyte transfusion dose and occurrence of the five different morbid events included in the composite morbidity outcome (fig. 2). Hospital-acquired infections increased dramatically up to a rate of 40% at 40 erythrocyte units and then plateaued at higher doses. Thrombotic events increased dramatically up to a rate of 50% at 50 units and then exhibited a similar plateau with higher doses. Renal, respiratory, and ischemic event rates increased gradually up to rates of 5 to 10% at an erythrocyte dose of 20 units but showed very little increase with higher doses. These findings suggest a difference in the dose–response for the various morbidities with respect to transfusion.
Fig. 2.
Event rates for five morbid outcomes are plotted according to the number of erythrocyte units transfused. In high-dose transfused patients, hospital-acquired infections and thrombotic events were four to five times more prevalent than renal, respiratory, or ischemic events. The incidence of infection increased with erythrocyte dose up to 40% and then plateaued. Thrombotic events increased up to a rate of 50% before reaching a plateau. Renal, respiratory, and ischemic event rates increased gradually up to rates of 5 to 10% at an erythrocyte dose of 20 units. RBC = erythrocyte.
Event rates for five morbid outcomes are plotted according to the number of erythrocyte units transfused. In high-dose transfused patients, hospital-acquired infections and thrombotic events were four to five times more prevalent than renal, respiratory, or ischemic events. The incidence of infection increased with erythrocyte dose up to 40% and then plateaued. Thrombotic events increased up to a rate of 50% before reaching a plateau. Renal, respiratory, and ischemic event rates increased gradually up to rates of 5 to 10% at an erythrocyte dose of 20 units. RBC = erythrocyte.
Fig. 2.
Event rates for five morbid outcomes are plotted according to the number of erythrocyte units transfused. In high-dose transfused patients, hospital-acquired infections and thrombotic events were four to five times more prevalent than renal, respiratory, or ischemic events. The incidence of infection increased with erythrocyte dose up to 40% and then plateaued. Thrombotic events increased up to a rate of 50% before reaching a plateau. Renal, respiratory, and ischemic event rates increased gradually up to rates of 5 to 10% at an erythrocyte dose of 20 units. RBC = erythrocyte.
×
Discussion
In this study, we found that for high- and very-high-dose transfused patients, overall morbidity increased in a dose-related curvilinear manner, with a 50% morbid event rate occurring at 10 or greater erythrocyte units. Infections and thrombotic events occurred four to five times more commonly than renal, respiratory, or ischemic events at the higher transfusion doses. Mortality also increased in a dose-related but in a linear manner (10% increase with every 10 erythrocyte units) and exceeded 50% after 50 erythrocyte units. Sicker patients (indicated by the Charlson comorbidity index), those with congestive heart failure, and patients who did not undergo surgery had worse clinical outcomes.
Previous retrospective studies have shown the incremental risk associated with increasing transfusion dose but have focused primarily on lower blood doses compared with those examined in our study. Turan et al.7  showed that 30-day mortality was greater in patients who received 5 or greater units of blood than in those who received 1 to 4 units. In addition, Yang et al.10  found that mortality increased with the increase in volume of erythrocyte transfusion for patients receiving 10 or greater units within 72 h. However, that study had only 52 patients who received more than 30 units. An additional study revealed a 39% mortality in patients who were massively transfused (10 or greater units),15  and Koch et al.8  showed a dose-dependent, risk-adjusted increase in postoperative morbidity and mortality in isolated patients who underwent coronary artery bypass graft surgery, focusing on the 0 to 10 unit range.
The novelty of our study lies in describing the outcomes in relation to transfusion dose in patients who received very high transfusion doses of 20, 40, 60, or even greater than 75 units of erythrocytes. Furthermore, we report dose–response curves illustrating the relative incidence of specific morbid events in these patients. In addition, a large cohort of high-dose transfused patients (n = 3,532) allowed us to identify the clinical characteristics that are predictive of adverse outcomes. The number of erythrocyte units was significantly associated with increased morbidity and mortality in both the risk-unadjusted and risk-adjusted models. Of note, patients with congestive heart failure had increased risk of mortality in the risk-adjusted model of high-dose transfused patients. Patients with poor ventricular function may be more susceptible to transfusion-associated circulatory overload than their healthy counterparts. This factor may account for the increased risk of morbidity and mortality among patients with congestive heart failure. Transfusion-associated circulatory overload is the second leading cause of transfusion-related death,16  with an incidence recently estimated at 3 to 5% of transfused patients,17  a rate higher than historically believed. Such evidence would suggest that a more cautious approach should be used in patients with poor cardiac function who require high-dose transfusion. It is also important to avoid severe anemia or hypovolemia in patients with poor ventricular function because these conditions can worsen heart failure.
Although the composite outcomes are commonly used in clinical studies, our findings were perhaps more revealing when the composite outcome was broken out into five individual morbid events. Surprisingly, infections and thrombotic events occurred not only with much greater frequency than other morbidities (renal, respiratory, and ischemic events), but they were also more dose-dependent with transfusion, showing steep curves from 0 units up to 40 to 50 units, followed by a plateau at higher doses. These findings are consistent with those of previous studies that indicate that transfusion-related immune suppression18  results in higher infection rates.19,20  Thrombotic events have also been associated with transfusion20–22  and are thought to be dose-dependent,23  but the mechanism is not entirely clear. High-dose transfusion has not been assessed previously with regard to thrombotic event rates. Our findings did show increased renal, respiratory, and ischemic event rates with increasing transfusion dose, but these events were much less common than infections and thrombosis and reached a plateau at approximately 20 erythrocyte units. These findings indicate that clinicians may need to be vigilant to prevent, diagnose, and treat infections and thrombosis in patients who undergo high-dose transfusion in order to improve outcome and reduce mortality.
Interestingly, surgical patients had lower mortality compared with nonsurgical patients across all blood doses, and this association remained in the risk-adjusted model for high-dose transfused patients. We hypothesize that this is because most surgical patients with massive bleeding are having surgery to correct and control the bleeding (splenectomy, trauma, aortic aneurysm repair, and many more). Therefore, these patients who have surgically correctable problems fare better than medical patients who are bleeding profusely (end-stage liver disease, gastrointestinal bleeding, and many more) that are not amenable to surgery or the patients not healthy enough to undergo surgery.
Allocation of blood, a precious resource, is a contentious point among medical providers. In one study, 3% of trauma patients who underwent massive transfusion accounted for 71% of all erythrocytes administered in a calendar year at the R Adams Cowley Shock Trauma Institute (Baltimore, Maryland).15  The quantification of survival at high doses of transfusion, which we have shown, can help with the allocation of these resources. After 50 units of transfusion, our data show that mortality was approximately 50%; therefore, our findings would discourage the practice of terminating erythrocyte transfusion due to futility at a set number of units, but rather to follow individual case-based clinical decision-making taking into account the patient’s clinical course. For instance, one patient in our cohort survived after receiving 129 units of erythrocytes.
The most significant limitations in our study stem from its retrospective nature. The primary findings, however, such as the dose–response for transfusion and outcomes, as well as the clinical predictors of adverse outcomes, should be valid, even with the retrospective design. Furthermore, we attempted risk adjustment using a multivariable analysis, although some residual confounding may remain, with sicker patients receiving higher-dose transfusions. Another potential limitation is that we analyzed transfusion data from each patient’s entire hospital stay. Therefore, a substantial number of the high- and even very-high-dose transfused patients may not have met the classic definition of massive transfusion (10 or greater units within 24 h). Nonetheless, we describe what we believe has not been previously shown—the dose–response curves over a wide range of transfusion volumes for in-hospital morbidity and mortality as well as the relative incidence of specific morbid events in relation to transfusion dose. Another potential limitation is the use of ICD-9 codes from an administrative database to assess morbid outcomes. Based on the nuances of coding charts, this method may not be as reliable as prospectively collected event rates for morbidity. Furthermore, the timing of these hospital-acquired morbidities relative to the transfusions cannot be established (e.g., we cannot specify that the transfusion occurred before the morbidity). To minimize this limitation, we did, however, exclude morbidity that was present on admission, which is a required data element at our institution. It should also be emphasized that our findings do not support a cause and effect relation between transfusion and outcomes, but rather an association between the two.
In summary, high-dose transfused patients are at high risk for hospital-acquired infections and thrombotic events, which were four to five times more prevalent than renal, respiratory, and ischemic events. Mortality increased in a linear manner, such that for every 10 erythrocyte units, mortality increased approximately 10%, and mortality exceeded 50% after 50 erythrocyte units. In the high-dose transfused cohort, those with high Charlson scores or congestive heart failure had worse clinical outcomes. Our findings may help guide treatment for high-dose transfusion patients by encouraging physicians to avoid volume overload in high-risk patients and to be vigilant for infectious and thrombotic complications. In addition, our data may allow ethical decisions to be made when considering the futility question of giving huge quantities of blood products to patients.
Acknowledgments
The authors thank Claire Levine, M.S., E.L.S. (Editorial Services, Department of Anesthesiology and Critical Care Medicine, Johns Hopkins Medicine, Baltimore, Maryland), for editorial assistance.
Support was provided solely from institutional and/or departmental sources.
Competing Interests
Dr. Frank has received consulting fees from Haemonetics (Braintree, Massachusetts), Medtronic (Minneapolis, Minnesota), and Biomet (Warsaw, Indiana), none of which have products discussed in this article. Dr. Ness has been a consultant for Terumo BCT (Lakewood, Colorado), which does not have products discussed in this article. The other authors declare no competing interests.
References
Pfuntner, A, Wier, LM, Stocks, C Most Frequent Procedures Performed in U.S. Hospitals, 2011: Statistical Brief #165, Healthcare Cost and Utilization Project (HCUP) Statistical Briefs. Rockville, MD, Agency for Healthcare Research and Quality. (2013).
Borgman, MA, Spinella, PC, Perkins, JG, Grathwohl, KW, Repine, T, Beekley, AC, Sebesta, J, Jenkins, D, Wade, CE, Holcomb, JB The ratio of blood products transfused affects mortality in patients receiving massive transfusions at a combat support hospital.. J Trauma. (2007). 63 805–13 [Article] [PubMed]
Holcomb, JB, Fox, EE, Wade, CE PROMMTT Study Group, The PRospective Observational Multicenter Major Trauma Transfusion (PROMMTT) study.. J Trauma Acute Care Surg. (2013). 75 S1–2 [Article] [PubMed]
Holcomb, JB, Tilley, BC, Baraniuk, S, Fox, EE, Wade, CE, Podbielski, JM, del Junco, DJ, Brasel, KJ, Bulger, EM, Callcut, RA, Cohen, MJ, Cotton, BA, Fabian, TC, Inaba, K, Kerby, JD, Muskat, P, O’Keeffe, T, Rizoli, S, Robinson, BR, Scalea, TM, Schreiber, MA, Stein, DM, Weinberg, JA, Callum, JL, Hess, JR, Matijevic, N, Miller, CN, Pittet, JF, Hoyt, DB, Pearson, GD, Leroux, B, van Belle, G 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). 313 471–82 [Article] [PubMed]
Boin, IF, Leonardi, MI, Luzo, AC, Cardoso, AR, Caruy, CA, Leonardi, LS Intraoperative massive transfusion decreases survival after liver transplantation.. Transplant Proc. (2008). 40 789–91 [Article] [PubMed]
Diaz-Martin, A, Escoresca-Ortega, AM, Hernandez-Caballero, C, Pena, M, Adsuar-Gómez, A, Herruzo-Aviles, A, Hinojosa-Perez, R Considerations regarding major bleeding after cardiac transplantation.. Transplant Proc. (2010). 42 3204–5 [Article] [PubMed]
Turan, A, Yang, D, Bonilla, A, Shiba, A, Sessler, DI, Saager, L, Kurz, A Morbidity and mortality after massive transfusion in patients undergoing non-cardiac surgery.. Can J Anaesth. (2013). 60 761–70 [Article] [PubMed]
Koch, CG, Li, L, Duncan, AI, Mihaljevic, T, Cosgrove, DM, Loop, FD, Starr, NJ, Blackstone, EH Morbidity and mortality risk associated with red blood cell and blood-component transfusion in isolated coronary artery bypass grafting.. Crit Care Med. (2006). 34 1608–16 [Article] [PubMed]
Rao, SV, Jollis, JG, Harrington, RA, Granger, CB, Newby, LK, Armstrong, PW, Moliterno, DJ, Lindblad, L, Pieper, K, Topol, EJ, Stamler, JS, Califf, RM Relationship of blood transfusion and clinical outcomes in patients with acute coronary syndromes.. JAMA. (2004). 292 1555–62 [Article] [PubMed]
Yang, JC, Sun, Y, Xu, CX, Dang, QL, Li, L, Xu, YG, Song, YJ, Yan, H Correlation between red blood cell transfusion volume and mortality in patients with massive blood transfusion: A large multicenter retrospective study.. Exp Ther Med. (2015). 9 137–42 [PubMed]
Collins, JA Problems associated with the massive transfusion of stored blood.. Surgery. (1974). 75 274–95 [PubMed]
Neal, MD, Marsh, A, Marino, R, Kautza, B, Raval, JS, Forsythe, RM, Marshall, GT, Sperry, JL Massive transfusion: An evidence-based review of recent developments.. Arch Surg. (2012). 147 563–71 [Article] [PubMed]
Frank, SM, Savage, WJ, Rothschild, JA, Rivers, RJ, Ness, PM, Paul, SL, Ulatowski, JA Variability in blood and blood component utilization as assessed by an anesthesia information management system.. Anesthesiology. (2012). 117 99–106 [Article] [PubMed]
Frank, SM, Wick, EC, Dezern, AE, Ness, PM, Wasey, JO, Pippa, AC, Dackiw, E, Resar, LM Risk-adjusted clinical outcomes in patients enrolled in a bloodless program.. Transfusion. (2014). 5410 Pt 2 2668–77 [Article] [PubMed]
Como, JJ, Dutton, RP, Scalea, TM, Edelman, BB, Hess, JR Blood transfusion rates in the care of acute trauma.. Transfusion. (2004). 44 809–13 [Article] [PubMed]
Popovsky, MA Pulmonary consequences of transfusion: TRALI and TACO.. Transfus Apher Sci. (2006). 34 243–4 [Article] [PubMed]
Clifford, L, Jia, Q, Yadav, H, Subramanian, A, Wilson, GA, Murphy, SP, Pathak, J, Schroeder, DR, Ereth, MH, Kor, DJ Characterizing the epidemiology of perioperative transfusion-associated circulatory overload.. Anesthesiology. (2015). 122 21–8 [Article] [PubMed]
Vamvakas, EC Possible mechanisms of allogeneic blood transfusion-associated postoperative infection.. Transfus Med Rev. (2002). 16 144–60 [Article] [PubMed]
Taylor, RW, O’Brien, J, Trottier, SJ, Manganaro, L, Cytron, M, Lesko, MF, Arnzen, K, Cappadoro, C, Fu, M, Plisco, MS, Sadaka, FG, Veremakis, C Red blood cell transfusions and nosocomial infections in critically ill patients.. Crit Care Med. (2006). 34 2302–8; quiz 2309 [Article] [PubMed]
Rohde, JM, Dimcheff, DE, Blumberg, N, Saint, S, Langa, KM, Kuhn, L, Hickner, A, Rogers, MA Health care-associated infection after red blood cell transfusion: A systematic review and meta-analysis.. JAMA. (2014). 311 1317–26 [Article] [PubMed]
Heal, JM, Phipps, RP, Blumberg, N One big unhappy family: Transfusion alloimmunization, thrombosis, and immune modulation/inflammation.. Transfusion. (2009). 49 1032–6 [Article] [PubMed]
Tan, TW, Farber, A, Hamburg, NM, Eberhardt, RT, Rybin, D, Doros, G, Eldrup-Jorgensen, J, Goodney, PP, Cronenwett, JL, Kalish, JA Vascular Study Group of New England, Blood transfusion for lower extremity bypass is associated with increased wound infection and graft thrombosis.. J Am Coll Surg. (2013). 216 1005–1014.e2; quiz 1031 [Article] [PubMed]
Kumar, MA, Boland, TA, Baiou, M, Moussouttas, M, Herman, JH, Bell, RD, Rosenwasser, RH, Kasner, SE, Dechant, VE Red blood cell transfusion increases the risk of thrombotic events in patients with subarachnoid hemorrhage.. Neurocrit Care. (2014). 20 84–90 [Article] [PubMed]
ICD-9 Codes Used to Define Morbid Outcomes
Fig. 1.
In-hospital morbidity and mortality rates according to the number of erythrocyte units transfused. In-hospital morbidity (a composite of all five morbid events shown in fig. 2) increased with erythrocyte dose in a curvilinear manner, reaching a 50% rate of morbidity at 10 or greater erythrocyte units. The slope was steepest up to 30 erythrocyte units, with an inflection point and plateau at higher doses. The formula defining the curve is y = 36.5 ln(x) + 10.4 (R2 = 0.962). Mortality increased in a linear manner with a slope close to 10, indicating that for each 10-erythrocyte unit increment, mortality increased approximately 10%. After transfusion of 50 units, mortality exceeded 50%. The formula defining the curve is y = 9.47 (x) − 10.56 (R2 = 0.99). RBC = erythrocyte.
In-hospital morbidity and mortality rates according to the number of erythrocyte units transfused. In-hospital morbidity (a composite of all five morbid events shown in fig. 2) increased with erythrocyte dose in a curvilinear manner, reaching a 50% rate of morbidity at 10 or greater erythrocyte units. The slope was steepest up to 30 erythrocyte units, with an inflection point and plateau at higher doses. The formula defining the curve is y = 36.5 ln(x) + 10.4 (R2 = 0.962). Mortality increased in a linear manner with a slope close to 10, indicating that for each 10-erythrocyte unit increment, mortality increased approximately 10%. After transfusion of 50 units, mortality exceeded 50%. The formula defining the curve is y = 9.47 (x) − 10.56 (R2 = 0.99). RBC = erythrocyte.
Fig. 1.
In-hospital morbidity and mortality rates according to the number of erythrocyte units transfused. In-hospital morbidity (a composite of all five morbid events shown in fig. 2) increased with erythrocyte dose in a curvilinear manner, reaching a 50% rate of morbidity at 10 or greater erythrocyte units. The slope was steepest up to 30 erythrocyte units, with an inflection point and plateau at higher doses. The formula defining the curve is y = 36.5 ln(x) + 10.4 (R2 = 0.962). Mortality increased in a linear manner with a slope close to 10, indicating that for each 10-erythrocyte unit increment, mortality increased approximately 10%. After transfusion of 50 units, mortality exceeded 50%. The formula defining the curve is y = 9.47 (x) − 10.56 (R2 = 0.99). RBC = erythrocyte.
×
Fig. 2.
Event rates for five morbid outcomes are plotted according to the number of erythrocyte units transfused. In high-dose transfused patients, hospital-acquired infections and thrombotic events were four to five times more prevalent than renal, respiratory, or ischemic events. The incidence of infection increased with erythrocyte dose up to 40% and then plateaued. Thrombotic events increased up to a rate of 50% before reaching a plateau. Renal, respiratory, and ischemic event rates increased gradually up to rates of 5 to 10% at an erythrocyte dose of 20 units. RBC = erythrocyte.
Event rates for five morbid outcomes are plotted according to the number of erythrocyte units transfused. In high-dose transfused patients, hospital-acquired infections and thrombotic events were four to five times more prevalent than renal, respiratory, or ischemic events. The incidence of infection increased with erythrocyte dose up to 40% and then plateaued. Thrombotic events increased up to a rate of 50% before reaching a plateau. Renal, respiratory, and ischemic event rates increased gradually up to rates of 5 to 10% at an erythrocyte dose of 20 units. RBC = erythrocyte.
Fig. 2.
Event rates for five morbid outcomes are plotted according to the number of erythrocyte units transfused. In high-dose transfused patients, hospital-acquired infections and thrombotic events were four to five times more prevalent than renal, respiratory, or ischemic events. The incidence of infection increased with erythrocyte dose up to 40% and then plateaued. Thrombotic events increased up to a rate of 50% before reaching a plateau. Renal, respiratory, and ischemic event rates increased gradually up to rates of 5 to 10% at an erythrocyte dose of 20 units. RBC = erythrocyte.
×
Table 1.
Patient Characteristics for Groups Defined by Units of Erythrocytes Transfused
Patient Characteristics for Groups Defined by Units of Erythrocytes Transfused×
Patient Characteristics for Groups Defined by Units of Erythrocytes Transfused
Table 1.
Patient Characteristics for Groups Defined by Units of Erythrocytes Transfused
Patient Characteristics for Groups Defined by Units of Erythrocytes Transfused×
×
Table 2.
Predictors of Morbidity* in High-dose Transfused Patients†
Predictors of Morbidity* in High-dose Transfused Patients†×
Predictors of Morbidity* in High-dose Transfused Patients†
Table 2.
Predictors of Morbidity* in High-dose Transfused Patients†
Predictors of Morbidity* in High-dose Transfused Patients†×
×
Table 3.
Predictors of Mortality* in High-dose Transfused Patients†
Predictors of Mortality* in High-dose Transfused Patients†×
Predictors of Mortality* in High-dose Transfused Patients†
Table 3.
Predictors of Mortality* in High-dose Transfused Patients†
Predictors of Mortality* in High-dose Transfused Patients†×
×