Secondary Logo

Journal Logo

Original Research Articles: Original Clinical Research Report

Transfusion of Red Blood Cells, Fresh Frozen Plasma, or Platelets Is Associated With Mortality and Infection After Cardiac Surgery in a Dose-Dependent Manner

Ming, Yue MMed*; Liu, Jing MMed; Zhang, Fengjiang MD, PhD*; Chen, Changwei MMed; Zhou, Li MD, PhD; Du, Lei MD, PhD; Yan, Min MD, PhD*

Author Information
doi: 10.1213/ANE.0000000000004528

Abstract

KEY POINTS

  • Question: What is the effect of transfusion of different blood products on mortality and morbidity in cardiac surgery?
  • Findings: Transfusion of red blood cells, fresh frozen plasma, or platelets was associated with an increased risk of mortality in a dose-dependent manner and transfusion of red blood cells or fresh frozen plasma or platelets (exceeding 2 units) was also associated with an increased risk of infection after cardiac surgery.
  • Meaning: This study suggests that the risk of mortality and infection after cardiac surgery is independently associated with the transfusion of blood products, so that reducing red blood cells, fresh frozen plasma, and platelets transfusions in cardiac surgery may be associated with improving outcomes.

Patients requiring a high volume of perioperative allogeneic red blood cells during cardiovascular surgery are known to be at increased risk of postoperative mortality and morbidity,1–6 especially infection.4,5,7,8 This has led to substantial efforts to reduce the need for red blood cell transfusion during surgery.9 In contrast, little is known about whether transfusions of platelets and/or fresh frozen plasma are associated with perioperative mortality and morbidity.10–13 This is primarily because platelets and fresh frozen plasma are almost always transfused with red blood cells,6 which confounds any attempt to determine the effects of individual blood products. Multiple studies have found that transfusion of the combination of fresh frozen plasma and red blood cells leads to significantly higher mortality and morbidity than transfusion of red blood cells on their own.14,15

The objective of this study was to determine the effects of transfusion of platelets and/or fresh frozen plasma on mortality and infection by adjusting the odds ratio to control for red blood cell transfusion and by using propensity score matching to control for confounding factors. Patients who underwent valve surgery and/or coronary artery bypass grafting from 2 hospitals were included to obtain a sufficiently large sample size for statistical analysis.

METHODS

Study Design

This trial was designed as a retrospective study to compare the effects of transfusions of red blood cells, fresh frozen plasma, and/or platelets on outcomes of elective cardiac surgery. The study was approved by the Ethics Committees of both West China Hospital in Chengdu, China (256/2017) and the Second Affiliated Hospital in Hangzhou, China (096/2017). The requirement for individual informed consent was waived by the ethics committees because of the retrospective study design. This manuscript adheres to the applicable Strengthening The Reporting of Observational Studies in Epidemiology (STROBE) guidelines (http://www.equator-network.org/).

We included data from patients over 18 years old undergoing valve surgery and/or coronary artery bypass grafting with cardiopulmonary bypass (CPB) from January 1, 2011 to June 30, 2017 at the West China Hospital, and from September 1, 2013 to June 30, 2017 at the Second Affiliated Hospital. Patients were excluded if they (1) required total aortic arch replacement surgery, (2) underwent cardiac tumor resection or emergency surgery, (3) died on the operating table, or (4) had inadequate medical records for analysis. Patient data were extracted from the Hospital Information System, Transfusion System, and the Laboratory Information System at West China Hospital, and from the “Do Care” System and Electronic Medical Record System at the Second Affiliated Hospital.

Anesthesia and CPB

The 2 participating hospitals generally followed similar strategies for anesthesia management and similar CPB protocols.16,17 In brief, anesthesia was induced with midazolam, sufentanil, and muscle relaxant, and maintained by infusion of remifentanil, inhalation of sevoflurane, intermittent muscle relaxant, and sufentanil. CPB was composed with a roll pump, a membrane oxygenator, a filter, and connecting tubers. CPB was primed with a colloid solution of 1000 mL and crystalline liquid of 500 mL. During CPB, blood flow was at 2.0–2.4 L/m2/min to maintain mean arterial pressure at 50–80 mm Hg. Cold 4:1 blood cardioplegia was used for heart arrest. Patients were maintained at a nasopharyngeal temperature of 30°C–33°C with moderate hemodilution. All patients received an initial heparin dose of 375 U/kg to achieve systemic anticoagulation, with additional heparin intermittently injected to maintain an activated clotting time of 480 seconds or more. After weaning from CPB, heparin was neutralized with protamine at a dose of 375 U/kg.

Cell Saver 5+ autologous blood recovery system (Haemonetics Corporation, Braintree, MA) was used for all patients undergoing CPB. Briefly, pericardial blood was salvaged by suction and returned to the CPB circuit. After surgery, the residual blood was collected into a bag containing sodium citrate, neutralized with protamine, and transfused into the patient as previously described.17,18

Blood Transfusion

Hemoglobin concentration was measured during and after surgery using a Cobas b 123 device (Roche, Basel, Switzerland) at West China Hospital, or a Cobas b 221 device at the Second Affiliated Hospital. Patients were transfused with red blood cells when their blood hemoglobin concentration fell below a threshold of 7 g/dL during CPB, 8 g/dL during surgery, or 9 g/dL in the intensive care unit. This last threshold was based on a report that a threshold of 7.5 g/dL was associated with 1.64-fold greater risk of death than 9.0 g/dL among patients in the United Kingdom undergoing nonemergency cardiac surgery.19

Patients were transfused with fresh frozen plasma (10–15 mL/kg) if the international normalized ratio was >1.4 or activated partial prothrombin time was >50 seconds. Patients were considered to transfuse with platelets if the platelet count fell below 50 × 109/L, as measured in the emergency laboratory.

Primary Outcomes

All-cause mortality was defined as patient deaths occurring between arrival at the intensive care unit and hospital discharge.20 Deaths were attributed to cardiogenic causes (including cardiogenic shock, refractory ventricular fibrillation, sudden cardiac arrest, acute myocardial infarction, heart failure, or cardiovascular hemorrhage21), respiratory causes (respiratory failure or pulmonary infection8), multiple-organ dysfunction syndrome (dysfunction in at least 2 organs according to the Sequential Organ Failure Assessment score22), infection,19 or gastrointestinal causes (gastrointestinal perforation, hemorrhage, or peritonitis19). Cardiogenic shock was defined as insufficient end-organ perfusion due to persistent hypotension arising from systolic blood pressure <80–90 mm Hg or mean arterial pressure 30 mm Hg lower than baseline as well as severe reduction in cardiac index (<1.8 L/min/m2 without support or <2.2 L/min/m2 with support).

Secondary Outcomes

Infection events included pneumonia, catheter-related bloodstream infection, wound infection, and sepsis. Patients were diagnosed with pneumonia if they had a positive sputum culture23 or met the following 4 criteria24: new or progressive pulmonary infiltrate, temperature >38°C, white blood cell count >12,000/mm3, and airway purulent sputum or secretion. Catheter-related bloodstream infection was diagnosed if central venous catheter tip colonization was detected and if the patient had at least 1 positive culture from peripheral blood.25 Wound infection was defined as an ASEPSIS (Additional treatment, Serous discharge, Erythema, Purulent exudates, Separation of the deep tissues, Isolation of bacteria, and duration of inpatient Stay) score >20.26 Patients were diagnosed with sepsis if antibiotics were required for suspected infection and if systemic inflammatory response syndrome was diagnosed during the preceding 24 hours. Systemic inflammatory response syndrome was defined as body temperature >38°C or <36°C, heart rate >90 beats/minute, breathing rate >20 breaths/min or partial pressure of carbon dioxide (Paco2) > 32 mm Hg, and white blood cell count >12,000/mm3 or < 4000/mm3.19,20

Statistical Analysis

All data were analyzed using SPSS 20.0 (IBM, Chicago, IL). Categorical variables were expressed as frequency (percentage), and intergroup differences were assessed for significance using the χ2 or Fisher exact tests. Data for continuous variables were expressed as mean ± standard deviation (SD) and tested for normality using the Shapiro-Wilk test. Intergroup differences on variables with no evidence against normality were assessed for significance using 1-way analysis of variance (ANOVA). Skewed data were expressed as median (interquartile range), and intergroup differences were assessed using the Wilcoxon rank-sum test. A P value <0.05 was considered statistically significant.

Risk factors of postoperative mortality and infection were identified using univariate logistic regression. Variables identified as P < .1 were included in the multivariable logistic regression model. The model was defined by backward selection using P < .05 as the criterion for variable retention and adjusted for analyzing the effect of blood products on outcomes.1,2,5,8

To investigate the effect of transfusing each blood product, we performed propensity score matching using a multivariable logistic regression model in which the independent variables were possible perioperative potential risk factors as identified through searches of EMBASE, PubMed, and Cochrane databases with the terms “risk factor,” “predictor,” “mortality,” “infection,” and “cardiac surgery.” We also included risk factors identified in the present study population.27 These factors included demographic characteristics (gender, ethnicity, age, body mass index, history of smoking, alcohol consumption), medical history (cardiac functional classification of New York Heart Association, hypertension, coronary heart disease, active endocarditis, history of heart failure, ejection fraction, pulmonary hypertension, stroke, diabetes mellitus, renal dysfunction), medications (antiplatelet drugs, anticoagulants, diuretics), preoperative examinations (hemoglobin and platelet count, abnormal international normalized ratio, fibrinogen, C-reactive protein), and surgical characteristics (CPB duration, repeat surgery, operative procedure). Patients with similar characteristics were matched in a 1:1 ratio using the greedy nearest-neighbor matching algorithm with a caliper width of 0.2. When we analyzed 1 product, the other 2 products were matched to minimize their effect on outcomes. Detailed information on matching is shown in Figure 1.

Figure 1.
Figure 1.:
Patient enrollment and blood product transfusions. Some patients required transfusion of more than 1 type of blood product. *For each blood product, we performed propensity score matching 4 times to minimize the effect of the other 2 blood products. For red blood cell transfusion, for example, the following pairs of patient subgroups were matched: “Only RBC versus nontransfusion,” “RBC + FFP versus only FFP,” “RBC + PLT versus only PLT,” and “RBC + FFP + PLT versus FFP + PLT.” Similarly, FFP and PLT were matched. FFP indicates fresh frozen plasma; PLT, platelets; RBC, red blood cells.

Using these propensity score-matched pairs of patients, we obtained 3 datasets comparing patients who received red blood cell transfusions or no red blood cell transfusions, fresh frozen plasma transfusions or no fresh frozen plasma transfusions, or platelets transfusions or no platelets transfusions. To ensure the patients were well-matched, we tested baseline characteristics between the patients within each dataset before and after matching. Characteristics found to be significantly different after matching were adjusted for as confounding variables. We then performed univariate and multivariable logistic regression analysis on the matched datasets to identify the risk of adverse outcomes associated with transfusing each type of blood product. To test the combined effect of all perioperative blood products on adverse outcomes, we performed multivariable logistic regression in which transfusions with red blood cells and/or fresh frozen plasma and/or platelets were combined into a unified variable (the total units of all 3 blood products).

Using G*Power 3.1.9.228 and assuming that the odds ratio for mortality among patients receiving blood transfusion was 1.77 relative to patients receiving no blood products,5 we would require a minimum sample of 5316 patients that received each type of blood product, to achieve a power of 0.95 and a type I error of 0.05.

RESULTS

Table 1. - Demographic and Perioperative Patient Characteristics
Characteristic Total
(n = 8238)
No Transfusion
(n = 3301)
Any TransfusionTable 1.
(n = 4937)
Demographic characteristics
 Male, n (%) 3487 (42.3) 1601 (48.5) 1886 (38.2)
 Ethnicity, Han Chinese, n (%) 8039 (97.6) 3222 (97.6) 4817 (97.6)
 Age (y) 52 (45.61) 49 (43.57) 55 (46.63)
 Body mass index (kg/m2) 22.90 ± 3.04 23.30 ± 3.02 22.62 ± 3.03
 History of smoking, n (%) 2301 (27.9) 1054 (31.9) 1247 (25.3)
 Alcohol consumption, n (%) 1525 (18.5) 718 (21.8) 807 (16.3)
Medical history
 NYHA class ≥ III, n (%) 6076 (73.8) 2362 (71.6) 3714 (75.2)
 Hypertension, n (%) 1344 (16.3) 413 (12.5) 931 (18.9)
 Coronary heart disease, n (%) 1065 (12.9) 192 (5.8) 873 (17.7)
 Active endocarditis, n (%) 196 (2.4) 48 (1.5) 148 (3.0)
 History of heart failure, n (%) 1523 (18.5) 586 (17.8) 937 (19.0)
 Ejection fraction (%) 61 ± 9 61 ± 9 61 ± 9
Pulmonary hypertension, n (%)
  Mild (26–35 mm Hg) 427 (5.2) 182 (5.8) 245 (5.2)
  Moderate (36–45 mm Hg) 710 (8.6) 301 (9.6) 409 (8.7)
  Severe (>45 mm Hg) 327 (4.0) 114 (3.6) 213 (4.5)
 Stroke, n (%) 355 (4.3) 121 (3.7) 234 (4.7)
 Diabetes mellitus, n (%) 531 (6.4) 119 (3.6) 412 (8.3)
 Renal dysfunction, n (%) 367 (4.5) 91 (2.8) 276 (5.6)
Medications
 Antiplatelet drugs, n (%) 328 (4.0) 71 (2.2) 257 (5.2)
 Anticoagulants, n (%) 1324 (16.1) 434 (13.1) 890 (18.0)
 Diuretics, n (%) 5727 (69.5) 2359 (71.5) 3368 (68.2)
Laboratory analyses
 Hemoglobin (g/dL) 12 (13.6, 14.6) 14 (13, 15) 13 (12, 14)
 Platelet count (109/L) 153 (116, 194) 158 (124, 196) 149 (110, 192)
 Abnormal INR, n (%) 1432 (17.4) 513 (15.7) 919 (18.9)
 Fibrinogen (mg/dL) 2.81 (2.37, 3.31) 2.71 (2.33, 3.19) 2.88 (2.42, 3.40)
 C-reactive protein (mg/L) 6.13 ± 13.54 5.37 ± 10.92 6.64 ± 15.02
Surgical characteristics
 CPB duration (min) 114 (89, 142) 106 (84, 132) 120 (94, 148)
 Repeat surgery, n (%) 98 (1.2) 15 (0.5) 83 (1.7)
 Operative procedure, n (%)
  Valve surgery 7287 (88.4) 3163 (95.8) 4124 (83.5)
  Coronary artery bypass grafting 800 (9.7) 106 (3.2) 694 (14.1)
  Combined surgery 151 (1.8) 32 (1.0) 119 (2.4)
Unless otherwise noted, values are mean ± standard deviation or median (interquartile range).
Abbreviations: CBP, cardiopulmonary bypass; INR, international normalized ratio; NYHA, New York Heart Association.
aTransfused with at least 1 type of blood product (red blood cells, fresh frozen plasma, or platelets).

A total of 8521 patients were screened, of whom 283 patients were excluded because they required ascending aortic replacement surgery (n = 131), cardiac tumor resection (n = 91), or emergency surgery (n = 10); because they died on the operating table (n = 19); or because their medical records were incomplete (n = 32). This last group included 15 with unclear transfusion data, 11 without baseline height or weight data, and 6 with insufficient postoperative information to determine infection events. In the end, 8238 patients were included in the final analysis (Figure 1). Of these, 7287 (88.4%) underwent valve surgery, 800 (9.7%) underwent coronary artery bypass grafting, and 151 (1.8%) underwent both procedures. Patient demographic and perioperative characteristics are shown in Table 1.

Blood Product Requirements

Among all patients, 4937 (59.9%) required transfusion of at least 1 type of blood product. Of these, 1848 (37.4%) received only red blood cells, 618 (12.5%) only fresh frozen plasma, and 319 (6.5%) only platelets. Another 1778 patients (36.0%) received 2 types of blood products, while 374 (7.6%) received all 3 types of blood products. In total, 3916 patients required red blood cells, 2466 required fresh frozen plasma, and 1081 required platelets.

Patients requiring red blood cells in this study consumed 16,414 units, corresponding to a mean of 1.99 units per person. Patients requiring fresh frozen plasma consumed a total of 8024.5 units (mean, 0.97 units per person), and those requiring platelets consumed 2560.5 units (mean, 0.31 units per person).

Outcomes and Risk Factors

Among all patients, 109 (1.3%) died, and 62 deaths (0.75%) were attributed to respiratory or circulatory failure, 27 (0.3%) to multiple-organ dysfunction syndrome, 4 (0.05%) to severe infection, and 16 (0.2%) to other causes (Table 2).

Table 2. - Patient Outcomes
Outcome Total
(n = 8238)
No Transfusion
(n = 3301)
Any TransfusionTable 2.
(n = 4937)
Mortality, n (%) 109 (1.3) 6 (0.18) 103 (2.0)
 Respiratory or circulatory failure 62 (0.75) 3 (0.09) 59 (1.2)
 Multiple-organ dysfunction syndrome 27 (0.3) 1 (0.03) 26 (0.5)
 Severe infection 4 (0.05) 0 (0) 4 (0.1)
 Other causesTable 2. 16 (0.2) 2 (0.06) 14 (0.2)
Infection, n (%) 812 (9.9) 157 (4.8) 655 (13.3)
 Pulmonary infection 751 (9.1) 136 (4.1) 615 (12.5)
 Catheter-related bloodstream infection 13 (0.2) 3 (0.1) 10 (0.2)
 Wound infection 25 (0.3) 4 (0.1) 21 (0.4)
 Sepsis 132 (1.6) 25 (0.8) 107 (2.2)
aTransfused with at least 1 type of blood product (red blood cells, fresh frozen plasma, or platelets).
bOther causes: 1 died of gastrointestinal hemorrhage; 15 died from unknown causes.

Patients who died were more likely to be older (61 vs 52 years; P < .01) and to have lower body mass index (22.11 ± 2.95 vs 22.91 ± 3.04; P = .007). Mortality was also significantly higher in patients who received transfusions than in those who did not, regardless of whether the transfusions were red blood cells (2.5% vs 0.2%; P < .01), fresh frozen plasma (3.3% vs 0.5%; P < .01), or platelets (4.7% vs 0.8%; P < .01). Multivariable logistic regression analysis identified the following independent risk factors of mortality: advanced age, decreased body mass index, lower preoperative hemoglobin, higher C-reactive protein, heart failure, pulmonary hypertension, diabetes mellitus, renal dysfunction and hepatic insufficiency, and longer CPB duration.

Infection occurred in 812 patients (9.9%) after surgery, including pulmonary infection (n = 751), catheter-related bloodstream infection (n = 13), wound infection (n = 25), and sepsis (n = 132; Table 2). Univariate and multivariable logistic regression analysis identified the following independent risk factors of postoperative infection: advanced age, high cardiac functional class, pulmonary hypertension, stroke, renal dysfunction, low preoperative hemoglobin, abnormal international normalized ratio, lower ejection fraction, preoperative medications (antiplatelet drugs, anticoagulants, digitoxin, diuretics), longer CPB duration, repeat surgery, and combination surgery.

Influence of Blood Product Transfusions on Outcomes

Transfusion of red blood cells and/or fresh frozen plasma and/or platelets was strongly associated with increased mortality and infection. In multivariable logistic regression, mortality risk was higher per transfused unit of red blood cells (odds ratio 1.18, 95% confidence interval [CI], 1.14–1.22), fresh frozen plasma (odds ratio 1.24, 95% CI, 1.18–1.30), or platelets (odds ratio 1.12, 95% CI, 1.07–1.18). Similarly, infection risk was higher per transfused unit of red blood cells (odds ratio 1.18, 95% CI, 1.15–1.21) or fresh frozen plasma (odds ratio 1.18, 95% CI, 1.14–1.22) (all P < .01, Supplemental Digital Content, Table 1, http://links.lww.com/AA/C960).

We performed propensity score matching to adjust for baseline differences before assessing the risk of poor outcomes associated with transfusing each type of blood product. We identified 2165 matched patient pairs to investigate the effect of red blood cell transfusion (Supplemental Digital Content, Table 2, http://links.lww.com/AA/C960). Matched patients receiving red blood cell transfusion showed higher incidence of mortality (1.7% vs 0.3%) and infection (11.8% vs 5.5%) than those who did not undergo red blood cell transfusion (P < .01). In other words, red blood cell transfusion was associated with a 6.26-fold increase in the risk of mortality and 2.28-fold increase in the risk of infection. We further adjusted for characteristics that were not well-matched between groups (age, cardiac functional class, preoperative hemoglobin, preoperative platelet count, preoperative fibrinogen) and found that red blood cell transfusion was still associated with a 5.83-fold increase in the risk of mortality and 2.31-fold increase in the risk of infection (Figure 2).

Figure 2.
Figure 2.:
Risk of all-cause mortality or infection in patients receiving (or not) transfusions of red blood cells, fresh frozen plasma, or platelets. Logistic regression was performed after propensity score matching and adjusted by poorly matched variables. #Adjusted by age, NYHA class, preoperative hemoglobin, preoperative platelet count, and preoperative fibrinogen. *Adjusted by ethnicity and preoperative fibrinogen. $Adjusted by body mass index, NYHA class, active endocarditis, antiplatelet drugs, diuretics, preoperative platelet count, and operative procedure. FFP indicates fresh frozen plasma; HL, higher limit; LL, lower limit; NYHA, New York Heart Association; OR, odds ratio; PLT, platelets; RBC, red blood cells.

We identified 2233 matched patient pairs to investigate the effects of fresh frozen plasma transfusion (Supplemental Digital Content, Table 3, http://links.lww.com/AA/C960). Patients receiving fresh frozen plasma showed higher incidence of mortality (2.9% vs 0.9%) and infection (15.0% vs 10.3%) than those who did not receive fresh frozen plasma (P < .01). Multivariable logistic regression showed that fresh frozen plasma was associated with a 3.5-fold increase in the risk of mortality and 1.57-fold increase in the risk of infection (Figure 2).

We identified 1043 matched patient pairs with platelets transfusions (Supplemental Digital Content, Table 4, http://links.lww.com/AA/C960). Mortality was higher in matched patients receiving platelets than in those who did not receive platelets (4.5% vs 1.6%; P < .01), but the rate of infection events was similar (16.3% vs 17.6%; P = .414). Multivariable regression showed that platelets transfusion was associated with a 2.78-fold increase in risk of mortality (95% CI, 1.56–4.95; P < .01), but not with an increase in risk of infection (odds ratio 1.00, 95% CI, 0.79–1.28; P = .993; Figure 2).

Combined Blood Product Transfusion and Outcomes

Figure 3.
Figure 3.:
Incidence of all-cause mortality and infection in patients receiving increasing units of red blood cells, fresh frozen plasma, platelets, or any blood product. FFP indicates fresh frozen plasma; PLT, platelets; RBC, red blood cells.
Figure 4.
Figure 4.:
Association of transfusion of total blood products with outcomes. *Adjusted by age, gender, body mass index, hypertension, diabetes mellitus, pulmonary hypertension, renal dysfunction, ejection fraction, preoperative hemoglobin, CPB duration, operative procedure. CPB indicates cardiopulmonary bypass; HL, higher limit; LL, lower limit; OR, odds ratio.

Risk of mortality or infection was associated in a stepwise fashion with the units of red blood cells, fresh frozen plasma, or platelets (Figure 3A–C, E–G). Next, we investigated the combined effect of transfusing red blood cells, fresh frozen plasma, and/or platelets on mortality and infection. Patients receiving transfusion of any blood product showed a higher rate of mortality than those not receiving any transfusion (2.0% vs 0.18%), as well as a higher rate of infection (13.3% vs 4.8%; both P < .01). The incidence of adverse outcomes was associated in a stepwise fashion with increasing transfusion (Figure 3D, H). Multivariable regression showed that the risk of mortality and infection was associated with the units of combined transfusion (Figure 4).

DISCUSSION

In this study, we found that red blood cells and fresh frozen plasma transfusion were independent risk factors of both mortality and infection in patients undergoing cardiac surgery. Platelets transfusion was also an independent risk factor for mortality, and transfusion of >2 units of platelets was associated with infection. We further found that combined transfusion of all 3 blood products was associated with increased risk of both mortality and infection in a dose-dependent manner.

Similar to the results in the present studies, others have found that red blood cell transfusion is associated, in a dose-dependent manner, with mortality29,30 and infection, including pneumonia7 and deep sternal wound infection.31 Indeed, even 1 unit of red blood cells can worsen outcomes and lengthen hospital stay.3 The present study extends this work by examining the risk of poor outcomes due to transfusion with other blood products (fresh frozen plasma, platelets), which typically have not been studied on their own because they are often transfused together with red blood cells. We found that platelets transfusion was associated with mortality but not of infection, unless transfusion volume exceeding 2 units. Consistent with this result, 1 study found that patients receiving platelets transfusion showed similar morbidity as propensity score-matched patients not receiving platelets transfusion, although that analysis did not adjust for fresh frozen plasma use.11 Another study found that platelets transfusion of 1 unit, without other blood products, was not associated with the risk of postoperative infection.10 The fact that smaller volumes of platelets transfusion do not increase the risk of infection may reflect the ability of platelets to strengthen immune defenses against bacterial and viral infection.32 This begs the question of why, in the present study, platelets transfusion of >2 units was associated with a higher risk of postoperative infection, which should be examined further.

We found that the risk of mortality per unit of blood product was quite similar for red blood cells, fresh frozen plasma, or platelets, so we combined them into 1 unified variable. Multivariable logistic regression showed that the risk of mortality and infection was associated dose-dependently with the total of blood products used. The model may help clinicians to predict adverse events after cardiac surgery. The findings of this study also highlight the importance of using blood conservation strategies such as transfusion triggers to avoid unnecessary transfusion33 of any blood product, whether red blood cells, fresh frozen plasma, or platelets. In addition, the results argue against transfusing more than 2 units of platelets. These findings agree with previous studies that reported that any quantity of blood transfusion decreases survival after cardiac surgery, with an especially pronounced risk of mortality for patients receiving 3 or more units of blood.5,34

Several well-designed randomized controlled studies on red blood cell transfusion triggers have given inconsistent results.35,36 We are unaware of multicenter studies to determine the effects of fresh frozen plasma and platelets triggers on outcomes, and strategies for reducing fresh frozen plasma and platelets transfusions are lacking. In this sense, the present study fills an important gap in the research literature by showing the importance of avoiding excess transfusion of any blood product.

The present study has several limitations due to its retrospective nature. One limitation is variability in patient baseline and surgical characteristics. We attempted to maximize the similarity between patient groups by propensity score matching, but some variables (such as age, preoperative platelet count, and preoperative fibrinogen) remained significantly different between groups. We adjusted the odds ratio based on these poorly matched variables, but the possibility of bias remains. Second, although we collected as much as information as possible on confounding factors affecting mortality and infection, there may be additional unknown variables not included in the study. Another limitation is that the international normalized ratio and fibrinogen levels or activity to determine the plasma transfusion could not be acquired at bedside during surgery, which may explain the higher rate of plasma transfusion here than in another study.35 Because the present study has not been externally validated, its generalizability to other patient populations and health care contexts needs to be confirmed. In particular, transfusion triggers can vary widely across medical centers,36–39 so the robustness of our results should be confirmed with different triggers. On the other hand, the transfusion rates in the present study are similar to those reported in large studies conducted in the United States,5,40 suggesting that the clinical situation at our 2 study sites may broadly represent the situation elsewhere.

In summary, we found that transfusion of red blood cells, fresh frozen plasma, or platelets was associated with mortality in a dose-dependent manner. Transfusion of red blood cells or fresh frozen plasma or platelets (exceeding 2 units) was also associated with an increased risk of infection after cardiac surgery. These results may provide a better model to predict prognosis after cardiac surgery, and they highlight the urgency of using strategies to reduce the use of not only red blood cells but also fresh frozen plasma and platelets in cardiac surgery.

DISCLOSURES

Name: Yue Ming, MMed.

Contribution: This author helped perform the study, complete the manuscript, and read and approved the final manuscript.

Name: Jing Liu, MMed.

Contribution: This author helped analyze the data and read and approved the final manuscript.

Name: Fengjiang Zhang, MD, PhD.

Contribution: This author helped check the data and read and approved the final manuscript.

Name: Changwei Chen, MMed.

Contribution: This author helped collect the data and read and approved the final manuscript.

Name: Li Zhou, MD, PhD.

Contribution: This author helped collect the data and read and approved the final manuscript.

Name: Lei Du, MD, PhD.

Contribution: This author helped review and modify the manuscript and read and approved the final manuscript.

Name: Min Yan, MD, PhD.

Contribution: This author helped design the study and read and approved the final manuscript.

This manuscript was handled by: Susan Goobie, MD, FRCPC.

FOOTNOTES

    REFERENCES

    1. Bhaskar B, Dulhunty J, Mullany DV, Fraser JF. Impact of blood product transfusion on short and long-term survival after cardiac surgery: more evidence. Ann Thorac Surg. 2012;94:460–467.
    2. Murphy GJ, Reeves BC, Rogers CA, Rizvi SI, Culliford L, Angelini GD. Increased mortality, postoperative morbidity, and cost after red blood cell transfusion in patients having cardiac surgery. Circulation. 2007;116:2544–2552.
    3. Crawford TC, Magruder JT, Fraser C, et al.; Investigators for the Maryland Cardiac Surgery Quality Initiative. Less is more: results of a statewide analysis of the impact of blood transfusion on coronary artery bypass grafting outcomes. Ann Thorac Surg. 2018;105:129–136.
    4. Horvath KA, Acker MA, Chang H, et al. Blood transfusion and infection after cardiac surgery. Ann Thorac Surg. 2013;95:2194–2201.
    5. 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.
    6. Surgenor SD, Kramer RS, Olmstead EM, et al.; Northern New England Cardiovascular Disease Study Group. The association of perioperative red blood cell transfusions and decreased long-term survival after cardiac surgery. Anesth Analg. 2009;108:1741–1746.
    7. Likosky DS, Paone G, Zhang M, et al. Red blood cell transfusions impact pneumonia rates after coronary artery bypass grafting. Ann Thorac Surg. 2015;100:794–800.
    8. Canet J, Gallart L, Gomar C, et al.; ARISCAT Group. Prediction of postoperative pulmonary complications in a population-based surgical cohort. Anesthesiology. 2010;113:1338–1350.
    9. Ferraris VA, Brown JR, Despotis GJ, et al.; Society of Thoracic Surgeons Blood Conservation Guideline Task F. 2011 update to the Society of Thoracic Surgeons and the Society of Cardiovascular Anesthesiologists blood conservation clinical practice guidelines. Ann Thorac Surg. 2011;91:944–982.
    10. van Hout FM, Hogervorst EK, Rosseel PM, et al. Does a platelet transfusion independently affect bleeding and adverse outcomes in cardiac surgery? Anesthesiology. 2017;126:441–449.
    11. McGrath T, Koch CG, Xu M, et al. Platelet transfusion in cardiac surgery does not confer increased risk for adverse morbid outcomes. Ann Thorac Surg. 2008;86:543–553.
    12. Ninkovic S, McQuilten Z, Gotmaker R, Newcomb AE, Cole-Sinclair MF. Platelet transfusion is not associated with increased mortality or morbidity in patients undergoing cardiac surgery. Transfusion. 2018;58:1218–1227.
    13. Wu B, Wang Y, Wang C, Cheng Y, Rong R. Intraoperative platelet transfusion is associated with increased postoperative sternal wound infections among type A aortic dissection patients after total arch replacement. Transfus Med. 2014;24:400–405.
    14. Ghazi L, Schwann TA, Engoren MC, Habib RH. Role of blood transfusion product type and amount in deep vein thrombosis after cardiac surgery. Thromb Res. 2015;136:1204–1210.
    15. Ahmed AB, Koster A, Lance M, Faraoni D; ESA VTE Guidelines Task Force. European guidelines on perioperative venous thromboembolism prophylaxis: cardiovascular and thoracic surgery. Eur J Anaesthesiol. 2018;35:84–89.
    16. Guo Y, Tang J, Du L, et al. Protamine dosage based on two titrations reduces blood loss after valve replacement surgery: a prospective, double-blinded, randomized study. Can J Cardiol. 2012;28:547–552.
    17. Zhou ZF, Jia XP, Sun K, et al. Mild volume acute normovolemic hemodilution is associated with lower intraoperative transfusion and postoperative pulmonary infection in patients undergoing cardiac surgery– a retrospective, propensity matching study. BMC Anesthesiology. 2017;17:1–9.
    18. Tan Z, Zhou L, Qin Z, et al. Low-dose sevoflurane may reduce blood loss and need for blood products after cardiac surgery: a prospective, randomized pilot study. Medicine (Baltimore). 2016;95:e3424.
    19. Casaer MP, Van den Berghe G. Liberal or restrictive transfusion after cardiac surgery. N Engl J Med. 2015;373:192.
    20. Hogervorst E, Rosseel P, van der Bom J, et al. Tolerance of intraoperative hemoglobin decrease during cardiac surgery. Transfusion. 2014;54:2696–2704.
    21. Hollenberg SM, Kavinsky CJ, Parrillo JE. Cardiogenic shock. Ann Intern Med. 1999;131:47–59.
    22. Bone RC, Sprung CL, Sibbald WJ. Definitions for sepsis and organ failure. Crit Care Med. 1992;20:724–726.
    23. Carson JL, Noveck H, Berlin JA, Gould SA. Mortality and morbidity in patients with very low postoperative Hb levels who decline blood transfusion. Transfusion. 2002;42:812–818.
    24. Hébert PC, Fergusson D, Blajchman MA, et al.; Leukoreduction Study Investigators. Clinical outcomes following institution of the Canadian universal leukoreduction program for red blood cell transfusions. JAMA. 2003;289:1941–1949.
    25. Wu M, Chen Y, Du B, Kang Y. Study protocol for a multicentre, randomised, controlled trial to assess the effectiveness of antimicrobial central venous catheters versus ordinary central venous catheters at reducing catheter related infections in critically ill Chinese patients. BMJ Open. 2017;7:e016564.
    26. Wilson AP, Treasure T, Sturridge MF, Grüneberg RN. A scoring method (ASEPSIS) for postoperative wound infections for use in clinical trials of antibiotic prophylaxis. Lancet. 1986;1:311–313.
    27. Huang D, Chen C, Ming Y, et al. Risk of massive blood product requirement in cardiac surgery: a large retrospective study from 2 heart centers. Medicine (Baltimore). 2019;98:e14219.
    28. Faul F, Erdfelder E, Buchner A, Lang AG. Statistical power analyses using G*Power 3.1: tests for correlation and regression analyses. Behav Res Methods. 2009;41:1149–1160.
    29. Delaney M, Stark PC, Suh M, et al. Massive transfusion in cardiac surgery: the impact of blood component ratios on clinical outcomes and survival. Anesth Analg. 2017;124:1777–1782.
    30. Sun Y, Jin ZK, Xu CX, et al. Investigation of the current situation of massive blood transfusion in different surgical departments a large multicenter study in China. Int J Clin Exp Med. 2015;8:9257–9265.
    31. Cutrell JB, Barros N, McBroom M, et al. Risk factors for deep sternal wound infection after cardiac surgery: influence of red blood cell transfusions and chronic infection. Am J Infect Control. 2016;44:1302–1309.
    32. Semple JW, Italiano JE Jr, Freedman J. Platelets and the immune continuum. Nat Rev Immunol. 2011;11:264–274.
    33. Schwann TA, Habib JR, Khalifeh JM, et al. Effects of blood transfusion on cause-specific late mortality after coronary artery bypass grafting-less is more. Ann Thorac Surg. 2016;102:465–473.
    34. Koch CG, Li L, Duncan AI, et al. Transfusion in coronary artery bypass grafting is associated with reduced long-term survival. Ann Thorac Surg. 2006;81:1650–1657.
    35. Møller A, Nielsen HB, Wetterslev J, Pedersen OB, Hellemann D, Shahidi S. Low vs high haemoglobin trigger for transfusion in vascular surgery: protocol for a randomised trial. Acta Anaesthesiol Scand. 2017;61:952–961.
    36. Koch CG, Sessler DI, Mascha EJ, et al. A randomized clinical trial of red blood cell transfusion triggers in cardiac surgery. Ann Thorac Surg. 2017;104:1243–1250.
    37. Mazer CD, Whitlock RP, Fergusson DA, et al.; TRICS Investigators and Perioperative Anesthesia Clinical Trials Group. Restrictive or liberal red-cell transfusion for cardiac surgery. N Engl J Med. 2017;377:2133–2144.
    38. Mazer CD, Whitlock RP, Fergusson DA, et al.; TRICS Investigators and Perioperative Anesthesia Clinical Trials Group. Six-month outcomes after restrictive or liberal transfusion for cardiac surgery. N Engl J Med. 2018;379:1224–1233.
    39. Murphy GJ, Pike K, Rogers CA, et al.; TITRe2 Investigators. Liberal or restrictive transfusion after cardiac surgery. N Engl J Med. 2015;372:997–1008.
    40. Goldberg J, Paugh TA, Dickinson TA, et al.; PERForm Registry and the Michigan Society of Thoracic and Cardiovascular Surgeons Quality Collaborative. Greater volume of acute normovolemic hemodilution may aid in reducing blood transfusions after cardiac surgery. Ann Thorac Surg. 2015;100:1581–1587.

    Supplemental Digital Content

    Copyright © 2019 International Anesthesia Research Society