0
Clinical Investigations: SURGERY |

Transfusion of Blood Components and Postoperative Infection in Patients Undergoing Cardiac Surgery* FREE TO VIEW

Santiago Ramón Leal-Noval, MD; María Dolores Rincón-Ferrari, MD; Andrés García-Curiel, MD; Angel Herruzo-Avilés, MD; Pedro Camacho-Laraña, MD; José Garnacho-Montero, MD; Rosario Amaya-Villar, MD
Author and Funding Information

*From the Critical Care Division (Drs. Leal-Noval, Rincón-Ferrari, Herruzo-Avilés, Camacho-Laraña, Garnacho-Montero, and Amaya-Villar) and the Microbiology Division (Dr. García-Curiel), Hospital Universitario “Virgen del Rocío,” Seville, Spain.

Correspondence to: Santiago Ramón Leal-Noval, MD, Servicio de Cuidados Críticos y Urgencias. Hospital Universitario “Virgen del Rocío,” Avda Manuel Siurot s/n, 41013. Seville, Spain; e-mail: sramon@cica.es



Chest. 2001;119(5):1461-1468. doi:10.1378/chest.119.5.1461
Text Size: A A A
Published online

Objective: To investigate the influence of blood derivatives on the acquisition of severe postoperative infection (SPI) in patients undergoing heart surgery.

Setting: The postoperative ICUs of a tertiary-level university hospital.

Design: A cohort study.

Methods: During a 4-year period, 738 patients, classified as patients with SPIs and patients without SPIs (non-SPI patients), were included in the study. We studied the influence of 36 variables on the development of SPI in general and individually for pneumonia, mediastinitis, and/or septicemia. The influence of the blood derivatives on infections was assessed for RBC concentrates, RBC and plasma, and RBC and platelets.

Results: Seventy patients (9.4%) were classified as having SPIs, and 668 (90.6%) were classified as not having SPIs. After multivariate analysis, the variables associated with SPI (incidence, 9.4%) were reintubation, sternal dehiscence, mechanical ventilation (MV) for ≥ 48 h, reintervention, neurologic dysfunction, transfusion of ≥ 4 U RBCs, and systemic arterial hypotension. The variables associated with nosocomial pneumonia (incidence, 5.9%) were reintubation, MV for ≥ 48 h, neurologic dysfunction, transfusion of ≥ 4 U blood components, and arterial hypotension. The variables associated with mediastinitis (incidence, 2.3%) were reintervention and sternal dehiscence, and those associated with sepsis (incidence, 1.6%) were reintubation, time of bypass ≥ 110 min, and MV for ≥ 48 h. The mortality rate (patients with SPI, 52.8%; non-SPI patients, 8.2%; p < 0.001) and mean (± SD) length of stay in the ICU (patients with SPI, 15.8 ± 12.9 days; non-SPI patients, 4.5 ± 4.4 days; p < 0.001) were greater for the infected patients. The transfused patients also had a greater mortality rate (13.3% vs 8.9%, respectively; p < 0.001) and a longer mean stay in the ICU (6.1 ± 7.2 days vs 3.7 ± 2.8 days, respectively; p < 0.01) than those not transfused.

Conclusion: The administration of blood derivatives, mainly RBCs, was associated in a dose-dependent manner with the development of SPIs, primarily nosocomial pneumonia.

Figures in this Article

Patient outcome after cardiac surgery is closely related to the development of postoperative infections. Nosocomial pneumonia, mediastinitis, bacteremia, and sepsis of any origin are frequent infectious processes with an important influence on outcome.1 The risk factors for the acquisition of postoperative infections have been studied generically and specifically for each type of infection.12 Broad-spectrum antibiotic therapy,1,3 age > 65 years,2 female sex,2 length of time receiving mechanical ventilation (MV),1 a high APACHE (acute physiology and chronic health evaluation) II score,4 and the development of multiorgan dysfunction1 all have been related to the development of nosocomial pneumonia. Obesity, the use of bilateral internal mammary arteries in coronary artery bypass grafting (CABG), and long length of surgery are among other factors closely linked to the appearance of postoperative mediastinitis.5Early bloodstream infection within 96 h of cardiopulmonary bypass surgery is related to pulmonary hypertension, diabetes, and the infusion of inotropic drugs.6

Blood transfusions are a frequent therapy in the ICUs. Patients received an average of 0.2 U/d in a multidisciplinary ICU, which increased to 1.3 U in cardiac surgery patients.7Blood transfusions appear to increase the susceptibility to infection in surgical patients. Seventeen of 19 retrospective studies found transfusion to be a significant factor and, frequently, the single best predictor of postoperative infection.8 In a number of studies, transfusion has been considered a risk factor for the development of mediastinitis,5,9 early bacteremia,6 increased mortality rate, and the length of stay in hospitals10 after cardiac surgery. There are, however, many other uncontrolled, confounding, or circumstantial variables that impede the adequate assessment of the role of transfusions in the development of postoperative infections.

To our knowledge, few studies have been designed to assess the influence of blood transfusions on the development of postoperative infections in patients undergoing cardiac surgery.10 Recent data suggest that these patients are transfused excessively11; therefore, a decreased number of transfusions could contribute to a reduction in morbidity and mortality. In a 1998 randomized controlled trial12carried out in patients undergoing cardiac surgery, nosocomial pneumonia was the most frequent infection and was found only when transfused RBC concentrates accounted for ≥ 3 U. In our hospital, a case control study13demonstrated that the transfusion of ≥ 4 U blood components was independently associated with the acquisition of nosocomial pneumonia. Intraoperative RBC transfusion may contribute to the inflammatory response after cardiac surgery and is associated with a worse postoperative recovery.14

This cohort study was designed to establish the influence of blood transfusions (ie, RBC, plasma, and/or platelets) on the development of severe postoperative infections (SPIs) in 738 patients undergoing cardiac surgery. Our working hypothesis was that blood transfusions contribute to the development of SPI.

Study Location and Patients

The study was conducted between June 1994 and June 1998 at the Hospital Universitario Virgen del Rocío of Seville, Spain, a 2,000-bed public teaching hospital. The institutional review board approved this study and waived the need for informed consent. The cardiac surgery carried out in this hospital includes CABG, valve replacement, the correction of congenital cardiopathies in adults and children, and cardiac transplantations. The same staff members (five cardiovascular surgeons, five anesthesiologists, and three intensive care specialists) performed all procedures throughout the study period. After surgery, the patients were admitted to the 40-bed ICU, 6 of which were assigned to the postoperative treatment of cardiothoracic and vascular surgery patients. All the patients received the same standard care. The surgical and anesthetic techniques, the type of cardiopulmonary bypass, and the treatment received in the ICU did not differ from ordinary procedures. After admission to the ICU, all of the patients required MV and were monitored by continuous ECG recordings. After verifying hemodynamic stability, the patients always were positioned at a 45° angle.

In most cases, patients received MV for < 8 h and were extubated when hemodynamically stable, with a Ramsay score of 2 to 3, arterial oxygen saturation of > 95% with a fraction of inspired oxygen of < 0.4, and no significant bleeding. Gastric protection was carried out routinely with famotidine (20 mg IV q12h) during the first 24 h after admission to the ICU, and if a patient required further treatment, famotidine was replaced by sucralfate (1 g po or by nasogastric tube q8h). When patients were in clinically stable condition, they were transferred from the ICU to a surgical ward, without any intermediate stay at a subacute unit.

Study Design

This study is a unicentric, cohort study in that all the data have been collected and stored in a specially designed database (Access, version 7 for Windows 95; Microsoft; Redmond, WA). All patients undergoing cardiac surgery were initially evaluable. Univariate and multivariate statistical analyses were performed using a statistical software package (SPSS, version 7.5; SPSS; Chicago, IL).

Patients with any of the following characteristics were excluded from the study: (1) a stay in the ICU of < 24 h; (2) infection before admission to the ICU; (3) cardiac surgery without cardiopulmonary bypass; (4) cardiac transplant recipients and patients receiving immunosuppressive therapy; (5) age < 16 years; and (6) hemoglobin level of < 11 g/dL and/or disorders of coagulation before surgery.

The principal aim of the study was to determine whether a relationship exists between the transfusion of blood components and SPI. Other variables that can adversely affect the final results also were included in the statistical analysis in order to differentiate the effect of the variable transfusion from the remaining variables.

Consequently, the influence of 36 variables, including transfusion, on the acquisition of SPI was assessed in those subjects with at least one of the following infections: nosocomial pneumonia; mediastinitis; and/or sepsis of unknown origin. Subsequently, the influence of the same variables on the development of nosocomial pneumonia, mediastinitis, and sepsis of undetermined origin was assessed on an individual basis. These variables were as follows.

  • Preoperative factors: age, sex, COPD (ie, COPD requiring medical therapy with bronchodilator drugs or corticosteroids), systemic arterial hypertension (requiring medication to lower BP), diabetes mellitus (requiring treatment with insulin or oral antidiabetic drugs), obesity (a 20% increase over the relative weight or a weight more than the 85th percentile of the body mass index), chronic renal failure (plasma creatinine ≥ 2 mg/dL), previous surgery, New York Heart Association (NYHA) preoperative score ≥ III, acute myocardial infarction before surgery (registered in the patients’ clinical records), and a cardiac ejection fraction before surgery of ≤ 40% (assessed by echocardiography and/or hemodynamic evaluation).

  • Intraoperative factors: the type of surgery (eg, CABG, valve replacement, or mixed and congenital cardiopathies in adults, such as atrial septal defects), the use of internal mammary arteries in patients undergoing CABG, the time of cardiopulmonary bypass circulation, the time of ischemia, urgent or elective surgery, perioperative myocardial infarction (suspected by clinical, ECG, and/or enzymatic criteria and confirmed by echocardiogram and/or pyrophosphate scintigraphy), and blood transfusions (RBC concentrates, plasma, platelets, or combinations of the three).

  • Postoperative factors: the presence of a catheter of thermodilution at the pulmonary artery, a catheter at the left atrium, pleural drainage, intra-aortic balloon pump, reintervention, MV for ≥ 48 h, the use of inotropic drugs (first used in the operating room or in the ICU), CNS dysfunction (focal or diffuse motor deficits without depressor medication of the CNS), postoperative cardiac failure (requiring inotropic drugs and/or vasodilator therapy in continuous perfusion), the transfusion of blood components (RBC concentrates, plasma, and platelets), the need for reintubation, the amount of mediastinal bleeding, arterial hypotension in the immediate postoperative period (mean arterial BP < 65 mm Hg for at least 1 h), fever during the first 24 h after surgery (temperature, > 38.5°C), sternal dehiscence, and an APACHE II score at admission to the ICU of ≥ 12. All these variables were considered before the development of nosocomial infection.

Infectious Surveillance and SPI

First-generation cephalosporins were used for antibiotic prophylaxis just before the onset of the surgical intervention and treatment was discontinued after the patient had been in the ICU for 24 h. Radial artery, pulmonary artery, and bladder catheters were discontinued in the ICU within the first 24 h, whereas mediastinal and pleural drainage was removed 48 h after admission to the ICU. Patients received a daily chest radiograph. Tracheal secretions (before extubation) and the tips of all intravascular catheters as well as an exudate of the sternotomy wound were systematically cultured. If the patient became febrile (body temperature, > 38°C), cultures of blood, urine, and tracheal secretions were taken. In addition, venous catheters were removed and cultured. If the origin of fever remained unknown, a CT scan was carried out to rule out mediastinitis or sinusitis as the cause of infection.

In the present study, only the following SPIs were considered: pneumonia; mediastinitis; or sepsis of unknown origin. In our experience, urinary tract infections and catheter-related bloodstream infections had a very low incidence because of our premature removal of bladder and venous catheters.

Five groups of patients were considered:

  1. patients with SPIs, ie, mediastinitis, nosocomial pneumonia, or sepsis of unknown origin;

  2. patients without SPIs (non-SPI patients), those without proven SPIs, or patients with temperatures < 38°C that disappeared after removing venous catheter and/or bladder catheter;

  3. patients with mediastinitis, ie, superficial and/or deep infection of the surgical wound, with positive findings in cultures obtained from the sternotomy wound, requiring reintervention for the treatment of the mediastinitis (aseptic dehiscence of the wound was not considered);

  4. patients with nosocomial pneumonia, ie, the presence of a new, persistent, or progressive lung infiltrate in the chest radiograph, and at least two of the following criteria: temperature≥ 38°C, leukocytosis ≥ 12,000 cells/μL, or leukopenia≤ 3,000 cells/μL, and purulent endotracheal secretion with a Gram’s stain showing > 25 neutrophils and < 10 epithelial cells per field. In all patients with a clinical suspicion of nosocomial pneumonia, bronchial secretion was collected for culturing, obtained with a protected pulmonary specimen brush (PSB), introduced by fiberoptic bronchoscopy through the endotracheal tube. The results of the brush culture were expressed in colony-forming units per milliliter. The diagnosis of nosocomial pneumonia was made when at least one of the following criteria was fulfilled: (1) positive quantitative culture of a sample of the lower respiratory tract obtained by PSB (≥ 103 colony-forming units per milliliter); (2) blood culture positive for the same microorganisms isolated in an approximate 48-h interval in samples obtained from the respiratory tract (by PSB); and (3) pleural culture positive for the same microorganisms isolated, in an approximate 48-h interval, in samples obtained from the respiratory tract (by PSB); and

  5. patients with sepsis of unknown origin and patients with sepsis or a sepsis syndrome.15 Only those patients in whom an evident infectious focus could not be determined, despite a scheduled clinical evaluation including cultures of blood, wound, tracheal secretions, and the tip of the venous catheter, were included.

Allogeneic Blood Transfusion Criteria for Transfusion

The criteria for transfusion were based on previously published regulations5,1617 acknowledged by the anesthesiologists, intensive care specialists, and surgeons involved in the care of these patients, and adapted by our institution (see “Appendix”). However, the decision on the transfusion requirement depended ultimately on the physician in charge of the patient.

The number of units of blood components related to the development of postoperative nosocomial infection was calculated in an ascendant and stepwise fashion, first for all blood components (ie, any combination of RBCs, plasma, and platelets) and subsequently for each blood component in particular. These variables underwent successive division in order to determine the value at which the differences were found to be significant.

Statistical Analysis

A univariate analysis was performed to identify those risk factors (36 in general, including transfusion) associated with SPI (ie, nosocomial pneumonia, mediastinitis, and/or sepsis of unknown origin). Subsequently, the influence of the same variables on the development of nosocomial pneumonia, mediastinitis, and sepsis of unknown origin was analyzed on an individual basis. Theχ 2 test was used for the comparison of dichotomized variables. The continuous variables, with a normal distribution, were compared applying Student’s t test and, if not normally distributed, they were compared using a Wilcoxon test. The comparisons were unpaired and two tiered for all the tests with statistical significance. The variables with p values ≤ 0.05 were included in a logistic regression model with stepwise elimination. The results of the univariate analysis were expressed as relative risk and confidence intervals. The results of the four multivariate analyses (ie, severe infection in general, nosocomial pneumonia, mediastinitis, and sepsis of unknown origin) were expressed as adjusted odds ratio and confidence intervals.

General Results

Seven hundred sixty-three patients were admitted, although, in the end, 738 were evaluated and 25 were excluded (surgery because of infectious endocarditis, 8 patients; not having received cardiopulmonary bypass circulation, 5 patients; and early death[ during the first 24 h] because of cardiac pump failure, 12 patients). The general characteristics of the group (738 patients) consisted of the following: mean (± SD) age, 58.4 ± 12 years; men, 61%; mean APACHE II score at admission to the ICU, 10.7 ± 4.7; arterial systemic hypertension, 32.6%; diabetes, 17.6%; obesity, 15.7%; COPD, 5.1%; chronic renal failure, 1.6%; cardiac failure of NYHA grade ≥ III, 29.6%; and a preoperative ejection fraction< 40%, 7.1% (Table 1 ).

Of these patients, 36.8% underwent coronary revascularization, 56.8% underwent surgery for valve replacement, 3.6% underwent surgery of congenital cardiopathies (eg, atrial septal defects) at an adult age, and 2.8% underwent mixed surgery (eg, valvular or revascularization). Elective surgery was performed in 97.7% of the cases.

Influence of the Variables on the Acquisition of SPI

Seventy patients (9.4%) fulfilled the requirements to be included in the infected group vs 668 patients (90.6%) in the noninfected group. In the univariate analysis, 17 variables showed a positive influence on the development of postoperative infection (Table 2 ). A logistic regression model was developed with these variables, pointing out seven of the following variables as being independently related to infection: reintubation; sternal dehiscence; MV for ≥ 48 h; reintervention; neurologic dysfunction; arterial hypotension; and the transfusion of ≥ 4 U blood components (Table 3 ).

Influence of the Variables on the Acquisition of Nosocomial Pneumonia

Forty-five patients (5.9%) received diagnoses of nosocomial pneumonia. In the univariate analysis, the following 12 variables showed a positive influence on the development of nosocomial pneumonia: reintubation; MV for ≥ 48 h; transfusion of ≥ 4 U RBC concentrates; transfusion of ≥ 4 U blood components; arterial hypotension; reintervention; and transfusion of ≥ 2 U plasma. Only the following five variables continued in the multivariate analysis: reintubation; MV for ≥ 48 h; neurologic dysfunction; transfusion of ≥ 4 U blood components; and arterial hypotension (Table 3).

Influence of the Variables on the Acquisition of Mediastinitis

Seventeen patients (2.3%) received diagnoses of mediastinitis. In the univariate analysis, a total of seven variables showed positive influence on the development of mediastinitis: sternal dehiscence; reintervention; reintubation; female sex; MV ≥ 48 h; postoperative cardiac failure; and transfusion ≥ 4 U of RBC concentrates. Only two variables appeared in the model of logistic regression: reintervention and sternal dehiscence (Table 3).

Influence of the Variables on the Acquisition of Sepsis of Unknown Origin

Twelve patients (1.6%) received diagnoses of sepsis of unknown origin. The univariate analysis encountered the following five variables with positive influences on the development of postoperative sepsis from an undetermined origin: reintubation; MV for ≥ 48 h; time of ischemia of ≥ 75 min; time of cardiopulmonary bypass circulation of ≥ 110 min; and APACHE II score ≥ 12. A logistic regression model was developed with the following three discriminating variables independently related to sepsis: reintubation; time of cardiopulmonary bypass circulation of ≥ 110 min; and MV for ≥ 48 h (Table 3).

Transfusion

Of the 738 patients, 592 patients (80.2%) underwent transfusion and received RBC concentrates. One hundred fifty-eight patients (21.4%) underwent transfusion with at least 1 U of plasma, and 89 patients (12%) underwent transfusion with at least 1 U of platelets. All the transfusions were performed before the development of nosocomial infection.

The transfusion of ≥ 4 U blood components (ie, any combination of RBC concentrates, plasma, or platelets) was associated with infection and pneumonia in the univariate analysis and only with infection in the multivariate analysis. The transfusion of ≥ 4 U RBC concentrates was associated with infection, nosocomial pneumonia, and mediastinitis (univariate analysis) and was significant only for infection and nosocomial pneumonia in the multivariate analysis. The transfusion of ≥ 2 U plasma or ≥ 1 U platelets is related to the development of infection and nosocomial pneumonia in the univariate analysis but not in the multivariate analysis.

A direct relationship (Fig 1 ) was established between the number of units transfused and the general rates of infections, pneumonia, and mediastinitis. The rates of infection, pneumonia, and APACHE II score increased considerably after the transfusion of ≥ 4 U blood components.

The patients who underwent transfusion experienced higher mortality rates and a longer mean stay in the ICU than patients without transfusion (mortality rate, 13.3% vs 8.9%, respectively[ p < 0.01]; mean stay in the ICU, 6.1 ± 7.2 days vs 3.7 ± 2.8 days, respectively [p < 0.01]).

Outcome

Ninety-two patients died, for a crude mortality rate of 12.5%. The mortality rate was 52.8% (37 of 70 patients) in the SPI group and 8.2% (55 of 668 patients) in the non-SPI group (p < 0.001). The mean length of stay in the ICU was 15.8 ± 12.9 days in the SPI group vs 4.5 ± 4.4 days in the non-SPI group (p < 0.001).

Our results suggest a strong, dose-dependent association between transfusion and infection. The transfusion of blood components in general (ie, any combination of RBC concentrates, plasma, or platelets), and of RBC concentrates in particular, were involved in the development of postoperative infection, mainly nosocomial pneumonia. The transfusion of plasma or platelets was related to infection, in general, and to pneumonia in the univariate analysis alone. A clear association between transfusion and infection could not be demonstrated in regard to mediastinitis and/or sepsis.

In critically ill patients, it has been hypothesized that measures aimed toward a maximal transportation of oxygen to the tissue aid in avoiding oxygen debt. When transfusions, fluids, or drugs increase oxygen delivery, the survival rate of high-risk surgical patients improves markedly.1819 However, this has not been verified by a number of studies, in that the increase of oxygen delivery to supranormal levels did not improve survival rates and, moreover, had deleterious effects.20A 1999 randomized controlled trial21was designed to determine whether a restrictive strategy of RBC concentrate transfusion (to maintain hemoglobin levels at 7 to 9 g/dL) and a liberal strategy RBC concentrate transfusion (to maintain hemoglobin levels at 10 to 12 g/dL) could produce equivalent results in critically ill patients. Results showed that death rates from all causes at 30 days were similar in the two groups. However, the rates were significantly lower with the restrictive transfusion strategy among patients who were less severely ill (APACHE II score, ≤ 20) and were < 55 years of age. Moreover, the restrictive strategy resulted in a relative decrease of 54% in the number of transfusions, and 33% of the patients assigned to this strategy did not receive any RBC transfusions. A study performed with patients undergoing bypass surgery did not find differences between the transfusion strategies.22In fact, transfusion abuse could lead to adverse results.2324 However, anemia also can increase the risk of death in critically ill patients with cardiac disease.25 These results have led to a considerable variation in transfusion policies among institutions in regard to patients who are undergoing cardiac surgery,6,11 with a high percentage of transfusions considered to be inappropriate. The implantation of transfusion criteria could decrease the rate of the patients who are transfused.5

The adverse effects of blood transfusions may be important, because they have been involved in the development of postoperative infections in patients undergoing cardiac surgery. Miholic et al3 studied the risk factors for the development of severe bacterial infections (ie, mediastinitis, pneumonia, septicemia, and prosthetic valve endocarditis) in 246 patients undergoing cardiac surgery. Many variables, such as the duration of surgery, reintervention because of bleeding, urgent surgery, intra-aortic balloon pumping, age, sex, and the professional status of the surgeon, were collected in the logistic regression model. The following three variables were found to be significant in the logistic regression analysis: the restoration of > 2,500 mL blood; reintervention; and the duration of the cardiopulmonary bypass. A similar study4 demonstrated that allogeneic transfusions were involved in the sternal infection of 2,579 patients undergoing cardiac surgery. Other studies have confirmed the relationship of the allogeneic transfusion with postoperative infection, mainly mediastinitis2627 and pneumonia.13 It has been acknowledged28 that transfusions can be related to early bacteremia. In a case control study of 7,928 patients undergoing cardiac surgery, the logistic regression analysis established that pulmonary hypertension, diabetes, inotropic infusions, and the number of units of blood that were transfused were related to early bacteremia during the first 96 h after cardiopulmonary bypass.

In our study, RBC concentrate transfusions were involved in the development of infection in general (multivariate analysis), pneumonia (multivariate analysis), and mediastinitis (univariate analysis). Plasma or platelet transfusions were related to infection and mediastinitis (univariate analysis).

The effect that a blood transfusion can have on the immune system is called immunomodulation and may have very important repercussions. It is estimated that the death rate due to a postoperative infection caused by this immunomodulation may exceed the death rate caused by the combination of all the other transfusion risks.29Measures directed toward decreasing the deleterious effects of immunomodulation could save between $6 and $12 billion dollars in hospital expenses, while reducing the number of postoperative infections.30 The mechanism of allogeneic transfusion-induced immunomodulation may involve altered cytokine regulation that contributes to the down-regulation of macrophage and T-cell functions and to the up-regulation of humoral immunity. Allogeneic transfusions decrease the cellular immune function of the monocytes, macrophages, T cells, and natural killer cells, down-regulating the antigen-presenting cells and T-cell function.29 This impaired cellular immune response is the mechanism that is proposed as being responsible for the high incidence of postoperative infections.31 The leukocytes present in transfused blood seem to have a primordial role in the suppression of cellular immunity.2930

As in other studies, the association of transfusion with infection was dose dependent with a threshold of ≥ 1 U for platelets,≥ 2 U for plasma, ≥ 4 U for RBC concentrates, and ≥ 4 U for all blood components (ie, any combination of RBC concentrates, plasma, or platelets). Although excessive bleeding (ie,≥ 800 mL) and severity of disease in the patient on admission to the ICU (ie, APACHE II score of ≥ 12) were implicated in the development of postoperative infection in the univariate analysis (Table 2), they were not involved in the multivariate analysis, suggesting that the association of the allogeneic transfusion with postoperative nosocomial infection is independent of bleeding and severity of disease in the patient. When the number of units transfused increased, the infection rate and the APACHE II score did also. The increase in the infection rate was, however, greater than that in the APACHE II score (Fig 1).

The transfusion of RBC concentrates was related fundamentally to the development of nosocomial pneumonia. This relationship has been documented previously.23,1213 In our study, the incidence of mediastinitis (2.3%) and of sepsis (1.6%) was low and the number of risk factors considered was high. So, it is probable that the number of patients necessary to demonstrate an association between allogeneic blood transfusion and mediastinitis and/or allogeneic blood transfusion and sepsis must be higher.

This study has important limitations. With an observational design, researchers can only control the effects of confounding factors that are known and are measurable, which can be introduced in the multivariate analysis.32 Unfortunately, this type of study does not measure the effects of other unknown modalities that can affect the final results. Therefore, it has been suggested that the immunomodulatory effects of allogeneic blood transfusion were not related to the transfusion itself,32but to other clinical variables not measured in the study. That led to the blood transfusion and, hence, dictated the final outcome. Transfusion would only be a marker of uncontrolled confounding factors. For these same reasons, the possible influence of transfusions on mortality may suffer important misinterpretations. A low number of patients transfused with plasma or platelets exist. The transfusion of these components always has been accompanied by the transfusion of RBC concentrates. Hence, it is difficult to assess whether the transfusion of plasma and/or platelets influences the acquisition of postoperative infection. Although APACHE II severity scoring performs well to predict mortality after cardiac surgery,3334 other severity scorings could be more appropriate for use with these patients. Other multiple factors relatedto the difficulty of surgery, personnel, equipment, manipulation, and length of stay may have been discarded involuntarily.

The practical consequences may be important. The perioperative transfusion of RBC concentrates is associated with nosocomial pneumonia. Our data also suggest that other blood components commonly used in patients undergoing cardiac surgery, such as plasma and/or platelets, could be involved. Until verification by new prospective studies, blood transfusions should be considered as a possible risk factor for postoperative infection.

Criteria for Blood Transfusion in Cardiac Surgical Patients
RBC Concentrates

  1. hemoglobin, ≤ 8 g/dL;

  2. hemoglobin, 8 to 10 g/dL in normovolemic patients, but with clinical signs of myocardial, cerebral, or respiratory dysfunction; and

  3. severe hemorrhage, ie, bleeding > 10 mL/kg in the first hour or 5 mL/kg in the first 3 h (averaged).

Fresh Frozen Plasma

Patients with severe hemorrhaging and:

  1. prothrombin time or partial thromboplastin time ≥ 1.5 times that of control subjects;

  2. massive transfusion of RBC concentrates; and

  3. previous treatment with coumarin derivatives and unscheduled surgery.

Platelets

Patients with severe hemorrhaging and:

  1. diffuse bleeding suggestive of platelet dysfunction;

  2. platelet count of < 100 × 109/L; and

  3. massive transfusion of RBC concentrates.

Abbreviations: APACHE = acute physiology and chronic health evaluation; CABG = coronary artery bypass grafting; MV = mechanical ventilation; NYHA = New York Heart Association; PSB = pulmonary specimen brush; SPI = severe postoperative infection

Table Graphic Jump Location
Table 1. General Characteristics of the Patients (n = 738)*
* 

Values given as percentage or mean ± SD (range). CPB =cardiopulmonary bypass; POMI = perioperative myocardial infarction.

 

Referring only to CABG.

 

Transfusion of at least 1 U of any type of blood component.

§ 

Transfusion of at least 1 U of this blood component.

Table Graphic Jump Location
Table 2. Significant Factors for the Acquisition of SPI*
* 

RR = relative risk; AMI = acute myocardial infarction. See Table 1 for abbreviations not used in text.

Table Graphic Jump Location
Table 3. Results of the Multivariate Analysis for the Different Entities Studied*
* 

AOR = adjusted odds ratio; CI = confidence interval. See Table 1 for abbreviations not used in text.

Figure Jump LinkFigure 1. Relationship between the units of blood transfused and the percentage of infections.Grahic Jump Location

We thank Professor Juan Muñoz of the School of Mathematics (University of Seville, Spain) for his cooperation in the study design and statistical analysis.

Landers, DF, Hill, GE, Wong, KC, et al (1996) Blood transfusion- induced immunomodulation.Anesth Analg82,187-204. [PubMed]
 
Murphy, PJ, Connery, C, Hicks, GL Homologous blood transfusion as a risk factor for postoperative infection after coronary artery bypass graft operations.Thorac Cardiovasc Surg1992;104,1092-1099
 
Miholic, J, Hudec, M, Domanig, E Risk factors for severe bacterial infections after valve replacement and aortocoronary bypass operations analysis of 246 cases by logistic regression.Ann Thorac Surg1985;40,224-228. [CrossRef] [PubMed]
 
Ottino, G, Paulis, R, Pansini, S Major sternal wound infection after open-heart surgery: a multi-variant analysis of risk factors in 2579 consecutive operative procedures.Ann Thorac Surg1987;44,173-179. [CrossRef] [PubMed]
 
Leal, SR, Jiménez, PI, Márquez, JA, et al Aportes de sangre a pacientes intervenidos de cirugía cardíaca: ¿pueden cambiarse los hábitos transfusionales?Med Clin (Barc)1996;106,368-371. [PubMed]
 
Goodnough, LT, Despostis, GJ, Hogue, CW, et al On the need for improved transfusion indicators in cardiac surgery.Ann Thorac Surg1995;60,473-480. [CrossRef] [PubMed]
 
Mezrow, CK, Bergstein, I, Tartter, PI Postoperative infections after autologous and homologous blood transfusion.Transfusion1992;32,27-30. [CrossRef] [PubMed]
 
Gu, YJ, deVries, AJ, Boonstra, PW, et al Leukocyte depletion results in improved lung function and reduced inflammatory response after cardiac surgeryThorac Cardiovasc Surg1996;112,494-500. [CrossRef]
 
Fernández, LA, MacSween, JM, Choong, K, et al Immunologic changes after blood transfusion in patients undergoing vascular surgery.Am J Surg1992;163,263-269. [CrossRef] [PubMed]
 
Blumberg, N, Heal, J Transfusion and recipient immune function.Arch Pathol Lab Med1989;113,246-253. [PubMed]
 
Goodnough, LT, Johnston, MFM, Toy, PTCY Transfusion Medicine Academic Award Group: the variability of transfusion practice in coronary artery bypass graft surgery.JAMA1991;265,86-90. [CrossRef] [PubMed]
 
Van Watering, LMG, Hermans, J, Houbiers, JGA, et al Beneficial effects of leukocyte depletion of transfused blood on postoperative complications in patients undergoing cardiac surgery: a randomized clinical trial.Circulation1998;97,562-568. [CrossRef] [PubMed]
 
Leal, SR, Márquez, JA, García-Curiel, A, et al Nosocomial pneumonia in patients undergoing heart surgery.Crit Care Med2000;28,935-940. [CrossRef] [PubMed]
 
Fransen, E, Maessen, J, Dentener, M, et al Impact of blood transfusion of inflammatory mediator release in patients undergoing cardiac surgery.Chest1999;116,1233-1239. [CrossRef] [PubMed]
 
Bone, RC The pathogenesis of sepsis.Ann Intern Med1991;115,457-469. [PubMed]
 
American College of Physicians.. Practice strategies for elective RBC transfusion.Ann Intern Med1992;116,403-406. [PubMed]
 
Consensus conference. Perioperative red blood cell transfusionJAMA1988;260,2700-2703. [CrossRef] [PubMed]
 
Shoemaker, WC, Appel, PL, Kram, HB, et al Prospective trial of supranormal values of survivors as therapeutics goals in high-risk surgical patients.Chest1988;94,1176-1186. [CrossRef] [PubMed]
 
Boyd, O, Grounds, RM, Bennett, ED A randomized clinical trial of the effect of deliberate perioperative increase of oxygen delivery on mortality in high-risk surgical patients.JAMA1993;270,2699-2707. [CrossRef] [PubMed]
 
Gattinoni, L, Brazzi, L, Pelosi, P, et al A trial of goal-oriented hemodynamic therapy in critically ill patientsN Engl J Med1995;333,1025-1032. [CrossRef] [PubMed]
 
Hébert, PC, Wells, G, Blajchman, MA, et al A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care.N Engl J Med1999;340,409-417. [CrossRef] [PubMed]
 
Johnson, RG, Thurer, RL, Kruskall, MS, et al Comparison of two transfusion strategies after elective operations for myocardial revascularization.Thorac Cardiovasc Surg1992;104,307-314
 
Spiess, BD, Ley, C, Body, SC, et al Hematocrit value on ICUs entry influences the frequency of Q-wave myocardial infarction after coronary artery bypass grafting.Thorac Cardiovasc Surg1998;116,460-467. [CrossRef]
 
Corwin, Hl, Parsonnet, KC, Gettinger, A RBC transfusion in the ICU: is there a reason?Chest1995;108,767-771. [CrossRef] [PubMed]
 
Hébert, PC, Wells, G, Tweeddale, M, et al Does transfusion practice affect mortality in critically ill patients?Am J Respir Crit Care Med1997;155,1618-1623. [PubMed]
 
El Oakley, R, Paul, E, Wong, PS, et al Mediastinitis in patients undergoing cardiopulmonary bypass: risk analysis and midterm resultsJ Cardiovasc Surg1997;38,595-600
 
Zacharias, A, Habib, RH Factors predisposing to median sternotomy complications: deep vs. superficial infection.Chest1996;110,1173-1178. [CrossRef] [PubMed]
 
Ryan, T, Mc Carthy, JF, Rady, MY, et al Early bloodstream infection after cardiopulmonary bypass: frequency rate, risk factors, and implicationsCrit Care Med1997;25,2009-2014. [CrossRef] [PubMed]
 
Blumberg, N Allogeneic transfusion and infection: economic and clinical implications.Semin Hematol1997;34(suppl),34-40
 
Blumberg, N, Heal, JM Immunomodulation by blood transfusion: an evolving scientific and clinical challenge.Am J Med1996;101,299-308. [CrossRef] [PubMed]
 
Leal, SR, de Luís, JC, Márquez, JA, et al Transfusion as a risk factor for infection in ICU.Clin Pulm Med1999;6,236-240. [CrossRef]
 
Vanvakas, EC Transfusion-associated cancer recurrence and postoperative infection: meta-analysis of randomized, controlled clinical trials.Transfusion1996;36,175-186. [CrossRef] [PubMed]
 
Martínez-Alario, J, Tuesta, ID, Plasencia, E, et al Mortality prediction in cardiac surgery patients: comparative performance of Parsonnet and general severity system.Circulation1999;99,2378-2382. [CrossRef] [PubMed]
 
Turner, JS, Morgan, CJ, Thakrar, B, et al Difficulties in predicting outcome in cardiac surgery patients.Crit Care Med1995;23,1843-1850. [CrossRef] [PubMed]
 

Figures

Figure Jump LinkFigure 1. Relationship between the units of blood transfused and the percentage of infections.Grahic Jump Location

Tables

Table Graphic Jump Location
Table 1. General Characteristics of the Patients (n = 738)*
* 

Values given as percentage or mean ± SD (range). CPB =cardiopulmonary bypass; POMI = perioperative myocardial infarction.

 

Referring only to CABG.

 

Transfusion of at least 1 U of any type of blood component.

§ 

Transfusion of at least 1 U of this blood component.

Table Graphic Jump Location
Table 2. Significant Factors for the Acquisition of SPI*
* 

RR = relative risk; AMI = acute myocardial infarction. See Table 1 for abbreviations not used in text.

Table Graphic Jump Location
Table 3. Results of the Multivariate Analysis for the Different Entities Studied*
* 

AOR = adjusted odds ratio; CI = confidence interval. See Table 1 for abbreviations not used in text.

References

Landers, DF, Hill, GE, Wong, KC, et al (1996) Blood transfusion- induced immunomodulation.Anesth Analg82,187-204. [PubMed]
 
Murphy, PJ, Connery, C, Hicks, GL Homologous blood transfusion as a risk factor for postoperative infection after coronary artery bypass graft operations.Thorac Cardiovasc Surg1992;104,1092-1099
 
Miholic, J, Hudec, M, Domanig, E Risk factors for severe bacterial infections after valve replacement and aortocoronary bypass operations analysis of 246 cases by logistic regression.Ann Thorac Surg1985;40,224-228. [CrossRef] [PubMed]
 
Ottino, G, Paulis, R, Pansini, S Major sternal wound infection after open-heart surgery: a multi-variant analysis of risk factors in 2579 consecutive operative procedures.Ann Thorac Surg1987;44,173-179. [CrossRef] [PubMed]
 
Leal, SR, Jiménez, PI, Márquez, JA, et al Aportes de sangre a pacientes intervenidos de cirugía cardíaca: ¿pueden cambiarse los hábitos transfusionales?Med Clin (Barc)1996;106,368-371. [PubMed]
 
Goodnough, LT, Despostis, GJ, Hogue, CW, et al On the need for improved transfusion indicators in cardiac surgery.Ann Thorac Surg1995;60,473-480. [CrossRef] [PubMed]
 
Mezrow, CK, Bergstein, I, Tartter, PI Postoperative infections after autologous and homologous blood transfusion.Transfusion1992;32,27-30. [CrossRef] [PubMed]
 
Gu, YJ, deVries, AJ, Boonstra, PW, et al Leukocyte depletion results in improved lung function and reduced inflammatory response after cardiac surgeryThorac Cardiovasc Surg1996;112,494-500. [CrossRef]
 
Fernández, LA, MacSween, JM, Choong, K, et al Immunologic changes after blood transfusion in patients undergoing vascular surgery.Am J Surg1992;163,263-269. [CrossRef] [PubMed]
 
Blumberg, N, Heal, J Transfusion and recipient immune function.Arch Pathol Lab Med1989;113,246-253. [PubMed]
 
Goodnough, LT, Johnston, MFM, Toy, PTCY Transfusion Medicine Academic Award Group: the variability of transfusion practice in coronary artery bypass graft surgery.JAMA1991;265,86-90. [CrossRef] [PubMed]
 
Van Watering, LMG, Hermans, J, Houbiers, JGA, et al Beneficial effects of leukocyte depletion of transfused blood on postoperative complications in patients undergoing cardiac surgery: a randomized clinical trial.Circulation1998;97,562-568. [CrossRef] [PubMed]
 
Leal, SR, Márquez, JA, García-Curiel, A, et al Nosocomial pneumonia in patients undergoing heart surgery.Crit Care Med2000;28,935-940. [CrossRef] [PubMed]
 
Fransen, E, Maessen, J, Dentener, M, et al Impact of blood transfusion of inflammatory mediator release in patients undergoing cardiac surgery.Chest1999;116,1233-1239. [CrossRef] [PubMed]
 
Bone, RC The pathogenesis of sepsis.Ann Intern Med1991;115,457-469. [PubMed]
 
American College of Physicians.. Practice strategies for elective RBC transfusion.Ann Intern Med1992;116,403-406. [PubMed]
 
Consensus conference. Perioperative red blood cell transfusionJAMA1988;260,2700-2703. [CrossRef] [PubMed]
 
Shoemaker, WC, Appel, PL, Kram, HB, et al Prospective trial of supranormal values of survivors as therapeutics goals in high-risk surgical patients.Chest1988;94,1176-1186. [CrossRef] [PubMed]
 
Boyd, O, Grounds, RM, Bennett, ED A randomized clinical trial of the effect of deliberate perioperative increase of oxygen delivery on mortality in high-risk surgical patients.JAMA1993;270,2699-2707. [CrossRef] [PubMed]
 
Gattinoni, L, Brazzi, L, Pelosi, P, et al A trial of goal-oriented hemodynamic therapy in critically ill patientsN Engl J Med1995;333,1025-1032. [CrossRef] [PubMed]
 
Hébert, PC, Wells, G, Blajchman, MA, et al A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care.N Engl J Med1999;340,409-417. [CrossRef] [PubMed]
 
Johnson, RG, Thurer, RL, Kruskall, MS, et al Comparison of two transfusion strategies after elective operations for myocardial revascularization.Thorac Cardiovasc Surg1992;104,307-314
 
Spiess, BD, Ley, C, Body, SC, et al Hematocrit value on ICUs entry influences the frequency of Q-wave myocardial infarction after coronary artery bypass grafting.Thorac Cardiovasc Surg1998;116,460-467. [CrossRef]
 
Corwin, Hl, Parsonnet, KC, Gettinger, A RBC transfusion in the ICU: is there a reason?Chest1995;108,767-771. [CrossRef] [PubMed]
 
Hébert, PC, Wells, G, Tweeddale, M, et al Does transfusion practice affect mortality in critically ill patients?Am J Respir Crit Care Med1997;155,1618-1623. [PubMed]
 
El Oakley, R, Paul, E, Wong, PS, et al Mediastinitis in patients undergoing cardiopulmonary bypass: risk analysis and midterm resultsJ Cardiovasc Surg1997;38,595-600
 
Zacharias, A, Habib, RH Factors predisposing to median sternotomy complications: deep vs. superficial infection.Chest1996;110,1173-1178. [CrossRef] [PubMed]
 
Ryan, T, Mc Carthy, JF, Rady, MY, et al Early bloodstream infection after cardiopulmonary bypass: frequency rate, risk factors, and implicationsCrit Care Med1997;25,2009-2014. [CrossRef] [PubMed]
 
Blumberg, N Allogeneic transfusion and infection: economic and clinical implications.Semin Hematol1997;34(suppl),34-40
 
Blumberg, N, Heal, JM Immunomodulation by blood transfusion: an evolving scientific and clinical challenge.Am J Med1996;101,299-308. [CrossRef] [PubMed]
 
Leal, SR, de Luís, JC, Márquez, JA, et al Transfusion as a risk factor for infection in ICU.Clin Pulm Med1999;6,236-240. [CrossRef]
 
Vanvakas, EC Transfusion-associated cancer recurrence and postoperative infection: meta-analysis of randomized, controlled clinical trials.Transfusion1996;36,175-186. [CrossRef] [PubMed]
 
Martínez-Alario, J, Tuesta, ID, Plasencia, E, et al Mortality prediction in cardiac surgery patients: comparative performance of Parsonnet and general severity system.Circulation1999;99,2378-2382. [CrossRef] [PubMed]
 
Turner, JS, Morgan, CJ, Thakrar, B, et al Difficulties in predicting outcome in cardiac surgery patients.Crit Care Med1995;23,1843-1850. [CrossRef] [PubMed]
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
Guidelines
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543