About the Authors
By Margaret Brown, RN, MSN, CCRN, and Patricia K. Whalen, RN. Ryder Trauma Center, University of Miami, Miami, Fla (MB) and Dartmouth-Hitchcock Medical Center, Lebanon, NH (PKW)


This supplement is sponsored by an educational grant from:






As many as 15 million blood transfusions are given each year in the United States.¹ They are administered routinely in the intensive care unit (ICU): in 1 study more than 50% of all patients admitted to the ICU received at least 1 unit of red blood cells (RBCs) during their stay.² Routine administration of RBC transfusions attests to their relative safety. Although many known complications such as graft-versus-host disease, hemolytic febrile allergic reactions, and transfusion-related acute lung injury have been thoroughly described,^3,4 our purpose is to examine the less well-defined risks of blood transfusion, especially as they may affect the critically ill patient.

Blood transfusions have been shown to increase the risk of postoperative bacterial infection in patients undergoing various types of surgery. Blood transfusions have also been implicated in higher in-hospital mortality rates. Transfusion-transmitted infections and bacterial infection from aged blood account for some of this increased risk. Transfusion-related immunosuppression, resulting in cancer recurrence and other complications, has been identified as a risk factor in some studies. These risks must be weighed against the potential benefit of transfusion for the patient.

A better understanding of the indications for and risks of transfusion may lead to a more measured approach to blood transfusions. Alternatives to allogeneic blood transfusion, such as use of pharmacological agents and autologous transfusion, will also be examined. We begin with a brief review of the pathophysiology of anemia and basic transfusion medicine.

Anemia
Anemia is the primary indication for transfusion of RBCs. Anemia occurs when the number of RBCs falls below normal, because of blood loss or deficient production of RBCs. Red blood cells carry hemoglobin, which enables them to carry oxygen from the lungs and deliver it through the body. The amount of hemoglobin in a blood sample can be used to determine if a patient is anemic. Normal hemoglobin levels are between 12 g/dL and 16 g/dL for women, and 14 g/dL and 18 g/dL for men.¹

The hematocrit—the proportion of RBCs in the total volume of blood—can also determine if a patient is anemic. A normal hematocrit value ranges from .37 to .48 for women, and .45 to .52 for men.¹

Typically, transfusions are administered at the “10/30 trigger” (hemoglobin, 10 g/dL; hematocrit, .30)2; however, no specific protocol exists for critically ill patients. Criteria for transfusion have been questioned in several recent studies.
When the number of RBCs or the amount of hemoglobin in them is reduced, the blood becomes oxygen deficient. The effects of oxygen deficiency due to anemia can be serious. Even mild to moderate anemia can have profound effects: tachycardia, systolic ejection murmur, exertional dyspnea, and palpitation may occur in addition to fatigue, weakness, and lightheadedness. More severe anemia can cause stroke or heart attack as diminished oxygen causes decreased perfusion to coronary and cerebral tissues.⁵

Pathophysiology of Anemia
When blood is lost, the body quickly pulls water from tissues outside the bloodstream in an attempt to keep the blood vessels filled. As a result, the blood is diluted and the proportion of RBCs drops. This dilution exacerbates the actual loss of RBCs.

Losing large amounts of blood suddenly can cause 2 problems. First, blood pressure falls because the amount of fluid remaining in the vessels is insufficient. Second, the body’s oxygen is diminished because the number of oxygen-carrying RBCs is diminished. Generally, the loss of oxygen delivery from decreased numbers of RBCs is compensated for by an increase in cardiac output. Immediate effects depend on the duration and volume of hemorrhage. Sudden loss of one third of blood volume may be fatal, but as much as two thirds may be lost slowly (over 24 hours or more) without such fatality.1 More common than a sudden drastic loss of blood, however, is chronic bleeding. Bleeding from stomach ulcers or colonic hemorrhage may not be obvious because the rate of blood loss is slower. When blood loss is slower and the resulting anemia is not as severe, the body may produce enough RBCs and shift fluids to compensate.

Phlebotomy for diagnostic testing is a major source of blood loss in critically ill patients. Smoller and Kruskall6 reported mean daily blood losses of 41.5 mL in their study of phlebotomy and its relationship to blood transfusion. This same study also found that patients in the ICU had phlebotomy twice as often, with about 3 times more blood drawn daily, than patients outside the ICU. Corwin et al⁷ found that 142 patients with ICU stays of more than 1 week had an average of 61 mL to 70 mL of blood drawn per day for diagnostic testing. With a mean stay of 19.6 days, this loss amounted to nearly a liter of blood per patient. There was a direct correlation between the amount of blood drawn for diagnostic testing and the amount of blood required for transfusion. The indication for transfusion was not as evident: no indication for transfusion (see Figure) was identified for 29% of transfusion events. A hematocrit that is lower than .25 was the indication in 19% of events. Nearly one third of all RBC transfusions were without a clear indication according to National Institutes of Health Consensus Conference guidelines (active bleeding or surgery, low hematocrit [.25], low cardiac output, myocardial infarction/ischemia, low oxygen transport, renal ischemia, and preoperative status).


x
Regardless of other indications, in this study all patients who received transfusion had hematocrits at or below .27. Patients who were not transfused had admission hematocrits significantly above those seen in transfused patients, and well above the .27 that was the apparent trigger in this ICU. These patients were less likely to experience a drop in hematocrit with phlebotomy or other blood loss.

Eventually, increased production of RBCs corrects anemia without intervention in many patients. The production of RBCs is regulated by a feedback loop. Erythropoietin (EPO), a glycoprotein hormone released primarily by the kidneys, stimulates production of RBCs in response to diminished arterial oxygen in the kidney. Under normal conditions, a drop in arterial oxygen would result in an increase in EPO production, and a resulting increase in RBCs.

This feedback loop may be blunted in critically ill patients, so that RBC production is not increased in response to anemia. Blunted EPO response may be caused by chronic renal failure, inflammatory disease, bone marrow damage, chemotherapy, or malnutrition.⁸ Compared to patients with iron-deficiency anemia, critically ill patients were found to have much lower EPO levels despite comparable hematocrit values.⁹ Low EPO levels and lack of RBC production were also noted in patients undergoing open heart surgery.¹⁰

Treatment of anemia caused by blood loss depends on how rapidly blood is being lost, the source of the bleeding, and the severity of the anemia. The source of bleeding must be found and the bleeding stopped. Red blood cell transfusion, in conjunction with hemostasis and treatment of shock, remains the treatment of choice for quickly restoring blood and intravascular volume.

Conditions That Increase the Risk of Anemia
Critically ill patients may have concomitant conditions that make them more prone to anemia, such as cancer, chronic renal failure, or human immunodeficiency virus (HIV). Surgery or trauma may also predispose patients to develop anemia. Sometimes no precise cause for anemia can be found. In critically ill patients, anemia is often referred to as the anemia of critical illness, which is a variant of anemia of chronic disease. Typically, anemia of critical illness is associated with underproduction of RBCs, inappropriately low concentrations of iron and circulating EPO, and high concentrations of ferritin.

In patients with cancer, anemia may be caused by blood loss, malnutrition, tumor metastases to the bone marrow, radiation, or chemotherapy. Patients with cancer may also experience hemolysis, in which RBCs are destroyed at a greater than normal rate. Chemotherapy causes anemia by suppressing either kidney or bone marrow function. Chemotherapy may disrupt regulation of endogenous EPO, presumably because of the action of various chemotherapeutic agents on renal cells that produce EPO.¹¹ Anemia in chronic renal failure is caused by inability of the kidney to produce sufficient EPO. Dietary deficiencies of iron, folic acid, and vitamin B₁₂, or excesses of aluminum in the body can also contribute to anemia in the patient with chronic renal failure. These patients may also suffer from anemia due to chronic blood loss.

Chronic anemia occurs in many patients with HIV, even those who are asymptomatic. The virus itself blunts EPO response, increasing the likelihood that anemia will develop.¹² As the disease progresses, virtually all patients with HIV become anemic. In addition to the virus, anemia can be caused by infectious disease and/or antiretroviral therapy. Anemia and blood transfusion are associated with increased mortality in patients with HIV.

Mechanics of Blood Transfusion
Blood Products
Depending on the reason for blood transfusion, physicians may order whole blood or a blood component, such as RBCs, platelets, blood clotting factors, or fresh-frozen plasma (the liquid part of blood). Ideally, transfusion consists only of the blood products required to meet the patient’s need, rather than whole blood.

Whole blood may be administered to help restore fluid volume and circulation in the case of profuse bleeding. It may also be used when a needed blood component is unavailable.

Packed RBCs are the most commonly transfused blood component. They are administered to restore the blood’s ability to carry oxygen, which may be diminished in patients with anemia or severe volume deficit caused by hemorrhage.

Platelets are administered to help restore the blood’s clotting ability. Thrombocytopenia (too few platelets) may cause severe and spontaneous bleeding.

Fresh-frozen plasma is an unconcentrated source of all clotting factors, and is used to treat bleeding caused by factor deficiencies and multifactor deficiency states such as liver failure and disseminated intravascular coagulation, and for urgent reversal of warfarin treatment.

Cryoprecipitated antihemophilic factor is a concentrate prepared from fresh-frozen platelets. Traditionally reserved for patients with bleeding disorders such as hemophilia and von Willebrand disease, it is also used as a source of fibrinogen.

Compatibility Testing
Before blood products are administered to patients, the products are tested for infectious diseases and typed (ABO testing). To prevent acute hemolytic transfusion reactions (AHTR), blood for transfusion should be of the same ABO type as the recipient’s. Symptoms of AHTR include an abrupt onset of discomfort and anxiety, usually within an hour of initiation of transfusion. Patients may have difficulty breathing, fever and chills, facial flushing, and severe pain, especially in the lumbar area. Evidence of shock may appear, with a rapid feeble pulse, cold clammy skin, dyspnea, a drop in blood pressure, and nausea and vomiting. If AHTR is suspected, the transfusion should be stopped immediately and supportive treatment begun. The first priority of therapy is to achieve and maintain adequate blood pressure and renal blood flow. Fluid therapy consists of intravenous administration of 0.9% sodium chloride solution. Antipyretics and antihistamines may also be used for symptomatic relief.

Patients should receive blood that matches their blood type. Type O RBCs may be used for any patient. In true emergency situations, or when the patient’s blood type is unknown, type O RBCs may be used (Table 1).



Rh typing determines whether the Rh factor is present (Rh-positive) or absent (Rh-negative). Rh-negative patients should always receive Rh-negative blood, except in life-threatening emergencies when Rh-negative blood is unavailable. Rh-positive patients may receive either Rh-positive or Rh-negative blood.

Antibody screening for unexpected antibodies is specific for red blood group antigens other than A and B. Antibody screening is performed routinely on pretransfusion samples from potential recipients, and on maternal blood prenatally. Early detection of these antibodies is critical, because they can cause hemolytic disease of newborns and serious transfusion reactions. For some patients who have received multiple transfusions over time, antibodies may make it difficult to find a compatible unit of blood for transfusion.

Indirect antiglobulin testing (indirect Coombs test) is performed to screen for unexpected anti-RBC antibodies. If an antibody is detected, its specificity is determined. Knowing the specificity of the antibody aids greatly in assessing clinical significance, selecting compatible blood, and managing complications.

Direct antiglobulin testing (direct Coombs test) identifies antibodies that have coated the patient’s RBCs. A positive result, when associated with clinical findings, may indicate autoimmune hemolytic anemia, a transfusion reaction, or other complications.

Transfusion Technique
Before a transfusion is started, blood is cross matched: a drop of the patient’s blood is mixed with the donor blood to confirm compatibility. The patient’s identification wristband, blood unit label, and compatibility test report must be checked to be sure that the blood is intended for the patient, that it is compatible, and that the component is the correct one. The test report should also be checked for special orders such as leukocyte-reduced or irradiated blood or premedications. These simple steps can eliminate some of the most serious complications of blood transfusion, such as AHTR.

An 18-gauge or larger needle is recommended to prevent damage to RBCs. Only 0.9% intravenous sodium chloride solution should be allowed into the blood bag or in the same tubing as the blood. Hypotonic solutions may lyse RBCs, and decrease RBC survival. Calcium in Ringer solution may cause clotting.¹

Transfusion of 1 unit of blood or blood product should be completed in 4 hours or less to avoid risk of bacterial growth. If transfusion must be given more slowly because of heart failure or hypervolemia, units should be separated into smaller aliquots or split units in the blood bank.

Close observation of the patient during transfusion is essential, especially during the first 15 minutes. Temperature, blood pressure, pulse, and respiratory rate should be noted. Periodic observation should be maintained during the transfusion.

Risks Associated With Blood Transfusion
Technical advances in laboratory testing of donor blood, and enforcement of strict donor-selection criteria have helped make transfusions today safer than ever. However, it is important to recognize that significant risks continue to be associated with blood transfusions. It may be several years before these risks are fully understood and described. What is known is that patients who receive allogeneic blood (blood from a donor) experience increased morbidity, and longer and more costly hospital stays.¹³

Part of this increased morbidity is the result of transfusion-transmitted infections such as hepatitis, HIV, cytomegalovirus, and human T-cell lymphotropic virus. Bacterial contamination of donor blood also presents a risk of infection. Bacterial infection from aged blood (blood stored more than 35 days) and transfusion-related immunosuppression have recently been foci of research. Patients with cancer, patients undergoing certain types of surgery, and critically ill patients face unique risks from transfusion, which are explored here.

Risk of Infection
Viral Infection
Hepatitis. Hepatitis is the most common transfusion-transmitted infection.¹⁴ These viruses have a long seronegative period when they cannot be detected by screening. Thus, donors may test negative for hepatitis, but in fact be infected with the virus. By the time hepatitis is detected, the donor blood may already be in use. High-risk factors for hepatitis can be identified through predonation screening questions that make certain patients ineligible to donate, helping decrease the chances of transmitting hepatitis. Hepatitis C accounts for more than 90% of transfusion-transmitted hepatitis cases, whereas 2% are attributed to Hepatitis B.15 Hepatitis can be transmitted by all blood components and most blood products, except albumin.¹⁴

HIV. Transfusion practices came under intense scrutiny in the early 1980s because of public health concerns about HIV-infected donor blood. Although less than 20 HIV cases per year are transfusion-related, this virus remains one of the most feared transfusion-transmitted infections for patients. Identification of high-risk behaviors among potential donors, and the use of sensitive enzyme immunosorbent assays have decreased the risk of HIV infection from donor blood. Transmission of HIV by blood transfusion occurs almost exclusively from blood that is collected when a donor is infected with HIV but has not yet produced antibodies. The length of this period has been estimated at 25 to 45 days.¹⁵ The risk of acquiring HIV from donor blood is about 1 in 493 000, which is much less than the risk of acquiring hepatitis C (1 in 103 000).¹⁷

Cytomegalovirus (CMV). This virus causes a significant number of infections as a result of blood transfusions. An estimated 60% of blood donors carry the virus.¹⁸ Transfusion of infected blood can cause CMV infection in the recipient. The virus is transmitted by all leukocyte-containing blood products, including whole blood and RBCs.

In critically ill patients, CMV is a major cause of increased morbidity and mortality. Infection with CMV may cause a transient but significant immunosuppression that has serious consequences for the critically ill patient: severe and possibly fatal disease such as pneumonia, gastroenteritis, and hepatitis.¹⁹

Posttransfusion infection with CMV can manifest within 2 to 4 weeks in immunocompetent patients but more quickly in critically ill patients. Symptoms include fever lasting 2 to 3 weeks, varying degrees of hepatitis, splenomegaly, and atypical lymphocytosis resembling that of mononucleosis.¹⁸

Human T-Cell Lymphotropic Virus Type I (HTLV-1). This is an oncogenic retrovirus that can cause adult T-cell lymphoma or leukemia, HTLV-associated myelopathy, and tropical spastic paraparesis. All donor blood is screened for HTLV-1 and HTLV-2 antibodies, with an estimated risk of false-negative results of 1 in 641 000.²⁰ The estimated risk of exposure to HTLV is 1.56 per million donations,²¹ which represents the probability that a unit was donated in the HTLV antibody-negative period. The average length of this seronegative period for HTLV is 36 to 72 days.

Bacterial Infection
Stored units of blood can become contaminated with bacteria and cause infection in the recipient of the transfused unit. The rate of fatal bacterial septicemia is estimated at 4 per 10 million transfusions.²² When it does occur, bacterial contamination can progress rapidly to endotoxic shock and death in immunocompetent and immunocompromised patients. The mortality rate from bacteremia associated with transfusion approaches 50%, even with proper antibiotic treatment.²²

Bacteria can enter blood products if there is inadequate skin disinfection at the phlebotomy site, asymptomatic bacteremia at the time of donation, minute leaks in blood containers, or contaminated containers.²¹ Several bacterial organisms, known as cold-loving or psychrophilic organisms, can proliferate under cold storage and use citrate—a common preservative for blood products—as a substrate. This group includes Pseudomonas, Yersinia, Campylobacter, and Serratia species, and some enteric bacteria.

Reports of life-threatening bouts of bacterial sepsis following transfusion of stored blood contaminated with various cold-loving organisms are not as uncommon as one would hope. The incidence of positive bacterial cultures from RBC and platelet units has been approximated at 0.3% for the past 15 years.²³ The concentration of bacteria is directly related to the length of storage.

The risk of bacterial infection is not limited to contaminated blood, however. In a retrospective analysis²⁴ of 9598 patients aged 60 years or older undergoing surgical repair of hip fractures, blood transfusion was associated with a 35% increase in risk of serious bacterial infection and a 52% increase in risk of pneumonia. The majority of patients in this study received at least 1 transfusion during their hospital stay. Serious bacterial infection occurred in 4.5% of patients, and 28.8% of these patients died in the hospital. Pneumonia complicated the course of 3.8% of patients, and urinary tract infection occurred in 12%. The investigators concluded that bacterial infection may be the most common life-threatening adverse effect of allogeneic blood transfusion.

Aged Blood
A typical 450-mL unit of whole blood, collected in a plastic bag containing an anticoagulant preservative, has a shelf life of 35 days. Red blood cells preserved with an adenine-saline solution may be stored for 42 days. A unit of packed RBCs contains platelets, white blood cells (WBCs), and plasma; however, the lifespan of WBCs and plasma is shorter than 42 days. As the WBCs die, they release enzymes that are toxic to cells. These mediators become significant after about 14 days of storage.²⁵

Complications from the use of aged blood include hypocalcemia caused by citrate binding, hyperkalemia caused by cell lysis, hyperammonia and hyperbilirubinemia due to breakdown of RBCs, and acidosis. In addition, use of older blood has been shown to be an independent risk factor for multiple organ failure (MOF).²⁶ Investigators identified 63 trauma patients at risk for developing MOF who had received 6 to 20 units of packed RBCs in the first 12 hours after injury. Of these 63 patients, 23 developed MOF. The mean age of blood transfused to the patients who developed MOF was significantly greater than that transfused to patients who did not develop MOF: 30.5 ±1.6 days versus 24 ±0.5 days. The mean number of units older than 14 and 21 days was also greater in patients who developed MOF. Mean age of blood, number of units older than 14 days, and number of units older than 21 days were found to be independent risk factors for MOF.

It is important to note that blood banks generally fill transfusion orders with the oldest units first. In theory, this practice would increase the risk of serious posttransfusion complications in certain patients.

In addition, older RBCs are less effective at unloading oxygen. One study of patients with sepsis on mechanical ventilation not only failed to demonstrate acute improvement in oxygen uptake after transfusion of 3 units of RBCs, but also found that patients who received RBCs older than 15 days developed evidence of splanchnic ischemia.²⁷

Risk of Immunosuppression
In addition to risk of infection, evidence of the profound negative effects of blood transfusion on the immune system seems to be accumulating. Although the underlying mechanisms of these effects are not completely understood, a number of immunomodulatory effects have been described.^28-30 These effects may persist for months or longer.

The immunosuppressive effects of blood transfusion were described first in the late 1970s. Patients who had received pretransplant blood transfusions were noted to experience decreased rejection and improved transplant outcomes. This observation was verified by evaluating the effect of planned blood transfusions on cellular immunity, as measured by lymphocyte responses. Within 1 week of transfusion, evidence of clear suppression was noted. This effect lasted for 4 weeks, but a second transfusion resulted in a more pronounced and long-lasting suppression.²⁸

Although the mechanism for this outcome was not completely clear, investigators theorized that the same immune effects that improved outcomes in patients undergoing kidney transplant might be impairing outcomes for nontransplant patients, especially those with repeated transfusions. Helper-suppressor (T4/T8) cell ratios and natural killer cell activity were measured in 4 groups: patients with sickle cell anemia who received monthly transfusions, patients with sickle cell anemia who received no transfusions, patients with hemophilia or severe von Willebrand disease, and normal patients.²⁹ The investigators found that patients with sickle cell anemia who received multiple transfusions had low T4/T8 ratios and low natural killer cell activity compared with controls.

To better describe the association between transfusion, immunosuppression, and postoperative infection, other studies have observed allogeneic transfusion versus autologous transfusion. The theory is that autologous blood transfusion, in which the patient’s own blood is used for transfusion, would not have the same immunosuppressive effects as donor blood. Although this theory had limitations in study design—a true control would mean withholding transfusions of any kind, and autologous blood donation was not always possible—the observations are compelling.

One study looked at 102 patients undergoing 109 spinal fusion procedures. In 60 of the procedures, patients received autologous blood only, in 24 they received at least 1 unit of allogeneic blood, and in 25 procedures they received no transfusions. Patients who received any allogeneic blood fared worse than those who received none. They had a higher rate of hospital-acquired infection, longer hospital stays, more days of fever, lengthier antibiotic treatments, longer duration of surgery, greater blood loss, and greater decrease in postoperative natural killer cells.³⁰ In this study, the most significant predictor of in-hospital infection was allogeneic transfusion.

Another study reached similar conclusions. Fifty recipients of autologous transfusion were matched and compared with 50 recipients of allogeneic transfusion. Investigators compared incidence of postoperative infection, postoperative leukocytosis, and febrile episodes. Patients in both groups had abnormally low hematocrits on admission. Patients who received allogeneic blood had longer procedures, greater blood loss, and received more units of blood. Cultures that were positive for bacteria were obtained for 16% of the allogeneic recipients and only 4% of the autologous recipients (P<.05). Seventeen patients who received allogeneic transfusions developed postoperative leukocytosis, compared with 12 in the autologous group.³¹

Much is yet to be learned about how blood transfusions suppress the immune system, but the evidence that this suppression takes place is troubling. Any suppression of immune response and increased risk of infectious complications are magnified in critically ill patients.

Tumor Recurrence
The association between transfusions and tumor recurrence has been described in many studies and in several different types of tumors. The results of these studies are not always clear-cut, preventing many investigators from reaching firm conclusions. The observed relationship between recurrence rate and number of allogeneic blood transfusions could reflect a population compromised by other factors such as worsening health. Results of these studies can suggest trends but cannot define cause and effect. Colorectal tumors, soft-tissue sarcoma, and breast cancer are some of the tumors that have been studied (Table 2).


x
The case against blood transfusions and cancer recurrence remains controversial. Available data do support an extremely conservative approach to transfusion in these patients.

Implications for Critically Ill Patients
For critically ill patients, the risk of blood transfusion–transmitted infection, eg, hepatitis, is not the greatest concern. Of greater consequence to critically ill patients is evidence that blood transfusion has profound negative effects on the immune system. As discussed, a variety of immunomodulatory effects associated with blood transfusion may persist for months or longer.

These effects may result from WBC components contained in RBC transfusions. Leukocyte reduction of transfused RBCs has been associated with a significant reduction in infection rates following colorectal surgery. Universal leukocyte reduction of transfused RBCs has been suggested as a way to reduce toxicity of RBC transfusion. However, the available data may not support the considerable expense involved with universal adoption of this change in transfusion practice, and many risks may still remain.

Currently, the best way to manage risks of transfusion is to adopt a conservative approach to blood transfusion in all patients. Alternatives to transfusion must be explored whenever possible.

Surgical Complications
Perioperative blood transfusions have been implicated as a risk factor for postoperative infection. These findings can be as controversial as those arising from studies of blood transfusion in patients with cancer. A retrospective analysis of various surgical procedures, including orthopedic, open heart, and colorectal surgeries, suggests that perioperative transfusions are associated with more wound infections and distant infections such as pneumonia and urinary tract infection. Results of these studies may not adequately account for confounding factors such as complications during surgery, presurgical patient status, and variations in surgical practice.

Some studies have attempted to control for some of these variables by looking at elective surgeries. Possibly the largest prospective study of transfusion requirements associated with total joint arthroplasty was conducted in the United States between September 1996 and June 1997.38 Three hundred thirty orthopedic surgeons participated, enrolling a total of 9482 patients undergoing total hip or knee replacement. Nearly half (46%) of these patients received an autologous or allogeneic transfusion. In this study, allogeneic transfusion was significantly associated with infection, fluid overload, and longer hospitalization.

Total Joint Replacement
Records from 84 patients who had hip replacement surgery and received 2 to 3 units of blood were evaluated.³⁹ Investigators looked specifically at whether transfusion with autologous blood or allogeneic blood affected outcome. Patients who received allogeneic blood had a 32% rate of proven or suspected infection, which was significantly greater than the 3% rate in patients who received autologous blood. Wound infection accounted for a fraction of these infections, suggesting that nonsurgical factors were involved. Accordingly, patients who received allogeneic blood had more days of antibiotic therapy and significantly longer hospital stays than patients who received autologous blood.

Spinal Surgery
The association among transfusion, quantitative immunological factors, and infection was evaluated in 102 patients undergoing 109 spinal fusion procedures.³⁰ As discussed previously, this study was the first to prospectively correlate laboratory measures of the immune system with postoperative bacterial infections in patients who received transfusions.

Allogeneic blood transfusion was significantly associated with the development of postoperative infection, longer hospital stays, more days of fever, lengthier antibiotic treatments, longer duration of surgery, greater blood loss, and greater postoperative drop in natural killer cells compared with no transfusion or autologous transfusion. A dose-response relationship between the development of infection and the number of units transfused was also observed. Of the patients with postoperative infection, the proportion of patients who received more than 2 units of blood during the hospitalization or at any time in the past was significantly higher than the proportion of patients who had received 1 or no units of blood. Patients who developed in-hospital infections had been exposed to more allogeneic blood than those who did not develop infection.

Two laboratory variables corresponded with a higher risk of infection that developed after the hospital stay. In patients who developed such infection, the number of natural killer cells dropped markedly on day 7 following surgery; however, hematocrit dropped less in the infected group on day 7 compared with the uninfected group. Overall, the drop in natural killer cells was strongly associated with allogeneic transfusion rather than postoperative infection.

Bowel Resection
A study evaluating the contribution, if any, of blood transfusion to the risk of postoperative complications in patients with Crohn disease who underwent bowel resection found the incidence of postoperative complications was significantly related to the number of units of blood transfused.⁴⁰ This study included 169 patients who were divided into 2 groups: 1 group received 2 or more units of blood, and 1 group received 1 unit or no blood at all. Septic complications developed in 26% of patients who had received 2 or more units of blood, compared to 8% of patients who received 1 unit or none.

Blood Management
Blood is a finite product, always in short supply. To ensure the availability of a plentiful supply of fresh blood, blood waste must be minimized. A conservative approach to transfusion based on physiological need, reduced phlebotomy loss, use of appropriate pharmacological agents to improve hemoglobin levels, and autologous donation will help maintain adequate supplies of blood while decreasing the risks of transfusion.

Transfusion Based on Physiological Need
Critically ill patients with anemia may be less able to compensate for diminished oxygen delivery, especially those with concomitant cardiovascular disease. As a normal adaptive response to increased oxygen demand or decreased oxygen supply, the heart increases blood flow through local and regional vasodilatory stimuli. When anemia limits oxygen supply, the coronary circulation may not be able to meet the increased demand. For patients who may be further compromised by myocardial dysfunction due to severe sepsis, a conservative approach to transfusion may result in higher morbidity and mortality.

One study evaluated outcomes in critically ill patients who were treated under restrictive and liberal transfusion strategies.⁴¹ This Canadian study included 838 critically ill patients with hemoglobin levels less than 9 g/dL within 72 hours of admission to the ICU. Four hundred eighteen patients were randomly assigned to the restrictive transfusion group and received transfusions only when hemoglobin levels fell below 7 g/dL. Hemoglobin level was maintained between 7 g/dL and 9 g/dL. In the liberal transfusion group, 420 patients received transfusion if hemoglobin levels fell below 10 g/dL. Hemoglobin level was maintained at 10 g/dL to 12 g/dL in this group.

Outcomes were mortality and severity of organ dysfunction. Although 30-day mortality was similar between the 2 groups (18.7% for restrictive, 23.3% for liberal), mortality was significantly lower in the restrictive group among patients who were less acutely ill and aged 55 years old or less with no evidence of clinically significant cardiac disease. This study did not determine an optimal hemoglobin level, but it showed the potential risks of a liberal transfusion policy.

In the absence of definitive guidelines, the goal remains to maintain a hemoglobin concentration that maximizes oxygen delivery and minimizes the need for transfusion.

Minimal Phlebotomy
Corwin et al⁷ found phlebotomy was a major contributing factor for transfusion in one ICU. In this study, a total of 176 liters of blood was drawn from 142 patients by phlebotomy for diagnostic purposes—an average of 65 mL per day. Patients were in the ICU for a week or more, with a mean stay of 19.6 days.⁷ This blood loss remained at a constant level throughout the ICU stay. Low hematocrit (<.25) was the only indication for an additional 19% of events. Patients who did not require blood transfusions had considerably less blood drawn than those who had transfusions.

Miller⁴² reviewed several studies that also investigated the problem of diagnostic blood loss in the ICU, and recommended strategies to reduce blood loss. Strategies such as use of small volume (pediatric) tubes, elimination of arterial line blood discard, timing of blood draws, and elimination of standing orders for laboratory tests can significantly reduce phlebotomy blood loss. Use of small-volume tubes alone was shown to decrease phlebotomy blood loss by 33% to 47%.^43,44

Two blood conservation systems designed to eliminate discard of blood are available on vascular devices used to draw blood.^45,46 These systems have potential for dramatically decreasing phlebotomy blood loss, especially in combination with in-line monitoring of laboratory parameters.

Pharmacological Agents
The search for a viable blood substitute has spurred imagination since the mid-19th century. Various agents, including hemoglobin dissolved in crystalloid solution and hemoglobin-lactated Ringer solution, have since been used with little success. Patients experienced renal and pulmonary failure after receiving these early blood substitutes. Investigation revealed that contamination of the solution with fragments (stroma) from the RBC wall triggered a transfusion reaction.⁴⁷

Stroma-free hemoglobin solutions that do not expose recipients to RBC antigens have been developed, and are currently in phase III clinical trials. (The Food and Drug Administration has not released any intravascular oxygen carriers for marketing.) These “oxygen therapeutics” include hemoglobin glutamer (Hemopure, Biopure Corp, Cambridge, Mass), Hemolink (Hemosol Inc, Toronto, Canada), and Polyheme (Northfield Labs, Evanston, Ill). Administered intravenously into the circulatory system, these RBC substitutes are intended to enhance oxygen delivery directly to body tissues.

Two major problems with these blood substitutes have yet to be solved. Although stroma-free hemoglobin binds oxygen readily, it does not release it when traveling through capillaries. Peripheral tissues may not receive needed oxygen. The second obstacle is the relatively short half-life (hours) of stroma-free hemoglobin. Over time, these obstacles may be overcome through chemical modification of the hemoglobin molecule.

The ultimate answer to zero-risk blood replacement may be in the development of artificial blood cells called neohemocytes. As yet, no substitute for human blood has been brought to market. Development issues remain, and it may be several years before a satisfactory product is available for testing.

Recombinant Human EPO. Although a viable blood substitute may not be available, recombinant human EPO (r-HuEPO, epoetin alfa) makes it possible to increase hemoglobin concentrations without transfusions. Ortho Biotech Inc (Raritan, NJ) markets epoetin alfa under the brand name Procrit. Epoetin alfa has the same effect as endogenous EPO, that is, to increase hemoglobin level.

Epoetin alfa has been prospectively studied in clinical trials with more than 10 000 patients in prospective trials and has been used widely and successfully in the treatment of various anemias—anemia associated with chronic renal failure, zidovudine-induced anemia in HIV, and chemotherapy-induced anemia—and in reduction of allogeneic blood transfusion in surgery patients. Epoetin alfa is contraindicated in patients with uncontrolled hypertension, known hypersensitivity to mammalian-cell derived products, and known hypersensitivity to albumin.

In prospective clinical trials of patients with cancer, therapy with epoetin alfa has been shown to reduce the need for blood transfusion, improve and maintain RBC levels, and improve quality of life.^48-50 Similarly, studies evaluating the use of epoetin alfa in patients undergoing urogenital surgery have found that epoetin alfa decreased the need for autologous or allogeneic blood, and maintained hemoglobin and hematocrit levels.^51,52 Results from some of these studies are summarized in Table 3.


x
Although studies are ongoing, it seems that critically ill patients may also benefit from treatment with epoetin alfa. In the only published study of epoetin alfa in critically ill patients, Corwin et al⁵⁷ found a dramatic reduction in the number of blood transfusions required. In this prospective, randomized, double-blind, placebo-controlled trial of 160 patients admitted to the ICU, patients were randomized to receive either epoetin alfa (n=80) or placebo (n=80) by subcutaneous injection. Injections were administered daily from ICU day 3 through ICU day 7, for a total of 5 days. Subsequent dosing was every other day to achieve a hematocrit of greater than .38. The outcomes of this study are described in Table 4.


x
The markedly positive response to epoetin alfa in these patients supports the theory that EPO response may be blunted in critically ill patients.

The optimal effective dose of epoetin alfa and timing of administration for treatment of anemia in critically ill patients have not yet been determined. Current ongoing clinical trials are examining a dose of 40 000 units given weekly. The body of evidence clearly shows a potential benefit for use of epoetin alfa in the ICU for reduction of transfusion and its risks.
Epoetin alfa is currently used for the treatment of anemia in patients with HIV, cancer, early renal insufficiency, and patients undergoing elective surgery.

Autologous Transfusion
One of the safest and most effective ways to treat blood loss is to give patients their own blood through preoperative donation, preoperative hemodilution, intraoperative salvage, or postoperative salvage of autologous blood. Replacement of lost blood with previously donated or salvaged shed blood from the patient is effective and eliminates virtually all transfusion-associated risks except errors associated with collection, storage, identification, and administration.

Preoperative Donation. Preoperative donation of autologous blood is ideal before elective procedures with a known potential for blood loss, such as orthopedic surgery. Advantages to preoperative donation include decreased use of banked blood, decrease in erythrocyte mass, and loss of erythrocytes during surgery, and stimulation of erythropoiesis.

However, limitations exist, such as patients who are not eligible because of low hemoglobin levels (<11 g/dL) or cardiac conditions. Patients must be selected carefully and should be screened using the same criteria as patients donating allogeneic blood.⁵⁸ Patients who did not meet the basic American Association of Blood Banks criteria for allogeneic donation, (hemoglobin level of more than 12.5 g/dL, aged 17 years or more, weight of 50 kg or more) have had higher rates of adverse reactions following autologous transfusion. Preoperative autologous donation is associated with additional costs, including those for collection, testing, and typing.

Preoperative Hemodilution. Normovolemic hemodilution is a form of autologous transfusion. Hemodilution is used extensively in cardiothoracic and urological procedures. During normovolemic hemodilution, the concentration of erythrocytes is diluted without reducing the intravascular volume. As many as 4 units of whole blood can be removed, stored temporarily, and replaced with crystalloid or colloid solution. Hemodynamic effects include increased cardiac output caused by decreased viscosity and increased venous return. Patients with cardiac disease may be adversely affected by hemodilution because of the increased demand on the heart. One study⁵⁹ found hemodilution as effective as autologous blood donation in reducing the need for allogeneic blood.

Intraoperative Salvage of Blood. Intraoperative salvage of shed blood from wound drainage, with subsequent reinfusion, is a method of blood replacement. Various systems are available to collect blood for salvage during surgery; some collect and reinfuse whole blood and others collect, wash, and reinfuse packed RBCs.

When whole blood is collected for reinfusion during surgery, it is suctioned from the wound into a collection and reinfusion device. The shed blood is anticoagulated with heparin or citrate-phosphate-dextrose solution, then filtered and reinfused. This shed blood contains erythrocytes, platelets, fibrinogen, and clotting factors.

The Cell Saver (Haemonetics, Braintree, Mass) is a device that not only collects and reinfuses shed blood, but also washes and spins blood to remove cellular debris and hemolytic byproducts. The shed blood is filtered and reinfused as packed red blood. The Cell Saver can also be used in the recovery room to wash and spin shed blood from suction catheters used to drain wounds. The rate of salvage of RBCs is 50% to 60% when these techniques are applied meticulously.⁶⁰

The efficacy of intraoperative, washed, autologous shed blood for transfusion in total joint arthroplasty was demonstrated by Semkiw et al.⁶¹ This type of blood replacement is recommended for orthopedic procedures in which blood loss is expected to be greater than 900 mL.

Systems that deliver intraoperative shed whole blood are simpler and less expensive to use than those that wash blood. Systems that do not wash blood collect blood through a suction device that adds heparin or citrate-phosphate-dextrose solution anticoagulant to a collection chamber.

Postoperative Salvage of Blood. Postoperative salvage of shed blood has also been shown to decrease the use of allogeneic blood. However, use of unwashed autologous shed blood from wound drainage for transfusions after orthopedic surgery is controversial. Wound drainage may be contaminated by fat particles, bone fragments, or methylmethacrylate monomer, in addition to vasoactive mediators, clotting factors, fibrin degradation products, and free hemoglobin, which are part of a healing wound.

Despite these factors, the clinical efficacy of postoperative autologous transfusion with unwashed shed blood has been shown in several studies; yet questions regarding the safety of this practice persist. Febrile reactions have been associated with reinfusion of unwashed autologous shed blood. Unwashed shed blood may contain soluble products such as vasoactive agents and free hemoglobin, which may cause adverse reactions such as hypotension or hyperthermia. Clotting factors in unwashed shed blood raise another issue. These clotting factors theoretically may cause clotting abnormalities with infusion of unwashed autologous blood; however, this effect has not been borne out in clinical trials.

Conclusion
Anemia is a serious condition, particularly in critically ill patients. These patients are more susceptible to the adverse effects of oxygen deficiency associated with even mild to moderate anemia. They may be less able to mount the appropriate response—increase in EPO production—in the face of a drop in arterial oxygen.

Transfusion remains the most common treatment for anemia in critically ill patients, regardless of etiology. Although accurate typing and testing of donor blood have made transfusions safer than ever, significant risks are associated with transfusions. The presence of equivocal findings about the risks of blood transfusion in certain patient groups makes it difficult to fully define these risks, however. It may be several years before we fully understand the nature of the risks from blood transfusion, but these potential risks require that we scrutinize our use of blood transfusions, and seek out alternatives when possible.

The risks of allogeneic blood transfusion may be minimized by reduction of blood loss during surgery, predonation of autologous blood, use of epoetin alfa, and use of small-volume tubes for phlebotomy. In any situation, a more measured approach to blood transfusions is likely to result in optimal patient care and outcome.

References
1. Beers MH, Berkow R, eds. The Merck Manual of Diagnosis and Therapy. 1999;17:854-856, 885-895, 996.
2. Corwin HL. Blood transfusions in the critically ill patient. Dis Mon. 1999;45:409-426.
3. Labovich TM. Transfusion therapy: nursing implications. Clin J Oncol Nurs. 1997;1:61-72.
4. Walter-Coleman S. Transfusion therapy for patients critically ill with cancer. AACN Clin Issues. 1996;7:37-45.
5. Bunn HF. Anemia. In: Wilson JD, Braunweld E, Isselbacher KJ, et al, eds. Harrison’s Principles of Internal Medicine. 12th ed. New York, NY: McGraw-Hill Inc; 1991;61:344-348.
6. Smoller BR, Kruskall MS. Phlebotomy for diagnostic laboratory tests in adults: pattern of use and effect on transfusion requirements. N Engl J Med. 1986;314:1233-1235.
7. Corwin HL, Parsonnet KC, Gettinger A. RBC transfusion in the ICU: is there a reason? Chest. 1995;108:767-771.
8. Krafte-Jacobs B. Anemia of critical illness and erythropoietin deficiency. Intensive Care Med. 1997;23:137-138.
9. Rogiers P, Zhang H, Leeman H, et al. Erythropoietin response is blunted in critically ill patients. Intensive Care Med. 1997;23:159-162.
10. Levine EA, Rosen AL, Sehgal LR, Gould SA, Egrie JC, Moss GS. Erythropoietin deficiency after coronary artery bypass procedures. Ann Thorac Surg. 1991;51:764-766.
11. Patterson WP, Reams GP. Renal toxicities of chemotherapy. Semin Oncol. 1992;19:521-528.
12. Coyle TE. Hematologic complications of human immunodeficiency virus infection and the acquired immunodeficiency syndrome. Med Clin North Am. 1997;81:449-470.
13. Vamvakas EC, Craven JH. Allogeneic blood transfusion, hospital charges, and length of hospitalization: a study of 487 consecutive patients undergoing colorectal cancer resection. Arch Pathol Lab Med. 1998;122:145-151.
14. Stites DP, Stobo JD, Wells JV, eds. Basic and Clinical Immunology. 6th ed. Norwalk, Conn: Appleton and Lange; 1987;19:310-312, 341.
15. Klein HG. Allogeneic transfusion risks in the surgical patient. Am J Surg. 1995;170(suppl):21S-26S.
16. Lackritz EM, Satten GA, Grasse-Aberle J, et al. Estimated risk of transmission of the human immunodeficiency virus by screened blood in the United States. N Engl J Med. 1995;333:1721-1725.
17. Schreiber GB, Busch MP, Kleinman SH, Korelitz JJ. The risk of transfusion-transmitted viral infections. N Engl J Med. 1996;334:1685-1690.
18. Beers MH, Berkow R, eds. The Merck Manual of Diagnosis and Therapy. Whitehouse Station, NJ: Merck Research Laboratories; 1999;17:1295-1296.
19. Waymack JP. Sequelae of blood transfusions. Infect Surg. July 1990:41-47.
20. Koll BS, Brown AE. Infections associated with blood-product transfusions in patients with cancer. Infect Med. 1993;10:12, 17-18.
21. Menitove J. Transfusion-transmitted infections: update. Semin Hematol. 1996;33:290-301.
22. Heltberg G, Silkov P, Gerner-Smidt HJ, et al. Nosocomial epidemic of Serratia marescens septicemia ascribed to contaminated blood transfusion bags. Transfusion. 1993;33:221-227.
23. Yomtovian R, Lazarus HM, Goodnough LT, et al. A prospective microbiologic surveillance program to detect and prevent the transfusion of bacterially contaminated platelets. Transfusion. 1993;33:902-909.
24. Carson JI, Altman DG, Duff A, et al. Risk of bacterial infection associated with allogeneic blood transfusion among patients undergoing hip fracture repair. Transfusion. 1999;39:694-700.
25. Silliman CC, Clay KL, Thurman GW, et al. Partial characterization of lipids that develop during routine storage of blood and prime the neutrophil NADPH oxidase. J Lab Clin Med. 1994;124:684-694.
26. Zallen G, Offner P, Moore E, et al. Age of transfused blood is an independent risk factor for postinjury multiple organ failure. Am J Surg. 1999;178:570-572.
27. Marik PE, Sibbaid WJ. Effect of stored blood transfusion on oxygen delivery in patients with sepsis. JAMA. 1993;269:3024-3029.
28. Fischer E, Lenhard V, Seifert P, Kluge A, Johannsen R. Blood transfusion-induced suppression of cellular immunity in man. Human Immunol. 1980;3:187-194.
29. Kaplan J, Sarnaik S, Gitlin J, Lusher J. Diminished helper/suppressor lymphocyte ratios and natural killer activity in recipients of repeated blood transfusions. Blood. 1984;64:308-310.
30. Triulzi DJ, Vanek K, Ryan DH, Blumberg N. A clinical and immunologic study of blood transfusion and postoperative infection in spinal surgery. Transfusion. 1992;32:517-524.
31. Mezrow CK, Bergstein I, Tartter PI. Postoperative infections following autologous and homologous blood transfusions. Transfusion. 1992;32:27-30.
32. Burrows L, Tartter P. Blood transfusions and colorectal cancer recurrence: a possible relationship. Transfusion. 1983;23:419.
33. Chung M, Steinmetz OK, Gordon PH. Perioperative transfusion and outcome after resection for colorectal carcinoma. Br J Surg. 1993;80:427-432.
34. Heiss MM, Mempel W, Delanoff C, et al. Blood transfusion-modulated tumor recurrence: first results of a randomized study of autologous versus allogeneic blood transfusion in colorectal cancer surgery. J Clin Oncol. 1994;12:1859-1867.
35. Busch ORC, Hop WCJ, Hoynck van Papendrecht MA, Marquet RL, Jeekel J. Blood transfusions and prognosis in colorectal cancer. N Engl J Med. 1993;328:1372-1376.
36. Rosenberg SA, Seipp CA, White DE, Wesly R. Perioperative blood transfusions are associated with increased rates of recurrence and decreased survival in patients with high-grade soft-tissue sarcomas of the extremities. J Clin Oncol. 1985;3:698-709.
37. Tartter PI, Burrows L, Papatestas AE, Lesnick G, Aufses AH. Perioperative blood transfusion has prognostic significance for breast cancer. Surgery. 1985;97:225-229.
38. Bierbaum BE, Callaghan JJ, Galante JO, Rubash HE, Tooms RE, Welch RB. An analysis of blood management in patients having a total hip or knee arthroplasty. J Bone Joint Surg Am. 1999;81:2-10.
39. Murphy P, Heal JM, Blumberg N. Infection or suspected infection after hip replacement surgery with autologous or homologous blood transfusions. Transfusion. 1991;31:212-217.
40. Tartter PI, Driefuss RM, Malon AM, Heimann TM, Aufses AH. Relationship of postoperative septic complications and blood transfusions in patients with Crohn’s disease. Am J Surg. 1988;155:43-48.
41. Hebert P, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J Med. 1999;340:409-417.
42. Miller KH. Decreasing diagnostic blood loss for patients in critical care units. Dimens Crit Care. 1999;18:46-54.
43. Foulke CE, Harlow DJ. Effective measures for reducing blood loss from diagnostic laboratory tests in intensive care unit patients. Crit Care Med. 1989;17:1143-1145.
44. Smoller BR, Kruskall MS, Horowitz GL. Reducing adult phlebotomy blood loss with the use of pediatric-sized blood collection tubes. Am J Clin Pathol. 1989;91:701-703.
45. Peruzzi WT, Parker MA, Lichthenthal PR, et al. A clinical evaluation of a blood conservation device in medical intensive care unit patients. Crit Care Med. 1993;21:501-506.
46. Silver MJ, Jubran H, Stein S, et al. Evaluation of a new blood-conserving arterial line system for intensive care units. Crit Care Med. 1993;21:507-511.
47. DeVenuto F, Friedman HI, Neville JR, et al. Appraisal of hemoglobin solution as a blood substitute. Surg Gynecol Obstet. 1979;149:417-436.
48. Abels RI, Larholt KM, Krantz KD, Bryant EC. Recombinant human erythropoietin (r-HuEPO) for the treatment of anemia of cancer. Oncologist. 1996;1:140-150.
49. Demetri GD, Kris M, Woods J, Degos L, Cella D. Quality of life benefit in chemotherapy patients treated with epoetin alfa is independent of disease response or tumor type: results from a prospective community oncology group. J Clin Oncol. 1998;16:3412-3425.
50. Glaspy J, Bukowski R, Steinberg D, Taylor C, Tchekmedyian S, Vadhan-Raj S. Impact of therapy with epoetin alfa on clinical outcomes in patients with nonmyeloid malignancies during cancer chemotherapy. J Clin Oncol. 1997;15:1218-1234.
51. Chun TY, Martin S, Lepor H. Preoperative recombinant human erythropoietin injection versus preoperative autologous blood donation in patients undergoing radical retropubic prostatectomy. Urology. 1997;50:727-731.
52. Monk TG, Goodnough LT, Andriole GL, Colberg JM, Lemon DJ, Marin KJ. Preoperative recombinant human erythropoietin therapy enhances the efficacy of acute normovolemic hemodilution [abstract]. Anesth Analg. 1995;80:S320.
53. Stowell CP, Chandler H, Jove M, Guilfoyle M, Wacholtz M. An open-label, randomized study to compare the safety and efficacy of perioperative epoetin alfa with preoperative autologous blood donation in total joint arthroplasty. Orthopedics. 1999;22(suppl):105s-112s.
54. Rosenblum N, Levine MA, Handler T, Lepor H. The role of preoperative epoetin alfa in men undergoing radical retropubic prostatectomy. J Urol. 2000;163:829-833.
55. D’Ambra MN, Gray RJ, Hillman R, et al. Effect of recombinant human erythropoietin on transfusion risk in coronary bypass patients. Ann Thorac Surg. 1997;64:1686-1693.
56. Laupacis A. Effectiveness of perioperative epoetin alfa in patients scheduled for elective hip surgery. Semin Hematol. 1997;33:51-53.
57. Corwin HL, Rodriguez RM, Pearl RG, Corwin MJ, Enny C, Gettinger A. Efficacy of recombinant erythropoietin in critically ill patients: a randomized, double-blind, placebo-controlled trial. Crit Care Med. 1999;27:2346-2350.
58. Lemos MJ, Healy WL. Current concepts review: blood transfusion in orthopedic operations. J Bone Joint Surg. 1996;78:1260-1270.
59. Ness PM, Bourke Dl, Walsh PC. A randomized trial of perioperative hemodilution versus transfusion of preoperatively deposited autologous blood in elective surgery. Transfusion. 1992;32:226-230.
60. McMurray MR, Birnbaum MA, Walter NE. Intraoperative autologous transfusion in primary and revision total hip arthroplasty. J Arthroplasty. 1990;5:61-65.
61. Semkiw LB, Schurman DJ, Goodman SB, Woolson ST. Postoperative blood salvage using the Cell Saver after total joint arthroplasty. J Bone Joint Surg. 1989;71:823-827.

Take The CE Test