Criticial Care Nurse
December 2002
Ronald G. Pearl
Anne Pohlman


Full Text Article:

Understanding and managing anemia in critically Ill patients

1. Understand the concepts underlying "transfusion triggers" and become familiar with both restrictive and liberal institutional transfusion policies.
2. Discuss differences in the compensatory mechanisms that occur in relatively healthy patients versus the critically ill during anemia and understand the impact on ICU patient management.

3. Differentiate subsets of anemic ICU patients who may have exceptional oxygen demands, eg, patients with myocardial ischemia, multiple organ failure, sepsis, and pulmonary edema, and describe potential complications of inadequate hemoglobin levels.

4. List the infectious and noninfectious risks of blood transfusion and in particular, the risks of administering older blood to ICU patients.

5. Identify the rationale for using epoetin alfa as a potential alternative to transfusion in the ICU and describe current research pertaining to epoetin alfa in the critically ill.

As the entire practice of medicine has evolved in the past few decades, so have transfusion practices. Two decades ago, a hematocrit between 0.20 and 0.25 was considered an urgent indication for transfusion, but at the turn of this century, maintaining a hematocrit at this level is considered to be "best-practice medicine." Statistics relating to transfusions, blood usage requirements, and availability are extensively reviewed elsewhere. (1,2) In brief, approximately 14 million units of blood are collected in the United States per year. Of these, approximately 12 million units are transfused, and 2 million units are not transfused within their useful storage life and must be discarded. (3) The demand for blood has been projected to shortly exceed supplies, primarily because donations have been decreasing or are increasingly rejected because of stringent donor screening standards. (4,5) Another factor that is likely to diminish future blood supplies is the fact that the pool of eligible, healthy donors is agin g. (1) To avoid critical blood shortages or cancellations of elective surgery, institutional transfusion guidelines have become more restrictive in recent years. A recent systematic review of the literature indicated that hemoglobin transfusion triggers currently vary between 70 g/L and 100 g/L (equivalent to 7.0 g/dL and 10.0 g/dL), and are most often between 80 g/L and 90 g/L. (6) In the intensive care unit (ICU), however, transfusion triggers have not been optimally defined, and the adoption of transfusion guidelines is highly variable. (7)

The cost of blood to hospitals and outpatient clinics is also increasing, and healthcare economics are sure to affect transfusion practices. (8) If one includes the costs associated with the inherent risks of transfusions, variable estimates of the cost per unit of blood are between $300 and $1000. (8-12) New technologies aimed at purification of the donor blood supply (13,14) and programs created to eliminate clerical errors and avoid transfusion reactions also contribute to the ever-increasing cost of blood and blood products. (15) Perioperative patients use two thirds of the total transfused units in the United States, (6) many of them in the ICU, and each patient who gets a transfusion receives on average 4 to 10 units. Thus, expenses attributed to acquiring blood and administering transfusions affect individual patients as well as the hospital's overall operating budget.


Anemia is prevalent in critically ill patients and is similar in clinical presentation to the anemia of chronic disease, characterized by inadequate secretion of endogenous erythropoietin in response to a decreasing hematocrit. Nearly two thirds of patients admitted to an ICU have hemoglobin levels of 120 g/L or less. Although intraoperative blood loss and gastrointestinal hemorrhage contribute to these statistics and are a frequent reason for administering transfusions, only 35% of transfusions administered in ICUs are associated with acute blood loss. (7) Similarly, in another multidisciplinary ICU, 65% of transfusions were administered because of nonacute blood loss, and transfusion requirements in the ICU were constant at 2 to 4 units per week. (16) Of the nonacute indications for transfusions, 63% were administered for low hematocrit alone or had no reason associated with the transfusion event. (16)

Phlebotomy for diagnostic tests can result in nosocomial anemia and accounts for approximately 30% to 50% of transfusions administered to ICU patients. (16,17) In a quantitative study conducted in an ICU in the United Kingdom, (18) phlebotomy for blood gas analyses in patients receiving mechanical ventilation accounted for a significant loss of blood compared with patients not receiving mechanical ventilation (P<.00 1). Mean blood loss was greatest (mean value = 55.2 mL/day) in a subgroup of patients who were in acute renal failure and undergoing renal replacement therapy; blood loss was significantly higher in that subgroup than in patients who were not in renal failure (P=.007). (18)

Decreased survival of red blood cells (RBCs) is another cause of anemia observed in critical care settings. Disseminated intravascular coagulopathy and hemolysis resulting from a number of diverse origins can produce anemia in critically ill patients. Examples include babesiosis, (19) HELLP syndrome (hemolysis, elevated liver enzymes and low platelet count), (20) and acute hepatitis. (21)

Anemia may also be caused by decreased production of RBCs, resulting from bone marrow diseases, hematopoietic injuries resulting from chemotherapy, nutritional ([B.sub.12], folate, and iron) deficiencies, inadequate endogenous erythropoietin production, or a blunted erythropoietin response to anemic challenge. In a large, randomized, prospective 3-center study comparing epoetin alfa with RBC transfusions in a multidisciplinary ICU setting, (10) a substudy was designed to evaluate the cause of anemia in critical illness. (22) Iron deficiency was defined as a ratio of iron to total iron binding capacity of less than 15% and a ferritin level of less than 100 ug/L. Folate, vitamin [B.sub.12], and iron deficiency occurred infrequently, and when grouped together, occurred in only 13% of patients. Most patients had low ratios of iron to total iron binding capacity but also had elevated ferritin levels consistent with laboratory values in patients with anemia of chronic disease. A blunted erythropoietin response asso ciated with low levels of reticulocytosis was a major contributor to the persistent anemia observed. Interestingly, a blunted response to endogenous erythropoietin or deficient erythropoietin production has been cited as the cause of anemia of chronic disease and is well documented in chronically ill patients. (23, 24)


Anemia will develop in approximately 75% of critically ill patients in the ICU, (25) and one third of patients admitted to an ICU are likely to receive a transfusion. (26) Although the risks of blood transfusion have been described elsewhere, (1,2,27,28) a brief summary of the most commonly encountered risks of transfusions is included to emphasize the particular impact of these risks on critically ill patients.

Noninfectious Complications

Fatal hemolytic reactions resulting from transfusions of ABO-incompatible blood are largely attributable to clerical error and should be preventable. (29,30) These errors still account for approximately 10 deaths per year in the United States, (1) and the true number may be significantly greater because of underreporting. A common noninfectious complication of transfusion is volume overload, occurring in 0.5% to 1% of transfusions. (2) Volume overload poses particular risks to patients with pulmonary edema or congestive heart failure. Transfusion-related acute lung injury is a complication estimated to occur in 1 of every 5000 transfusions. Because the characteristics of transfusion-related acute lung injury resemble those of adult respiratory distress syndrome (ie, dyspnea and hypoxia due to noncardiogenic pulmonary edema), this syndrome may be underreported. (1) Febrile reactions associated with the infusion of white cell antigens present in RBC units occur in 1% of patients receiving transfusions. (31) How ever, many more febrile reactions related to transfusions in the ICU may go unrecognized as such.

Infectious Complications

The viral safety of transfused blood products in the United States depends on the careful selection of donors and viral screening of donations. Despite selection and screening efforts, a threat to the blood supply is still posed by seronegative donors who donate during the infectious period when they are undergoing seroconversion. Schreiber et al (32) estimated the risk of viral transmission from blood donated during this infectious period as follows: for human immunodeficiency virus, 1 in 493 000; for human Tlymphotrophic virus, 1 in 641 000; for hepatitis C virus, 1 in 103 000; and for hepatitis B virus, 1 in 63 000. They calculated the aggregate risk from any of these viruses at 1 in 34 000 per unit transfused.

In addition to viral contaminants that can be tested for, current blood screening methods do not exclude the potential for donations of blood contaminated with hepatitis A virus or human parvovirus B19. Almost 40% of persons in the United States are seropositive for hepatitis A virus, indicating prior infection. (33) For human parvovirus B19, the range of expected viremia in donors has varied between 1 in 3300 and 1 in 40 000 in published studies. (34) Other infectious agents that could contaminate the blood supply include cytomegalovirus (35) and the agents that cause Creutzfeldt-Jakob disease, bovine spongiform encephalopathy, (36) and malaria. (37) Bacterial infections from blood transfusions, most frequently Yersinia enterocolitica, also pose risks, and fatal outcomes have been reported. (1,38)

Multiple published studies and reviews describe an increased risk of infection associated with transfusions, and these reports often refer to patients who are critically ill or postoperative. (38-43) Data generated from Project Impact (sponsored by the Society of Critical Care Medicine) indicate that ICU patients who receive transfusions have rates of nosocomial infection that are nearly triple the rate in the general ICU population (5.9% compared with 15.6%). (39) Compared with ICU patients who do not receive transfusions, this rate represents a more than 5-fold increase. For each unit of packed RBCs (PRBCs) administered, the odds of a nosocomial infection developing increased by a factor of 1.5. In a prospective study of 740 patients with colorectal cancer undergoing elective resection, multivariate analysis indicated that the risk of death was significantly greater in patients who developed infection after transfusion than in patients who neither received transfusions nor developed infections. The combinat ion of infection and transfusion in this study led to a particularly poor prognosis. (44) Houbiers et a1 (45) observed that the risk of infection developing can increase disproportionately with the number of units transfused. In their study of colorectal cancer patients undergoing surgery, a 60% increase in risk was noted if patients received 1 to 3 units of blood and the risk was more than tripled if greater than 3 units were administered. (45)

Immunologic Consequences of Blood Transfusions

Unwanted allogeneic leukocytes in transfused blood have been associated with immune modulation that can exacerbate infectious outcomes and lead to other adverse postoperative sequelae, including graft-versus-host disease (46) and impaired wound healing. (47) Suppression of natural killer cell activity in colorectal cancer patients (48) and tumor recurrence have been linked to transfusions, (49) although recent reports (50,51) suggest that no link exists between cancer recurrence and the presence of allogeneic leukocytes in blood products.

In a randomized study of 914 cardiac surgery patients who required transfusions, (52) removal of leukocytes from PRBCs by filtration resulted in significantly reduced infection rates (from 23% to 17%) and lower 60-day mortality (from 8% to 4%) when more than 3 units of blood were transfused. Fresh filtered PRBCs and PRBCs stored after filtration yielded similar results. These authors speculated that the presence of activated inflammatory cytokines in PRBCs can enhance the systemic inflammatory response. Because critically ill patients in an ICU are at high risk of sepsis, (53) transfusions with unfiltered PRBCs may potentiate adverse outcomes. In spite of these observations, there are currently no requirements in the United States for PRBCs to be filtered before administration.

Aged Blood and Risks to the Critically Ill

Blood that has been collected and stored has a limited shelf life, and requirements for storage of whole blood differ from the requirements for storage of its individual components. Following collection, the shelf life of RBCs ranges from 21 to 42 days and depends on the storage solution, processing, and system used for storage. (15) There are observed differences between "fresh RBCs" and RBCs near their expiration date that may cause deleterious effects if older RBCs are administered to a critically ill patient. Parameters that change during the shelf life of RBC units include the following: increased fragility (decreased deformability) of RBC membrane due to oxidation, diminished levels of 2,3-diphosphoglyceric acid impairing oxygen delivery, decreasing pH (acidosis), diminished oxygencarrying capacity, decreased number of viable cells per unit, and an increase in inflammatory cytokines released by contaminating leukocytes. (54-56) Clinical evidence also suggests that the frequency of nosocomial infections increases with the length of RBC storage. (57)

Any or all of these characteristics of older blood can adversely affect the anticipated benefit of a transfusion in a critically ill patient. In a retrospective study of 31 patients admitted to an ICU with sepsis, Purdy et al (58) examined the effect of the age of the PRBC units administered on mortality. Survivors received a greater proportion of PRBCs less than 10 days old (85%), whereas 76% of nonsurvivors received blood more than 20 days old. A highly significant negative correlation (r=-0.73, P<.001) was noted between survivors and the age of PRBC units transfused, with increased mortality being associated with the administration of older blood. Although this study was preliminary and limited because of its retrospective design, it suggests that older blood may be harmful. In a prospective study of 23 ICU patients with sepsis (55) designed to determine the effect of gastrointestinal and whole body oxygen uptake, no beneficial effect of older transfused blood could be demonstrated. Patients received an av erage of 3 units of RBCs. Changes in splanchnic (visceral) oxygen availability were measured by tonometry to detect intramucosal gastric pH, and systemic oxygen uptake was measured by calorimetry. A significant inverse relationship was found between the change in intramucosal gastric pH and the age of the RBCs transfused. Patients who received RBCs stored for more than 15 days had a consistent decrease in intramucosal gastric pH. Acidosis of the gastrointestinal mucosa was reversed if newer blood was infused. Although oxygen delivery was increased, oxygen consumption was not. The authors postulated that the most likely cause of the observed ischemia was splanchnic microcirculatory obstruction by poorly deformable RBCs present in the older stored blood.


A transfusion trigger may be defined as the hemoglobin or hematocrit value below which the risks of decreased oxygen-carrying capacity outweigh the risks of transfusion. (59) The trigger to initiate an order to administer an RBC transfusion is not necessarily the same as the "optimal" hematocrit; moreover, for a critically ill patient, the optimal hematocrit and the hematocrit that triggers ordering a transfusion may be very different. Achieving an appropriate balance of benefit and risk for a transfusion trigger is challenging in any setting, but particularly in the ICU, because tissue oxygen demand may be increased and stored RBCs may pose additional risks to critically ill patients. (58)

In deciding on appropriate transfusion algorithms, the traditional consideration has been the balance between oxygen delivery ([Do.sub.2]) and oxygen consumption ([Vo.sub.2]) (60) (Figure 1). As [Do.sub.2] decreases, the point of critical oxygen delivery (COD) is reached when complete [Vo.sub.2] cannot be sustained. Below this point, anaerobic metabolism begins and lactic acidosis ensues. (61) As anemia progresses in otherwise healthy persons, compensatory mechanisms are recruited and a certain degree of anemia may be tolerated. Decreased blood viscosity leading to improvements in cardiac output and stroke volume is the major compensatory mechanism; increased efficiency of oxygen extraction within the microcirculation is another. (62)

As the hematocrit decreases from its normal range (about 0.45) down to 0.30 and 0.20, the observed increase in cardiac output is proportionally greater than the decrease in hemoglobin level. (63) In order to maintain adequate tissue oxygenation, [DO.sub.2] peaks at a hematocrit of 0.30 and does not decrease to "normal" [DO.sub.2] values until the hematocrit reaches approximately 0.20. Experimental and clinical data suggest that the COD is in the range of 8 to 10 mL/kg per minute, which can be maintained at a hemoglobin level of 40 g/L. (60) The [SVO.sub.2] (mixed venous oxygen saturation) in this case is low (approximately 35 mm Hg), but still acceptable. In a review of reports about patients who refuse transfusions because of religious beliefs, survival predominated as long as the hemoglobin level did not decrease below 50 g/L. (64)

In relatively healthy patients or healthy persons, anemia is apparently well tolerated, at least for short periods. In a study of healthy volunteers (n = 21) and healthy patients about to undergo surgery (n= 11), Weiskopf et al (65) induced acute normovolemic hemodilution to a hemoglobin level of 50 g/L and observed no adverse consequences. As expected, the cardiac index (calculated as cardiac output in liters per minute divided by body surface area in square meters) increased from 3.05 to 5.71, heart rate increased from 58/min to 92/min, and the systemic vascular resistance index decreased 2.5-fold. As evidence that these compensatory mechanisms were effective in these healthy persons, [DO.sub.2] decreased only slightly from 13.5 to 10.7 mL/kg per minute and [VO.sub.2] was adequately maintained (increased from 3.07 to 3.42 mL/kg per minute). Neither lactic acidosis nor myocardial ischemia was observed. These observations were acknowledged by Weiskopf et al (65) to apply only to healthy subjects at rest and c ould not be extrapolated to postoperative patients or patients with acute respiratory distress syndrome, sepsis, or other hypermetabolic states that demand increased [DO.sub.2] to keep up with increased [VO.sub.2].

The compensatory mechanisms just outlined that are recruited to fulfill oxygen demands in otherwise healthy persons may not operate efficiently or at all in critically ill patients. For example, the major compensatory mechanism of increased cardiac output cannot be achieved in a patient with pulmonary edema. These patients are frequently fluid restricted or treated with diuretics such that their cardiac output is normal or slightly below normal. (66) In patients with cardiac disease, or in those who are at risk of myocardial infarction, the progression of anemia can present particular concerns and eventually become life threatening. In the normal coronary circulation, extraction of oxygen by the myocardium is more efficient than in the general microcirculation, approaching an extraction ratio of 50% compared with 25% elsewhere. (63,67) Because the normal myocardium is efficient, and extracts near-maximal concentrations of oxygen under resting conditions, the margin for incremental increase is narrow. The norm al coronary circulation responds to moderate anemia by undergoing vasodilatation; however, in a patient with coronary stenosis, this compensatory route may not be available. Indeed, even animals with normal hemoglobin levels with moderate to high-grade coronary stenosis show more myocardial dysfunction and ischemia than normal controls (60)

A number of published studies suggest that myocardial ischemia or other morbid cardiac events occur when the hematocrit decreases below 0.28. (69-71) Anemia beyond the second postoperative day and a transfusion requirement of more than 10 units of RBCs were predictors of postoperative morbidity and mortality in elderly patients undergoing cardiac surgery. (72) In a retrospective study, Hebert et al (73) noted that mortality increased as hemoglobin levels declined below 100 g/L in an overall cohort of ICU patients. The subset of severely ill patients with cardiovascular disease, or those who had undergone high-risk cardiovascular procedures, had a higher mortality rate if they did not receive RBC transfusions. (73) In a subsequent prospective study, (74) these investigators quantified a strong trend toward increased mortality in a subgroup of cardiac patients when the transfusion policy was restrictive, with transfusions withheld until the hemoglobin level decreased to 70 g/L or less.

Blood transfusion practices used in a large cohort of elderly patients (n = 78974) with acute myocardial infarction were evaluated by Wu et al. (75) The Medicare database was used to explore relationships between hematocrit at admission, blood transfusions, 30-day mortality, and development of in-hospital outcomes, including congestive heart failure, shock, and length of hospitalization. Elderly patients with a lower hematocrit at admission had an increased risk of cardiac arrest, shock, and congestive heart failure and also had longer hospital stays. Crude 30-day mortality rates increased proportionately with declining hematocrit values. Transfusions were administered to 4.7% of patients, and the likelihood of transfusion correlated strongly with low hematocrit at admission. Other predictors of transfusion, in order of decreasing importance, were renal insufficiency, recent trauma or surgery, female sex, history of internal bleeding, primary reperfusion therapy on admission, mean arterial pressure less than 70 mm Hg, prior angioplasty or congestive heart failure, and higher score on the Acute Physiology and Chronic Health Evaluation (APACHE) II. The type of hospital, treatment facility, or physician specialties were also factors determining whether transfusions were administered. Transfusion was associated with lower mortality in elderly patients whose initial hematocrit was 0.33 or less, but higher mortality was observed if the hematocrit was 0.36 or more. More dramatic positive outcomes were observed when transfusions were given to severely anemic elderly patients (ie, with hematocrits [less than or equal to] 0.24), but benefit was observed only in those with hematocrits up to 0.33.


Widely accepted standards for administering RBC transfusions in an ICU do not exist. The design and conduct of a large prospective study. (Transfusion Requirements in Critical Care) by Hebert et al has been reviewed in an earlier supplement. (2) The results of this study (76) raise important issues to consider in the context of managing anemia in subsets of critically ill patients and are therefore summarized in Table 1. Briefly, ICU patients who were randomly assigned to receive transfusions under a restrictive policy (hemoglobin level <70 g/L) showed trends toward better overall outcomes than ICU patients whose transfusion trigger was more liberal (hemoglobin level TRANSFUSION AND ANEMIA MANAGEMENT RESEARCH IN THE ICU

Transfusion and anemia management practices have changed considerably in the past 2 decades, and it is therefore meaningful to provide historical perspective by first describing results of a classic retrospective chart review conducted at Dartmouth-Hitchcock Medical Center by Corwin et al. (16) During the year 1990, all patients who stayed in a multidisciplinary ICU for more than 1 week were included in the review; 142 of 609 patients met entry criteria and of these, 121 (85%) received at least 1 RBC transfusion. A total of 1147 RBC units were transfused; the mean number of units per patient receiving a transfusion was 9.5 [+ or -] 0.8, which is surprisingly high by today's standards. Blood transfusion guidelines and conservation efforts have changed practice in the 12 years since the study by Corwin et al was conducted, and 2 additional larger trials have recently been completed. The first is the ABC study (Anemia and Blood Transfusion in Critical Care), which included 3534 patients in 146 institutions in We stern Europe. (77) The ABC study produced very similar results to those of the GRIT study (78) done in the United States (GRIT is not an acronym, rather, it is a mnemonic term for hematocrit and critical care).

The CRIT study was a multicenter, observational trial initiated in 213 hospitals after the results of the prospective trial conducted by Hebert et al (76) were published. The objectives of this study were to assess the prevalence of anemia and to describe current transfusion practices in critically ill patients admitted to ICUs. Clinical outcomes in response to RBC transfusions were to be assessed. A total of 4892 patients (mean hemoglobin level 110 [+ or -] 24 g/L on day 2 of ICU stay) were enrolled. Anemia persisted and worsened throughout the ICU stay, and 44% of patients received at least 1 RBC transfusion. RBC units were, on average, 21 days old. The mean number of units transfused per patient was 4.6 [+ or -] 4.9, considerably lower than that reported by Corwin et al (16) 12 years earlier. Transfusion practices were consistent across age groups and admitting diagnoses. The mean hemoglobin level before transfusion was 86[+ or -] 17 g/L. Of interest in GRIT was the observation that patients who were in t he ICU longer (7 days) were more likely (P<.001) to receive transfusions (63%) than were patients who were in the ICU for shorter periods (33%). Conversely, receiving more transfusions was a predictor of poorer outcome as evidenced by increases in organ dysfunction, longer duration of mechanical ventilation, longer stay in the ICU, and longer stay in the hospital. Patients with lower minimum hemoglobin levels also had longer durations of mechanical ventilation and longer ICU and hospital stays.


Because of the shortages in blood supply and the many inherent risks associated with transfusions, significant research effort has been aimed at demonstrating effectiveness and reduced mortality when transfusion triggers are conservative and restrictive. (79) As we have seen, several studies have successfully demonstrated that decreasing the use of transfusions is desirable, but certain populations of critically ill patients, particularly those with severe cardiac disease, sepsis, and higher APACHE II scores, will have a poor prognosis if hemoglobin level is not adequately maintained. Other strategies aimed at elevating hemoglobin to levels that will support oxygen demand, including recombinant human erythropoietin (rHuEPO, epoetin alfa) therapy, might prove to be safer and more effective than RBC transfusions.

Epoetin Alfa in the ICU

Anemia in critically ill patients may, in many cases, be attributed to a blunted erythropoietin response, similar to that seen in patients with anemia of chronic disease. In patients who are not severely ill and have hematocrits greater than 0.30, endogenous erythropoietin levels are generally 20 IU/L or less. (23) These levels normally increase in response to a physiological feedback loop as anemia progresses. When hematocrit is between 0.26 and 0.30, a doubling of endogenous erythropoietin level to 40 IU/L is observed, and serum level of erythropoietin can increase dramatically from 100 IU/L to as high as 1000 IU/L as hematocrit decreases below 0.25. (80) Evidence of a blunted erythropoietin response was seen in patients evaluated in the 3-center substudy described earlier. (22) These patients had erythropoietin levels that were an order of magnitude lower than expected for the degree of anemia and, additionally, reticulocyte counts did not increase with the increased erythropoietin levels. Similar results have been demonstrated by others. For example, a randomized, open-label study of 36 critically ill anemic patients was conducted specifically to evaluate endogenous erythropoietin production and the response to epoetin alfa in an ICU. (81) All patients received folic acid (1 mg daily). As interventions, 12 patients received intravenous iron for 14 days (iron group), 12 patients received epoetin alfa (300 IU/kg) subcutaneously on days 1, 3, 5, 7, and 9, plus iron for 14 days (epoetin alfa group), and 12 received only folic acid (control group). Similar to the laboratory characteristics of patients in the previous study, these patients had low iron, low total iron-binding capacity, high ferritin concentration, and erythropoietin levels at baseline that were inappropriately low for the degree of anemia present. After treatment, those in the iron group did not improve their reticulocyte count. In contrast, those in the epoetin alfa group showed a significant increase in reticulocytes compared with both the contro l (P<.05) and iron (P<.05) groups. This effect was evident on day 6, and persisted for approximately 1 week after the epoetin alfa therapy was discontinued. An increase in serum transferrin receptor concentration paralleled the reticulocytosis, suggesting that erythropoiesis was occurring.

The group at Stanford participated in a randomized, double-blind, placebo-controlled study designed to evaluate whether epoetin alfa therapy decreases the need for blood transfusions in IGU patients. (10) Two other centers, Dartmouth-Hitchcock and the Naval Medical Center in San Diego, also participated. A total of 160 patients were enrolled; 80 patients were randomly assigned to each of 2 treatment groups: one group received 300 IU/kg epoetin alfa and the other group received placebo. Beginning on day 3 of the ICU stay, study drug (epoetin alfa or placebo) was administered by subcutaneous injection every day for 5 days, then every other day for a total of 2 weeks. For patients who were in the ICU longer than 2 weeks, study drug was continued every other day until ICU discharge. All patients were followed up for 42 days from the date of randomization, unless death occurred earlier. Baseline characteristics and pretransfusion hematocrits were similar in the 2 groups (Table 3), as was the primary ICU diagnosis. A significant reduction (P= .002) in the total number of RBC units transfused was observed in the epoetin alfa group (166 units) compared with the placebo group (306 units). Patients in the placebo group continued to require transfusions throughout the study, whereas in the epoetin alfa group, transfusions were nearly eliminated toward the end of the follow-up period. In the epoetin alfa group, 45% of patients received a transfusion or died between days 8 and 42, compared with 55% of patients in the placebo group (relative risk, 0.8; 95% confidence interval [CI], 0.6, 1.1). Despite the epoetin alfa group receiving, on average, about one-half the number of transfusions per patient (2.1 transfusions per patient) as compared with the placebo group (3.8 transfusions per patient), the average final hematocrit was higher by approximately .035 (P<.01). The percentage change in reticulocyte count was also significantly higher in the epoetin alfa group than in the placebo group (P <.001). No significant differences b etween groups were observed in mortality or in the number of adverse events.

A much larger study, enrolling 1302 patients at 65 sites, known as EPO2, has recently been completed. This study was designed to be a pivotal trial with the goal of determining whether epoetin alfa can reduce the number of transfusions in a much larger population than the previous work. A more convenient dosage schedule of epoetin alfa was used: 40 000 units subcutaneously administered once a week beginning on day 3 of ICU stay. Oral iron supplementation was used. The transfusion "trigger" was left to the discretion of the attending physician, provided that hematocrit was 0.27 or less, or some other major indication for transfusion was present (eg, active bleeding or myocardial ischemia). The end points of the study included the following: achievement of transfusion independence through day 28 of the study, cumulative blood transfusion requirements, mortality, changes in hemoglobin, and other analyses including hospital and ICU lengths of stay and number of days on mechanical ventilation. Although additional analyses are still pending, the data showed a reduction in the percentage of patients receiving transfusions (60.4% vs 50.5%; P<.001) and the number of RBC units transfused per patient (2.4 vs 3.0, P<.001). It is expected that these results will define the role of epoetin alfa use in the IGU as we approach the year 2003.


Although anemia is apparently tolerated in most patients, particularly those who are relatively healthy, the ICU population must be thought of differently. Anemia in the ICU may be due to acute blood loss, phlebotomy, or to the presence of inflammatory disease. The anemia in critically ill patients resembles anemia of chronic disease, which is believed to result from a poor endogenous erythropoietin response or erythropoietin deficiency. The risks of blood transfusions are many and ICU patients may not tolerate infusions of older, stored blood. Nonetheless, hemoglobin levels at or above 100 g/L may be important for oxygen delivery to vital organs, especially in critically ill patients with increased oxygen demands. The appropriate transfusion trigger for critically ill patients in this setting remains unknown. Blood transfusions in the IGU may not improve outcomes, and numerous studies have been published to suggest the contrary, that transfusions may actually worsen patients' outcomes in certaln ICU settings . Recombinant human erythropoietin (rHuEPO, epoetin alfa) has been shown to reduce transfusion needs and increase hemoglobin levels in multiple settings and now, it appears to also do so in the ICU.



Figure 1 Hemoglobin and oxygen delivery. [Do.sub.2] = CO x [Cao.sub.2] = CO x 1.37 x Hb x [Sao.sub.2] [Vo.sub.2] = CO x ([Cao.sub.2] - [Cvo.sub.2]) Critical oxygen delivery is the [Do.sub.2] below which [Vo.sub.2] decreases [Cao.sub.2] indicates arterial oxygen concentration; [Cvo.sub.2], venous oxygen concentration; CO, cardiac output; [co.sub.2], oxygen delivery; Hb, hemoglobin; [Sao.sub.2], arterial oxygen saturation; [Vo.sub.2], oxygen consumption. Table 1 Summary of results from the Transfusion Requirements in Critical Care study (76) Strategy Restrictive Liberal Parameter description transfusion transfusion trigger <70 g/L trigger <100 g/L Enrollment and baseline characteristics No. of patients randomized 418 420 No. of withdrawals 5 4 Baseline hemoglobin level, mean (SD), g/L 82 (7) 82 (7) Results: primary intent-to-trat analysis Daily hemoglobin level, mean (SD), g/L (P<.01) 85 (7) 107 (7) No. of transfusions, mean (SD) (P<.01) 2.6 (4.1) 5.6 (5.3) No transfusions, % of patients (P<.01) 33 0 30-day mortality, % of patients (P=.11) 18.7 23.3 Hospital mortality, 28.1 % of patients patients (P=.05) 22.2 16.2 Intensive care unit 13.4 mortality, % of patients (P=.29) 8.8 (4.4) Organ failure score, 8.3 (4.6) mean (SD) (P=.10) 30-day mortality: healthier, younger subgroup APACHE II score [less than or equal to]20 (n = 424), % of patients (P=.03) 8.7 16.1 <55 years old (n = 334), % of patients (P=.02) 5.7 13.0 30-day mortality: at-risk subgroup Primary or secondary cardiac disease 20.5 22.9 (n = 326), % of patients (P=.69) Severe infections and septic shock (n = 218), 22.8 29.7 % of patients (P=.36) Trauma (n = 200), 10.0 8.8 % of patients (P=.81) APACHE indicates Acute Physiology and Chronic Health Evaluation. Table 2 Summary of results from cardiovascular disease subsets of the Transfusion Requirements in Critical Care study (74) Strategy Restrictive Liberal transfusion transfusion Parameter description trigger <70 g/L trigger <100 g/L Enrollment and baseline characteristics No. of patients evaluated in 160 197 subset * Age, mean (SD), years 64.0 (14.1) 65.6 (13.7) Diuretic use, % of patients 43 58 (P<.01) Epidural anesthetic medications, % 8 2 of patients (P<.01) APACHE II score, mean (SD) 22.7 (7.6) 23.1 (8.2) Results from cardiovascular disease subset (all) Daily hemoglobin level, mean (SD), 85 (6.2) 103 (6.7) g/L (P< .01) No. of transfusions, mean (SD) 2.4 (4.1) 5.2 (5.0) (P<.01) 30-day mortality, % of patients 23 23 (P=1) Hospital mortality, % of patients 27 28 (P=.81) ICU mortality, % of patients 19 16 (P=.49) Mean organ failure score, mean 11.1 (7.6) 11.9 (7.9) (SD) (P=.39) Outcomes in subset of patients with confirmed severe heart disease No. of patients evaluated in 111 146 subset + 30-day mortality, % of patients 26 21 (P=.38) Hospital mortality, % of patients 29 27 (P=.78) ICU mortality, % of patients 23 17 (P=.27) Organ failure score, mean (SD) 11.8 (8.2) 11.6 (7.5) (P=.80) APACHE indicates Acute Physiology and Chronic Health Evaluation. * Subset represents 357 patients (43%) from population described in Table 1 with the following primary cardiac diagnoses in decreasing order of frequency: ischemic heart disease (with or without coronary artery bypass graft), peripheral vascular disease, cardiogenic shock, congestive heart failure, cardiac arrest, arrhythmia, valvular heart disease, hypertension, pulmonary embolus/right ventricular failure, or other not specified. There were no withdrawals from this subset, and all patients were followed up for at least of 30 days. + Subset represents 257 patients with confirmed severe ischemic heart disease, severe peripheral vascular disease, or severe comorbid cardiac disease. Table 3 Baseline characteristics and study outcomes of recombinant human erythropoietin (epoetin alfa) in the intensive care unit trial (10) Treatment group Characteristics and outcomes Epoetin alfa (n=80) Baseline characteristics * % Male Age, mean (SD), years 55 APACHE II score, mean (SD) 59.1 (20.5) Hematocrit, mean (SD) 18.2 (5.3) Reticulocyte count, mean (SD) 0.30 (0.04) Serum erythropoietin level, U/L 0.017 (0.011) Pretransfusion hematocrit (%) 39 (40) Study outcomes 0.27 (0.04) Total units of red blood cells transfused + Change in hematocrit (baseline to final), mean (95% CI) ++ 166 Final hematocrit, mean (SD) ss 0.05 (0.04, 0.06) Change in reticulocyte count (baseline to final), mean (95% C1) ++ 0.35 (0.06) 0.025 (0.019, 0.030) Treatment group Characteristics and outcomes Placebo (n=80) Baseline characteristics * % Male Age, mean (SD), years 53 APACHE II score, mean (SD) 60.4 (18.8) Hematocrit, mean (SD) 18.4 (5.4) Reticulocyte count, mean (SD) 0.30 (0.04) Serum erythropoietin level, U/L 0.017 (0.011) Pretransfusion hematocrit (%) 30 (29) Study outcomes 0.27 (0.04) Total units of red blood cells transfused + Change in hematocrit (baseline to final), mean (95% CI) ++ 305 Final hematocrit, mean (SD) ss 0.01 (0.003, 0.01) Change in reticulocyte count 0.32 (0.04) (baseline to final), mean (95% 0.008 (0.003, 0.013) C1) ++ APACHE indicates Acute Physiology and Chronic Health Evaluation. * No significant differences observed. + p<.002. ++ p<.001. (ss)p<.01. References

(1.) Goodnough LT, Brecher ME, Kanter MH, AuBuchon JP. Transfusion medicine: first of two parts--blood transfusion. N Engl J Med. 1999;340:438-447.

(2.) Pohlman AS, Carven JH, Lindsay K. Conserving blood in the intensive care unit. Crit Care Nurse. December 2001;21 (suppl).

(3.) National Blood Data Resource Center. Report on Blood Collection and Transfusion in the United States in 1999. Available at: Accessed August 15, 2002.

(4.) Wallace EL, Churchill WH, Surgenor DM, Cho GS, McGurk S. Collection and transfusion of blood and blood components in the United States, 1994. Transfusion. 1998;38:625-636.

(5.) Hernandez R. Citing Disease, F.D.A. Panel Backs Blood Donor Curbs. New York Times. June 29,2001:3.

(6.) Carson JL, Hill S, Carless P, Hebert P, Henry D. Transfusion triggers: a systematic review of the literature. Transfus Med Rev. 2002;16:187-199.

(7.) Hebert PC, Wells G, Martin C, et al. Variation in red cell transfusion practice in the intensive care unit: a multicentre cohort study. Crit Care. 1999;3:57-63.

(8.) Goodnough LT, Bodner MS. Martin JW. Blood transfusion and blood conservation: cost and utilization issues. Am J Med Qual. 1994:9:172-183.

(9.) Cremieux P-Y, Barrett B, Anderson K, Slavin MB. Cost of outpatient blood transfusion in cancer patients. J Clin Oncol. 2000;18:2755-2761.

(10.) Corwin HL, Gettinger A, Rodriguez RM, et al. Efficacy of recombinant human erythropoietin in the critically ill patient: a randomized, double-blind, placebo-controlled trial. Crit Care Med. 1999:27:2346-2350.

(11.) Spence RK. The cost of transfusion and alternatives. Surgery. 2002; 131:236-237.

(12.) Thurer RL. Blood transfusion in cardiac surgery. Can J Anaesth. 2001;48:S6-12.

(13.) AuBuchon JP, Pickard CA, Herschel LH, et al. Production of pathogen-inactivated RBC concentrates using PEN 110 chemistry: a phase I clinical study. Transfusion. 2002;42:146-152.

(14.) Hambleton J, Viele M, Rios J, et al. RBCs treated with Helinx[TM] pathogen inactivation have recovery and half-life comparable to conventional RBCs in a randomized crossover trial. Presented at the 7th Annual Congress of the European Hematology Association (EHA). Florence, Italy, June 2002.

(15.) American Association of Blood Banks. Standards for Blood Banks and Transfusion Services. 20th ed. Bethesda, Md: American Association of Blood Banks; 2000.

(16.) Corwin HL, Parsonnet KC, Gettinger A. RBC transfusion in the ICU: is there a reason? Chest. 1995;108:767-771.

(17.) 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.

(18.) Andrews T, Waterman H, Hillier V. Blood gas analysis: a study of blood loss in intensive care. J Adv Nurs. 1999;30:851-857.

(19.) Cheng D, Yakobi-Shvlli R, Fernandez J. Life-threatening hypotension from babesiosis hemolysis. Am J Emerg Med. 2002;20:367.

(20.) Woods JB, Blake PG, Perry KG Jr, Magann EF, Martin RW, Martin JN Jr. Ascites: a portent of cardiopulmonary complications in the preeclamptic patient with the syndrome of hemolysis, elevated liver enzymes, and low platelets. Obstet Gynecol. 1992:80:87-91.

(21.) Abid S, Khan AH. Severe hemolysis and renal failure in glucose-6-phosphate dehydrogenase deficient patients with hepatitis E. Am J Gastroenterol. 2002;97:1544-1547.

(22.) Rodriguez RM, Corwin HL, Gettinger A, Corwin MJ, Gubler D, Pearl RG. Nutritional deficiencies and blunted erythropoietin response as causes of the anemia of critical illness. J Crit Care. 2001;16:36-41.

(23.) Greendyke RM, Sharma K, Gifford FR. Serum levels of erythropoietin and selected other cytokines in patients with anemia of chronic disease. Am J Clin Pathol. 1994:101:338-341.

(24.) Krantz SB. Pathogenesis and treatment of the anemia of chronic disease. Am J Med Sci. 1094:307:353-359.

(25.) von Ahsen N, Muller C, Serke S, Frei U, Eckardt K-U. Important role of nondiagnosticblood loss blunted erythropoietic response in the anemia of medical intensive care patients. Crit Care Med. 1999;27:2630-2639.

(26.) Hebert PC, Wells G, Marshall J, et al, Transfusion requirements in critical care: a pilot study. JAMA. 1995;273:1439-1444.

(27.) Sazama K, DeChristopher PJ, Dodd R, er al. Practice parameter for the recognition, management, and prevention of adverse consequences of blood transfusion. Arch Pathol Lab Med. 2000:124:61-70.

(28.) Perrotta PL, Snyder EL. Non-infectious complications of transfusion therapy. Blood Rev. 2001;15:69-83.

(29.) Linden JV, Wagner K, Voytovich AE, Sheehan J. Transfusion errors in New York State: an analysis of 10 years' experience. Transfusion. 2000:40:1207-1213.

(30.) Williamson LM, Lowe S, Love EM, et al. Serious hazards of transfusion (SHOT) initiative: analysis of the first two annual reports. BMJ. 1999:319:16-19.

(31.) Bordin JO, Heddle NM, Blajchman MA. Biologic effects of leukocytes present in transfused cellular blood products. Blood. 1994;84:1703-1721.

(32.) Schreiber GB, Busch MP, Kleinman SH, Korelitz JJ. The risk of transfusion-transmitted viral infections. N Engl J Med. 1996;334:1685- 1690.

(33.) Koff RS. Seroepidemiology of hepatitis A in the United States. J Infect Dis. 1995;171:S19-S23.

(34.) Prowse C, Ludlam CA, Yap PL. Human parvovirus B19 and blood products. Vox Sang. 1997:72:1-10.

(35.) Blajchman MA, Goldman M, Freedman JJ, Sher GD. Proceedings of a consensus conference: prevention of post-transfusion CMV in the era of universal leukoreduction. Transfus Med Rev. 2001:15:1-20.

(36.) Vamvakas EC. Risk of transmission of Creutzfeldt-Jakob disease by transfusion of blood, plasma, and plasma derivatives. J Clin Apheresis. 1999:14:135-143.

(37.) Mungai M, Tegtmeier G, Chamberland M, Parise M. Transfusion-transmitted malaria in the United States from 1963 through 1999. N Engl J Med. 2001:344:1973-1978.

(38.) Kuehnert MJ, Roth VR, Haley NR, et al. Transfusion-transmitted bacterial infection the United States, 1998 through 2000. Transfusion. 2001:41:1493-1499.

(39.) Manganaro LA, O'Brien JA, Parkar ND, Taylor RW, Trottier SJ, Veremakis C. The impact of packed red blood cell transfusions on nosocomial infection rates in the critically ill patient [abstract]. Crit Care Med 2000:28:A103.

(40.) Carson JL, 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.

(41.) Leal-Noval SR, Rincon-Ferrari MD, Garcia-Curiel A, et al. Transfusion of blood components and postoperative infection in patients undergoing cardiac surgery. Chest. 2001:119:1461-1468.

(42.) 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.

(43.) Offner PJ. Moore EE, Biffl WL, Johnson IL, Silliman CC. Increased rate of infection associated with transfusion of old blood after severe injury. Arch Surg. 2002:137:711-717.

(44.) Mynster T, Christensen IJ, Moesgaard F, Nielsen HJ. Effects of the combination of blood transfusion and postoperative infectious complications on prognosis after surgery for colorectal cancer, Br J Surg. 2000:87:1553-1562.

(45.) Houbiers JGA, Brand A, van de Watering LMG, et al. Randomised controlled trial comparing transfusion of leucocyte-depleted or buffy-coat-depleted blood in surgery for colorectal cancer. Lancer. 1994;344:573-578.

(46.) Vogelsang GB. Transfusion-associated graft-versus-host disease in nonimmunocompromised hosts. Transfusion. 1990:30:101-103.

(47.) Tadros T, Wobbes T, Hendriks T. Opposite effects of interleukin-2 on normal and transfusion-suppressed healing of experimental intestinal anastomoses. Ann Surg. 1993:218:800-808.

(48.) Jensen LS, Kissmeyer-Nielsen P, Wolff B, Qvist N. Randomised comparison of leucocyte-depleted versus buffy-coat-poor blood transfusion and complications after colorectal surgery. Lancer. 1996;348:841-845.

(49.) Parrott NR, Lennard TWJ, Taylor RMR, Proud G, Shenton BK, Johnston IDA. Effect of perioperative blood transfusion on recurrence of colorectal cancer. Br J Surg 1986:73:970-973.

(50.) van de Watering LMG, Brand A, Houbiers JGA, Kranenbarg WMK, Hermans J, van de Velde CJH. Perioperative blood transfusions, with or without allogeneic leucocytes, relate to survival, not to cancer recurrence. Br J Surg. 2001;88:267-272.

(51.) Jakobsen EB, Eickhoff JH, Andersen J, Lundvall L, Stenderup JK. Perioperative blood transfusion and recurrence and death after resection for cancer of the colon and rectum. Scand J Gastroenterol. 1990;25:435-442.

(52.) van de 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. Circulation. 1998;97:562-568.

(53.) Dahlke MB. Red blood cell transfusion therapy. Med Clin North Am. 1984;68:639-656.

(54.) Mynster T, Dybkjaer E, Reimert CM, et al. Prestorage leukofiltration of whole blood and SAGM blood prevents extracellular bloactive substance accumulation. Inflamm Res. 1999:48:363-368.

(55.) Marik PE, Sibbald WJ. Effect of stored-blood transfusion on oxygen delivery in patients with sepsis. JAMA 1993:269:3024-3029.

(56.) Daly JF, ed. Dailey's Notes on Blood. 3rd ed. Arlington, Mass: Medical Consulting Group: 1996.

(57.) Vamvakas EC, Carven JH. Transfusion and postoperative pneumonia in coronary artery bypass graft surgery: effect of the length of storage of transfused red cells. Transfusion. 1999;39:701-710.

(58.) Purdy FR, Tweeddale MG, Merrick PM. Association of mortality with age of blood transfused in septic ICU patients. Can J Anaesth. 1997;44:1256-1261.

(59.) Clark P, Mintz PD. Transfusion triggers for blood components. Curr Opin Hematol. 2001;8:387-391.

(60.) Maizes JS, Murtuza M, Kvetan V. Oxygen transport and utilization. Respir Care Clin N Am. 2000;6:473-500.

(61.) Greenburg AG. A physiologic basis for red blood cell transfusion decisions. Am J Surg. 1995; 170:44S-48S.

(62.) Eckardt KU. Anemia in critical illness. Wien Kiln Wochenschr. 2001;113:84-89.

(63.) Jan K-M, Chien S. Effect of hematocrit variations on coronary hemodynamics and oxygen utilization. Am J Physiol. 1977;233:H106-H113.

(64.) Viele MK, Weiskopf RB. What can we learn about the need for transfusion from patients who refuse blood? The experience with Jehovah's Witnesses. Transfusion. 1994;34:396-401.

(65.) Weiskopf RB, Viele MK, Feiner J, et al. Human cardiovascular and metabolic response to acute, severe isovolemic anemia. JAMA. 1998; 279:217-221.

(66.) Braunwald E, Fauci AS, Kasper DL, Hauser SL, Longo DL, Jameson JL, eds. Harrison's Principles of Internal Medicine. 15th ed. New York, NY: McGraw-Hill; 2001.

(67.) Wilkerson DK, Rosen AL, Sehgal LR, Gould SA, Sehgal HL, Moss GS. Limits of cardiac compensation in anemic baboons. Surgery. 1988;103:665-670.

(68.) Hagl S, Heimisch W, Meisner H, Erben R, Baum M, Mendler N. The effect of hemodilution on regional myocardial function in the presence of coronary stenosis. Basic Res Cardiol. 1977;72:344-364.

(69.) Nelson AH, Fleisher LA, Rosenbaum SH. Relationship between post-operative anemia and cardiac morbidity in high-risk vascular patients in the intensive care unit. Crit Care Med. 1993;21:860-866.

(70.) Carson JL, Spence RK, Poses RM, Bonavita G. Severity of anaemia and operative mortality and morbidity. Lancet. 1988;1:727-729.

(71.) Carson JL, Duff RM, Berlin JA, et al. Effect of anaemia and cardiovascular disease on surgical mortality and morbidity. Lancet. 1996;348:1055-1060.

(72.) Rady MY, Ryan T, Starr NJ. Perioperative determinants of morbidity and mortality in elderly patients undergoing cardiac surgery. Crit Care Med. 1998;26:225-235.

(73.) Hebert PC, Wells G, Tweeddale M, et al. Does transfusion practice affect mortality in critically ill patients? Am J Respir Crit Care Med. 1997;155:1618-1623.

(74.) Hebert PC, Yetisir E, Martin C, et al. Is a low transfusion threshold safe in critically ill patients with cardiovascular diseases? Crit Care Med. 2001;29:227-234.

(75.) Wu W-C, Rathore SS, Wang Y, Radford MJ, Krumholz HM. Blood transfusion in elderly patients with acute myocardial infarction. N Engl J Med. 2001;345:1230-1236.

(76.) Hebert PC, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. N Engl J Med. 1999;340:409-417.

(77.) Vincent JL, Baron JF, Gattinoni L, Reinhart K, Thijs L, Webb A. Anaemia and blood transfusion in critical care (ABC) survey: organ function and mortality in transfused patients [abstract]. Intensive Care Med. 2001;27(suppl 2).

(78.) Corwin HL, Abraham E, Fink MP, et al. Anemia and blood transfusion in the critically ill: current clinical practice in the US: the CRIT study [abstract]. Crit Care Med. 2001;29:A2.

(79.) Hill SR, Carless PA, Henry DA, et al. Transfusion thresholds and other strategies for guiding allogeneic red blood cell transfusion [CD]. San Diego, Calif: Cochrane Lib. 2002;CD002042.

(80.) Krafte-Jacobs B, Levetown ML, Bray GL, Ruttimann UE, Pollack MM. Erythropoietin response to critical illness. Crit Care Med. 1994;22:821-826.

(81.) van Iperen CE, Galliard CAM, Kraaijenhagen RJ, Braam BG, Marx JJM, van de WA. Response of erythropoiesis and iron metabolism to recombinant human erythropoietin in intensive care unit patients. Crit Care Med. 2000;28:2773-2778.


Ms Pohlman: The primary causes of anemia are blood loss or the abnormal or insufficient production of RBCs that ultimately impair the ability of RBCs to deliver oxygen to the tissues. In the ICU, minimization of phlebotomy, optimization of blood Conservation therapy, and application of recent research to bedside management can result in better patient outcomes. What explains the recent interest in the treatment of anemia in ICU patients?

Dr Pearl: Interest in anemia and avoiding blood transfusions in the ICU dates back to the 1970s, when it was recognized that hepatitis could be transmitted via the blood supply. Interest continued into the 1980s and 1990s, when transmission of AIDS via blood became highly publicized. Research interest is now focused on the immunosuppressive risks of blood transfusions with increased morbidity and the ability to develop effective strategies for avoiding transfusions, such as decreasing phlebotomy, changing transfusion triggers, and using epoetin alfa in the ICU.

Question: What is the current status of clinical research pertaining to the use of blood substitutes in the ICU?

Dr Pearl: For about the past 15 years, blood substitutes have been undergoing development and clinical testing. Although the available compounds have improved, artificial hemoglobin and hemoglobin-based blood substitutes are still not as safe and effective as RBCs for delivering oxygen to tissues. Most substitutes under development have altered oxygen delivery characteristics. For example, these compounds may have nitric oxide binding effects that can lead to systemic and pulmonary hypertension, decreases in microcirculatory flow to vital organs, and pancreatitis. Extracellular hemoglobin, that is, hemoglobin outside of RBCs, is readily filtered from the bloodstream into the kidneys or other tissues. In general, currently existing blood substitutes have short half-lives Lasting several hours up to 1 to 2 days. Since the goal of an RBC transfusion, particularly in the ICU, is to maintain hemoglobin above critical levels for longer periods up to several weeks, pharmacokinetic problems with blood substitutes mus t first be resolved. Outcomes studies following administration of hemoglobin substitutes have not shown benefit, or have shown adverse effects, particularly in the trauma population. Effective blood substitutes will undoubtedly be forthcoming, but are most likely at least 5 to 10 years away.

Question: As blood shortages are anticipated and nurses are implementing conservation practices, what can be done to remove or avoid administrative roadblocks when ordering blood alternatives? For example, hematology consults are required in some institutions when epoetin alfa is requested for an ICU patient who has no underlying hematological disease or symptoms.

Dr Pearl: The ICU nurse is very much in control of patient care, especially with regard to the amount of blood drawn for lab tests and also when suggesting and ordering alternative interventions. Clinical judgment must be used when "transfusion triggers" are used in daily practice and ICU nurses have a large influence on whether a blood transfusion is ultimately ordered. ICU nurses identify which patients might need transfusions; for those patients who are anticipated to have extended IGU stays, alternative interventions should be suggested. If hospital administrators and decision makers can be educated to understand that patients who have a long IGU stay will eventually require transfusions, the administrative burden might be reduced.

Question: What recommendations can be made about new clinical pathways pertaining to medical ICU or trauma patients with multiple organ system involvement, that is, are there any standards or guidelines for the optimal use of epoetin alfa?

Dr Pearl: First, try to limit anemia and the need for blood transfusion in all patients by attempting to decrease phlebotomy and the amount of blood discarded with each draw. With regard to epoetin alfa use, our studies have demonstrated the ability of epoetin alfa to decrease the need for blood transfusion in the ICU. The data from the EPO1 study strongly suggest that patients who stay in the IGU for a long period will eventually require blood transfusions and therefore would benefit from epoetin alfa therapy. It may be difficult to predict on day 1 of an ICU stay which patients will have an extended stay, but by day 3, it becomes easier to determine who is at high risk of requiring a blood transfusion. In the EPO1 study, epoetin alfa administration began on day 3. However, the earlier that epoetin alfa therapy is initiated, the sooner an effect is seen. Ideally, if patients can be identified early, as in the trauma patient with multiple organ system involvement who is certain to be confined to the ICU for s everal weeks, treatment should begin early. In contrast, a patient with mild stroke and no organ system involvement might not develop anemia during the course of the ICU stay and treatment can be withheld or delayed. A common institutional practice is that when the hematocrit begins to decrease, when transfusions appear to be imminent, or when patients appear to be at risk of a long ICU stay, epoetin alfa therapy should be initiated.

Question: When a unit of blood is administered, an immediate change in the hemoglobin and hematocrit can be detected. What markers reflect the early effects of epoetin alfa?

Dr Pearl: In concept, the effects of epoetin alfa are reflected in the development of an increased reticulocyte count. ICU patients, like patients with anemia of chronic disease, do not respond to a decreasing hematocrit with an increased reticulocyte count, and anemia progresses. Patients treated with epoetin alfa develop an increased reticulocyte count after about 4 to 5 days. Hematocrit may not increase in parallel, especially if there is ongoing blood loss from phlebotomy or wound drainage. However, fewer blood transfusions will be needed and the hematocrit will gradually increase over a period of time.

Question: Which types of cardiovascular disease would indicate administering a blood transfusion when the hematocrit is between 0.27 and 0.30?

Dr Pearl: The data generated from the study by Hebert and colleagues resulted in 2 papers that reached different conclusions. In the first, patients had cardiovascular disorders, but they did not necessarily have known coronary disease. Hebert's data would suggest that these patients tolerate anemia relatively well, and most practitioners would withhold transfusion until hemoglobin level decreased to 85 g/L or less. In the second, an analysis of patients was focused on patients who actually had identified coronary artery disease; those patients were likely to benefit when the hemoglobin was maintained above 100 g/L.

Question: Aging or stored blood appears to have a significant effect on mortality, but it is an institutional practice to use oldest blood first. Is it feasible to change the process so that ICU patients could receive newer blood?

Dr Pearl: Ideally, only fresh or new blood would be used for all transfusions, but, even with current practices to use older blood first, statistics indicate that 2 out of 14 million units per year outdate. The only ways to prevent having more blood outdate is to use the oldest blood first or to improve the efficiency of the blood banking systems. There are models that propose expanding the geographical areas served by individual blood banks so that blood inventories are better utilized, but for now, this cannot be done. Technologies are under development aimed at preventing blood storage defects, known as "rejuvenating solutions." There are also some very intriguing data suggesting that the problem of old blood pertains only to ICU patients. In patients who develop SIRS, (systemic inflammatory response syndrome), interactions between certain elements in stored blood have been observed; for example, cytokines and adhesion molecules may have pathological consequences in ICU patients, but not in healthier patie nts.

Question: What is the cost of epoetin alfa therapy and are there any added costs to blood when leukocytes are removed? Second, are there toxicities associated with either therapy?

Dr Pearl: The cost of epoetin alfa therapy is approximately 1 cent per unit, and if administered as in the EPO2 study (40 000 units weekly), the cost would be $400 per dose. One of the issues is how many doses a patient may need and whether it is appropriate to stop administering epoetin alfa when the patient leaves the ICU. Leukodepletion adds approximately $30 to $50 per unit for transfusion. With regard to adverse effects of leukodepletion, the use of white cell filters has been associated with severe hypotension in patients treated with ACE inhibitors. Adverse events were carefully monitored in the EPO1 study, and they occurred at similar rates in both groups.

Question: What are the specific indications for iron administration, knowing that access to and absorption by the gastrointestinal tract can pose a particular challenge in ICU patients?

Dr Pearl: In the design of both the EPO1 and the EPO2 studies, extensive discussions took place about the indications for iron and route of administration. Oral iron is poorly absorbed and may be ineffective. There are data to suggest that excess iron is detrimental; specifically, iron is known to support bacterial growth and can accelerate free radical--mediated reactions, which may aggravate adult respiratory distress syndrome, septic shock, and other conditions. In the EPO1 study, parenteral iron was given to only 1 patient, who had a documented iron deficiency and was unable to respond to oral iron. Clinicians should consider the results of specific laboratory tests relating to iron levels, in particular, serum ferritin level, when determining whether iron should be administered and the dose that should be used.

COPYRIGHT 2002 American Association of Critical-Care Nurses