key: cord-009417-458rrhcm
authors: Luce, Judith A.
title: Use of Blood Components in the Intensive Care Unit
date: 2009-05-15
journal: Critical Care Medicine
DOI: 10.1016/b978-032304841-5.50082-0
sha: 
doc_id: 9417
cord_uid: 458rrhcm

nan

Most patients admitted to an intensive care unit (ICU) require the administration of one or more blood components during their stay. Such patients exhibit great diversity in conditions necessitating care in the ICU, age, underlying medical problems, and integrity of physiologic compensatory mechanisms. All these patients, however, share the need for optimized oxygen-carrying capacity and tissue perfusion. Ongoing blood loss resulting from injuries, surgical wounds, invasive monitoring equipment, and blood sampling requirements, coupled with inadequate marrow function and, in some, red cell destruction, makes red cell transfusion a necessity for many ICU patients. Additionally, many patients are susceptible to the development of hemostatic disorders requiring the administration of such blood components as plasma, cryoprecipitate, or platelet concentrates.

Blood components should be considered drugs because they exert potent therapeutic responses yet are also capable of causing signifi cant adverse effects. The Food and Drug Administration (FDA) regulates blood component preparation, testing, and administration. 1 Unlike pharmaceutical agents, however, blood components have fewer objective indications for use and no therapeutic index relating dose to safety. It is not as simple to monitor the effi cacy and continuing need for a blood component as it is to determine the blood level of a drug. In addition, the risks associated with transfusion cannot be known in advance and may be lethal; such risks include medical errors, as well as infectious and immunologic hazards. Unlike pharmaceutical agents, these prescribed products require documentation of patient consent and indication for use.

Although the American blood supply is now safer than ever before, zero-risk transfusion is not achievable, even if blood components could be sterilized. The process of donor selection and screening has become increasingly stringent, an evolution that began in response to the welldefi ned risks of transfusion-transmitted hepatitis and human immunodefi ciency virus (HIV) infection. Although the value of maximizing recipient safety is unarguable, increasing donor selectivity has its price. As more tests are added and more conditions placed on the donor, the number of usable donations has declined. This trend has led to occasional regional and seasonal blood shortages and, rarely, outright inability to provide certain blood components. Clinicians who prescribe blood components must be aware of these uncertainties in availability and contribute by using blood products appropriately while the national blood banking system seeks strategies to ensure an adequate, safe blood supply.

Donor screening strategies to ensure recipient safety take several forms. 1, 2 American blood donors are voluntary donors; cash payment was eliminated in the 1970s after studies linked professional donors with transmission of hepatitis. Confi dential questionnaires were initiated to limit transmission of HIV and hepatitis and to allow voluntary self-exclusion and involuntary exclusion of donors who pose an increased risk of transmitting infectious agents. Multiple specifi c serologic and biochemical tests are performed to detect the potential for transmission of HIV and other retroviruses, hepatitis, and syphilis. Any donor who indicates high-risk behavior or who tests repeatedly positive is placed on a permanent deferral list.

Some patients may insist on blood obtained from relatives or friends. This practice is termed directed or designated donation. These selected donors must undergo the same rigorous questioning and testing as volunteer donors. Some studies have found an increased frequency of hepatitis markers in the blood of directed donors when compared with blood drawn from unselected volunteers, but others suggest that designated donors may be no different from new volunteers. 3, 4 There continues to be no consensus about whether directed donors are, as a group, as safe as volunteer donors. 5, 6 Institutional policies about the acceptability and processing of directed donations vary widely. In any case, supporting ICU patients who require large-volume transfusion with directed donations is unlikely to be advantageous or practical.

The basic principle of blood component therapy is prescription of the specifi c blood product needed to meet the patient's requirement. A single whole blood (WB) donation can be separated into its composite parts, or components, which can be distributed to several recipients with differing physiologic needs. Component therapy thus meets the clinical requirements of increased safety, efficacy, and conservation of limited resources. As the variety of blood product components increases, however, the complexity of transfusion medicine also increases.

A WB donation is typically separated into red blood cells (RBCs), a platelet concentrate, and fresh frozen plasma (FFP) within hours of its collection. The plasma may be further processed into cryoprecipitate and supernatant (cryopoor) plasma.

One unit of WB measures approximately 500 mL, including 63 mL of citrate anticoagulant/preservative solution. Each unit of WB supplies about 200 mL of RBCs and 300 mL of plasma for volume replacement. WB is refrigerated for 21 to 35 days, depending on the preservative used. After less than 24 hours of refrigerated storage in this preservative and bag system, platelet and granulocyte function is lost. With further storage, levels of the "labile" coagulation factors V and VIII decrease. 7 Some blood centers offer modifi ed WB, which is produced by removal of the platelet or cryoprecipitate fraction and return of the supernatant plasma to the red cells. This permits provision of the more labile components to patients with specifi c needs, with the remainder forming a product having a composition essentially the same as cold-stored WB. However, the growing need for specialized blood components has resulted in processing the majority of blood donations into components, thus limiting the availability of WB and modifi ed WB.

RBCs, or in common usage, "packed" red cells (PRBCs), are the blood component most commonly transfused to increase red cell mass. PRBCs are derived from the centrifugation or sedimentation of WB and removal of most of the plasma/anticoagulant solution. If collected into citrate-phosphate-dextrose-adenine solution, the volume is approximately 250 mL, the hematocrit (Hct) is 70% to 80%, and the storage life is 35 days. Extended additive solutions permit storage up to 42 days but increase the volume to 300 mL and decrease the Hct to 60%. These extended storage units are commonly used and easier to transfuse because of lower viscosity, but they may pose a problem because of their larger volume.

The transfusion of leukocyte-reduced RBCs may benefi t certain patients. Transfusion of blood components containing leukocytes may lead to febrile reactions, a greater propensity for alloimmunization, platelet alloimmunization, and transmission of pathogens carried by leukocytes, such as cytomegalovirus (CMV). Leukocyte reduction, as defi ned by the FDA, requires fi ltration of the blood component by a special fi lter. 1 Filtration may be performed either at the time of blood donation and processing or later at the time of transfusion ("bedside fi ltration"). Filtration before storage conveys the benefi t of removing white blood cells (WBCs) before they can deteriorate and elaborate cytokines and other unwanted substances during storage. 8 Because of proven and theoretical benefi ts of leukocyte reduction of blood components (discussed later in the section covering the adverse effects of transfusion), many European countries and Canada require that all transfusions be leukocyte reduced, a process called universal leukoreduction (ULR). Some institutions in the United States have also made that decision, but either method of leukocyte reduction adds signifi cantly to the cost of each transfusion ($25 to $30), and the benefi ts of this measure when applied globally have yet to be quantifi ed. 9 Washing PRBCs involves recentrifuging to remove the plasma/preservative solution from the unit. However, washing may take an hour or more, limits subsequent storage time, and causes some loss of RBCs. Washing is also not an effective method of leukoreduction. There are very few indications for the use of washed RBCs, although some recipients with plasma reactions may benefi t. PRBCs can be frozen in cryoprotective solution and stored for extended periods. Frozen RBCs are generally limited to units of special value, such as those with a rare RBC antigen profi le or autologous blood donations that need to be stored for future use. A rare-donor registry of frozen PRBCs exists to assist in providing blood to patients with complex or multiple alloantibodies to red cell antigens. Signifi cant advanced planning is necessary to acquire and thaw frozen PRBCs for transfusion, thus limiting their use in acute situations.

WB and PRBCs suffer some cell loss during storage. The current technology of bag and preservative solutions attempts to optimize cell quality and quantity by using strict criteria to determine the length of allowable storage time. Nonetheless, as red cell metabolism decreases progressively, a "storage lesion" results, 10 with accumulation of a variety of undesirable substances and loss of cellular function. Over time in storage, a slow rise in the concentration of potassium, lactate, aspartate aminotransferase, lactate dehydrogenase, ammonia, phosphate, and free hemoglobin and a slow decrease in pH and bicarbonate concentration occur. Cytokines and infl ammatory mediators such as interleukin-1, interleukin-6 and tumor necrosis factor also accumulate. The pH of freshly stored blood in citrate solution is 7.16, which declines to approximately 6.73 at the end of the unit's shelf life. As potassium leaks from red cells during storage, levels as high as 25 mEq/L may result. However, each unit transfused supplies at most 7 mEq of potassium, which is well tolerated under most circumstances.

During the storage period there is also a progressive decrease in RBC-associated 2,3-diphosphoglycerate (2,3-DPG) and adenosine triphosphate (ATP). 10 A decrease in 2,3-DPG increases the affi nity of hemoglobin for oxygen, which shifts the oxygen dissociation curve to the left and decreases oxygen delivery to tissues. There is little evidence, however, that this transient increase in oxygen affi nity has clinical importance. After infusion, 2,3-DPG gradually increases as the transfused red cells circulate, with 25% recovery in 8 hours and full replacement by 24 hours. 11 Decreased ATP during storage diminishes the viability of red cells after transfusion and is one of the chief factors limiting storage time. There is no currently available storage or rejuvenation solution that optimizes these cellular constituents.

The majority of blood transfusions are in the form of PRBCs, the component indicated for normovolemic patients or those for whom intravascular volume constraints are necessary. The use of WB may be desirable for patients who require both increased oxygen-carrying capacity and volume resuscitation because of a large and ongoing hemorrhage; however, the availability of WB is generally limited. Resuscitation is effectively achieved with the use of PRBCs and crystalloid solutions. Each unit of PRBCs or WB is expected to raise the hemoglobin level by 1 g/dL and the Hct by 3% in stable, nonbleeding, average-sized adults. Although some studies have demonstrated a slight superiority of fresh WB over components when used during cardiac surgery in selected patients, 12 the benefi ts of fresh blood remain controversial, and current testing and processing requirements limit general availability.

Despite a long tradition of transfusion of RBCs in critically ill patients, the precise indications for transfusion remain a source of controversy, and specifi c transfusion practices may vary widely among clinicians. Before the major randomized studies of RBC transfusion policies, a survey of transfusion practice showed that about half of ICU patients were receiving red cell transfusions, 13 and another showed that if the ICU stay was longer than a week, the rate of transfusion was 85%. 14 The total number of transfusions was high, and ICU practice was characterized by high rates of transfusion. 15 The reasons for the controversies are clear: RBCs should be transfused only to enhance tissue oxygen delivery, but the underlying physiology of anemia, the complex adaptations to anemia, and the potential advantages and disadvantages to particular groups of patients are not as well understood. Compensatory mechanisms for acute and chronic anemia are diverse and complex. 16, 17 All work in concert to maintain oxygenation within the microcirculation. Cardiovascular adjustments leading to increased cardiac output include decreased afterload and increased preload resulting from changes in vascular tone, increased myocardial contractility, and elevated heart rate. Lowered blood viscosity permits improved fl ow of erythrocytes within capillaries. Blood fl ow is redistributed to favor critical organs with higher oxygen extraction. Pulmonary mechanisms, though contributing relatively little to shortterm oxygenation demands, exert potent effects on related metabolic variables. Finally, the hemoglobin molecule can undergo biochemical and conformational changes to enhance unloading of oxygen at the capillary level. All these mechanisms contribute to an "oxygen reserve" capacity that exceeds baseline requirements by approximately fourfold. 16 No experimental model exists that encompasses the diversity of physiologic compensations for hypoxia. Experiments carried out in animals and case reports in patients refusing transfusion indicate that an extremely low Hct is tolerated if tissue perfusion is adequate. [18] [19] [20] Certain objective, though indirect, measurements of tissue oxygenation exist and are available to clinicians caring for patients monitored invasively in the ICU. Mixed venous oxygen content (Pv O 2 ) and cardiac output can be measured in patients undergoing pulmonary artery catheterization; arterial oxygen content can also be measured directly. The oxygen extraction ratio (ER) can be calculated directly, and in the presence of normal or high cardiac output it is a measure of tissue oxygen extraction and, indirectly, the adequacy of tissue oxygen delivery. The total body ER at baseline is about 25%. A falling Pv O 2 and an ER increasing to greater than 50% have been proposed as indicators of the need for red cell transfusion. 21 There have been only 10 randomized trials of transfusion policy in the ICU, and only 1 of them was large enough to draw specifi c, statistically signifi cant conclusions. 22 The Canadian Critical Care Trials Group compared a liberal (target hemoglobin, 10 to 12 g/dL) with a restrictive (target hemoglobin, 7 to 9 g/dL) red cell transfusion policy in patients stratifi ed for disease severity. At 30 days from randomization, the restrictive strategy was at least as good as, if not better than (P = .11) the liberal strategy, and overall hospital mortality was signifi cantly lower in the restrictive strategy group (P = .05). For patients younger than 55 years and for patients with lower (<20) APACHE (Acute Physiology, Age, and Chronic Health Evaluation) II scores, the restrictive strategy was clearly superior. In addition, liberal transfusion was not associated with shorter ICU stays, less organ failure, or shorter hospital stays; longer mechanical ventilation times and cardiac events were more frequent in the liberal strategy group. A later subgroup analysis of patients with cardiovascular disease, though small enough to have statistical doubt, suggested that a more liberal transfusion strategy was probably appropriate for patients with severe ischemic coronary disease. 23 This observation has some support in experimental studies of the effects of anemia in laboratory animals with coronary occlusion. 24 The Canadian study has highlighted the many and complex issues involved in transfusion decision making in the ICU.

Since publication of the Canadian study, several large reports have examined the use of red cell transfusions in critical care units. Vincent and colleagues 25 surveyed European ICUs and found that the transfusion rate in 3534 patients was 37% during the ICU stay and 12.7% after the stay. The mean pretransfusion hemoglobin level was 8.4 g/ dL. Corwin and colleagues 26 studied 284 ICUs in the United States a year later and found great similarity: nearly 50% of patients received transfusions, and the mean threshold hemoglobin level was 8.6 g/dL. A single large Scottish teaching hospital reported a more parsimonious practice: the rate of transfusion was still 52% in its ICU patients, but the total volume of blood used was slightly smaller and the mean pretransfusion hemoglobin level was only 7.8 g/dL. 27 All these authors have concluded that ICU practice has not fully embraced the guidelines of the Canadian clinical trial. In contrast, 18 hospitals in Australia and New Zealand have reported on transfusion in 1808 consecutive ICU admissions, and although the authors found a median pretransfusion hemoglobin concentration of 8.2 g/dL, the rate of transfusion was lower, at only 19.7% of patients, 60% of whom were bleeding. 28 The "inappropriate" transfusion rate was 3%. The authors speculate that the practitioners may have been infl uenced by publication of the Canadian study and their own regional survey of transfusion practices. Nonetheless, they agree that full implementation of the Canadian guidelines in their clinical setting might be controversial.

The literature on RBC transfusion in the setting of surgery, particularly surgery with the use of blood products, is growing. A mounting body of data illustrate the human tolerance of a low Hct during and after surgery. A recent randomized trial of RBC transfusion strategy in orthopedic surgery demonstrated no signifi cant differences in outcome between a restrictive (8 g/dL) and a liberal (10 g/dL) transfusion threshold and included monitoring for silent myocardial ischemia preoperatively and postoperatively. 29 Provided that adequate perfusion of the microcirculation is maintained, purposeful maintenance of a low Hct during surgery, a technique called normovolemic hemodilution, 30 can be a powerful tool in minimizing blood loss and the attendant need for red cell transfusion. Table 80 -1 summarizes guidelines proposed by the National Institutes of Health, 31 the American Society of Anesthesiologists, 32 and the American College of Physicians 33 relative to the transfusion of RBCs. These guidelines have been provided with the intent of establishing parameters, not with the intent of substituting for the individual clinician's judgment. The art of medical decision making in transfusion, as in other areas of medicine, lies in determination of the appropriate treatment for the individual patient.

A platelet concentrate (random-donor platelets) is obtained by centrifugation from a unit of donated WB. Each unit contains a minimum of 55 � 10 10 platelets suspended in about 50 mL of plasma. Platelets are stored at room temperature to avoid loss of function from refrigeration and are constantly agitated to maximize gas exchange. The length of storage varies with the container used, but most systems permit 5-day storage. Because of this limited storage time and the increasing demand for this component, platelets are often subject to supply shortages. Some loss of viability and platelet numbers occurs during storage, but 5-day-old platelets still effect hemostasis. Once the bags are entered for pooling before transfusion, the platelets must be administered within 4 hours.

Each unit of platelets is expected to increase the platelet count by 10 � 10 9 /L in a typical 70-kg adult. The usual dose is 6 units, or 1 U/10 kg of body weight. A 1-hour post-transfusion platelet count should be obtained to determine the adequacy of response. The following equation, which relates platelet number and body size to the post-transfusion increment, can be used to assess the effectiveness of the transfusion: ABO-compatible platelets are desirable but not essential. When ABO-mismatched platelets are given, removal of some of the incompatible plasma can be carried out at the time of pooling for transfusion. Likewise, volume reduction may be necessary for patients at risk for fl uid overload from the 300 to 500 mL of plasma present in 6 to 10 units of platelets. Nonetheless, the remaining plasma is a good source of stable coagulation factors and contains diminished but still potentially benefi cial amounts of factors V and VIII. There is no contraindication to the use of Rh-positive platelets in Rh-negative patients; if given to women with future childbearing potential, Rh immune globulin (RhIG) may be used prophylactically against the small risk of Rh alloimmunization from red cells that may be contained in the platelet concentrate.

Plateletpheresis (common terms: single-donor platelets, apheresis platelets) involves separating and removing platelets from one donor by cytapheresis during a 1 1 / 2 -to 2-hour procedure on an automated device and then retransfusing the remainder of the blood back into the donor. Each collection contains an equivalent of 6 to 10 units of platelet concentrates. Single-donor platelets are suspended in about 300 mL of plasma, so the same ABO and volume considerations discussed earlier pertain. Single-donor platelets offer the clear benefi t of reducing the risk of multiple-donor exposure to the recipient. Single-donor platelets may also be the only available alternative for recipients who have been alloimmunized by previous platelet transfusions because they may be human leukocyte antigen (HLA) or platelet antigen matched to the recipient. The use of apheresis platelets now exceeds the use of pooled random-donor platelets; however, use of this product in emergency situations is limited by the availability of volunteer donors. 35

Platelet transfusions are indicated for patients bleeding because of thrombocytopenia or functional platelet defects. 36 Guidelines for transfusion continue to evolve, and the current guidelines merely provide a desirable range for platelet counts, assuming normal platelet func-tion (Table 80 -2). There is ample evidence that bleeding medical or surgical patients with platelet counts of 50 � 10 9 /L or above will not benefi t from transfusion if thrombocytopenia is the only abnormality. For critical invasive procedures in which even a small amount of bleeding could lead to loss of vital organ function or death, maintaining the platelet count at 50 � 10 9 /L or greater is typically preferred. The presence of other factors that diminish platelet function, such as certain drugs, foreign intravascular devices (e.g., intra-aortic balloon pump or membrane oxygenator), infection, or uremia, may alter this requirement upward. Patients at risk for small but strategically important hemorrhage, such as neurosurgical patients, may need to be maintained at counts of 80 to 100 � 10 9 /L.

Patients without hemorrhage who have platelet counts of 5 � 10 9 /L or lower appear to be at increased risk for signifi cant hemorrhage. Indications for transfusion to patients with counts above 10 � 10 9 /L are less well established; thus, the majority of guidelines propose prophylactic platelet transfusion to prevent hemorrhage at a threshold of 10 � 10 9 /L. The bleeding time is not a useful procedure in this situation because it is usually prolonged at counts below 80 � 10 9 /L, may be insuffi ciently reproducible, and correlates poorly with the risk for bleeding. 37 Patients undergoing cardiac bypass surgery experience a drop in platelet count and often acquire a transient platelet functional defect from damage associated with the bypass apparatus. 38 Most patients do not experience platelet-associated bleeding, however, so prophylactic transfusion in the absence of bleeding is not warranted. In a patient who continues to bleed postoperatively, more likely causes are a localized, surgically correctable lesion or failure to reverse heparinization. If these conditions are excluded, empiric transfusion of platelets may be justifi ed.

Patients thrombocytopenic by virtue of immunologic destructive processes such as idiopathic thrombocytopenic purpura (ITP) receive little benefi t from platelet transfusions because the transfused platelets are rapidly removed from the circulation. In the event of life-threatening hemorrhage or an extensive surgical procedure, transfusion may prove benefi cial for its short-term effect. Transfusion may be accomplished effectively by pretreatment with high-dose immunoglobulin or high-dose anti-D antiserum (RhIG). 39, 40 Platelet transfusion has been reported to be deleterious in thrombotic thrombocytopenic purpura (TTP), 41 in the related hemolytic-uremic syndrome, and in heparin-induced thrombocytopenia. Cautious administration, in cases of life-threatening thrombocytopenic bleeding only, is prudent.

Prophylactic platelet transfusion for thrombocytopenia secondary to underproduction remains controversial. The common practice of transfusion to maintain the platelet count above 20 � 10 9 /L derives from data published in 1962, which demonstrated an increase in spontaneous bleeding in leukemic patients at that level. 42 However, critical evaluation of the data reveals that serious hemorrhage was not greatly increased until counts fell to 5 � 10 9 /L or lower and that these patients received aspirin for fever, which might have compromised platelet function and enhanced the bleeding. A somewhat more recent study quantitating stool blood loss in aplastic anemia patients defi ned a bleeding threshold at platelet counts of 5 to 10 � 10 9 /L. 43 A prospective study of a more conservative transfusion protocol found that major bleeding episodes occurred on 1.9% of days with counts of less than 10 � 10 9 /L and on only 0.07% of days with counts of 10 to 20 � 10 9 /L. 44 The trigger for prophylactic platelet transfusion in the 5 to 10 � 10 9 /L range, however, applies primarily to stable thrombocytopenic patients. Factors such as fever, use of anticoagulant or antiplatelet drugs, and invasive procedures must be considered when generating a treatment plan for individual patients. Patients experiencing rapid drops in platelet count may be at greater risk than those at steady state and thus may benefi t from transfusion at higher counts. Benefi ts to the patient with more judicious use of platelet transfusion include decreased donor exposure, which lessens the risk of transfusion-transmitted disease; fewer febrile and allergic reactions that may complicate the hospital course; and the potential delay or prevention of alloimmunization to HLA and platelet antigens. 45 The development of refractoriness to platelet transfusions is a serious event heralded by a falling CCI. Poor response to platelet transfusions can be seen in patients with other reasons for platelet consumption, including splenomegaly, fever, trauma and crush injury, burns, disseminated intravascular coagulation (DIC), concomitant drugs, or transfusion of platelets of substandard quality. 46 These factors should be sought and corrected if possible. Alloimmunization is characterized by the development of anti-HLA or platelet-specifi c antibodies, with resultant immune platelet destruction. As many as 70% of patients receiving multiple red cell or platelet transfusions become immunized. 45 Leukocyte depletion of transfused components can prevent or delay this phenomenon, but it is important to use leukoreduced components early in the course of transfusion therapy. 45, 47 When patients fail to achieve expected increments after platelet transfusion, provision of ABO-specifi c platelet concentrates that are less than 48 hours old may improve the response. If no improvement is seen and the aforementioned medical conditions are excluded, the patient should be screened for HLA antibodies or be HLA typed and provided with HLA-compatible single-donor platelets. Alternatively, platelet crossmatching with the patient's serum can be carried out. There is no advantage to unmatched singledonor platelets in this situation.

Standard FFP is prepared by centrifugation of WB and is frozen within 8 hours of blood donation. 1,2 FFP may be stored frozen for 1 year. The usual volume is about 250 mL, depending on the donor's Hct. The most common method of thawing before transfusion is soaking in a 37° C water bath, which requires about 30 to 45 minutes. Once thawed, FFP can be stored refrigerated for a maximum of 24 hours. When prepared and stored in this manner, FFP supplies all the constituents in the amounts normally present in circulating plasma, including stable and labile coagulation factors, complement, albumin, and globulins. By convention, the coagulation factors are present in concentrations of 1 U/mL. Crossmatching to the recipient is not performed, but FFP must be ABO compatible.

Standard FFP is as likely to transmit hepatitis, HIV, and most other transfusion-related infections as cellular components are. New FFP products have recently been introduced in response to concern about the transmission of infectious diseases. One such product is solventdetergent-treated FFP. 48 Solvent-detergent treatment is a means of viral inactivation that removes the infectivity of lipid-enveloped viruses, such as hepatitis B and C and HIV. Because the product is derived from pooled plasma, with as many as 2500 donors in each lot, it has the potential to actually increase recipient exposure to pathogens not inactivated by the solvent-detergent method, such as hepatitis A and parvovirus B19, and be more vulnerable to any newly emerging non-lipid-enveloped agent. A variety of other techniques for reducing pathogen exposure in FFP have been developed, including exposure to low pH or vapor heating and treatment with ultraviolet irradiation, gamma irradiation, or psoralens and light to inactivate pathogens by inducing DNA damage. 49 Because none of the FFP products is entirely free from the risk of disease transmission or other adverse effects and because infection-reducing modifi cations add significantly to the cost of the components, FFP should be used judiciously. 50 It should be administered only to provide coagulation factors or plasma proteins that cannot be obtained from safer sources. FFP is commonly used to treat bleeding patients with acquired defi ciency of multiple coagulation factors, as in liver disease, DIC, or dilutional coagulopathy, or to treat patients with congenital defi ciency of a coagulation factor or other protein for which concentrates or safer sources do not exist. FFP may be indicated for emergency reversal of the coagulopathy induced by warfarin anticoagulants when more concentrated products are not available or for the provision of protein C or S in patients who are defi cient and suffering acute thrombosis. FFP should be administered as boluses as rapidly as feasible so that the resulting factor levels allow hemostasis. The use of FFP infusions without adequate bolus administration is not helpful. FFP should not be used for volume expansion or wound healing or as a nutritional source of protein. FFP does not reverse anticoagulation induced by heparin and in theory might exacerbate bleeding by supplying more antithrombin, heparin's cofactor.

Prophylactic administration of FFP does not improve patient outcome in the setting of massive transfusion or cardiac surgery unless there is bleeding with an associated documented coagulation abnormality. 51, 52 Patients do not usually bleed as a result of coagulation factor insuffi ciency when the international normalized ratio (INR) is less than about 2.0, and even then the results are not always predictable. 53 The partial thromboplastin time (PTT) is not useful in predicting procedural bleeding risk. 54 FFP is often requested prophylactically before an invasive procedure when the patient exhibits mild prolongation in coagulation studies. Most of these procedures may be carried out safely without transfusing FFP. 53, 55 FFP is probably the most misused blood component, as illustrated by retrospective surveys. 56 Coagulation factors are normally present in the blood far in excess of the minimum levels required for hemostasis. As little as 10% of the normal plasma concentration of several factors will effect hemostasis. Conversely, FFP treatment of acquired multiple defi ciencies, as in hepatic failure, is often ineffective because many patients cannot tolerate the infusion volumes required to achieve hemostatic levels of coagulation factors, even transiently. 57 The plasma half-life of transfused factor VII is only 2 to 6 hours. It may be impossible to administer suffi cient FFP every few hours without encountering intravascular volume overload. Finally, in some instances, transfusion of seemingly adequate volumes may still fail to correct the coagulopathy. 58 Careful documentation of both the need for FFP and the adequacy and outcomes of therapy is essential. 59

Cryoprecipitate is manufactured by thawing and centrifuging FFP below 6º C and resuspending the precipitated proteins in about 15 mL of supernatant plasma. 1,2 Each bag is a concentrated source of factor VIII (80 to 120 units), von Willebrand factor (vWF) (50% of original plasma content), fi brinogen (250 mg), factor XIII (30% of original plasma content), and fi bronectin. Cryoprecipitate offers the advantage of transfusing more specifi c protein and less total volume than the equivalent dose of FFP does. It has been used to treat patients with inherited coagulopathies, such as hemophilia A, von Willebrand disease, or factor XIII defi ciency. In the critical care setting, it is more commonly used to replenish fi brinogen, especially in bleeding patients with hypofi brinogenemia caused by dilutional or consumptive coagulopathy. Cryoprecipitate also reportedly improves hemostasis in uremic patients, presumably by reversing the functional platelet defect, 60 but desmopressin acetate (DDAVP) 61 or conjugated estrogens exert similar effects and should be used preferentially to avoid potential transfusion-transmitted disease.

The usual dose of cryoprecipitate to treat hypofi brinogenemia is 10 bags/units to start, then 6 to 10 bags/units every 8 hours or as necessary to keep the fi brinogen level above 100 mg/dL. Each bag/unit of cryoprecipitate carries a risk of disease transmission equivalent to that of 1 unit of blood. For this reason, commercial factor VIII concentrates, recombinant or treated to inactivate viruses, are preferred over cryoprecipitate for treating hemophilia A patients.

Immune serum globulin (IG), RhIG, and hyperimmune globulins for diseases such as hepatitis B and varicella zoster are obtained by fractionation of pooled plasma, followed by chromatography, delipidation, and other steps to remove aggregates and infectious agents. Intravenous IG (IVIG) is available in solution or lyophilized form, with protein content varying by mode of preparation. The available products vary slightly in the amounts of IgA and IgM contained in them, which are mostly present in only trace quantities.

IG preparations can be used to provide passive antibody prophylaxis or to supply IG in certain immunodeficiency states. Hyperimmune globulins may be used to treat active infections in immunosuppressed hosts. Recent applications have exploited IG's immunomodulatory effects in treating a wide variety of disorders with an immune basis. The specifi c mechanism of action of IVIG in such conditions has not yet been identifi ed, but possibilities include interference with macrophage Fc receptor function, neutralization of anti-idiotypic antibodies, and interference with the incorporation of activated complement fragments into immune complexes. A recent review more completely discusses the effects of IVIG on the immune system and its potential uses. 62 RhIG is prepared from pools of plasma obtained from donors sensitized to the red cell antigen D from the Rh group. The standard-dose vial contains primarily IgG anti-D, with a protein content of 300 µg in 1 mL. This dose will protect against 15 mL of D + red cells or 30 mL of WB. 63 RhIG carries no risk of virus transmission. Although RhIG is used primarily in obstetrics, it may also be indicated to prevent alloimmunization in Rh-negative patients receiving small amounts of Rh-positive red cells, as in platelet concentrates. Routine prophylaxis against large numbers of red cells, as in a unit of Rh-positive WB or PRBCs given by accident to an Rh-negative recipient, is not reliable and usually involves the administration of large amounts, but instances of its effective use in these circumstances have been reported. Higher doses of intravenous RhIG have been used in the treatment of ITP.

Plasma-derived colloids include human serum albumin (HSA), available in 5% and 25% solutions, and plasma protein fraction (PPF), available in a 5% solution. Both are derived from pooled donor plasma but are essentially pathogen-free. HSA is composed of at least 96% albumin, whereas PPF is subjected to fewer purifi cation steps and contains at least 83% albumin, with correspondingly more globulins. The 5% solutions are iso-oncotic, whereas the 25% solution of HSA is hyperoncotic and requires infusion with crystalloid solutions.

Potential clinical indications for colloid solutions include hypovolemic shock, hypotension associated with hypoproteinemia in patients with liver failure or protein-losing conditions, as a replacement solution in plasma exchange or exchange transfusion, and to facilitate diuresis in fl uidoverloaded hypoproteinemic patients. Albumin solutions are not indicated as a nutritional source to raise serum albumin. Their use in some indications, particularly for resuscitation, has become controversial, and pulmonary edema has been reported in association with their infusion. 64 Although albumin solutions are reasonably safe products to administer, expense and limited availability restrict their use. Anaphylactic reactions have been reported in less than 0.1% of recipients. The use of PPF has been associated with severe hypotensive episodes, with Hageman factor fragments or prekallikrein activator being demonstrated, 65 thus making PPF a less desirable resuscitation fl uid and contraindicated in cardiac surgery.

Granulocyte concentrates for transfusion are obtained from a single donor by cytapheresis methods, which generally involve the administration of hydroxyethyl starch and corticosteroids to the donor to improve granulocyte yield. Granulocyte colony-stimulating factor (G-CSF) has been added to some collection regimens and increases both cell counts and granulocyte survival substantially. Each collection should contain at least 10 10 granulocytes 1,2 and is suspended in approximately 200 mL of plasma. A signifi cant number of red cells are present, so crossmatching for the recipient is required. Because of the potential risk for graft-versus-host disease (GVHD), granulocytes are usually collected from HLA-matched donors. Granulocytes are stored at room temperature and must be transfused within 24 hours of collection, although sooner is better because of rapid deterioration of the cells.

Patients who may benefi t from granulocyte transfusions include those who are neutropenic (absolute neutrophil count of less than 0.5 � 10 9 /L) and those who are unresponsive to appropriate antibiotic treatment but in whom bone marrow recovery is expected to occur. A course of therapy generally involves daily infusion for 4 to 7 days. Granulocytes have been used for progressive fungal infections in immunosuppressed granulocytopenic patients, in patients with defective leukocytes (e.g., chronic granulomatous disease), and in the neonatal ICU for neonatal sepsis. Randomized trials had suggested that granulocyte transfusions under these circumstances can reduce mortality, but such trials have not been conducted for more than 2 decades. 66 Effective antibiotic regimens and the signifi cant adverse effects associated with the use of granulocyte concentrates, including pulmonary insuffi ciency related to alloimmunization and CMV infection, have limited their use in recent years.

The decision to transfuse blood components, like any therapeutic maneuver, must be made with full awareness of the potential risk to the recipient, as well as the expected benefi ts. Public expectations of a zero-risk blood supply help raise the acuity of physicians' decisions. For some patients, the benefi t from transfusion is so obvious that the associated risks pale in comparison to the consequences of withholding transfusion. However, the clinician's knowledge of the incidence and management of adverse reactions to transfusion is vital, not only to ensure the best patient care but also to provide appropriate patient education and true informed consent.

Almost every patient who receives an allogeneic blood transfusion will experience some adverse reaction if such universal effects as immunomodulation and bone marrow suppression are considered. Measurable reactions to transfusion occur in about 20% of patients; more serious adverse responses may be expected in only 1% to 2% of transfusions. 67 The nature of these adverse reactions ranges from those that are common but clinically unimportant to those that may cause signifi cant morbidity or death (Table 80-3) .

Transfusion in the ICU is a common and often lightly regarded event. However, because the signs and symptoms of severe, life-threatening reactions are frequently indistinguishable from those of troublesome, but less signifi cant reactions, every transfused patient who experiences a signifi cant change in condition, such as an elevation in temperature, change in pulse or blood pressure, dyspnea, or pain, must be promptly and fully evaluated to identify the cause of the reaction and to institute treatment when necessary. The basic approach to all acute reactions should be to maintain a high index of suspicion for acute hemolytic reactions by stopping the transfusion immediately, maintaining venous access with intravenous fl uids, and informing the blood bank laboratory immediately so that the appropriate transfusion reaction protocol can be in stituted and post-transfusion specimens obtained. Early recognition of severe transfusion reactions may be lifesaving.

The most feared reaction to blood transfusion is intravascular hemolysis, caused by the recipient's complementfi xing antibodies attaching to donor RBCs with resultant RBC lysis. ABO incompatibility is most often implicated in these incidents. Intravascular hemolysis is still the single most common acute cause of fatalities associated with the transfusion episode. 68 In addition to hemolysis, complement activation stimulates the release of infl ammatory mediators and cytokines and thereby leads to hypotension and vascular collapse. Activation of the coagulation system may result in DIC. Acute renal failure may also occur, presumably on the basis of immune complex interactions. Morbidity and mortality are directly related to the quantity of incompatible blood transfused, which is why prompt recognition and cessation of transfusion cannot be overemphasized. Misidentifi cation of the patient, or "clerical error," at any time beginning with the process of specimen acquisition through release of the unit and initiation of infusion is the major cause of acute intravascular hemolysis. 69, 70 This reaction is more likely to occur in critical care settings, such as the ICU, operating room, and emergency department, than anywhere else in the hospital. It is far preferable to transfuse uncrossmatched group O red cells than to chance ABO incompatibility caused by improper patient and specimen identifi cation procedures.

The most common clinical sign of hemolysis is fever, with or without chills. 71 Other common signs and symptoms include back or fl ank pain, anxiety, nausea, lightheadedness, dyspnea, and hemodynamic instability. In a comatose or anesthetized patient, many of these symptoms will not be evident; therefore, signs such as hypotension, hemoglobinuria, and diffuse oozing from puncture sites or incisions may be the only notable features.

Immediate management of hemolytic transfusion reactions must include cessation of the transfusion; the remainder of care is supportive. Rapid verifi cation of patient and unit identifi cation must be made, not only to confi rm the suspected reaction but also to prevent a second patient from receiving a reciprocally incompatible unit if a clerical error has been made. Desired end points of supportive care include maintenance of blood pressure, high urine output, and support of coagulopathy or further blood loss. Steroids, heparin, or other specifi c pharmacologic interventions have no role in treatment.

Anaphylactic reactions to blood transfusions are fortunately rare but may be life-threatening. The usual cause is recipient antibody to a component of plasma that the patient lacks, most commonly antibody to IgA in IgAdefi cient individuals. 72 Signs and symptoms include severe malaise and anxiety, fl ushing, dizziness, dyspnea, bronchospasm, abdominal pain, vomiting, diarrhea, hypotension, and eventually shock. Fever and hemolysis do not occur. Management includes immediate cessation of transfusion and standard therapy for anaphylaxis. If anti-IgA antibodies are determined to be the cause of this reaction, the patient must receive blood components donated by IgA-defi cient individuals or, if unavailable, specially prepared washed RBCs and platelet concentrates. Plasmaderived preparations, such as albumin, and IG contain varying amounts of IgA and pose a substantial risk in these patients.

Febrile nonhemolytic reactions (FNHRs) are the most commonly occurring immediate transfusion reaction. These reactions are annoying to the clinician, patient, and transfusion service alike in that they can cause signifi cant discomfort and, because they share certain manifestations with acute hemolytic reactions, must be investigated in every instance. FNHRs occur in approximately 0.5% to 1.0% of transfusion episodes. 73 The etiologic factors are probably complex and multiple, but many reactions are caused by the release of cytokines and pyrogens, either within the transfused unit of blood or as a result of recipient antibodies to donor leukocytes. Clinical signs include fever, with or without chills, usually beginning 1 to 2 hours after the start of the transfusion but occasionally delayed up to 4 to 6 hours. Multiparous women and patients who are multiply transfused are particularly prone to FNHRs. The transfusion must be stopped and the appropriate transfusion reaction evaluation instituted. Antipyretics such as acetaminophen may be administered. Though commonly used, antihistamines such as diphenhydramine are neither preventive nor therapeutic. Once acute hemolysis is excluded, transfusion of a new unit may be instituted. Most patients will not experience a second such reaction. 73 If repeated reactions become problematic, leukocyte-depleted blood components may be supplied. The implementation of ULR results in a reduction in the frequency of all fevers seen after transfusion by only about 12%. 74

Hives and pruritus are relatively common adverse effects of transfusion. 68 They are a hypersensitivity reaction localized to the skin, and their cause is unknown but may include both donor and recipient characteristics. These reactions consist of localized or generalized urticaria beginning shortly after the start of transfusion without other signs or symptoms of anaphylaxis or hemolysis. The transfusion should be temporarily stopped, and antihistamines may be administered. If the hives resolve in a short time, the same unit of blood may be cautiously restarted. If repeated urticarial reactions occur, premedication with antihistamines may be effective, or blood components washed to remove plasma may be required.

Intravascular volume overexpansion is particularly likely to occur in critical care patients with limited cardiac reserve. Aside from the inherent volume of the blood components, the intravenous normal saline concurrently administered adds to the volume load. Unfortunately, normal saline solution is the only intravenous fl uid that may be administered with blood components. With careful attention to transfusion requirements and the use of volume reduction maneuvers available to the transfusion service, volume overload can be minimized in most instances. The frequency of this complication of transfusion is not reported.

Delayed hemolysis is an uncommon but probably underrecognized reaction to transfusion that results from the stimulation of a primary or secondary (anamnestic) recipient antibody response to foreign RBC antigens. These antibodies are undetected at the time of transfusion but increase after transfusion in a manner analogous to the vaccination "booster" effect. These reactions typically occur 3 to 14 days after transfusion but are unrecognized because of the lack of a clear temporal association with transfusion. Fever, chills, and an unexplained decline in Hct are the usual signs. 75 Transient elevation in bilirubin and lactate dehydrogenase may also occur. The diagnosis is established by a positive direct antiglobulin (Coombs) test resulting from recipient antibody coating donor RBCs. The antibody may be identifi ed by eluting it from the RBCs or by demonstrating it within the recipient's serum. The specifi city of the antibody is often against such RBC antigens as the Rh family, Kidd, Duffy, or Kell systems. Hemolysis may not occur, but if it does, it is likely to be extravascular and only rarely causes renal failure or DIC.

Prevention of these reactions is diffi cult. Alloimmunization to foreign RBC antigens occurs in approximately 1% of transfusions. 67 Detection of delayed antibodies is the purpose for requiring a new blood bank specimen every 72 hours if the patient has recently been transfused. Permanent transfusion records should record the occurrence of delayed antibodies, even though they may not be apparent at a later crossmatch. Access to transfusion databases is critical for the care of patients with a past history of transfusion.

Transfusion-related acute lung injury (TRALI) is an uncommon (0.02%) 76 but serious adverse effect of transfusion that has only recently been gaining recognition. Similar reactions have been called pulmonary leukoagglutinin reaction or noncardiogenic pulmonary edema. These reactions consist of acute respiratory distress syndrome (ARDS), which develops 1 to 6 hours after transfusion. Signs and symptoms include bilateral pulmonary infi ltrates, hypoxemia, fever, and occasionally hypotension. Monitored patients are found to have normal or low pulmonary wedge pressure and central venous pressure, as contrasted with patients experiencing volume overload. If adequate respiratory support and oxygenation are established promptly, spontaneous resolution generally occurs within 1 to 4 days. Deaths have nonetheless occurred, particularly with a delay in diagnosis. 77, 78 Episodes of TRALI appear to have several possible causative mechanisms. Some cases may be caused by donor antibodies reacting with recipient neutrophil or HLA antigens. 79 Plasma factors related to blood storage have also been implicated, such as lipid substances from deterioration of donor cell membranes that prime recipient neu-trophils, which then damage the pulmonary vasculature and lead to increased capillary permeability and an ARDSlike syndrome. 80 Other clinical factors may contribute to increased risk, such as cardiac bypass surgery or other procedures. In the antibody model at least, the implicated antibody is unique to the donor and the affl icted recipient will probably not experience another such reaction, provided that the recipient is not exposed to the same donor. TRALI is undoubtedly under-recognized in the critical care setting and may frequently be confused with fl uid overload or cardiogenic pulmonary edema.

Transfusion-associated GVHD (TA-GVHD) is a welldocumented, but probably under-recognized, highly lethal immunologic complication of blood transfusion. 81 Immunocompromised patients infused with blood components containing viable donor lymphocytes are at risk for engraftment of the allogeneic lymphocytes and ensuing rejection of recipient (host) tissues. Transfusion recipients who are at highest risk include neonates, especially the very premature, bone marrow and organ transplant recipients, and leukemia and lymphoma patients. TA-GVHD has also been reported in patients after cardiac surgery who received designated donor blood from relatives; presumably, the HLA antigenic differences between donor and recipient were insuffi cient to stimulate a recipient immune response but suffi cient to elicit a donor immune response. 82 The onset of TA-GVHD is usually within 8 to 30 days after transfusion, and it is manifested as fever and rash, followed by diarrhea and evidence of liver and bone marrow injury. TA-GVHD differs from that seen in bone marrow transplantation (BMT) by its involvement of the marrow and by far greater mortality. Treatment is largely ineffective, and mortality exceeds 90%.

Irradiation of blood components at 25 Gy prevents TA-GVHD by eliminating the donor lymphocyte mitogenic response. All cellular blood components should be irradiated before transfusion to high-risk patients. The functions of the cellular components of blood are unaffected, although damage to RBC membranes limits postirradiation storage of PRBCs. 83 Blood donated by a relative for any patient should be irradiated, as should HLA-matched or crossmatched platelet products.

Allogeneic blood transfusion has been shown to modulate and suppress the recipient's immune response, an effect fi rst noted with kidney transplantation. 84 Immunosuppression in a critical care setting is generally undesirable, but whether transfusion has a signifi cant impact is debated. Ongoing clinical issues center around two areas of controversy: the putative association between blood transfusion and increased numbers of postoperative infections and increased and more rapid rates of tumor recurrence in surgical oncology patients with certain malignancies. There has been no resolution of either issue despite a few prospective trials having been performed. The largest pro-spective trial of colorectal cancer resection, for example, is negative, 85 but a meta-analysis of the extant data suggests that an adverse effect on recurrence does exist. 86 Similarly, most of the randomized trials of postoperative or critical care unit infections are too small to indicate an effect of transfusion, but all point in the direction of an adverse effect. 87, 88 Controversy will continue until larger randomized trials are conducted.

The precise mechanism of the immunosuppression induced by allogeneic transfusion has not yet been delineated, and several mechanisms may be involved. 89 Alterations identifi ed in laboratory and clinical transfusion recipients have included depression of the T-helper/Tsuppressor lymphocyte ratio, decreased natural killer cell activity, diminished interleukin-2 generation, formation of anti-idiotype antibodies, impairment of phagocytic cell function, and chronic persistence of donor lymphocytes (microchimerism), suggestive of low-level GVHD. Difficulties in analysis of human data arise because patients requiring blood transfusions have conditions that themselves induce immune changes. There is some evidence, bolstered by the results of two large clinical trials, to suggest that leukocyte reduction of blood components reduces or eliminates this immunosuppressive effect. 90 Proponents of this viewpoint argue that for this reason, ULR would benefi t most patients receiving blood transfusions and lead to fewer infections, tumor recurrences, and other related putative risks of transfusion, all potentially resulting in saving lives and cost. Prospective trials will be extremely important. 91

Public awareness of transfusion-associated acquired immunodefi ciency syndrome (AIDS) has done more to revolutionize transfusion practice than any other transfusion risk by resulting in more conservative blood use, more stringent donor selection criteria, and improved screening tests. The result is that viral transmission rates are now diffi cult to measure, and the risk of transfusionrelated infectious diseases is lower than ever. 92 The current best estimate is that 3 to 4 units per 10,000 will transmit some kind of infection 93 if agents such as CMV or Epstein-Barr virus are included. Bacterial infection has become the most common infectious risk thanks to increasingly sensitive donor screening tests, including nucleic acid testing (NAT) to detect viral DNA or RNA, which has shortened the infectious period and reduced the risk for post-transfusion hepatitis (PTH) and other viral infections.

Several fatalities are reported yearly from the transfusion of blood components contaminated with viable, proliferating bacteria, with or without the accumulation of endotoxin. 94 Platelet concentrates, because they must be stored at room temperature, are particularly prone to bacterial growth, with a reported incidence of 6 in 10,000 transfusions. 95 Organisms isolated from platelets and implicated in fatal transfusion reactions include Staphylococcus and Streptococcus species and gram-negative bacilli. Fatalities resulting from bacterial contamination of refrigerated RBCs have occurred as well and more often involve cryophilic bacteria. RBC transfusions contaminated by Yersinia enterocolitica have been consistently reported for a decade. 96 Transfusion reactions caused by bacterial or endotoxin contamination are fortunately quite rare, but mortality exceeds 60%.

Signs and symptoms of reactions caused by microbial contamination overlap those of hemolytic transfusion reactions and consist primarily of fever and hypotension, along with other signs of endotoxic shock. If recognized promptly, a Gram stain of the implicated unit can be prepared immediately and, if positive, appropriate antibiotic and supportive therapy instituted. Autologous blood components may also be contaminated at the time of collection; therefore, reactions occurring in patients who are receiving their own blood should not be dismissed but instead should be evaluated as fully as though the patients had received allogeneic blood.

The success of viral screening measures is most clearly illustrated by the fall in the risk for PTH over the past 2 decades. Although PTH continues to be a signifi cant cause of morbidity and mortality, the nature of PTH has changed through the years with the stepwise institution of various donor screening measures. The elimination of paid donors in 1972 and the successive introduction of immunologic tests for hepatitis B have resulted in a steady reduction in the rates of PTH caused by hepatitis B virus (HBV) to approximately 17 per million units of transfused blood products. Although about 30% to 40% of HBV transmissions will result in acute hepatitis, chronic HBV infection develops in less than 10% of such patients. In contrast, the risk for chronic hepatitis C virus (HCV) infection after transfusion is higher, nearly 50%, and the long-term risk for cirrhosis-or hepatocellular carcinoma-related mortality is about 15% over more than 20 years after PTH secondary to HCV. 97, 98 The clinical course of hepatitis A is generally milder, and the lack of a chronic carrier state means that with donor screening for symptoms of the acute illness, the risk of transmission is much lower, estimated at less than one in a million units. 99 The prevalence of hepatitis B surface antigenemia among fi rst-time blood donors is 0.7%, and the prevalence of hepatitis C antibodies in donors is approximately 0.1% to 0.5%. At this time, given the sensitivity of current screening assays, including the latest generation of enzyme immunoassays (EIAs) and NAT, the current risk of PTH resulting from HCV is believed to be about 1 in 150,000 or less. 100 Although HBV is still implicated in PTH (attributable to the seronegative "window" period in newly infected donors), the risk of transfusion-associated hepatitis B is about 1 in 200,000 units. 100

Retroviruses, RNA-based viruses characterized by their reverse transcriptase and integration into the host genome, and lentiviruses, a subset of retroviruses, are ubiquitous in animals and were initially identifi ed in humans in the early 1980s. Those known to be capable of transmission by transfusion are HIV-1, HIV-2, and human T-cell leukemia/lymphoma virus (HTLV) I and II.

Transfusion-associated AIDS was initially reported in late 1982. 101 The fi rst report of an associated viral agent did not appear until late in 1983, and in March 1985 the screening enzyme-linked immunosorbent assay (ELISA) to detect antibody to HIV-1 was licensed and immediately incorporated into the blood-screening process. Improved confi dential donor screening appeared to decrease the risk of infectious units appearing in the donor pool. 102, 103 The discovery that heat treatment reduced transmission resulted in a reduction in transmission by plasma products, especially to persons with hemophilia. Clinical AIDS developed in more than 90% of recipients of infected blood products, and the vast majority succumbed to the disease. Removal of donor units with seropositivity by ELISA was insuffi cient to prevent transmission of HIV-1; several hundred cases were reported annually after introduction of the ELISA test. Subsequent development of an assay for the p24 antigen and then NAT has lowered the risk of transfusion-associated HIV-1 infection to less than one in a million (see Table 80 -3).

Despite donor screening and sensitive assays, including EIA, NAT, and p24 antigen, an extremely small, but fi nite risk of HIV-1 transmission by screened blood transfusions remains. This risk is largely due to the seronegative "window" period experienced by newly infected donors, which is estimated to be an average of 16 days. 100 A second retrovirus, HIV-2, fi rst described in residents of countries in West Africa and subsequently detected in migrants to western Europe, causes an immunodeficiency syndrome similar to that caused by HIV-1. Although very few cases of HIV-2 have been reported in the United States 104, 105 and there have been no reported transfusion-transmitted cases, experience with other retroviruses suggests that screening may prevent the majority of potential transmission. Therefore, donated blood is now screened by an assay for the presence of antibody to HIV-2.

The retrovirus HTLV-I is the causative agent of adult T-cell leukemia (ATL) and is strongly implicated in the chronic, progressive neurologic disorder termed tropical spastic paraparesis or HTLV-I-associated myelopathy (TSP/HAM). HTLV-II has been linked to hairy cell leukemia, but no transfusion-transmitted cases have been reported. The virus exhibits strong serologic crossreactivity with HTLV-I such that screening assays fail to distinguish between antibodies to either virus.

Transfusion-transmitted HTLV-I has been demonstrated. 106 TSP/HAM has developed in a small percentage of infected transfusion recipients, but no transfusionassociated cases of ATL have been seen. Approximately 0.025% of donors in the United States are seropositive for HTLV-I and HTLV-II 107 ; further testing reveals the majority of them to be HTLV-II. Donated blood is currently screened for antibodies to HTLV-I and HTLV-II. The estimated risk of HTLV transmission by screened negative blood is believed to be 1 in 250,000 to 2 million.

CMV is a human herpesvirus that establishes latent infection in the host's tissues, particularly leukocytes, and is transmitted by all cellular blood components. 108 Seropositivity, or the presence of antibody, denotes previous exposure to the virus but does not confer protective immunity. Secondary reinfection or reactivation of latent infection can occur. Antibodies to CMV persist for life and serve as a marker indicating the potential for transmission of live virus.

Immunocompetent recipients of transfused CMVpositive blood experience minimal morbidity and mortality. The majority are asymptomatic, whereas a heterophile-negative mononucleosis syndrome may develop in a few. Immunocompromised patients, however, may suffer life-threatening manifestations such as severe interstitial pneumonitis, gastroenteritis, hepatitis, or disseminated disease. Several groups of patients are at particular risk (Box 80-1), 109 and these patients should receive blood incapable of transmitting the virus. Other patients may benefi t from CMV-negative blood as well, such as seronegative solid organ transplant recipients or autologous BMT patients. Screening of donated blood for CMV is not routinely done but can be performed quickly if necessary. Because the prevalence of donor seropositivity is quite high in some regions (50% to 70%), CMVseronegative blood may not be readily available. Blood that is leukocyte depleted ("CMV safe") may be as effective as seronegative blood in the prevention of CMV transmission, although a recent meta-analysis of clinical trials comparing the two methods suggests that CMVnegative blood products might have a slight advantage over leukocyte-depleted products. 110

Many blood-borne parasites may be transmitted by transfusion, although this is a rare occurrence in the United States because of donor screening questions and the low endemicity of implicated agents. 111 Changing immigration patterns and worldwide travel, however, make transfusion-transmitted parasites an increasing concern.

On a worldwide basis, malaria is the most important transfusion-transmitted infective organism, although only about three cases occur in the United States each year. Such infections are manifested by delayed fever, chills,

Seronegative pregnant women Seronegative premature infants weighing less than 1200 g Seronegative allogeneic or autologous bone marrow transplant recipients Seronegative transplant recipients of seronegative organs diaphoresis, and hemolysis, often masked by underlying medical conditions. Fatalities have occurred. Babesiosis, a tick-borne disease, is endemic in regions of the United States, especially the northeast, with a seroprevalence of about 4%. Transfusion-transmitted cases have been reported, with asplenic or immunocompromised patients being particularly susceptible. With increases in the number of Latin American immigrants to the United States, American trypanosomiasis (Chagas' disease), which is endemic in Latin American countries, has emerged as a potential pathogen. Other parasitic diseases that have been transmitted by transfusion include toxoplasmosis, leishmaniasis, and Lyme disease.

Parvovirus B19 has now been recognized as a pathogen capable of transmission by transfusion, with typical clinical fi ndings and the potential for severe hematologic complications. Cases of Epstein-Barr virus infection with a typical mononucleosis-like illness have been reported after transfusion. West Nile virus has also been transmitted by transfusion. H2N1 infl uenza, severe acute respiratory syndrome (SARS), and other new viral infections should be capable of transmission by transfusion, although cases have not been reported and the prevalence of asymptomatic disease is unknown. A rising area of concern is the transmission of prion disease, either Jacob-Creutzfeldt disease or bovine spongiform encephalopathy (BSE). Donor referral criteria were implemented in 1987 for these diseases, and transmission of BSE has been reported in the United Kingdom.

Massive transfusion is defi ned as the administration of blood components in excess of one blood volume within a 24-hour period. In an average adult (70 kg), this represents approximately 10 units of WB or equivalent PRBCs, crystalloid solution, and other components. Massive transfusion, especially in the range of 20 or more units of blood products, causes complications not generally seen in usual transfusion practice: accumulation of undesirable substances present within banked blood and dilutional depletion of normal blood constituents that are lacking in stored units. Trauma victims, surgical patients undergoing extensive procedures, and patients with vascular or coagulation disorders may be massively transfused in the critical care setting. Survival of the massive transfusion episode is determined more by the nature and degree of the patient's injuries or medical conditions than by the transfusions themselves, but the presence of adverse effects of massive transfusion can complicate patients' courses in the ICU. Transfusion of large quantities of stored blood defi cient in functional platelets often results in hemostatic defects or outright thrombocytopenia. Circulating platelets consistently decrease in inverse proportion to the amount of blood administered, with the hemostatically signifi cant level of 50 � 10 9 /L reached after 20 U. 112, 113 Functional defects have also been noted, and the bleeding time is prolonged. 114 Despite these laboratory changes, severe diffuse bleeding develops in less than 20% of massively transfused patients, and no laboratory studies predict those who will. Prophylactic platelet transfusion has not been shown to be of benefi t. 115 Platelet counts may return to hemostatically effective levels quickly in patients with normal marrow function.

Currently, resuscitation of massively bleeding patients is most often accomplished with PRBCs in combination with crystalloid solution. This should result in hemodilution to about 60% of normal plasma factor levels after the transfusion of about 10 units; this factor level can effect normal hemostasis. In reality, however, crystalloids may be given in excess of PRBCs, so after 10 units is transfused, less plasma protein may remain. Bleeding is unlikely until prothrombin time (PT)/INR and PTT prolongations exceed 1.5 to 1.8 times the midpoint normal range, the equivalent of an INR approaching 2.0. 113 As with platelets, prophylactic administration of FFP has not proved effective in preventing diffuse bleeding. 116 Thus, the decision to transfuse should be made on an individual basis, as determined by the presence of bleeding or unacceptable risk in patients with documented abnormalities in coagulation.

One new area of controversy in the treatment of patients with massive hemorrhage is the use of recombinant activated factor VII. This new agent was created for the treatment of hemophiliac patients with high titers of antibodies to factor VIII, which makes them unable to benefi t from transfusion of recombinant factor VIII. Activated factor VII bypasses that problem by binding to tissue factor and directly activating thrombin and hence generating fi brin. 117 It is extremely expensive, has a short half-life, and carries a risk of inducing pathologic thrombosis, with potentially grave consequences. Nevertheless, in numerous case reports, this new agent appears to potentially be benefi cial if used early in the resuscitation of massively injured patients. Unfortunately, its unsupervised use has also resulted in thrombotic complications and relative lack of success, both of which suggest that carefully controlled clinical trials are appropriate. 118 Blood preservative solutions contain excess citrate, which anticoagulates stored blood by binding ionized calcium. WB contains approximately 1.8 g of citrate/citric acid per unit in the plasma fraction. Patients with normal liver function can metabolize the citrate load in 1 unit of WB in 5 minutes, but hepatic impairment may extend removal to 15 minutes or longer. Toxicity may result when citrate is administered in excess of the metabolic rate, thereby causing a decrease in ionized calcium levels. 119 Although paresthesias, cramps, and myoclonus accompany citrate excess, the chief danger of hypocalcemia is depression of myocardial contractility and potential prolongation of the QT interval. Because the effects of citrate are transient and the use of PRBCs containing little residual citrated plasma is far more common than massive transfusion with WB, routine administration of calcium is not indicated; clinically signifi cant rebound hypercalcemia may result. Calcium infusion should be limited to hypoperfused patients with hepatic or cardiac failure who manifest citrate toxicity, and careful monitoring is essential.

As potassium leaks from RBCs during storage, up to 7 mEq of extracellular potassium may accumulate in each unit. However, dangerous levels of potassium rarely develop in adults from stored blood; the potassium level is more likely to be determined by the patient's acid-base status. 117 Studies of massively transfused patients have demonstrated a wide range of potassium levels, with hypokalemia seen as frequently as hyperkalemia. Because of the many physiologic mechanisms altered during resuscitation, including those of the respiratory, renal, cardiac, and hepatic systems, it is impossible to predict the net effect of massive transfusion on serum potassium levels.

The pH of banked blood drops during storage, from 7.16 at the time of collection to as low as 6.73 after several weeks of storage. Administration of large quantities of acidic blood, together with the metabolic acidosis common in these patients before resuscitation, would lead one to expect worsening acidosis as the outcome of massive transfusion. However, patients are more likely to exhibit metabolic alkalosis at the end of the transfusion episode, 120,121 partly because of improved tissue perfusion and the metabolism of citrate and lactate to bicarbonate. Patients in renal failure may be unable to handle the bicarbonate load and require dialysis. Acidosis persisting after transfusion suggests inadequate tissue perfusion. 119 Empiric administration of bicarbonate to counter the acid load is not warranted and may contribute to the deleterious effects of hypercapnia in patients with impaired ventilation.

As discussed previously, the level of RBC-associated 2,3-DPG in banked blood declines during storage, which increases the affi nity of hemoglobin for oxygen and thereby results in decreased oxygen off-loaded to tissues. Even in massively transfused patients, it has been diffi cult to document a clinical impact of this shift, and no reliable method for restoring red cell 2,3-DPG has been developed.

WB and PRBCs are stored at approximately 4º C and require 30 to 45 minutes to warm to room temperature. Elective transfusions at standard fl ow rates are tolerated without the need to warm the blood; however, core body temperature, measured by esophageal probe, can fall to 30º C or lower with the administration of large volumes of cold blood over a period of 1 to 2 hours. 122 Adverse effects of hypothermia include a decreased heart rate and myocardial contractility, cardiac arrhythmias, increased affi nity of hemoglobin for oxygen resulting in decreased tissue oxygen delivery, DIC, and impaired ability to metabolize the citrate load of stored blood. Both blood warmers and patient warming may be instituted during massive transfusion, and patient core temperature should be monitored during such resuscitative efforts.

Whether massive transfusion in and of itself is a cause of ARDS is another source of controversy. There are certainly theoretical reasons why massive transfusion might precipitate ARDS: all cellular transfusions contain damaged or activated WBCs, cell membranes, aggregated platelets, and microthrombi, all of which are capable of lodging in and damaging pulmonary capillaries. Despite this possibility, neither microfi ltration of transfusions nor routine leukocyte depletion has shown a signifi cant impact on the incidence of ARDS in massively transfused patients. 123 Certainly, other causes of ARDS exist in patients who undergo massive transfusion, and the possibility of volume overload and TRALI should be considered in the evaluation of patients with hypoxia and diffuse pulmonary infi ltrates after massive transfusion. Management of such patients is supportive, consistent with the overall management of massive transfusion. 124, 125 Autoimmune Hemolytic Anemia Patients with autoimmune hemolytic anemia (AIHA) have an autoantibody, usually of broad specifi city, that fi xes itself to their RBCs and triggers extravascular immune-mediated destruction. Patients with AIHA have a positive direct antiglobulin test 126 (DAT, commonly known as the Coombs test) and varying degrees of he molysis, and their autoantibodies cause agglutination of RBCs from all donors during crossmatching. If the hemolysis is brisk, patients may require red cell transfusion to support oxygen needs before medical management of the AIHA is effective. Hence, transfusion is diffi cult because agglutination during crossmatching interferes with proper defi nition of compatible units of RBCs and because the transfused RBCs are themselves subject to the same immune hemolysis as the host RBCs. Many blood banks have methods for depletion of autoantibodies from the recipient's plasma and elution of antibodies from RBCs to arrive at a proper crossmatch. 127 Although such crossmatches are time consuming and not generally available on an emergency basis, they can be lifesaving. Criteria for transfusion should remain the same as for other recipients.

RBCs are crossmatched for red cell antigens in the ABO and Rh 0 (D) group and for other red cell antigens when antibodies are present. However, there are several hundred other red cell antigens in the human family, and with repeated transfusion recipients may become alloimmunized to other antigens. Generally, alloimmunization occurs in approximately 1% of transfusions, 68 but the prevalence of alloantibodies is higher in chronically transfused, relatively immunocompetent patients, especially African Americans, whose distribution of red cell antigens has signifi cant variation from the white population. Alloimmunization rates of 30% or higher may be found in chronically transfused patients with hemoglobinopathies who have not received RBCs matched to potent minor antigens such as Kell, Duffy, and Lewis. Alloimmunization may present diffi culties in crossmatching of blood, to the point that compatible blood must be obtained from raredonor registries, if at all. Other patients present unresolved serologic problems in that the alloantibody is never precisely identifi ed yet the majority of blood available for transfusion is incompatible. The delay engendered by working with multiple or unidentifi ed antibodies may be unacceptable in some critical care situations in which the need for oxygen-carrying capacity leaves no choice but to transfuse incompatible blood. The behavior of these antibodies in the laboratory may assist in predicting the clinical outcome of the incompatible transfusion. 128 Special procedures such as clearance studies, 129 fl ow cytometry 130 and in vivo crossmatching (cautious administration of a small aliquot of blood, with subsequent observation of serum and urine for evidence of hemolysis) are useful if time permits.

Emergency transfusion of type O, Rh-negative uncrossmatched blood is generally reserved for the resuscitation of trauma patients, for whom the delay in crossmatching may be life-threatening. The risks of alloimmunization are generally accepted as low. Even Rh-positive type O RBCs may be used because rates of alloimmunization to Rh 0 (D) are low under the circumstances of emergency transfusion. 128, 131 

DIC can present the clinician with diffi cult therapeutic choices. This common disorder in critically ill patients may be manifested as severe hemorrhage or thrombosis. Therapy is primarily directed at alleviating the cause and supporting the patient. Supportive therapy includes the transfusion of components needed to correct the bleeding diathesis caused by the consumption of platelets and fi brinogen, in addition to PRBCs to restore oxygencarrying capacity. Platelets and fi brinogen (as cryoprecipitate) are the most useful components needed to repair the coagulopathy, but their use risks merely "fueling the fi re" and increasing the microthrombosis of DIC. Heparin anticoagulation is controversial 132, 133 and may increase the risk of bleeding, especially if depleted factors are not replenished. No defi nitive clinical trials have endorsed the routine use of heparin, and randomized trials of other components and coagulation inhibitors have uniformly been negative. In general, the use of heparin and antifi brinolytic agents has been confi ned to the most severe and protracted cases of DIC. 134

Cirrhotic patients or those with fulminant hepatic failure have a variety of hemostatic disorders that complicate transfusion management of a bleeding patient. 135 Hepatic synthesis of coagulation factors may be markedly diminished, thereby necessitating replacement by FFP or cryoprecipitate. Patterns of factor diminution may vary between acute hepatic necrosis and chronic cirrhosis. 136 Associated hemodynamic alterations may make it impossible to administer the volumes required for effective hemostasis, however, and any effect is transient. The use of factor concentrates or antifi brinolytic agents may precipitate thrombosis. Activation of fi brinolysis and decreased clearance of activated factors may produce or mimic chronic DIC, thus further exacerbating the factor defi ciencies and impairing coagulation. Abnormal platelet function and thrombocytopenia may contribute to the coagulopathy of liver disease, with concomitant splenomegaly reducing the effectiveness of platelet transfusions.

Bleeding in uremic patients is exacerbated by an acquired platelet defect, in part secondary to dialyzable circulating molecules soluble in platelet membranes. Plateletassociated vWF and plasma high-molecular-weight vWF multimers have also been shown to be decreased, 137 which may explain the benefi t shown by DDAVP 138 and cryoprecipitate in shortening the bleeding time and improving hemostasis in some uremic patients. Raising the Hct by red cell transfusion in anemic patients has also been shown to shorten the bleeding time, presumably as a result of blood vessel wall-laminar blood fl ow interaction. Transfusion of platelets in the absence of thrombocytopenia is unlikely to be of benefi t because the transfused platelets rapidly become dysfunctional. More aggressive hemodialysis is the most widely accepted method of reducing platelet dysfunction.

BMT patients are vulnerable to the severe infectious and toxic side effects of ablative treatment and hence may be cared for in critical care units. These patients may have intensive red cell and platelet transfusion requirements and need specialized products such as CMV-negative and irradiated blood components. A blood bank problem uniquely encountered in BMT is the need to switch the patient's ABO group because of an ABO-mismatched transplant, thus necessitating an exchange transfusion of red cells and plasma-containing products (i.e., platelet concentrates) of differing ABO type to avoid hemolysis of donor and recipient cells. BMT patients may also manifest an increased rate of delayed hemolytic reactions 139 as donor "passenger" lymphocytes recognize recipient or transfused red cell antigens. Patients should be monitored particularly closely between days 10 and 20 after a minormismatched allogeneic transplant, and aggressive transfusion should be undertaken if the hemoglobin level falls and the DAT result becomes positive.

The safest transfusion is one that is not given. Therefore, alternatives to blood component therapy continue to be sought and are valuable adjuncts in some instances. It is possible to limit homologous blood exposure by the appropriate use of pharmacologic agents that promote hemostasis and the administration of recombinant hematopoietic growth factors or biologic growth modifi ers to stimulate marrow hematopoiesis.

Only one substitute for RBC transfusions has been approved in the United States, a polyfl uorocarbon oxygen carrier with signifi cant limitations as a blood substitute. 140 Other preparations that have been explored in clinical trials are cell-free hemoglobin solutions cross-linked or polymerized by chemical manipulation to prevent rapid clearance from the circulation. They are intended to provide short-term oxygen-carrying capacity for acutely ill patients and have the advantage of not requiring crossmatching or infection control. Although these proposed products may have a longer shelf-life and are easier to transport, their drawbacks are many. Most have a circulatory half-life of only about 24 hours. The oxygen dissociation curve for these substitutes is also frequently not favorable: either a high FIO 2 is required to "load" these molecules or they are less likely to deliver oxygen efficiently at lower PO 2 levels. 141 Because the hemoglobin source is reclaimed bovine or human red cells, it is unlikely that patients who do not accept blood components because of their religious beliefs (Jehovah's Witnesses) will accept these types of hemoglobin solutions. One product in development uses recombinant technology to generate hemoglobin, and it is hoped that this solution may be acceptable to these patients.

The licensed perfl uorocarbon solutions have failed to demonstrate any utility as intravascular oxygen carriers because of their unfavorable P-50 (oxygen half-saturation pressure) and oxygen off-loading characteristics. They are fi nding limited application in regional oxygenation during angioplasty or stent placement procedures and a more novel use in "liquid ventilation." This involves the ventilation of intubated patients experiencing severe pulmonary compromise with superoxygenated perfl uorocarbon solutions in place of oxygen-enriched air. 142

The synthetic vasopressin analogue DDAVP increases plasma factor VIII : c and promotes the release of vWF from endothelial stores. 143 DDAVP has provided effective hemostasis in bleeding patients with mild hemophilia A and type I von Willebrand's disease and has been used as prophylaxis for patients undergoing surgery. DDAVP reportedly improves platelet function in some patients with qualitative platelet disorders associated with uremia, 136 cirrhosis, and aspirin ingestion. Studies of its effi cacy in cardiopulmonary bypass procedures are confl icting, but a subset of these patients may benefi t. The chief drawback to its use is tachyphylaxis, which develops in essentially all cases after short-term repeated administration.

The lysine analogues ε-aminocaproic acid and tranexamic acid inhibit fi brinolysis by blocking the binding of plasminogen and plasmin to fi brin. These antifi brinolytic agents may decrease bleeding and thus the need for homologous blood components in patients with hemophilia, thrombocytopenia, and systemic fi brinolysis. A novel and effective use of tranexamic acid involves administration as a mouthwash in preparation for oral surgery in patients with hemophilia or those receiving oral anticoagulant therapy. 144 The most serious side effect of these agents when systemically administered is thrombosis; thus, it is important to use them appropriately and monitor the patient carefully during their use.

Aprotinin is a naturally occurring bovine serine protease inhibitor that acts on plasma serine proteases such as plasmin, kallikrein, trypsin, and some coagulation proteins. Aprotinin has been shown to reduce blood loss in patients undergoing cardiopulmonary bypass surgery 145 by inhibiting fi brinolysis and preventing platelet damage. However, more recent reports of renal injury and longterm mortality may mean an end to its use. 146 Aprotinin has been used extensively in liver transplantation, which involves high blood loss. Repeated administration poses the risk of anaphylaxis and renal dysfunction.

When time permits, vitamin K is the preferred agent to reverse the coagulopathy induced by oral anticoagulants. Normalization of the PT can be seen in as few as 6 to 12 hours. Additionally, selected cirrhotic patients may exhibit improvement in the PT when treated with therapeutic doses of vitamin K. Many patients in critical care units exhibit a prolonged PT, especially if dietary supplements are limited and broad-spectrum antibiotic therapy is given. Vitamin K is a safe and effective agent for reversing this effect.

Recombinant erythropoietin (EPO) has dramatically reduced the red cell transfusion requirements of patients in chronic renal failure. EPO also has applications in the adjunctive treatment of the anemia of premature infants and the anemia of chronic disease, especially rheumatoid arthritis, cancer, and AIDS. Studies of its effi cacy in reducing perioperative red cell transfusion requirements by increasing the yield of predeposited autologous blood or stimulating bone marrow synthesis after surgery have shown benefi t in reducing blood transfusion, although preoperative planning and autologous deposits are required. 147 In contrast and probably because the impact of EPO is not immediate, the effi cacy of EPO in the ICU is unproven and awaits the results of large clinical trials.

Recombinant growth factors such as granulocytemacrophage colony-stimulating factor (GM-CSF) and G-CSF stimulate marrow production of leukocytes by enhancing several different granulocyte and macrophage functions. These agents are fi nding application in reducing the neutropenic period in BMT and cancer chemotherapy by increasing the leukocyte count in hypoproliferative marrow conditions. These myeloid growth factors are replacing granulocyte transfusions for their few remaining indications.

Cell salvage equipment has been in clinical use for several decades, and although cell salvage is clearly capable of rescuing otherwise "lost" red cells, its full impact on transfusions has been poorly documented. Cell salvage generally consists of collection of shed blood from a clean, uncontaminated operating fi eld, followed by removal of the cellular elements and retransfusion into the patient. Cell salvage has been used both intraoperatively and postoperatively, especially in cardiac surgery. Although the clinical studies of cell salvage have many fl aws, the overall success of this therapy in reducing transfusion has resulted in its wide application. 148 Risks include bacterial contamination, febrile reactions, triggering of DIC, and coagulopathy as a result of dilution. When combined with acute intraoperative hemodilution, this technology is also potentially cost saving. 149

The word apheresis is derived from the Greek aphairein, "to take away"; thus, therapeutic hemapheresis is performed to remove unwanted plasma constituents (plasmapheresis) or blood cells (cytapheresis). Automated cell separators use centrifugation or membrane fi ltration to remove and concentrate the selected blood element. Many of the same devices used to prepare apheresis blood components for transfusion are used to perform patient procedures, so therapeutic apheresis is often administered under the auspices of the transfusion medicine service. Rapid removal of plasma or cells may fi nd several applications in intensive care practice (Box 80-2).

The goal of plasmapheresis, or plasma exchange (PE), is to remove or reduce the levels of an undesirable plasma constituent or, alternatively, by means of plasma replacement, to supply a missing substance. The agent to be removed by PE is thought to be an autoantibody in some of the neurologic, renal, or hematologic conditions treated in this manner. 150 Immunomodulation by PE is another explanation for its effect, a theory indirectly supported by the equivalent effi cacy of IVIG therapy for several of these disorders. 151 PE for the amelioration of hyperviscosity from either excess IgM in Waldenström's macroglobulinemia or excess Ig in multiple myeloma is an effective temporizing measure in the treatment of these conditions. 152 Plasmapheresis with PE is the standard therapy for TTP. 153 Unfortunately, few controlled trials of PE exist, although anecdotal reports abound. PE is seldom the defi nitive treatment of most of these conditions and is used most appropriately as a short-term adjunct to other medical modalities. The kinetics of PE predicts that a one-volume exchange removes 65% of a given plasma constituent if the blood volume does not change or additional synthesis or mobilization of the substance does not occur. Two or three volume exchanges remove 87% and 95%, respectively. Highly protein-bound, intravascularly concentrated substances are most effi ciently removed, whereas substances with a large volume of distribution such as IgG, active synthesis, or large extravascular stores are removed at less than predicted rates. The usual short-term intense course of PE schedules fi ve one-volume exchanges (approximately 3 L in normal-sized adults) over a 7-day period. The appropriate replacement fl uid in most conditions is an albumin-saline mixture, which provides oncotic support without the risk of disease transmission borne by FFP. PE in patients with TTP uses replacement with FFP to supply the plasma protease that is consumed during the disease.

Side effects of PE are relatively common (10% to 30% of procedures) but generally minor and are related to vascular access, temporary discomfort, or vasomotor symptoms. 154 Patient death is rarely due to the procedure itself but is largely of cardiopulmonary causes. Plasma proteins such as coagulation factors, immunoglobulins, and complement will be removed by PE, and laboratory test results of coagulation and electrolytes may be deranged in the hours after PE. Clinical bleeding is rarely observed. Most coagulation factors do not fall below hemostatic levels and recover within hours, with the exception of fi brinogen, which may require several days for complete replenishment.

Leukapheresis may be required to urgently reduce the WBC count in patients with acute myeloid or lymphoblastic leukemia or chronic myelogenous leukemia with peripheral counts of 100 � 10 9 /L or greater. Each procedure is expected to drop the count by a third, but the effect is short lived. Leukapheresis should be reserved for use only as an adjunct to chemotherapy in patients with pulmonary or cerebral leukostasis or for cytoreduction before chemotherapy in patients at risk for severe tumor lysis syndrome.

Plateletpheresis may be benefi cial as short-term therapy in patients with symptomatic thrombocythemia manifested as cerebral or myocardial ischemia, pulmonary emboli, or gastrointestinal bleeding. Each procedure should effect a 50% reduction in the platelet count. Cytotoxic therapy should be started concomitantly as the defi nitive treatment.

Litigation related to blood transfusion has become prominent, particularly after the epidemic of transfusionassociated AIDS. 155 Most states regulate blood banking and medical practice, but blood products are regarded as

Symptomatic hyperviscosity Thrombotic thrombocytopenic purpura Neurologic diseases: myasthenia gravis, Guillain-Barré syndrome Uncontrolled systemic vasculitis with critical end-organ injury

Symptomatic leukocytosis Symptomatic thrombocythemia Sickle cell anemia crisis (pulmonary or central nervous system manifestations) a service, not as a commodity, so standard product liability does not pertain to blood components. 156 However, negligence in the course of preparing, testing, transferring, crossmatching, or administering blood products is still a potential cause for legal action. Every clinician who orders transfusions must be aware that blood components, like drugs, are approved for specifi c uses and that the indications should be clearly documented in the medical record.

The informed consent of the patient is an important area of potential liability. The Joint Commission on Accreditation of Healthcare Organizations (JCAHO) has required written patient consent for blood transfusions since 1996. What constitutes adequate informed consent and who is responsible for advising the patient are still in contention. Elements of informed consent include an understanding of the need for transfusion, its risks and benefi ts, and the alternatives, including the risk of not undergoing transfusion, as well as the opportunity to ask questions. Whether the clinician documents informed consent with an individual progress note in the patient record or with a standardized form is generally established as institutional policy. Similarly, institutions vary with respect to policies for consenting adults who are temporarily incompetent, such as sedated patients in the ICU.

A competent adult patient may refuse blood transfusion, and Jehovah's Witnesses commonly do so for religious reasons. Case law is clear in upholding this right of the patient, 157 which extends to care given at such time as the patient may become incompetent (i.e., comatose) after such refusal was expressed before becoming incompetent. Courts will usually order a lifesaving transfusion for minors. Exceptions have been made in the case of some "emancipated minors" who are at the age of reason. Most states have evoked a "special interest" in the welfare of a fetus in ordering transfusions to pregnant women.

The advent of sentinel event reviews and other quality management procedures for patient safety has had an impact on transfusion practice as well. Procedures for patient identifi cation before surgical procedures, including devices such as bar code readers, have also been applied to transfusion practice. However, annual sentinel event reviews reporting transfusion errors have remained constant according to JCAHO records. 158 

� Blood components should be prescribed like drugs.

Appropriate blood component therapy requires that the specifi c blood product needed for a clear indication be prescribed, with avoidance of a formulaic approach.

� Red blood cells should be transfused only to increase oxygen-carrying capacity. Transfusion decisions should be based on individual patient physiology. The majority of patients with hemoglobin levels greater than 60 or 70 g/L will not require transfusion unless they have limited cardiopulmonary reserve or active bleeding.

� Platelet transfusions are indicated for patients who are bleeding because of thrombocytopenia or functional platelet defects. Guidelines for platelet transfusion are also conservative. Prophylactic platelet transfusion remains controversial and is not warranted in many situations.

� Fresh frozen plasma is indicated for the repletion of coagulation factors in bleeding patients defi cient in those factors or to provide specifi c plasma proteins that cannot be obtained from safer sources.

� Cryoprecipitate is a concentrated source of fi brinogen and selected coagulation factors. Cryoprecipitate may be more helpful in correcting the hypofi brinogenemia of dilutional or consumptive coagulopathy than fresh frozen plasma.

� Adverse reactions to blood components occur in 1% to 2% of transfusion episodes. Adherence to routine protocols for the evaluation of transfusion reactions may save lives.

� Acute hemolytic reactions are the leading cause of immediate transfusion fatalities. Prevention of these reactions requires strict adherence to transfusion and patient identifi cation procedures.

� Transmission of infectious agents by transfusion has been markedly reduced, and bacterial infection is now the most common infectious complication of transfusion.

� Adverse effects unique to massive transfusion are likely to occur in the ICU and complicate the management of critically ill or severely injured patients. Component therapy for such patients should remain conservative. The emerging role of activated factor VII in the treatment of these patients requires further evaluation.

� Informed consent for blood transfusion is a standard of practice. A competent adult has the legal right to refuse blood transfusion. Consent in critically ill patients remains subject to individual institution policies.

Department of Health and Human Services, Food and Drug Administration: The Code of Federal Regulations, 21 CFR Parts 600, 606, 640

Standards for Blood Banks and Transfusion Services

Markers for transfusiontransmitted disease in different groups of blood donors

Comparative safety of units donated by autologous, designated and allogeneic (homologous) donors

Directed blood donations: Con

Goldfi nger D: Directed blood donations: Pro

Shelf-life of bank blood and stored plasma with special reference to coagulation factors

Generation of cytokines in red cell concentrates during storage is prevented by prestorage white cell reduction

Universal WBC reduction: The case for and against

Chemical and hematological changes in stored CPDA-1 blood

Restoration in vitro of erythrocyte adenosine triphosphate, 2,3-diphosphoglycerate, potassium ion, and sodium ion concentrations following the transfusion of acid-citrate-dextrose stored human red blood cells

Comparison of the hemostatic effects of fresh whole blood, stored whole blood, and components after open heart surgery in children

A practice guideline and decision aide for blood transfusion

RBC transfusion in the ICU: Is there a reason?

Descriptive analysis of critical care units in the United States: Patient characteristics and intensive care unit utilization

Oxygen transport in man

Physiologic aspects of anemia

Oxygen extraction ratio: A valid indicator of myocardial metabolism in anemia

Human cardiovascular and metabolic response to acute, severe isovolemic anemia

Transfusion guidelines for cardiovascular surgery: Lessons learned from operations in Jehovah's Witnesses

Physiologic effects of acute anemia: Implications for a reduced transfusion trigger

A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care

Is a low transfusion threshold safe in critically ill patients with cardiovascular diseases?

Oxygen extraction ratio: A valid indicator of transfusion need in limited coronary vascular reserve?

for the ABC Investigators: Anemia and blood transfusion in critically ill patients

The CRIT study: Anemia and blood transfusion in the critically ill-current clinical practice in the United States

Red cell transfusion practice following the Transfusion Requirements in Critical Care (TRICC) study: Prospective observational cohort study in a large UK intensive care unit

Appropriateness of red blood cell transfusion in Australasian intensive care practice

Silent myocardial ischaemia and haemoglobin concentration: A randomized controlled trial of transfusion strategy in lower limb arthroplasty

Mathematical analysis of isovolemic hemodilution indicates that it can decrease the need for allogeneic blood transfusion

Guidelines for perioperative red blood cell transfusions

American Society of Anesthesiologists Task Force: Practice guidelines for blood component therapy

Prudent strategies for elective red blood cell transfusion

Platelet transfusion therapy. Onehour posttransfusion increments are valuable in predicting the need for HLA-matched preparations

Volunteer donor apheresis

National Institutes of Health Consensus Conference: Platelet transfusion therapy

The bleeding time as a screening test for evaluation of platelet function

Changes in blood coagulation during and following cardiopulmonary bypass: Lack of correlation with clinical bleeding

Gamma globulin for idiopathic thrombocytopenic purpura

Intravenous anti-D treatment of immune thrombocytopenic purpura: Experience in 272 patients

Hazard of platelet transfusion in thrombotic thrombocytopenic purpura

Mantel N: The quantitative relation between platelet count and hemorrhage in patients with acute leukemia

Controversies in platelet transfusion therapy

Safety of stringent prophylactic platelet transfusion policy for patients with acute leukemia

The natural history of alloimmunization to platelets

Clinical factors infl uencing the effi cacy of pooled platelet transfusions

Optimizing platelet transfusion therapy

Current status of solvent/detergenttreated frozen plasma

Update on pathogen reduction technology for therapeutic plasma: An overview

National Institutes of Health Consensus Conference: Fresh frozen plasma: Indications and risks

The role of prophylactic fresh frozen plasma in decreasing blood loss and correcting coagulopathy in cardiac surgery: A systematic review

Hemostasis testing during massive blood replacement: A study of 172 cases

Should plasma be transfused prophylactically before invasive procedures?

Screening for the risk for bleeding or thrombosis

Lack of increased bleeding after liver biopsy in patients with mild hemostatic abnormalities

Why is fresh-frozen plasma transfused?

Effect of plasma transfusions on the prothrombin time and clotting factors in liver disease

Clotting factor levels and the risk of diffuse microvascular bleeding in the massively transfused patient

Fresh frozen plasma and platelet transfusion for nonbleeding patients in the intensive care unit: Benefi t or harm?

Treatment of the bleeding tendency in uremia with cryoprecipitate

Desmopressin: A nontransfusional form of treatment for congenital and acquired bleeding disorders

Clinical uses of intravenous immunoglobulin

American College of Obstetricians and Gynecologists: Prevention of D Isoimmunization, Technical Bulletin No 147

Human albumin solution for resuscitation and volume expansion in critically ill patients

Hypotension associated with prekallikrein activator (Hageman-factor fragments) in plasma protein fraction

Granulocyte transfusions for treating infections in patients with neutropenia or neutrophil dysfunction

Special report: Transfusion risks

Transfusion errors: Scope of the problem, consequences, and solutions

Transfusion errors in New York State: An analysis of 10 years' experience

JCAHO: Blood transfusion errors: Preventing future occurrences: Available at

Hemolytic transfusion reaction

Transfusion reactions associated with anti-IgA antibodies: Report of four cases and review of the literature

Febrile transfusion reaction: What blood component should be given next?

Clinical outcomes following institution of the Canadian universal leukoreduction program for red blood cell transfusions

Delayed hemolytic transfusion reaction: An immunologic hazard of blood transfusion

Transfusion-related acute lung injury and pulmonary edema in critically ill patients: A retrospective study

for the NHLBI Working Group on TRALI: Transfusion-related acute lung injury: Defi nition and review

Transfusion-associated acute lung injury (TRALI): Clinical presentation, treatment and prognosis

Transfusion-related acute lung injury caused by two donors with antihuman leucocyte antigen class II antibodies: A look-back investigation

for the TRALI Consensus Panel: Proceedings of a consensus conference: Towards an understanding of TRALI

Graft-versushost disease: New directions for a persistent problem

Survey of transfusion-associated graft-versus-host disease in immunocompetent recipients

The effect of prestorage irradiation on post-transfusion red cell survival

Improvement of kidney-graft survival with increased numbers of blood transfusion

Blood transfusion-modulated tumor recurrence: First results of a randomized study of autologous versus allogeneic blood transfusion in colorectal cancer surgery

Transfusion-associated cancer recurrence and postoperative infection: Meta-analysis of randomized, controlled clinical trials

Transfusion practice and nosocomial infection: Assessing the evidence

Transfusion increases the risk of postoperative infection after cardiovascular surgery

Immunosuppressive effects of blood transfusion

Transfusion of leukoreduced red blood cells may decrease postoperative infections: Two meta-analyses of randomized controlled trials

Transfusion immunomodulation or TRIM: What does it mean clinically?

Risks of blood transfusion

Transfusiontransmitted cytomegalovirus and Epstein-Barr virus diseases

Current status of microbial contamination of blood components: Summary of a conference

Septic reactions to platelet transfusions: A persistent problem

Red blood cell transfusions contaminated with Yersinia enterocolitica-United States, 1991-1997, and initiation of a national study to detect bacteria-associated transfusion reactions

Routes of infection, viremia, and liver disease in blood donors found to have hepatitis C infection

Clinical outcomes after transfusionassociated hepatitis C

Adverse consequences of blood transfusion: Quantitative risk estimates

Stramer SL: Current prevalence and incidence of infectious disease markers and estimated window-period risk in the American Red Cross blood donor population

Possible transfusionassociated acquired immune defi ciency syndrome (AIDS): California

Impact of explicit questions about high-risk activities on donor attitudes and donor referral patterns. Results in two community blood centers

The effectiveness of the confi dential unit exclusion option

Human immunodefi ciency virus type 2 infection in the United States: Epidemiology, diagnosis, and public health implications

Update: HIV-2 infection among blood and plasma donors-United States

Transmission of human Tlymphotropic virus types I and II by blood transfusion

A prospective study of transmission by transfusion of HTLV-I and risk factors associated with seroconversion

Post-transfusion cytomegalovirus infections

Reducing the risk for transfusion-transmitted cytomegalovirus infection

Is white blood cell reduction equivalent to antibody screening in preventing transmission of cytomegalovirus by transfusion? A review of the literature and meta-analysis

Transmission of parasitic infections by blood transfusion

Hemostasis in massively transfused trauma patients

Laboratory hemostatic abnormalities in massively transfused patients given red blood cells and crystalloid

Serial changes in primary hemostasis after massive transfusion

Prophylactic platelet administration during massive transfusion

Clotting factor levels and the risk of diffuse rnicrovascular bleeding in the massively transfused patient

Potential role of recombinant factor VIIa as a hemostatic agent

Recombinant factor VIIa: Unregulated continuous use in patients with bleeding and coagulopathy dues not alter mortality and outcome

Massive blood replacement: Correlation of ionized calcium, citrate, and hydrogen ion concentration

Potassium levels, acid-base balance and massive blood replacement

Acid-base status of seriously wounded combat casualties: Resuscitation with stored blood

Blood temperature: A critical factor in massive transfusion

An in vivo evaluation of microaggregate blood fi ltration during total hip replacement

Massive transfusion as a risk factor for acute lung injury: Association or causation?

Guidelines on the management of massive blood loss

Autoimmune hemolytic anemia

Approaches to selecting blood for transfusion to patients with autoimmune hemolytic anemia

The clinical implications of platelet transfusions associated with ABO or Rh(D) incompatibility

Survival curves of incompatible red cells: An analytical review

Isotype-specifi c detection of ABO blood group antibodies using a novel fl ow cytometric method

Use of Rh positive blood in emergency situations

Pharmacologic agents in the management of bleeding disorders

Disseminated intravascular coagulation. Approach to treatment

The pathogenesis and management of disseminated intravascular coagulation

Coagulation disorders in liver disease

New insights into haemostasis in liver failure

Plasma and platelet von Willebrand factor defects in uremia

Deamino-8-D-arginine vasopressin shortens the bleeding time in uremia

Donor-derived red blood cell antibodies and immune hemolysis after allogeneic bone marrow transplantation

Fluosol-DA as a red-cell substitute in acute anemia

The prospect for red cell substitutes

Low-dose perfl uorocarbon: A revival for partial liquid ventilation?

Response of factor VIII/von Willebrand factor to DDAVP in healthy subjects and patients with haemophilia A and von Willebrand's disease

Management of oral bleeding in haemophiliac patients

Amelioration of the bleeding tendency of preoperative aspirin after aortocoronary bypass grafting

for Investigators of the Multicenter Study of Perioperative Ischemia Research Group: Mortality associated with aprotinin during 5 years following coronary artery bypass graft surgery

Does the use of erythropoietin reduce the risk of exposure to allogeneic blood transfusion in cardiac surgery? A systematic review and meta-analysis

Cell salvage for minimizing perioperative allogeneic blood transfusion

Cost-effectiveness of cell salvage and alternative methods of minimizing perioperative allogeneic blood transfusion: A systematic review and economic model

Plasmapheresis in nephrology: An update

National Institutes of Health Consensus Conference: The utility of therapeutic plasmapheresis for neurological disorders

Correction of hyperviscosity by apheresis

Improved survival in thrombotic thrombocytopenic purpura-hemolytic uremic syndrome

Therapeutic plasma exchange as a nephrological procedure: A singlecenter experience

A review of transfusion-associated AIDS litigation: 1984 through 1993

Legal, fi nancial, and public health consequences of HIV contamination of blood and blood products in the 1980s and 1990s

Legal aspects of transfusion of Jehovah's Witnesses

Joint Commission on Accreditation of Hospitals and Healthcare Organizations: Sentinel event statistics: Available at