Packed Red Blood Cell Transfusion

The normal blood volume is 7 to 8% of ideal body weight. This corresponds to a hemoglobin (Hb) level of 14 to 16 g/dL and a hematocrit of 40 to 45%. Transfusion of red blood cells (RBCs) can restore circulating blood volume and oxygen-carrying capacity as described by the formula

The body has many adaptive responses to increase oxygen delivery in the face of anemia ( Table 81-1 ). It may be advantageous to increase oxygen (O ₂ ) delivery by increasing the oxygen saturation or Hb concentration because increasing cardiac output can increase myocardial oxygen demand and may precipitate ischemia in patients with coronary artery disease.

Anecdote and historical practices have dictated that the ideal Hb/hematocrit value in hospitalized patients is 10 g/dL or 30%. The basis for this claim lies in part on rheologic calculations suggesting that it is associated with an optimal balance between oxygen-carrying capacity (where high is better) and viscosity (where low is better). Such a balance would minimize cardiac work and maintain peripheral oxygen delivery. As recently as the 1990s, this recommendation was supported, in part, by two large retrospective studies in Jehovah’s Witness populations that showed a significant increase in perioperative mortality if the preoperative Hb was 6 g/dL as opposed to 12 g/dL (odds ratio of 2.5 for each gram that the postoperative Hb was less than 8 g/dL; Table 81-2 ). The risk of death was highest in patients with known cardiovascular disease.

A series of trials in postoperative patients have called into question the validity of these retrospective studies and suggested that the transfusion threshold should be individualized based on documentation of end-organ hypoxia. Two randomized studies of postcardiac surgery patients found no difference in morbidity in those randomized to a liberal versus restrictive transfusion strategy. Although the study by Hajjar et al. found a dose-dependent increase in morbidity after transfusion, the study by Murphy et al. found an increase in mortality but no change in morbidity in those randomized to the restrictive transfusion arm of the study. Reasons for this finding are uncertain considering that the mean difference in Hb level between the two arms was only 1 g/dL. Thus, although a restrictive strategy for blood transfusion after cardiac surgery may be desirable, the actual Hb trigger has yet to be determined. Another randomized study of elderly patients undergoing total hip arthroplasty similarly found no difference in morbidity or mortality in those randomized to an Hb transfusion threshold of 10 g/dL versus 8 g/dL, whereas a retrospective study in a similar patient population found a significant increase in perioperative morbidity and no change in mortality.

In a single prospective, randomized, blinded study, blood transfusion, used as part of a “sepsis bundle,” was found to improve survival in patients with septic shock whose hemodynamic parameters did not correct with intravenous fluids. Because the interventions in this study were delivered as a bundle, though, it is not possible to determine the relative impact of transfusion alone on outcome. More recently, an adequately powered, randomized, prospective study of critically ill patients with septic shock found no difference in mortality or need for ongoing critical care interventions, such as mechanical ventilation, vasopressor use, or renal replacement therapy, between patients assigned to transfusion at an Hb trigger of 7 g/dL versus 9 g/dL. Moreover, there was no change in the results in the subgroups of patients older than 70 years or those with known cardiovascular disease, although those with acute coronary syndrome (ACS) were excluded from this study. Reasons to account for this may be explained by the findings of three smaller randomized studies that found that PRBC transfusion does not improve oxygen delivery or uptake in septic patients in the ICU.

Many recent studies have addressed the role of PRBC transfusion in asymptomatic, hemodynamically stable, nonbleeding, anemic critically ill patients. A single randomized, blinded, prospective study in 1999 and several subsequent observational studies found that patients who are transfused above an Hb value of 7 g/dL have either the same or better outcomes than those who are transfused to an Hb value of 10 g/dL. These findings are consistent with many other studies and one meta-analysis that also documented an increased risk of infection after PRBC transfusion. Other studies have documented an increased risk of death after RBC transfusion. On the basis of these studies, current guidelines regarding PRBC transfusion in critically ill but asymptomatic and resuscitated (i.e., hemodynamically normal) patients call for an Hb transfusion trigger of 7 g/dL ( Tables 81-3 and 81-4 ).

There are no randomized studies evaluating a threshold for PRBC transfusion in patients with unstable angina or ACS, but a large post hoc analysis from the combined patient pool of three studies that were originally designed to evaluate efficacy of antiplatelet agents in those with myocardial ischemia found a significant increase in the hazard ratio in patients who were transfused to a hematocrit greater than 25%. This finding has been corroborated in multiple other retrospective studies. Conversely, a recent retrospective study of a highly select cohort of patients with ACS suggested a decrease in mortality for those transfused for an Hb trigger of 9 g/dL. The authors caution, though, that the cohort enrolled in the study was not representative of typical patients with ACS; therefore the results of the trial may not be generalizable to the overall population of patients with ACS. Thus, although there are insufficient data upon which to firmly recommend a transfusion threshold below an Hb level of 10 g/dL in this cohort, transfusion to an Hb level between 8 and 10 g/dL may be reasonable in most patients, and the need for transfusion to a higher Hb level may be better reserved for patients with evidence of ongoing end-organ ischemia.

Transfusion of blood products carries many risks. These include transmission of blood-borne pathogens, transfusion-associated circulatory overload (TACO), transfusion-related acute lung injury (TRALI), and transfusion-related immunomodulation (TRIM). Clinically significant transfusion reaction is rare under current guidelines and is most commonly due to clerical error. Interestingly, this adverse event is rarely seen in exsanguinating patients. Although the reason for this is uncertain, it is likely due to alterations in the immune system resulting from severe injury massive transfusion. TRALI and TRIM are most likely variants of the same disorder—an exaggerated inflammatory response and an altered or deranged immune system due to transfusion of foreign protein—and may explain the increased risk for infection. TRALI may result from local (pulmonary) inflammation whereas TRIM may represent systemic immune derangement. Both entities are likely underreported because of a lack of unique diagnostic criteria and adequately designed studies aimed to address their incidence.

TRALI is defined as noncardiogenic pulmonary edema that occurs within 4 to 6 hours of transfusion. It has a reported incidence of 1:5000 to 1:10,000 transfusions and is most common after transfusion of plasma. TRIM is best exemplified by reports showing the association between PRBC transfusion and infection ^∗

∗ References .

Plasma Transfusion

The plasma portion of donated whole blood contains most of the necessary clotting factors of the coagulation cascade. Although there are decreased concentrations of factors V, VII, and VIII because of degradation and fibrinogen (factor I) because of dilution, spontaneous hemorrhage rarely occurs with factor concentrations greater than 25%. Higher levels are needed, however, to arrest hemorrhage. Plasma is dosed as 10 to 15 mL/kg (ideal body weight), and generally 4 units will result in 40% to 60% factor recovery. It should be noted that transfusion of 5 units of random-donor platelets or 1 unit of single-donor platelets also results in the transfusion of 1 unit equivalent of plasma because platelets are suspended in plasma. Plasma is commonly used in the ICU to rapidly treat coagulopathy with concomitant hemorrhage or in anticipation of an invasive procedure in a patient with coagulopathy.

Warfarin is an oral anticoagulant that is very commonly used to prevent thromboembolic disease from various causes. A retrospective study found that each 30-min delay in administration of the first unit of plasma decreases the odds of correction of warfarin-induced coagulopathy by 20% in patients with intracerebral bleeding, underscoring the need for rapid and accurate reversal of the drug in hemorrhaging patients. Because the speed with which plasma can be administered is limited by its supply and the time required to thaw and prepare the product, use of PCCs to quickly restore clotting ability is becoming increasingly common. PCC provides a concentrated source of three or four vitamin-K-dependent coagulation factors. PCCs are stored in a lyophilized state and only require reconstitution. A type/screen is not needed, the product does not need to be thawed, and total volume of drug to be administered is less than 100 mL, thereby making administration significantly faster and with no risk of TACO. Multiple international organizations, including the American College of Chest Physicians, recommend a combination of PCC and vitamin K for emergency anticoagulation reversal. A 2013 study by Sarode evaluated the efficacy and safety of PCC compared with plasma in patients on vitamin K antagonists presenting with major bleeding. Rapid international normalized ratio reduction was achieved in 62% of patients receiving PCC versus 10% of patients receiving plasma, demonstrating PCC superiority. The safety profile was similar between groups.

There is wide variability in the manner in which physicians utilize plasma (fresh frozen plasma [FFP]) in nonbleeding patients with coagulopathy. Many physicians use FFP prophylactically to reverse coagulopathy in nonbleeding patients despite published guidelines recommending against this and an unknown risk-benefit ratio. Others cite mild coagulopathy as a reason to use FFP as a volume expander in nonbleeding volume-depleted patients. To date, there are no universally agreed-upon guidelines for use of FFP in nonbleeding patients. Suggested indications and dosing are shown in Table 81-5 .

Transfusion of plasma has the same risks as transfusion of RBCs, but the incidence of adverse events is higher for all possible complications. The most frequent adverse event associated with plasma transfusion is TRALI. Recent theory postulates that this reflects variability in plasma protein (and presumably antibody) content in the fluid being transfused. This proposed mechanism is supported by a randomized, blinded, crossover study that found that this risk is higher after transfusion of plasma obtained from multiparous women. A retrospective study found a 3-fold higher relative risk of infection in critically ill surgical patients who received FFP, a finding that is consistent with the risk of infection after PRBC transfusion. Hemolytic transfusion reactions also are possible after transfusion of plasma because plasma contains variable titers of anti-A and anti-B antibody.

Cryoprecipitate Transfusion

Cryoprecipitate is the precipitated fraction obtained from thawing FFP at 4° C. This method of isolation means that cryoprecipitate is pooled from the FFP obtained from multiple donors. Cryoprecipitate is rich in factor VIII, von Willebrand factor, factor XIII, and fibronectin. Most importantly, it is the only blood component that contains concentrated fibrinogen; thus the main indication for its use is treatment of coagulopathy due to hypofibrinogenemia. Therefore it may be useful in the management of disseminated intravascular coagulation (DIC) with hemorrhage and in reversal of thrombolytic agents ( Table 81-6 ). Although an adequate dose of plasma can replete fibrinogen, hypofibrinogenemia can be reversed more quickly using cryoprecipitate. Cryoprecipitate is dosed as a 10-pack transfusion, in which each 10-pack raises the fibrinogen level 75%. Bleeding patients with known von Willebrand deficiency should also receive cryoprecipitate to optimize platelet function whereas nonbleeding patients with this disorder can be treated with DDAVP.

Risks associated with transfusion of cryoprecipitate are the same as those reported for the other blood components. However, the incidence of TRALI and TRIM is probably lower than that associated with transfusion of plasma because the total volume of cryoprecipitate transfused is much less than plasma, minimizing the recipient’s exposure to foreign protein antigen. The risk of transmission of blood-borne pathogens, though, may be higher because of the pooled nature of this product. There are no well-designed studies assessing outcomes or adverse events related to transfusion of cryoprecipitate.

Platelet Transfusion

Platelet transfusion is less common than RBC or plasma transfusion. The most common indication for platelet transfusion is decreased production followed by increased destruction of cells. In the critically ill population, in which DIC is more prevalent, increased utilization of platelets can also lead to thrombocytopenia. Although the absolute platelet count may not correlate with function or ability to form a stable clot, it is generally agreed that spontaneous bleeding can occur with platelet counts less than 10,000 cells/μL. Although not validated in studies, many clinicians recommend that a minimum platelet count of 50,000 cells/μL should be maintained, if possible, for patients at significant risk of bleeding (e.g., trauma, postoperative patients, or those about to undergo an invasive procedure associated with a significant risk of hemorrhage), and a target of 80,000 to 100,000 cells/μL is recommended for patients who are actively bleeding or at risk for intracranial hemorrhage.

Although the platelet count can be determined easily and quickly, there is no reliable method to test platelet function. A possible exception is thromboelastography (TEG), one of two available viscoelastographic means of assessing clot formation and lysis. Limited evidence from observational data suggests that TEG is able to diagnose platelet dysfunction after trauma. Unfortunately, effects on blood product transfusion, mortality, and other outcomes remain unproven in randomized trials.

There are no studies that can be used to recommend timing and volume of platelet transfusion in nonbleeding critically ill patients. Furthermore, although there are no good studies to determine the effect that use of aspirin or nonsteroidal anti-inflammatory agents has on hemorrhage after injury, a review of the literature suggests that use of aspirin may worsen intracranial hemorrhage after traumatic brain injury. An open-label, ex vivo study in volunteers showed that platelet transfusion can reverse the platelet dysfunction caused by clopidrogel, and platelet transfusion may be prudent in patients with traumatic brain injury who were prescribed antiplatelet medications, including nonsteroidal anti-inflammatory agents. The efficacy of platelet transfusion to reverse the effects of antiplatelet medications for other causes of hemorrhage remains speculative. As previously noted, because platelets are suspended in plasma, platelet transfusion also has the added risks and benefits associated with plasma transfusion.