The ultimate goal of PRBC transfusion is to increase oxygen delivery. Oxygen delivery is described by the formula: O₂ delivery = cardiac output × O₂ content where O₂ content = 1.34 × hemoglobin × O₂ saturation + 0.003 × Pao₂. On the basis of this equation, oxygen delivery can be increased only by increasing cardiac output or the oxygen content of blood. Most often, attempts are made to increase O₂ delivery by increasing the oxygen saturation or hemoglobin. Increasing cardiac output beyond that seen in injured patients may result in increased myocardial oxygen consumption, which can cause demand ischemia in patients with coronary artery disease.

Indications for PRBC transfusion can be based on signs/symptoms of severe hemorrhagic shock or anemia, physiologic endpoints of hypoperfusion, or blood count. The need for transfusion can also be predicted using several scoring systems, although this is less useful during the actual resuscitation. Signs and symptoms of severe shock or anemia include ongoing bleeding, tachycardia, hypotension, oliguria, or altered mental status. Physiologic endpoints of hypoperfusion include lactic acidosis, base deficit, or decreased central/mixed venous oxygen saturation (Svo₂).

The Advanced Trauma Life Support manual of the American College of Surgeons categorizes severity of hemorrhage into four classes. Class I and II hemorrhage can be treated with crystalloid resuscitation alone and do not require PRBC transfusion; whereas, class III and IV hemorrhage (>30% loss of total blood volume or >1,500 mL) require PRBC as a key component of resuscitation. Such patients will be tachycardic (pulse >120 beats per minute), hypotensive, tachypneic, anxious, or confused.³ Crystalloid resuscitation of severely anemic patients will not improve oxygen delivery and may actually worsen it by further diluting the RBC concentration. Furthermore, stored RBCs rapidly deplete 2,3 diphosphoglycerate (DPG) and can require up to 24 hours to replete their stores following transfusion.⁴ During this time, there is a left shift of the oxyhemoglobin disassociation curve resulting in impaired offloading of oxygen from hemoglobin to the tissues. Therefore, transfusion should be used liberally during initial resuscitation and the physician should not wait for symptoms to develop before the start of transfusion in patients who may be bleeding because the benefit of restoring oxygen content may not be realized for many hours. Criteria for transfusion should be more stringent once the patient has been resuscitated and hemorrhage either excluded or controlled. Indications and techniques of massive transfusion will be discussed separately.

Physiologic criteria for transfusion are based on laboratory indices of end-organ ischemia and include lactic acidosis, base deficit, and decreased Svo₂. These tests are needed because they are more sensitive markers of hemorrhage and impaired oxygen delivery than vital signs alone.⁵ In a series of studies, Davis et al. showed a dose-response curve between the magnitude of base deficit on admission and blood volume needed for resuscitation.⁶,⁷ More recently, elevated lactate levels have been shown to be as reliable as the base deficit for predicting the need for transfusion with the added advantage of being more predictive of mortality than base deficit alone.⁸,⁹,¹⁰ There have been few studies evaluating the role of Svo₂ in predicting the need for transfusion, and these studies have had conflicting results.⁸,¹¹ Whereas the role of Svo₂ in monitoring the cardiovascular status of patients in the intensive care unit has been validated, its role in resuscitation of acutely injured patients remains uncertain. Moreover, various single center studies have yielded conflicting results regarding the efficacy of blood transfusion in increasing tissue oxygen delivery in various trauma populations.¹²,¹³,¹⁴,¹⁵ Because of the many detrimental effects of transfusion (discussed in the subsequent text), a consistent increase in tissue oxygen delivery has not been demonstrated in trauma patients, and Svo₂ alone is not an indication for transfusion following hemorrhage.

Historically, the transfusion trigger based solely on hemoglobin value was a level of 10 g per dL or a hematocrit of 30%. This was the value at which the blood was felt to have the highest oxygen-carrying capacity while also having the lowest viscosity, thereby decreasing cardiac work while maintaining peripheral oxygen delivery.¹⁶ This practice was validated, in part, by studies on Jehovah’s Witnesses showing that perioperative mortality increases significantly for each gram of hemoglobin <8 and lessens
if the preoperative hemoglobin is >12.¹⁷,¹⁸ However, more recently, several well-designed multicenter studies have shown either no benefit or an increase in morbidity and mortality when asymptomatic, nonbleeding, critically ill patients are transfused above a hemoglobin value of 7 g per dL.¹⁹,²⁰,²¹ This finding was corroborated through multiple single-center studies in various types of trauma patients, where blood transfusion was shown to be an independent predictor of increased length of stay, infection, morbidity, and mortality.¹³,¹⁵,²²,²³,²⁴,²⁵,²⁶,²⁷,²⁸ On the basis of these studies, the current recommendation in asymptomatic, nonbleeding, anemic patients who have been resuscitated and are hemodynamically stable is to transfuse to maintain a hemoglobin concentration >7 g per dL unless there is evidence of ongoing end-organ ischemia.

Multiple scoring systems, such as the prehospital index, trauma score, revised trauma score, and injury severity score can be used to assess severity of injury and predict the need for blood transfusion. Although exact calculation of these scores is often cumbersome during the actual resuscitation, knowledge of the approximate degree of severity of injury and score can help the trauma surgeon predict the need for transfusion and mobilize necessary resources. Table 1 shows the various scoring systems and their variables. West et al. showed that 70% of patients with a trauma score ≤ 14 required a blood transfusion and 90% of patients with a score >14 did not.²⁹ Starr et al. showed that both the revised trauma score and the injury severity score predict the need for transfusion in patients with severe pelvic fractures.³⁰ The relationship between injury severity score and blood transfusion was further validated by Como³¹ and more recently by Holcomb et al.² in a study of soldiers in Iraq.

In most reports, early transfusion of severely injured and bleeding patients is lifesaving and a standard of emergency care. Ultimately, the trauma surgeon must combine the mechanism of injury, probability of hemorrhage, vital signs, and laboratory tests in deciding whether to initiate transfusion. Although transfusion of blood is beneficial in bleeding patients and in those with severe anemia (hemoglobin <7 g per dL), as discussed in the subsequent text, it is also associated with late complications. Table 2 lists the possible benefits and detrimental effects of RBC transfusion. The trauma surgeon must weigh these risks and benefits in deciding if the patient meets indications for transfusion.

As previously noted, blood transfusion carries risk and has been shown to increase morbidity and mortality independent of other factors in various patient populations, including trauma patients. Although transfusion reaction and transmission of blood-borne pathogen are rare, the cumulative risk of transmission of blood-borne pathogen can be significant following massive transfusion. Table 3 lists the incidence of blood-borne disease transmission following transfusion of 1 unit of PRBC. In addition, other adverse events are now being reported with increasing
frequency and include transfusion-related immunomodulation (TRIM), transfusion-associated volume overload, transfusion-related acute lung injury (TRALI), and possibly end-organ ischemia due to sludging. To date, the exact cause underlying these observations has not been elucidated, although many hypotheses are proposed. It is known, however, that the risks of transfusion are cumulative and may be related to cotransfusion of leukocytes and/or to storage time of the blood before transfusion. Furthermore, it has been suggested that cotransfusion of other soluble proteins, such as human leukocyte antigen (HLA) or antibody in blood or blood products may be responsible for some of the detrimental effects noted, especially TRIM and TRALI. It is possible that recipient leukocytes are activated through the interaction between the donor’s antibody and the recipient’s tissue resulting in a potent inflammatory response, particularly in the lung.

Immunomodulation following transfusion was first described more than 25 years ago when recipients of cadaveric renal transplants were noted to have a lower incidence of rejection if they received a unit of random donor PRBC in the perioperative period.³³,³⁴,³⁵ Most likely this immunosuppressive effect also accounts for the reason that transfused patients are also more likely to develop wound infection and pneumonia in the perioperative period.¹⁹,²⁵,³⁶,³⁷ In a meta-analysis by Hill et al., the odds ratio of infection following transfusion in the trauma cohort was 5.2 (confidence interval 5.03-5.43). The immunosuppressive effects of blood transfusion were further described in trauma patients with the observation that donor leukocytes can be found in the peripheral blood of transfused patients years after the transfusion itself.²⁷ Although the clinical significance of this finding is still unknown, this suggests that engraftment of donor stem cells and tolerance of donor leukocytes is occurringbecause of immunomodulation resulting from transfusion. This phenomenon has not been reported in transfused medical patients or those undergoing elective surgery. It is possible that the immunomodulatory effects of the trauma itself predispose some patients to develop this chimeric state. Lastly, increasing age of the blood transfused has also been shown to cause derangement of the immune system by both causing the generation of inflammatory mediators and also increasing the incidence of wound infection.²⁶ This is especially relevant to trauma centers which tend to transfuse blood that is older than that in low-volume, nontrauma hospitals.

Transfusion-associated volume overload can occur in patients with preexisting cardiomyopathy or depressed cardiac function due to trauma or medications. Because blood products are colloids, they are able to draw fluid from the interstitial space into the intravascular space and are therefore very potent expanders of intravascular volume. It is generally taught that the ratio of the volume of colloid infused to the volume of colloid retained in the intravascular space is approximately 1:1 whereas this ratio is 3:1 for crystalloid. This means that patients with impaired cardiac function are at risk of developing cardiogenic pulmonary edema following robust colloidbased resuscitation. Similarly, patients may also develop anasarca without pulmonary edema, impeding healing of surgical wounds as well as efforts to mobilize the patient.³⁸,³⁹,⁴⁰ Complications related to massive transfusion are discussed separately in the subsequent text and are also listed in Table 4.

TRALI is increasingly diagnosed, although its true incidence is probably still underreported because of lack of consensus on its definition and overlap of signs and symptoms with adult respiratory distress syndrome (ARDS). TRALI is defined as noncardiogenic pulmonary edema occurring <4 hours after the start of transfusion. It is distinguished from ARDS by its temporal association with transfusion, but can be difficult to diagnose in patients with preexisting cardiac or pulmonary failure, such as those
with significant pulmonary contusion or ARDS. As with TRIM, the mechanism underlying this disorder has not yet been described but current hypotheses under consideration include transfusion of leukocytes (even with leukoreduced blood), transfusion of HLA protein or other soluble foreign proteins, and/or transfusion of cellular or fatty debris.⁴¹ It is possible that TRALI may be a localized manifestation of TRIM with acute, localized inflammation of the lung due to emboli of foreign antigen. Treatment involves supportive care only and most cases resolve quickly.