Red Blood Cell Transfusion

The normal blood volume is 7% to 8% of predicted body weight (PBW). This corresponds to a total blood volume of ~70 mL/kg PBW (≈4.9 L for a 70 kg patient) with a hemoglobin volume of ~30 mL/kg and plasma volume of ~40 mL/kg. This corresponds to a normal hematocrit of 40% to 45% and a normal hemoglobin (Hgb) of 14 to 16 g/dL. Transfusion of red blood cells can help restore both circulating blood volume and oxygen carrying capacity as described by Equations 1 and 2 in Box 9.2, Chapter 9.

The body has many adaptive responses to increase oxygen delivery in the face of anemia (Box 19.1). Clinicians can increase O₂ delivery by increasing the oxygen saturation or hemoglobin concentration or cardiac output. However, the latter results in increased myocardial oxygen consumption. Although this increases oxygen delivery acutely, a profound or prolonged increase in demand may precipitate ischemia in patients with underlying coronary artery disease.

Historical practice dictated that the ideal target for hemoglobin and hematocrit in hospitalized patients should be 10 g/dL and 30%, respectively. The basis for this target lies in part on rheologic calculations suggesting that this was the level at which there was an optimal balance between oxygen carrying capacity (where high is better) and viscosity (where low is better). Such a balance theoretically would minimize cardiac work while maintaining peripheral oxygen delivery. As recently as the 1990s, this recommendation was supported, in part, by retrospective studies.

General ICU Patients

Given the need to balance the harmful sequelae of transfusion with the potential benefits of red blood cells in oxygen delivery, Hebert and co-workers reported in 1999 the results of a multicenter, randomized study, the Transfusion Requirements in Critical Care (TRICC) trial, which assessed the clinical impact of a restrictive transfusion strategy versus a traditional transfusion strategy in ICU patients. The restrictive strategy’s transfusion threshold was Hgb < 7 g/dL, whereas the liberal’s threshold was Hgb < 10 g/dL. When patients in both arms were transfused PRBCs, the volume administered was one unit. Although 30-day mortality (the study’s primary outcome) was lower in the restricted transfusion group versus the traditional strategy, it was not significantly different (18.7% versus 23.3%, P = 0.11). However, patients who were transfused by the restrictive strategy had a lower hospital mortality compared to those in the traditional transfusion group (22.2% versus 28.1%, P = 0.05). Not only were no adverse outcomes associated with the restrictive strategy, the restrictive strategy utilized fewer PRBCs.

Based on the findings of the TRICC trial, as a general rule, hemodynamically stable and asymptomatic patients in the ICU should not be transfused until their Hgb drops < 7 g/dL, at which point they should be transfused a single unit of PRBCs (Table 19.1). After this one unit is given, the patient’s Hgb should be rechecked to determine whether another transfusion is necessary to maintain the hemoglobin level at ≥ 7 g/dL.

Although the TRICC protocol’s Hgb threshold of < 7 g/dL for PRBC transfusion has been shown to benefit a population of patients, individuals may manifest varying degrees of tolerance to anemia. Should an anemic patient develop anginal pain, electrocardiographic (ECG) changes, or other signs/symptoms of inadequate oxygen delivery, the patient should be transfused, even if his or her Hgb is > 7 g/dL. Conversely, young and otherwise healthy patients may tolerate a Hgb < 7 g/dL. Though the safe and prudent lower limit for this patient population has not yet been defined, both the American Society of Anesthesiologists and the American Red Cross have published guidelines suggesting a lower limit of Hgb 6 g/dL in asymptomatic patients.

Stable Cardiovascular Disease

In patients undergoing surgery, coronary artery disease appears to reduce the tolerance for anemia, with mortality increasing with greater levels of anemia. For this reason, many espouse a higher transfusion threshold in patients with cardiovascular disease. On the other hand, ~20% of the patients enrolled in the TRICC trial had clinically significant cardiac disease, and no difference in mortality was observed between those transfused with a restrictive versus liberal strategy (20.5% and 22.9%, respectively; P = 0.69). Similarly, a large, randomized clinical trial (n = 2016) evaluating transfusion strategies in high-risk hip fracture patients with cardiovascular disease found no benefit from a liberal (Hgb < 10 g/dL) strategy versus a more restrictive one (Hgb < 8 or symptomatic) in end points of mortality, walking independently, and myocardial infarction.

A 2012 guideline published by the American Association of Blood Banks (AABB) reflected these findings, recommending that asymptomatic patients be transfused at Hgb of 7 to 8 g/dL. Although general recommendations for transfusion triggers can be made (see Table 19.1), an appreciation of a particular patient’s physiology and symptoms, as well as an understanding of the inherent risks with transfusion, must be considered.

Acute Coronary Syndrome

For patients with acute coronary syndromes (ACS), there is very little clinical evidence to guide the transfusion threshold, as the TRICC trial excluded these patients and—despite the common presentation of ACS with anemia—a randomized, clinical trial has yet to be conducted. Myocardial oxygen delivery is dependent on coronary artery flow, and one of the physiologic compensations for anemia is coronary artery vasodilatation. Additionally, increased cardiac output (i.e., increased myocardial work) is another adaptation to anemia. For these reasons, the clinical rationale for transfusion in these patients is that an increase in oxygen carrying capacity should improve myocardial oxygenation in the background of acute coronary events.

Although a single study found that, in older adults, transfusion at a Hgb 10 g/dL was associated with improved survival, two other studies found that in patients with non-ST segment elevation myocardial infarctions (non-STEMI), a Hgb threshold of ~8 g/dL was associated with a trend toward improved outcomes. In all three of these trials, transfusion of patients who were not anemic was associated with increased mortality. Although the specific threshold remains to be determined, patients with ACS should be transfused to maintain a higher Hgb than patients with stable coronary artery disease (see Table 19.1). Until a randomized clinical trial provides better guidance, using a transfusion threshold of 8 to 10 g/dL seems reasonable.

Early Septic Shock

The care of patients with early septic shock was dramatically altered by the paradigm-shifting 2001 publication by Rivers and co-workers. This study found that a multifactorial intervention of hemodynamic goals, including volume resuscitation, vasopressor support, inotropic support, and blood transfusion, dramatically improved survival when administered to patients with severe septic shock within the first 6 hours of presentation. Because blood transfusion was one of several components administered (as a “sepsis bundle”), one cannot ascertain any independent effect that it may have had on outcome. Although many practitioners may transfuse hemodynamically unstable or acidemic patients to a Hgb of 10 mg/dL, the appropriate Hgb level for this patient population has not been determined prospectively. However, Angus and co-workers are conducting a multicenter, randomized controlled clinical trial, the Protocolized Care for Early Septic Shock (ProCESS) study (http://clinicaltrials.gov/ct2/show/NCT00510835, accessed August 2, 2012), which is addressing this important question and whose results should help to provide further guidance.

Beyond the early phase of septic shock, multiple, well-designed studies have failed to show that the transfusion of packed red cells can independently improve oxygen consumption or end-organ oxygen utilization in patients with early sepsis. Additionally, in late septic shock (> 24 hours after presentation), transfusion has not been shown to improve organ perfusion or oxygen consumption by multiple techniques, including gastric tonometry, sublingual microvascular studies, or indirect calorimetry. Because of the immunosuppression associated with transfusion and its strong association with the development of acute lung injury (ALI) and ARDS (with sepsis being the most common etiology of ALI/ARDS), the administration of packed red cells may be harmful. Some clinicians argue, pending the results of the ProCESS study noted previously, that a Hgb threshold of < 10 gm/dL for transfusing the septic patient should not be used (see Table 19.1).