Chapter 16. Anesthetic Considerations for the Patient with Anemia and Coagulation Disorders

1. Anemia is common in perioperative patients.

2. Hemoglobin (Hb) concentration or hematocrit (HCT) level can be used to rapidly assess the severity of anemia.

3. Treatment of anemia should be based on the physiology and etiology of anemia. Maintenance and restoration of normovolemia and cardiac output (CO) are necessary but insufficient aims in treating anemia.

4. Tachycardia and hypotension are important clinical signs of hypovolemia and anemia, but compensatory increases in heart rate and CO may be impeded by an insufficient cardiac reserve or anesthetic-induced sympathectomy.

5. Consideration of the physiologic signs and laboratory evidence for inadequate tissue oxygen delivery is mandatory before making a decision to transfuse red blood cells (RBCs).

6. Evidence-based outcomes supporting a specific transfusion trigger level of Hb or HCT do not yet exist for perioperative patients; however, the limited present information suggests that Hb levels as low as 7 to 8 g/dL may be as safe as higher levels of Hb in cardiac surgical and critically ill patients without clinical evidence of ischemia. Further clinical investigations are needed to guide transfusion decisions in critically ill patients.

7. Alternatives to transfusion of allogeneic RBCs are available and should be integrated into a blood-conservation strategy for selected surgical patients.

8. Goals of the perioperative management of patients with sickle cell disease are focused on clinical measures to avoid precipitating a vaso-occlusive crisis and include avoiding hypoxia, hypothermia, and dehydration.

9. Implementation of standard or exchange transfusions for sickle cell patients with the goal of reducing Hb S concentration to less than 30% to 40% can be helpful to reduce the incidence of a perioperative vaso-occlusive crisis.

Anemia is a common blood disorder of perioperative patients.1 The primary physiologic consequence of severe anemia to the surgical patient is inadequate tissue oxygen delivery, which may lead to tissue hypoxia, biochemical imbalances, organ dysfunction, and ultimately organ damage.2 Mismanagement of the anemic surgical patient can adversely affect perioperative outcomes.3 Understanding the laboratory techniques used to assess anemia, the various anemia classifications, and the physiology and appropriate treatment of anemia permits the anesthesiologist to provide better perioperative care for anemic patients.

Anemia Defined and Measured

Anemia is defined as a reduction in the total red cell mass (RCM). Both hematocrit (HCT) level and hemoglobin (Hb) concentration measurements reflect the body's RCM but do not define it. The HCT level, defined as the fractional volume of sampled blood that erythrocytes occupy, is an indirect measurement of the body's RCM (Fig. 16-1). The HCT is a simple, commonly used test to indirectly assess the severity of anemia as well as estimate whole-blood viscosity, oxygen-carrying capacity, and RCM.4 Hb is the predominant protein component of blood and serves as the major carrier transporting oxygen, carbon dioxide, and some nitric oxide.4,5 Hb concentration is a directly measured value that is commonly used to indirectly assess RCM. An isotopic dilution assay of tagged red cells can provide a more accurate assessment of RCM, but is not commonly used in clinical medicine because of the complex, costly, and impractical nature of performing a test that can require the use of radioisotopes.4

Although measurements of Hb and HCT values generally provide reliable estimates of a patient's RCM, they inaccurately reflect the total circulating RCM under certain conditions.6,7 The importance of considering the measured HCT and Hb values in the context of overall blood volume cannot be overemphasized. Hb and HCT values are influenced by dynamic changes of plasma volume and thus can be subject to errors. Examples of when Hb or HCT values misrepresent the RCM occur in clinical scenarios involving major, acute blood loss. In this situation the concurrently obtained circulating Hb and HCT values will overestimate the RCM because there has not been sufficient time to redistribute fluid from the interstitial space into the intravascular compartment. The HCT and Hb values will then spuriously indicate an increased RCM. A second example of a failure of the HCT and Hb values to adequately reflect RCM occurs during pregnancy. During normal pregnancy both the RCM and the plasma volume expand. The plasma volume increases more rapidly and to a greater extent than the RCM and creates an apparent anemia termed the physiologic anemia of pregnancy.7 Beginning in the first trimester of pregnancy, the reduced HCT value does not represent a decreased RCM. Dehydration is a third example of when HCT and Hb values can misrepresent the RCM. In the dehydrated patient, an apparently normal or elevated HCT or Hb value can mislead the clinician to erroneously overestimating the patient's RCM.6

There are 2 necessary considerations when attempting to correctly identify anemia in an individual patient. First, the laboratory values of Hb and/or HCT must be compared with standard age/sex normalized reference values (Table 16-1).8 Second, the patient's overall plasma volume status must be assessed to determine whether the patient is significantly plasma volume contracted (ie, dehydration, acute bleeding) or expanded (ie, anemia of pregnancy or volume overload).

Prevalence of Anemia in Surgical Patients

Large studies reveal prevalence of anemia in the US population of 4.4% to 5.9%; however, the prevalence of anemia in surgical populations ranges from 5% to 75%.1,9 Anemia may be produced or exacerbated in the perioperative period.10

Red Blood Cell Maturation

Normal red blood cells (RBCs) mature by proliferation and differentiation of bone marrow precursor stem cells (ie, erythropoiesis) (Fig. 16-2). A mature RBC will circulate for a period of approximately 100 to 120 days, after which time it is removed from the circulation by the body's reticuloendothelial system. The body replenishes the total RCM with the goal of achieving a steady state depending on the rate at which RBCs are being produced and removed (or lost). The kidney plays a vital role in this process by producing the hormone erythropoietin. This process is up-regulated when the kidney responds to decreased oxygen delivery due to a contracted RCM (Fig. 16-3). Optimal erythropoiesis occurs when the body possesses sufficient substrates (eg, folate, iron, vitamin B12, and other nutrients) to produce new erythrocytes.8

Classifications of Anemia

Anemia is classified as either (1) relative or (2) absolute on the basis of determining the total RCM. Relative anemia is characterized by a normal total RCM; this is a condition that does not represent a hematologic disorder but rather a disturbance in plasma volume regulation.11 The absolute anemias (ie, those characterized by a decreased RCM) are divided into 3 classes: hemorrhage-related, hypoproliferative, and hyperproliferative anemias. Additionally, the anemias are classified as either primary or secondary. Secondary anemias result from a separate disease process that precipitates the decreased RCM. Many RBC-related laboratory parameters may help determine the etiology of anemia, including the reticulocyte count, mean corpuscular volume, and RBC size distribution. Examination of the peripheral smear using light microscopy can also aid in determining etiology. This chapter provides an introduction to various RBC parameters enabling classification of anemia rather than a comprehensive review of all known anemia diagnoses and their associated RBC parameter alterations. The association of these parameters with the various classes of anemia is summarized in Tables 16-2, 16-3, and 16-4.8

Hemorrhage-Related Anemia

Anemias due to acute or chronic loss of RCM are frequently encountered by the anesthesiologist.1 The anesthetic implications for the various classes of anemias are discussed in a later section of this chapter (The Decision to Transfuse Red Blood Cells).

Hypoproliferative Anemia

Hypoproliferative anemias are common in surgical patients and are due to impaired RBC production. These anemias result from an inability to synthesize an adequate number of erythrocytes in response to a physiologic stimulus (eg, increased plasma erythropoietin levels). Hypoproliferative anemias (ie, absolute reticulocyte count <75,000/μL) are commonly due to acquired nutritional deficiencies (eg, iron deficiency anemia) and/or systemic disease (eg, anemia of chronic disease) (Table 16-2).8

Hyperproliferative Anemia

Hyperproliferative anemias (ie, absolute reticulocyte count >100,000/μL) are commonly termed hemolytic anemias. These anemias occur due to premature destruction of RBCs and may be attributable to a host of disorders, either congenital or acquired, such as sickle cell anemia, β-thalassemia major, glucose-6-phosphate dehydrogenase deficiency, autoimmune hemolytic anemia, or microangiopathic hemolytic anemia. Sickle cell anemia has important implications for anesthetic management and is discussed in a later section of this chapter (Table 16-2).8

Anesthesia and Anemia

The perioperative management of anemia frequently requires a decision of whether to transfuse RBCs or to consider transfusion alternatives (eg, use RBC salvage techniques). Making the best transfusion decision for an individual patient involves considering the balance of risks and advantages of transfusing allogeneic and/or autologous RBCs. We do not have prospective, randomized controlled trials to guide clinicians in making appropriate perioperative transfusion decisions. However, based on retrospective studies, it is likely that transfusing critically ill patients (especially with blood stored >14 days) is independently associated with significantly greater morbidity (ie, increased incidence of infection, multiorgan dysfunction syndrome, acute respiratory distress syndrome) and mortality.3,12

Hb concentration and HCT level are important predictors of transfusion risk; therefore, it is prudent to continue using these simple tests in concert with wise clinical judgment to make a transfusion decision for an individual patient.13 Given the lack of supportive clinical trial data to guide our transfusion decisions, it is of paramount importance to fully understand the physiology of anemia. The subsequent sections of this chapter review the physiology of anemia, present a practical approach to determining an individual patient's transfusion threshold that is supported by the current studies, discuss alternatives to RBC transfusion and risks of RBC transfusion, and describe the special anesthetic considerations for patients with sickle cell disease.

Physiology of Anemia

The physiologic response to anemia should be considered from at least 2 vantage points: acute anemia and chronic anemia. The physiologic response to acute anemia, most often due to hemorrhage, will depend on the extent and rate of acute blood loss. There are several normal physiologic responses to acute anemia that are related by the common goal of increasing blood flow and oxygen delivery to the peripheral tissues. Baroreceptor reflexes help to mediate the initial response to acute blood loss; chemoreceptors do not play a significant role.14 Heart rate and minute alveolar ventilation initially increase after hemorrhage in the absence of volume replacement. Hyperventilation and tachycardia serve to increase cardiac output (CO) that in turn increases blood flow to tissues. The increased heart rate augments CO more directly than hyperventilation, which increases CO by augmenting right heart filling.15 Acute hyperventilation reduces PaCO2, increasing pHa, and both of these changes increase the oxygen saturation of arterial Hb via the Bohr effect; these changes favor a higher affinity of oxygen for the Hb molecule and decrease oxygen unloading at the tissues.14,16 The reduced PaCO2 causes a left shift in the oxyhemoglobin dissociation curve, which augments oxygen uptake in the pulmonary capillaries but ultimately may decrease the amount of oxygen unloaded to tissues.17 (Fig. 16-4). The increased CO that accompanies this left shift of the oxyhemoglobin dissociation curve may compensate for the higher affinity of oxygen for the Hb molecule. Insufficient cardiac reserve to increase CO and/or existing vascular dysfunction may worsen tissue hypoxia by interfering with compensatory mechanisms.5,18

Recent insights suggest that nitric oxide (NO) works in concert with the lungs and RBCs to provide additional regulation of both pulmonary and tissue blood flow augmenting O2 delivery. NO is a free radical that is known to be a potent dilator of arterial microvasculature. NO is carried to hypoxic tissue beds by RBCs and nitrite and can serve to increase tissue blood flow.5

Concurrent with the rapid baroreceptor-mediated response to acute blood loss is the release of catecholamines, angiotension II, and vasoactive hormones.14 These biochemical mediators increase systemic vascular resistance and systemic blood pressure. The increase in systemic vascular resistance is selective; blood flow is increased to a lesser extent to gut, skin, muscle, and renal tissue than to the heart and brain, so these vital organs receive preferential blood flow.18 The increase in catecholamines also confers a positive inotropic and chronotropic effect on the heart that further augments CO.

Decreased intravascular volume stimulates the renin–angiotension–aldosterone axis, contributing to increasing CO and improving tissue oxygen delivery by water retention. Water retention augments cardiac preload.18 Acute hemorrhage is associated with redistribution of water from the extravascular space into the intravascular space, accompanied by a decreased Hb concentration. As intravascular volume is replenished, Hb concentration and HCT level decline, reducing blood viscosity. Redistribution of water occurs rapidly with severe hemorrhage; a significantly lower Hb concentration is measurable within 2 minutes after an acute loss of 40% of the circulating blood volume. Complete fluid redistribution, however, takes hours. The rapid redistribution of water is attributable to the decreased hydrostatic pressure of the intravascular compartment and an increased plasma oncotic pressure due to acute mobilization of albumin from the interstitial matrix as well as increased hepatic production of glucose and amino acids.19 The resulting lower blood viscosity serves to reduce resistance to blood flow and increase venous return to the right heart; these changes help to maintain or increase CO. Reduced resistance to blood flow reduces myocardial work and oxygen consumption despite the overall increased CO.18

If the rate of blood loss is too severe for these compensatory changes to restore blood flow to tissues, or cardiac performance cannot be increased due to preexisting cardiac pathology (eg, cardiomyopathy), then unmet tissue oxygen needs will result in lactic acid production and decreased pHa. Under these conditions acidosis reduces the Hb–oxygen affinity augmenting the delivery of oxygen to hypoxic tissues (ie, there is a right shift of the oxyhemoglobin dissociation curve)20 (Fig. 16-4).

Another compensatory mechanism for acute anemia occurs within the erythrocyte; RBCs can increase their intracellular concentration of 2,3-diphosphoglycerate (2,3-DPG), a product of anaerobic metabolism within RBCs. Increased levels of 2,3-DPG reduce the affinity of Hb for oxygen and result in decreased oxygen uptake in the pulmonary capillaries but result in increased tissue oxygen delivery (Fig. 16-4).20 Stored allogeneic RBCs have lower 2,3-DPG levels immediately after transfusion (ie, there is a left shift of the oxyhemoglobin curve) and may not readily unload oxygen to tissues. It may take days for 2,3-DPG levels to increase in transfused red cells.

By linking NO and O2 flux, RBCs effectively couple arterial vessel diameter (ie, resistance) to O2 availability in the lung and to O2 demand in the periphery. The NO is carried from the lungs via RBCs to peripheral tissues, where regional blood flow is regulated by selective release of NO to tissues requiring increased amounts of O2 (Fig. 16-5). RBCs regulate the existing form of NO (eg, NO+, NO−, or S-nitroso-hemoglobin [SNO-Hb]), protecting it from degradation during transport to peripheral tissue beds. Specifically, the heme and cysteine thiol portions of Hb within RBCs react with NO free radicals. NO reacts with soluble guanylyl cyclase (sGC) in smooth muscle cells, inducing the rapid production of cyclic guanylate monophosphate (cGMP), which results in vasodilation (ie, increased blood flow). Relevant to transfusion physiology is the fact that stored allogeneic blood with usual blood bank collection methods is completely depleted of SNO-Hb in less than 1 day, raising a concern that the complex storage lesion associated with banked blood causes depletion of SNO-Hb.5

Subsequent to these physiologic compensations (eg, hyperventilation, decreased PaCO2, increased 2,3-DPG levels, regulation of selective NO deployment, and ultimately an increased CO), there is increased erythropoiesis (Fig. 16-3). Increased erythropoietin concentration and an increase in the circulating reticulocyte count can be detected as little as 2 days after a hemorrhage event. Blood Hb concentration begins to increase within 7 days after a major hemorrhage. This continuum of physiologic responses occurs in the healthy surgical patient. An individual patient's ability to mount these physiologic responses largely depends on the patient's ability to increase CO and augment tissue oxygen delivery.18 The compensatory physiologic response to acute, unreplaced volume loss may be reduced during general anesthesia, emphasizing the importance of maintaining intravascular volume.21 The sympathetic stimulatory response to acute anemia can be impaired during general anesthesia because the heart rate does not increase to the same degree as in unanesthetized patients, and the systemic vascular resistance is lower.22 The inability to appropriately increase heart rate and augment CO during general anesthesia or due to medications (eg, β-blockers) must also be considered when determining an individual patient's transfusion threshold. Anesthetized patients with a marked impairment of their compensatory responses to anemia may benefit from treatment with more liberal transfusions despite limited data supporting this approach.

Chronic anemia is associated with expansion of plasma volume, hyperventilation, and an increased CO.23 In deciding whether to perform transfusion on a chronically anemic patient, the clinician should inquire about the signs and symptoms of anemia (eg, fatigue, decreased exercise capacity, and increased frequency of angina in patients with ischemic cardiac disease). For anesthetized patients who are chronically anemic but have been partially volume repleted, electrocardiographic evidence of myocardial ischemia or a systemic acidosis should prompt the clinician to transfuse erythrocytes. However, rapid transfusion of physiologically compensated chronically anemic patients can precipitate congestive heart failure due to intravascular volume overload, even in patients with normal myocardial function.

In forming an understanding of the physiology of anemia, one must consider oxygen delivery (DO2) to the body, oxygen consumption (VO2), and the relationship between them (VO2/DO2, or the oxygen extraction ratio). DO2 is calculated by the formula:

where CO is cardiac output, SaO2 % arterial oxygen saturation, k1 is the oxygen carrying capacity of Hb and equals 1.34 mL·g−1, [Hb] is the hemoglobin concentration, k2 is the plasma oxygen solution coefficient at body temperature and equals 0.23 mL·L−1·kPa−1, and PaO2 the partial pressure of oxygen in arterial blood.13 VO2 is determined by multiplying the oxygen content difference between the arterial and venous blood by CO. Unfortunately all the data required for these calculations are not routinely available to clinicians. When these data are available, they can aid in determining the transfusion trigger point. Additional information on physiologic transfusion triggers will be provided in a later section of this chapter.

Crystalloid versus Colloid for Non-Erythrocyte Fluid Replacement

Options to replace intravascular fluid loss include crystalloidal and colloidal based solutions. Crystalloidal solutions are composed of solutes with a molecular weight below 30 kDa, whereas colloidal solutions usually contain molecules greater than 30 kDa in size.24 Crystalloidal solutions (eg, normal saline, lactated Ringer's, hypertonic saline) possess a lower viscosity compared with the more viscous colloidal solutions (eg, blood, 5% albumin, hydroxyethyl starch [HES]), and although the crystalloidal solutions can be delivered more rapidly, contemporary rapid infusion devices make flow rate differences negligible (Fig. 16-6). In practice, roughly 2 to 5 times greater volumes of crystalloid are required to obtain the same degree of volume expansion as that obtained with an equivalent volume of colloid, a fact attributed to the greater oncotic pressure of colloids relative to crystalloids.25

Crystalloids do not present a risk of infectious disease transmission and are significantly less expensive than colloidal solutions (Table 16-5). Although administration of large volumes of crystalloid can produce a dilutional coagulopathy, HES colloidal solutions incur a risk of anaphylactoid reactions, coagulopathy, renal impairment, protracted pruritus, and a trend toward higher mortality rates in patients receiving large doses.26,27 Human albumin represents an expensive colloidal solution that does not produce coagulopathy and renal failure. Finfer et al28 established the safety of albumin over normal saline for volume expansion in the critically ill. Subjects in their trial required roughly 1.4 times more total fluid as crystalloid when compared with albumin to reach equal resuscitation endpoints. However, it is possible that albumin-treated patients with traumatic brain injury had a higher mortality, whereas those subjects with severe sepsis and septic shock suffered a higher mortality rate in the saline group.28

Gelatin solutions are known to be associated with a high incidence of anaphylactoid reactions. They are derived from bovine collagen and are not currently available in the United States. Dextran colloidal solutions (eg, dextran-40 and dextran-70, referring to their average molecular weight in kilodaltons) are produced by the hydroxylation of polysaccharides and produce various effects on coagulation. Dextrans decrease platelet adhesion, induce fibrinolysis, decrease fibrinogen levels, and lower blood viscosity. Dextran solutions have been linked to the development of renal failure in hypovolemic patients and are associated with a high incidence of anaphylactoid reactions.29

A newly available starch-based colloidal solution is 6% HES 130/0.4 in 0.9% sodium chloride (Voluven). It was expected that this lower-molecular-weight starch would produce fewer adverse effects and provide a more cost-effective alternative to albumin; however, these hopes remain unproven.27 At present all the synthetic colloids can produce important noxious side effects, and the search continues for a safe and cost-effective alternative to infusing human albumin.

The Decision to Transfuse Red Blood Cells

Clinical Indicators of Organ Ischemia

After establishing normovolemia, the anesthesiologist should review the patient's vital signs and physiologic data and assess the patient's ability to compensate for anemia while weighing the risks (eg, immunosuppression, infection, transfusion-related acute lung injury, transfusion mismatch) and benefits (eg, relieving organ ischemia) of administering a blood transfusion.3,30 Consideration should be given to the patient's ability to increase cardiac output, given the important role that augmented systemic oxygen delivery plays in compensating for anemia. Physiologic indications of inadequate tissue oxygenation include electrocardiogram (ECG) or echocardiographic evidence of myocardial ischemia (eg, new ST-segment elevation/depression, regional wall motion abnormalities, or spectral Doppler new evidence of significant diastolic dysfunction), systemic acidosis related to enhanced lactate production, an increased calculated oxygen extraction ratio (VO2/DO2) (ie, normal VO2/DO2 is 20%-30%; a VO2/DO2 >50% may indicate inadequate tissue oxygenation), a reduced mixed venous oxygen saturation (SvO2) (ie, SvO2 <50%), a low mixed venous oxygen partial pressure (PvO2) (ie, PvO2 <32 mm Hg), a 25% or greater decrease in baseline regional cerebral oximetry values (rSO2), a reduced local tissue oxygen saturation (StO2), or a persistent tachycardia unrelated to hypovolemia and/or inadequate anesthesia (Table 16-6).12 Often these measurements reflecting tissue hypoxia are not available; therefore, clinicians often consider systemic acidemia in the setting of tachycardia and hypotension without other apparent causes to indicate the need to transfuse RBCs.

Clinical Trial Data on Transfusion Triggers

To date no appropriately powered, prospective, randomized controlled trials have been reported that definitively compare a liberal versus restrictive RBC transfusion strategy on the morbidity and mortality of critically ill patients. One multicenter, prospective, randomized, controlled study sought to effectively address this question but was stopped for administrative reasons. This study, the Transfusion Requirements in Critical Care (TRICC) trial, enrolled 838 (<20% of projected patient enrollment) intensive care unit (ICU) patients randomized to either a restrictive (ie, [Hb] 7.0-9.0 g/dL) or liberal (ie, [Hb] 10.0-12.0 g/dL) transfusion strategy. The overall 30-day mortality was lower in the restrictive transfusion strategy group (18.7% vs 23.3%), but this difference was not statistically significant (p = 0.11). An a priori subgroup analysis of subjects who were less acutely ill and older than 55 years of age demonstrated a significantly lower mortality in the restrictive transfusion group. Although the TRICC trial is the largest study to date, it was stopped prematurely and thus did not answer the question of an appropriate transfusion strategy in ICU patients. The authors concluded that a restrictive approach to RBC transfusion was as effective and potentially superior to a liberal transfusion strategy for ICU patients, with the possible exception of patients with unstable angina and acute myocardial infarction.31 Unfortunately, the TRICC trial was conducted before leukoreduction of transfused blood became a common practice. A review of the importance of leukoreduction suggests that it may improve the clinical outcomes of patients undergoing transfusion.32 An additional weakness of the TRICC trial is the fact that it did not include cardiac surgical patients, a group requiring more blood transfusion than any other patient group.

More recently, the results of a single-center, prospective, randomized, controlled non-inferiority trial assessing the effects of 2 transfusion strategies (liberal [HCT ≥ 30%] vs restrictive [HCT ≥ 24%]) in a group (n = 502) of cardiac surgical patients has been completed. The primary outcome measure was the occurrence of the composite end point of all-cause 30-day mortality and severe morbidity. The occurrence of the primary end point was similar in both groups (liberal 10% vs restrictive 11%; p = 0.85). Independent of group assignment, the number of RBC units transfused is an independent risk factor for death or clinical complications at 30 days (hazard ratio for each additional RBC unit transfused 1.2 [95% confidence interval, 1.1-1.4]; p = 0.002). Unfortunately, blood was not leukoreduced, a common practice in US blood banks, and the median storage time for the transfused RBCs was only 3 days, dramatically shorter than most US centers (ie, average is approximately 14 days). The authors concluded that among cardiac surgical patients, the use of a restrictive perioperative transfusion strategy as compared with a more liberal strategy resulted in non-inferior rates of their composite outcomes.33

A third, slightly smaller (n = 428), prospective, randomized, controlled trial of a restrictive (transfused if Hb <8 g/dL) versus liberal (transfused if Hb <9 g/dL) transfusion strategy in cardiac surgical patients was published over a decade ago.34 The authors did not report any significant differences in morbidity, mortality, or anemia-related fatigue self-assessment scores and concluded that a transfusion threshold of 8 mg/dL did not adversely affect the outcome. Because the conservative group was transfused with significantly less blood (0.9 ± 1.5 vs 1.4 ± 1.8 RBC units, restrictive vs liberal; p = 0.005), they concluded that using 8 mg/dL as a transfusion threshold would conserve RBCs without increasing the patient's risk.34 A number of smaller trials failed to provide any additional insights into the question of determining a transfusion threshold.3

Presently, a large, multicenter, National Heart, Lung, and Blood Institute–sponsored, prospective clinical study comparing a liberal transfusion strategy versus a symptom-driven strategy is occurring. Named the FOCUS trial (The Transfusion Trigger Trial for Functional Outcomes in Cardiovascular Patients Undergoing Surgical Hip Fracture Repair), the primary end point is to determine whether a 10 g/dL transfusion strategy is associated with an improved ability to walk 10 ft without assistance at 60 days after hip surgery as compared with a symptom-driven transfusion strategy. The goal of trial enrollment is 2600 subjects, and greater than 2000 subjects are enrolled presently. The FOCUS trial has a number of secondary end points, including assessing the composite outcome of 30-day mortality, myocardial infarction, pneumonia, stroke, and thromboembolism.35,36

Several large retrospective studies assessing transfusion-related outcomes in varied patient groups have been completed.37-39 In a study by Corwin et al,37 the number of transfused RBC units was found to be an independent risk factor for hospital length of stay, mortality, and a number of complications. In another large study by Vincent et al,38 mortality was significantly higher in patients with similar levels of organ dysfunction in whom RBCs were transfused. A large (n = 8787) observational study of elderly, orthopedic surgical patients by Carson et al39 found that postoperative transfusion did not influence 30- or 90-day mortality after adjusting for the transfusion strategy (ie, transfusion trigger between 8 and 10 g/dL−1 vs < 8 g/dL−1), cardiovascular disease, and other risk factors.39 Similar to reported prospective studies, these larger retrospective studies suggest a clear relationship between more RBC transfusions and a worse outcome. However, it remains debatable whether patients who received transfusions were equivalently ill.

Our distillation of peer-reviewed, threshold-related transfusion literature suggests that no critical hemoglobin concentration for moderately anemic patients should be used, but rather clinical signs and symptoms of organ ischemia should be assessed in each patient in whom a transfusion of RBCs is considered. The available literature appears to consistently show independent relationships between the number of RBC units transfused and a worse clinical outcome when other measured variables are taken into account and controlled for, regardless of the transfusion treatment strategy. Furthermore, drawing from the limited body of available evidence, it appears to be just as safe to accept Hb concentrations of 7 or 8 mg/dL in critically ill patients, including those with cardiovascular disease, provided they are not exhibiting signs or symptoms of end organ ischemia. Additional prospective investigations are needed to shed light on these important clinical questions.

Undesirable Effects of Red Blood Cell Transfusion

Retrospective studies report that transfusion of allogeneic RBCs is associated with increased morbidity and mortality in both cardiac and noncardiac surgical populations.3 Perioperative transfusion is associated with undesirable outcomes including prolonged hospital length of stay, infections, immunosuppression, and acute transfusion reactions.40 Presently, human error, or a mistransfusion, accounts for the second highest incidence of adverse transfusion effects, whereas the transfusion-associated viral infectious risk remains remarkably low (Table 16-7).41

More than 200 cases of transfusion-associated graft-versus-host disease have been reported, with a near 90% fatality rate. Irradiation of blood products can prevent this problem, but irradiation induces detrimental effects on RBCs.42 Allogeneic blood transfusion is associated with an increased risk of tumor recurrence in patients previously treated for cancer, and this appears to be a dose-dependent effect.43

Transfusion-related acute lung injury (TRALI) is the leading cause of transfusion-related mortality Estimated TRALI incidence is once for every 3000 to 70,000 plasma-rich components transfused and 1 per 50,000 units of low plasma volume components transfused.44,45 Large-scale efforts to decrease the risk of TRALI have been initiated.46 TRALI is an antibody-mediated reaction characterized by the acute onset of respiratory distress within 4 hours of transfusion. Typical manifestations include severe hypoxemia, hypotension, fever, and diffuse pulmonary edema. Patients with TRALI require immediate aggressive respiratory support, with the majority requiring intubation and mechanical ventilation. Despite the severity of symptoms, permanent lung injury does not occur. The majority of patients return to their baseline respiratory function within 48 hours of symptom onset and have total resolution of pulmonary edema within 4 days.46 Present measures to mitigate against the rising TRALI incidence appear to be reducing the incidence.45

Other potentially fatal complications of allogeneic blood transfusion include hemolytic transfusion reactions due to antigen-antibody interactions, transfusion-associated circulatory overload (TACO), anaphylaxis, and post-transfusion purpura due to platelet-specific antigens.47 Administration of large volumes of stored blood can also lead to hyperkalemia, hypoglycemia, hypothermia, and citrate-induced hypocalcemia.

Transfusion Alternatives

Acute Normovolemic Hemodilution

Acute normovolemic hemodilution (ANH) involves intraoperative removal of a predetermined quantity of the patient's blood (ie, up to 50% of a healthy patient's total circulating RCM can be safely removed) while intravascular volume is concurrently replaced. Because the RCM is decreased before intraoperative bleeding begins, there is a smaller net loss of RCM per unit blood loss by surgical hemorrhage. Additionally, the benefit of transfusing the shed autologous whole blood, including functional platelets and coagulation proteins, is available to the patient at any point during or after the operative procedure.22 Individuals with religious prohibitions to RBC transfusion (eg, Jehovah's Witnesses) provide a group of patients who can often benefit from ANH if the shed blood is maintained in a parallel circuit connected to the patient.

Red Blood Cell Substitutes

Further development and testing of RBC substitute–based transfusion solutions is needed to learn whether they can provide a safe alternative to erythrocyte transfusion. Extensive laboratory and clinical research has been done with Hb-based oxygen-carrying (HBOC) solutions, but despite 50 years of laboratory and clinical research efforts with HBOC solutions, there is presently no US Food and Drug Administration (FDA)–approved product available for clinical use. Recently, the FDA convened a workshop to review the issues affecting the development of blood substitutes. Problems impeding the development of HBOC solutions include renal toxicity, gastrointestinal symptoms, systemic arterial hypertension, coronary vasoconstriction, pulmonary hypertension, short half-life of the transfused molecules, methemoglobin production, and a significant systemic arterial inflammatory response occurring several days after transfusion. The vasoconstrictor effects of these molecules appear to be related to their small size, tendency to migrate into the interstitium, and ability to scavenge nitric oxide produced by vascular endothelium. Perfluorocarbons are a non–Hb-based RBC substitute, but these synthetic solutions are also not available for clinical use in the United States.48

Red Blood Cell Salvage

Red cell salvage techniques are associated with a significant reduction in allogeneic transfusion requirements and a cost benefit for some patients.47 Autologous RBC salvage involves the collection of blood from a wound site, processing (eg, centrifugal washing and/or filtering through a 20-μm filter) of scavenged blood, and reinfusion. Centrifugal washing concentrates the salvaged blood to allow the infusion of an autologous product with a high HCT level (eg, 45%-65%) and removes undesirable components of shed blood (eg, free Hb, tissue debris, fibrin split products). Cell salvage requires specialized equipment and disposable devices (Fig. 16-7). Processed autologous blood infuses normally biconcave disc-shaped erythrocytes suspended in normal saline. In contrast, allogeneic RBCs develop an abnormal echinocyte shape after 14 days. This conformational change impairs the cells' ability to transit capillary beds49 (Fig. 16-8). Centrifugal washing removes coagulation proteins, although some leukocytes and platelets are autotransfused.

Alternatively, collected shed autologous blood is passed through a filter with 20-μm pores and re-transfused without centrifugal washing.47 Red cell salvage with centrifugal washing within certain guidelines has been studied and found to be generally safe; however, current recommendations of The Society of Thoracic Surgeons and the Society of Cardiovascular Anesthesiologists do not allow re-transfusion of shed blood without centrifugal washing.50 Infection at the operative site or the use of microfibrillar collagen materials are contraindications to using RBC salvage techniques.47

Malignancy has been regarded as a contraindication to the use of a cell saver and transfusing autologous blood products. However, recent experience with the cell saver in cases of prostate cancer resection did not show an increased cancer recurrence rate and effectively reduced the need for autologous transfusion. A small prospective observational study of the use of the cell saver for orthotopic liver transplantation for hepatocellular carcinoma has been reported.43 No large-scale randomized controlled trials have examined the safety of using the cell saver in cases of malignancy; thus this practice is usually avoided.

The use of the cell saver in cases of enteric perforation is avoided. One prospective study investigated the use of the cell saver with a leukocyte depletion filter with the goal of reducing bacterial contamination of salvaged blood.43 No large-scale randomized controlled trials have been conducted to investigate the safety of using the cell saver in the setting of blood contamination; thus this use should be avoided.

Known risks incurred with the use of cell salvage include nonimmune hemolysis, air embolism, febrile nonhemolytic transfusion reactions, coagulopathy, and contamination with drugs, infectious agents, and cleansing solutions. Incomplete blood washing can lead to administration of activated leukocytes, cytokines, and other microaggregates. Risks of coagulopathy are minimal in patients with blood loss less than 3 L, and technological advances have reduced other risks.43

It is currently recommended to use cell saver technology for surgical cases if anticipated blood loss is more than 1 L or more than 20% of estimated blood volume. Other situations of clear benefit are for patients with a low preoperative Hb level at risk for bleeding, patients with multiple antibodies or rare blood types, and patients who object to receiving allogeneic blood.43,47 A recent systematic review reports that processing cardiotomy blood in cardiac surgery patients through the cell saver can reduce postoperative neurocognitive dysfunction, most likely through the reduction in lipid related microembolization.47

Antifibrinolytic Therapy

Antifibrinolytic therapy (eg, aprotinin, and the lysine analogues [ϵ-aminocaproic acid and tranexamic acid]) can reduce blood loss in patients at risk of major hemorrhage (eg, cardiac surgical patients, orthotopic liver transplant patients, selected orthopedic surgery patients).51,52 These agents inhibit fibrinolysis (ie, inhibit conversion of plasminogen to plasmin or directly inhibit the action of plasmin on fibrin) that often accompanies major surgery (Fig. 16-9). The results of several prospective, randomized controlled trials indicate that aprotinin therapy can reduce transfusion requirements as well as the incidence of surgical reexploration for hemorrhage. However, recent data report that aprotinin use is associated with a significantly higher morbidity and mortality in high-risk cardiac surgical patients and resulted in the removal of aprotinin in 2007 from the world market.54-56 Aprotinin is presently only available for investigational use and requires local institutional review board approval as well as individual patient consent before administration. The lysine analogues have been proven effective in blood conservation for selected patient populations.51,55 Tranexamic acid use has been associated with an increased incidence of perioperative seizures in cardiac surgical patients.54 The lysine analogues remain in widespread clinical use.

Additional Adjuncts to Promote Blood Conservation

There are data suggesting that the administration of a human, pooled fibrinogen concentrate (Riastap) may be beneficial for perioperative patients who develop coagulopathic bleeding. In a prospective, randomized, placebo-controlled study of 20 general surgery patients who developed a hydroxyethyl starch infusion–induced dilutional coagulopathy and were treated with either a fixed dose of fibrinogen concentrate (FC) or a placebo, Fenger-Ericksen et al57 demonstrated that FC-treated patients exhibited improved maximal clot firmness (MCF), a rotational thromboelastometric (ROTEM)–derived parameter indicating the improved strength and quality of a blood clot, as compared with patients treated with a placebo. They also demonstrated that FC-treated patients were less likely to require blood transfusion as compared with patients in the placebo group (p = 0.023). In another small (n = 20), prospective, randomized controlled study conducted in cardiac surgical patients, the ability of prophylactic FC administration to reduce postoperative bleeding was investigated by Karlsson et al.58 The investigators did not observe any difference in the primary end point (adverse events and early coronary graft occlusion as assessed by computed tomography) between the treatment group and a control group. In addition there was less postoperative bleeding in the FC-treated group (565 ± 150 vs 830 ± 268 mL/12 h; p = 0.010). Additional clinical investigation is required of this product before it is appropriate to use it to reduce surgical bleeding.

Special Anesthetic Considerations for Patients with Sickle Cell Anemia

Sickle cell disease (SCD) is prevalent in the African American population. Patients with this hemoglobinopathy can exhibit both anemia and a functional defect of their circulating Hb. Hb is composed of globin chains combined with a heme molecule. Four distinct types of globin chains are normally produced, including alpha (α), beta (β), gamma (γ), and sigma (σ). Normally humans produce 3 types of Hb using various combinations of the 4 distinct globin chains, including Hb A (α2β2), Hb F (α2γ2), and Hb A2 (α2σ2). In early infancy, 97% percent of the circulating Hb exists as Hb A. Hemoglobinopathies result from decreased synthesis of normal globin chains (eg, thalassemia syndromes) or, in SCD, a functional defect of the Hb complex involving a qualitative defect in globin synthesis.59

Pathophysiology of Sickle Cell Disease

The pathophysiology of SCD and sickle cell trait (SCT) is due to a substitution of the amino acid valine for glutamine at the sixth amino acid location of the β-globin chain, resulting in production of Hb S. In humans with both chromosomes directing Hb S synthesis (homozygous SS disease), the normal Hb A is substituted by Hb S. Increased production of Hb A2 and Hb F is minimally effective in alleviating the pathophysiologic changes of SCD. A major percentage (80%-95%) of the synthesized Hb is Hb S in patients afflicted with homozygous SS disease.60,61 SCT represents the heterozygous state and is produced by 1 sickle cell gene locus and 1 normal gene locus. SCT results in few clinical problems, despite the Hb S level rising to nearly 40% of circulating Hb.60 Among African Americans in the United States, approximately 0.4% exhibit homozygous SCD status, with a prevalence of 6% to 8% heterozygotes.61

There is a spectrum of abnormal tissue oxygen delivery resulting from the pathophysiologic mechanisms of SCD. Patients with SCD exhibit an abnormal conformation of the deoxygenated Hb S molecule, reducing the solubility of the molecule and resulting in intra-erythrocytic precipitation of Hb S.62 The precipitated intraerythrocytic deoxy-Hb S alters red cell function and damages intracellular RBC contents and the cell membrane. These cellular changes result in reduced capillary blood flow and early cell destruction. The normally pliable, biconcave disc-shaped erythrocyte transmutes into a smaller, less compliant, characteristically sickle-shaped erythrocyte (Fig. 16-10).63 Erythrocytes of patients with SCD have a reduced circulating life span of approximately 10 to 20 days, contributing to the anemia.64 Sickled erythrocytes increase the viscosity of blood and enhance thrombus formation to produce a vaso-occlusive crisis.61 The goal of SCD treatment is to prevent initiation of sickling, because this process is irreversible once intracellular Hb S has precipitated.61 There is mounting evidence that SCD is not simply the result of erythrocyte sickling, as there is widespread evidence of progressive inflammatory vascular damage and a widespread vasculopathy. It has been confirmed that in addition to sickling of erythrocytes, the oxygenated form of hemoglobin S has a highly unstable structure, leading to accelerated breakdown of the molecule and the release of toxic heme and iron compounds. In addition, the accelerated auto-oxidation of hemoglobin to methemoglobin leads to increased production of hemichromes. The net result of these processes is widespread oxidative damage to the RBC membrane and adhesion of iron-laden–oxidized erythrocytes to vascular endothelium, leading to endothelial damage and dysfunction.61

A vaso-occlusive crisis can produce damage to many organs (eg, renal papillary necrosis, renal medullary infarction, retinal or vitreous hemorrhage, aseptic necrosis of the femoral head, splenic infarction).64-66 Additionally, central nervous system and pulmonary infarctions are common and related to both chronic inflammation and emboli of sickled cells and platelet aggregates formed in the vasculature. Alterations of cellular, plasma, and vascular components of hemostasis and impaired vasomotor regulation contribute to vaso-occlusion, with sickling being a secondary exacerbating effect.64

Treatment of Sickle Cell Disease

Patients with SCD should be treated to prevent perioperative hypoxia, dehydration, acidosis, and hypothermia, which can result in a vaso-occlusive crisis. Preventive measures are the mainstay of treatment for SCD patients.67 General anesthesia per se does not overtly increase the risk of sickled erythrocytes.64

One approach to avoid a perioperative vaso-occlusive crisis is to reduce the percentage of Hb S to less than 30% to 40% before surgery by exchange and standard blood transfusions.64,67 As with other chronic anemic states, there is physiologic compensation for anemia that results in an increased intravascular volume and augmented CO. The increase in intravascular volume places the anemic SCD patient at risk for volume overload if they are rapidly transfused with allogeneic blood; thus a cautious approach to transfusion is mandated. The benefit of transfusion can persist for weeks after the postoperative period, as evidenced by a reduced incidence of sickling episodes. Perioperative mortality in SCD patients is most commonly caused by a fulminant infection related to immunosuppression associated with prior splenic infarctions.67

Key Points

1. Given the broad range of congenital and acquired disease states as well as the multiple pharmacotherapies producing a hypocoagulable state, serious attention to these pathologies and drugs is warranted to avoid the adverse effects of transfusions, excessive perioperative bleeding, and depletion of community blood bank resources.

2. The management of patients with hypercoagulable disorders is a responsibility of the anesthesiologist because many of these patients require perioperative management of their anticoagulant regimens.

3. Hematology consultation is vital to obtain for the perioperative management of patients with complex hypocoagulable and hypercoagulable disorders in order to avoid unnecessary perioperative morbidity and mortality. After surgery, many of these patients will be managed by a hematologist.

The human coagulation system maintains a delicate balance between hemostasis and the distribution of fluid blood to the tissues. Pathologic changes within the coagulation system may result in either a hypocoagulable state (ie, a tendency to hemorrhage inappropriately) or a hypercoagulable state (ie, a tendency to inappropriately form thrombus). The coagulation system relies on the appropriate level and function of several blood components to allow normal clotting. The process of blood coagulation relies on the interaction of platelets, vascular endothelium, vascular smooth muscle, soluble coagulation proteins, and various local and systemic biochemical mediators to produce a hemostatic clot.68

Hypocoagulable States

It is crucial to recognize and treat a preexisting hypocoagulable state to avoid massive hemorrhage and potentially unnecessary blood product transfusion. Optimal treatment of a hypocoagulable disorder is best accomplished by selectively transfusing the blood component(s) that is/are deficient (eg, transfuse factor VIII concentrate into a hemophiliac requiring surgery rather than fresh-frozen plasma). Hypocoagulable states are usually either congenital or acquired; both can occur simultaneously in the same patient. Knowledge of both the coagulation cascade and the biologic half-life of a particular coagulation element are needed to effectively treat many hypocoagulable disorders (Fig. 16-9, Table 16-8). It is of paramount importance to combine a history and physical examination with appropriate laboratory testing to discover congenital and/or acquired hypocoagulable/hypercoagulable disorders in surgical patients. After direct questioning of whether a patient has a known coagulation disorder, one must ask specific questions to elicit the presence or absence of these disorders (Table 16-9). If a patient answers "yes" to any of these questions, inquire regarding the details and frequency of the problem. Depending on the complexity of the presenting hypocoagulable or hypercoagulable state, a hematology consultation may be indicated.68

Factor VIII Deficiency (Hemophilia A)

Patients with a factor VIII deficiency, or hemophilia A, can require urgent surgical intervention for bleeding episodes (eg, craniotomy to decompress an intracranial hemorrhage or fasciotomy to relieve a compartment syndrome). Hemophilia A is an X-linked recessive trait that occurs with a frequency of 1 in 5000 male births.69

Individuals with hemophilia A manifest a range of bleeding symptoms, with a severity inversely correlated to their plasma activity level of factor VIII. Normally, factor VIII activity ranges from 50% to 150% (or 0.5-1.5 U/mL). Mild hemophilia is present in patients with factor VIII activity levels greater than 5%. Moderate hemophilia is present in patients with factor VIII activity levels from 1% to 5%, and severe hemophilia symptoms manifest at levels less than 1% (or 0.01 U/mL).69

Spontaneous hemarthroses are the hallmark of severe hemophilia. Recurrent joint hemorrhage produces painful degeneration of the cartilage (hemophilic arthropathy) that eventually can destroy the joint (Fig. 16-11).70 Hemophiliacs with 1% to 5% of factor VIII activity do not generally have spontaneous bleeding events, but trauma and surgery can precipitate a hemorrhage.

Intracranial and intracerebral bleeding are the most common causes of death in hemophiliacs. Approximately half of the intracranial hemorrhages in hemophiliacs occur spontaneously. Gastrointestinal and oropharyngeal bleeding may also occur. Oropharyngeal bleeding is of great importance because it can lead to airway compromise. In severe hemophilia, recurrent hemorrhagic episodes may produce an enlarged pseudotumor that can invade local soft tissue and bone. Operative removal of pseudotumors is associated with a high mortality despite the provision of recommended perioperative therapies.71

In the United States, the diagnosis and management of hemophilia A is achieved with an activated partial thromboplastin time (aPTT)–based 1-stage assay. The aPTT is abnormally prolonged in all hemophiliacs. The prothrombin time (PT), thrombin time (TT), bleeding time, platelet count, and platelet function are usually normal in hemophiliacs.69

Hemophilia A is usually treated by administration of factor VIII concentrate. The goal of chronic therapy is to maintain the trough level of factor VIII activity at approximately 3% of normal to reduce the incidence of spontaneous intracranial hemorrhage and minimize spontaneous intra-articular hemorrhage.69 Plasma-derived and recombinant factor VIII concentrates are available, and both provide equally effective treatment (Table 16-10). The treatment goal for acute, life-threatening bleeding episodes (eg, intracranial hemorrhage or limb-threatening hemorrhage) is to attain a factor VIII activity of 50% to 100% of normal, as estimated by the aPTT-based 1-stage assay.

One anesthetic consideration for patients with hemophilia A involves increasing preoperative factor VIII activity levels to approximately 100% of normal. Additionally, it is reasonable to avoid nonsteroidal anti-inflammatory drugs (eg, ibuprofen or ketorolac) because they can precipitate spontaneous, gross hematuria in hemophiliacs. Patients requiring a lumbar puncture should first be restored to approximately 100% of normal factor VIII activity levels; outcome data regarding spinal anesthesia in hemophiliacs are not available, but a similar therapeutic goal appears warranted if the benefits of spinal anesthesia outweigh the risks. Administration of factor VIII concentrates to hemophiliacs with active pharyngeal hemorrhage may not provide acute airway protection; thus elective intubation may be needed to protect against life-threatening airway obstruction.69 In bleeding hemophiliacs or those requiring prophylaxis who have developed inhibitors to factor VIII treatment, recombinant activated factor VII (NovoSeven) is indicated. Additionally, a complete history and physical examination will reveal whether the hemophiliac has developed narcotic tolerance associated with prior treatment of painful complications related to the disease (eg, hemophilic arthropathy or ureteral blood clots).

Factor IX Deficiency (Hemophilia B)

Factor IX deficiency, also termed hemophilia B, or Christmas disease, results in a clinical syndrome that is similar to hemophilia A. Hemophilia B is transmitted as an X-linked recessive trait and occurs in 1 of 30,000 male births. The clinical manifestations of hemophilia B are similar to those described for hemophilia A. Mild, moderate, and severe forms of hemophilia B are defined by the same criteria used for hemophilia A (ie, severe hemophilia B is defined by a factor IX activity level of less than 1% [or 0.01 U/mL]). Severity of disease correlates with the factor IX plasma activity; patients with less than 1% activity will exhibit spontaneous hemarthroses, intramuscular hemorrhages, and gross hematuria as symptoms. The diagnosis of hemophilia B is also made with an aPTT-based 1-stage assay. The aPTT will be prolonged in patients with hemophilia B but will correct when the patient's plasma is mixed in equal parts with plasma from a normal individual.69

Hemophilia B is treated by administration of concentrated factor IX. Factor IX concentrates are available as either recombinant or plasma-derived preparations. The plasma half-life of transfused or endogenous factor IX is approximately 18 hours. The same levels of factor activity replacement are targeted as described in the preceding section on hemophilia A. Development of inhibitory antibodies can complicate the treatment of both factor VIII and IX deficiency. Antibodies occur more commonly in patients treated with ultra-high-purity factor concentrates (ie, the incidence is approximately 1%-4% in patients with severe factor IX deficiency).69 Recombinant activated factor VII (rVIIa or NovoSeven) is FDA approved for the treatment of hemophilia B in patients with inhibitors to factor IX.72 The anesthetic implications for factor IX deficiency are analogous to those of hemophilia A.

von Willebrand Disease

von Willebrand disease (vWD) is a disorder of platelet function caused by a deficiency of normally functioning von Willebrand factor (vWF). vWD can result in a hypocoagulable state that manifests as bleeding episodes.73 vWD is an autosomal-dominant congenital disorder with an equal male-to-female distribution ratio occurring with an incidence of approximately 1%.

vWF plays 2 major roles in the hemostatic process. First, vWF acts as a bridge between platelet receptors and exposed collagen in disrupted endothelium. Platelet binding to exposed collagen in disrupted endothelium initially occurs via a glycoprotein Ib-vWF and collagen interaction and is followed by firmer adhesion of the platelet to the collagen molecule via a glycoprotein IIb/IIIa-vWF and collagen interaction. vWF enables platelets to create a primary plug in disrupted endothelium and commence the hemostatic process at the site of vessel injury. vWF also acts as a bridge between platelet receptors, specifically between glycoprotein IIb/IIIa receptors, to allow platelet aggregation.73 Second, vWF circulates in the plasma to create a complex bound to factor VIII and provides protection for factor VIII against proteolytic cleavage.73,74

There is a broad range of bleeding severity associated with vWD. Individuals with moderate or severe vWD tend to bleed abnormally in childhood or young adulthood. The majority of affected individuals have only mild to moderate disease. Symptoms of vWD include easy bruising, epistaxis, and bleeding from mucous membranes. Menorrhagia and life-threatening gastrointestinal bleeding occur in vWD. Surgical procedures can also be associated with excessive bleeding in vWD patients.73

The diagnosis of vWD is complex and is made with laboratory tests, including vWF antigen, vWF activity, factor VIII activity, and bleeding time. These tests are followed by a second series of assays to determine the classification of vWD and include a ristocetin-induced platelet aggregation assay as well as vWF multimer studies.73

The treatment of vWD depends on the type of vWD that is diagnosed (Tables 16-11 and 16-12). Use of nonsteroidal anti-inflammatory drugs can further impair preexisting compromised platelet function and thus are contraindicated in vWD. The correlation of the degree of diagnostic laboratory test abnormalities in vWD with the bleeding severity is poor, and therefore empirical goals of treatment are pursued. The treatment goal for the perioperative period should be to increase the activity of vWF and factor VIII to 50% to 100% of normal, an analogous goal to the usual treatment of bleeding episodes in vWD.73

vWF and factor VIII activity levels are increased in patients with vWD by IV administration of DDAVP (1-desamino-8-D-arginine-vasopressin or desmopressin) and giving vWF-rich plasma-derived blood products. The use of topical hemostatic agents (eg, fibrin glue, Gel-foam, or thrombin-soaked Surgicel) will also promote hemostasis in vWD patients.73 Alternatively, intravenous (IV) vWF concentrates (eg, Humate P) can be given to patients who do not respond positively to DDAVP therapy. Pregnant patients with vWD merit special considerations. Although evidence-based recommendations for central neuraxial blockade in vWD do not exist, consideration should be given to the fact that causing a spinal or epidural hematoma may be permanently debilitating. vWF levels increase 2 to 3 times above the baseline during the second and third trimesters of a pregnancy and may provide some protection from bleeding complications. The increased levels should be verified by selected assays because an increase in vWF is not universal, and some patients with qualitative vWF defects will not correct their vWF activity during pregnancy. DDAVP can be administered to vWD patients in active labor who are known to have a positive response to this therapy. vWF replacement therapy can be given to patients who are still bleeding after DDAVP administration. Patients who are known nonresponders to DDAVP should be treated with vWF concentrates. Lysine analog antifibrinolytic agents (eg, ε-aminocaproic acid and tranexamic acid) have been successfully used as both adjuncts and monotherapy in some patients with vWD-related hemorrhage.73

Rare Factor Deficiencies

The individual deficiencies of factors V, VII, X, and XI have all been described and are known to be rare. The reader is referred to in-depth subject matter on these disorders if further information is needed, because covering these hypocoagulable disorders is beyond the scope of this text.75 Table 16-8 provides some detailed information on these factor deficiencies and their treatment.

Contact Factors Deficiency

Factor XII, high-molecular-weight kininogen, and prekallikrein constitute the contact factors of the coagulation cascade (Fig. 16-9); definitive diagnosis is made with a specific assay for the deficient contact factor. Contact factor deficiencies are not associated with bleeding, although these individuals demonstrate a prolonged aPTT. No specific perioperative treatment is required for these disorders.75

Factor XIII Deficiency

Factor XIII is important for the structural stabilization of a fibrin clot; it creates cross-linking peptide bonds between the fibrin strands in a blood clot (Fig. 16-9). Blood clots that lack cross-linked peptide bonds are unstable, permeable to blood, and susceptible to fibrinolysis and provide a poor framework for wound healing. Although 3 different forms of congenital factor XIII deficiency have been described, each is extremely rare. All share the tendency for excessive perioperative bleeding, as well as a lifelong history of bleeding problems when factor XIII activity is less than 1%. Hemorrhage may be seen from birth and throughout life. Although most bleeding episodes are triggered by trauma, spontaneous intracranial bleeding can occur.75,77

Definitive diagnosis of factor XIII deficiency is made with a specific clot solubility assay and an assay to quantify the level of factor XIII deficiency. An acquired version of factor XIII deficiency exists that involves the generation of factor XIII autoantibody.75,77

The treatment of factor XIII deficiency is accomplished by administration of fresh-frozen plasma or cryoprecipitate. In Europe, factor XIII concentrate (Fibrogammin-P) is available to treat this disorder. Normal hemostasis occurs with factor XIII levels of greater than 5%. Preoperative preparation of individuals with factor XIII activity levels of less than 5% is to give 2 to 3 mL/kg of fresh-frozen plasma or 1 unit of cryoprecipitate per 10 to 20 kg of body weight. The long plasma half-life of factor XIII (9-10 days) usually obviates a need to re-dose. Lifelong supplementation of factor XIII is indicated in individuals with very low levels to prevent intracranial hemorrhage, even when surgery is not planned. The treatment of individuals who have developed inhibitory antibodies to factor XIII is complex and may involve exchange transfusion, administration of platelets (ie, platelets are known to contain factor XIII), or administration of immunosuppressants (eg, intravenous gamma globulin, steroids, and cyclophosphamide).75,77

Afibrinogenemia

Fibrinogen serves as the precursor molecule for the proteinaceous fibrin scaffolding of a blood clot. A congenital or acquired deficiency of fibrinogen can produce clinically important bleeding disorders, especially in the surgical patient. Congenital homozygous afibrinogenemia is a rare disorder that is associated with a varying severity of bleeding problems (eg, easy bruising, mucosal hemorrhage, hematuria, hemarthroses, hemopericardium, hemoperitoneum, menometrorrhagia, obstetrical complications, intracerebral hemorrhage, and spontaneous splenic rupture). These patients have no detectable serum fibrinogen and a markedly prolonged PT, aPTT, and TT; occasional thrombocytopenia; and a plasma fibrinogen assay reveals no detectable fibrinogen.75

Preoperative fibrinogen levels must be restored to 50 to 100 mg/dL by infusing cryoprecipitate or human-derived pooled fibrinogen concentrate (Riastap). Five to ten units of cryoprecipitate are needed to increase fibrinogen levels to 50 to 100 mg/dL. A single unit of cryoprecipitate contains approximately 250 to 300 mg of fibrinogen and is expected to raise the plasma fibrinogen level approximately 10 mg/dL. Fibrinogen levels will persist for 2 to 4 days after infusion. The following formula is used to determine the necessary dose of fibrinogen concentrate:

or, when the fibrinogen level is unknown, administer 70 mg/kg body weight of fibrinogen concentrate.78 Operations associated with postoperative hemorrhage result in decreased fibrinogen levels so they should be monitored perioperatively on a daily basis. Thrombosis has been reported in patients with fibrinogen levels normalized via cryoprecipitate infusion.75,79

Dysfibrinogenemia

Many variants of congenital dysfibrinogenemia exist. They share the common feature of abnormal conversion of fibrinogen to fibrin. Most individuals with dysfibrinogenemias are heterozygous and usually produce nearly 50% of normal plasma fibrinogen levels, enabling effective hemostasis. Some dysfibrinogenemia variants result in fibrinogen levels well below normal or despite near normal levels of fibrinogen are associated with bleeding problems if the abnormal variant affects the function of the fibrinogen.75

Clinical manifestations of dysfibrinogenemia include the following: thrombosis (17%), mild bleeding after trauma (20%), both thrombotic and hemorrhagic manifestations (20%), no hemorrhagic phenomenon, or asymptomatic (43%). Affected patients who demonstrate bleeding problems usually manifest with soft tissue bleeding, easy bruising, menorrhagia, and most commonly perioperative bleeding. Individuals who manifest thromboses can experience either venous (eg, deep venous thrombosis, pulmonary embolism) or arterial involvement (eg, thrombosis of the aorta and carotid arteries). Individuals with thrombotic manifestations can possess concurrent disorders of other portions of the coagulation system (eg, factor V Leiden mutation or other thrombophilias) contributing to their propensity to inappropriately clot. Poor surgical wound healing can be associated with dysfibrinogenemia.75

Standard laboratory tests of the coagulation system (eg, PT, aPTT, TT) are usually prolonged despite normal levels of fibrinogen due to a functional defect in the circulating fibrinogen. The reptilase time is often prolonged and fibrinogen immunoelectrophoresis using agarose gels can reveal an abnormal pattern of protein migration. Acquired dysfibrinogenemia associated with hepatic disease can be differentiated from congenital dysfibrinogenemia because it will demonstrate decreased levels of other clotting factors that are synthesized by the liver.

Symptomatic dysfibrinogenemia is treated by administering cryoprecipitate in a manner similar to that described in the Afibrinogenemia section of this chapter (ie, the goal is to restore functional fibrinogen levels to 50-100 mg/dL). Fresh-frozen plasma or fibrinogen concentrates (eg, Riastap) may also be administered to restore functional levels of fibrinogen to normal. The thrombotic variants of dysfibrinogenemia require treatment with intravenous anticoagulation with unfractionated heparin and eventual conversion to oral anticoagulation.78

Prothrombin (Factor II) Deficiency

Prothrombin deficiency consists of either a quantitative (hypoprothrombinemia) or qualitative (dysprothrombinemia) defect. Prothrombin is normally converted to thrombin (factor IIa) (Fig. 16-9). Thrombin is needed for the conversion of fibrinogen to fibrin, which serves as the protein scaffold of a blood clot. Dysprothrombinemia is an extremely rare disorder that can be either homozygous, heterozygous, or compound heterozygous.75

Heterozygous individuals generally have prothrombin activity levels of 50%, normal quantitative levels of prothrombin, and are asymptomatic or have only minor bleeding symptoms (eg, may develop bleeding only after surgical procedures). Homozygous or compound heterozygous individuals demonstrate a lifelong history of bleeding.75

Dysprothrombinemia is definitively diagnosed using an assay of functional prothrombin activity. PT and aPTT are prolonged; both will correct when mixed 1:1 with normal plasma. Hepatic function may cause acquired dysprothrombinemia. Acquired dysprothrombinemia is distinct in that multiple hepatic synthetic defects are present in the absence of a history of bleeding problems.75

Symptomatic dysprothrombinemia may be treated by administration of fresh-frozen plasma at a loading dose of 15 to 20 mL/kg followed by 3 mL/kg every 12 to 24 hours (ie, an appropriate regimen for individuals with severe bleeding), administration of prothrombin complex concentrates with a loading dose of 20 U/kg prothrombin followed by 5 U/kg every 24 hours, with care not to exceed these doses to avoid the risk of thromboembolism, or plasma exchange transfusion to restore normal levels of prothrombin.75

Drug-Induced Hypocoagulable States

Thrombotic complications associated with surgery and percutaneous coronary interventions have stimulated the development of intravenous, subcutaneous, and oral anticoagulant drugs. The 2 general classes of anticoagulant drugs include antiplatelet and antithrombotic agents.79,80

Knowing the pharmacodynamic and pharmacokinetic profiles of drugs that affect the coagulation system is necessary to make safe decisions regarding the timing of surgical procedures. Safe conduct of regional anesthesia requires knowledge of how these drugs affect coagulation. An anticoagulant drug's half-life may not predict the duration of its effects on the coagulation system if it irreversibly inhibits one or more aspects of the coagulation system. For example, aspirin has an elimination half-life of 15 to 20 minutes but irreversibly inhibits platelet activation through a selective enzymatic pathway.81 Consideration of the route of elimination and the individual patient's level of organ function is necessary to predict a drug's effect. For instance, compromised renal function will prolong the elimination half-life of a drug that relies on the kidney to clear its active form. In some instances, suppression or supplementation of components of the coagulation system will be necessary to allow the safe conduct of an invasive procedure (ie, patients who are systemically anticoagulated with unfractionated heparin may require the administration of protamine to neutralize the effects of the heparin-antithrombin complex on factor II and activated factor X).

The Antithrombotic Agents

The antithrombotic agents include 5 subcategories of drugs: unfractionated heparin (UFH), low-molecular-weight heparin (LMWH), activated factor X inhibitor, direct thrombin inhibitors, and coumarins. Antithrombotic agents are used to both treat and prophylax against thrombotic events.79

UFH, a large molecule that is extracted either from porcine intestine or bovine lung, exerts its antithrombotic effect primarily via an interaction with antithrombin (Fig. 16-9). The heparin-antithrombin complex permits an approximately 1000-fold increase in the ability of heparin to inhibit factor II (prothrombin) activity and activated factor X activity.82 The result of UFH's action is suppression of the coagulation cascade, as reflected in a dose-dependent prolongation of both the aPTT and the activated clotting time (ACT). The half-life of the anticoagulation effect of UFH as estimated by the aPTT is approximately 1.5 hours irrespective of the heparin dose. In contrast, the functional half-life varies with the magnitude of the administered dose from approximately 40 to 150 minutes, with larger doses yielding a longer duration. UFH's elimination half-life is not affected by renal dysfunction because the drug is cleared from the plasma by transfer to the extravascular space, most likely to the reticuloendothelial system. UFH dosage is calculated according to actual body weight; depending on the anticoagulation goal, different dosing schemes are used (eg, initiation of cardiopulmonary bypass requires a bolus of approximately 300 U/kg, whereas treatment of a venous thrombosis requires an IV bolus of approximately 5000 units plus an infusion).82 Laboratory monitoring is required (eg, the aPTT, ACT, whole-blood heparin concentration, or anti–factor Xa activity) because some individuals exhibit insensitivity to heparin.83 An elapsed time of 4 hours after IV heparin therapy cessation may be required to achieve an aPTT of less than 35 seconds, indicating abatement of UFH's anticoagulant effects, before safely initiating an elective surgical procedure or attempting central neuraxial blockade. Alternatively, IV protamine sulfate can be administered to rapidly neutralize UFH's anticoagulant effect. In practice, protamine is given in a range of 0.7 to 1.3 mg: 100 units of UFH administered, or alternatively the protamine dose can be titrated to effect by giving small doses (25-50 mg) and checking the ACT or aPTT to assess whether values have normalized. Protamine is best administered slowly to prevent systemic hypotension and/or pulmonary hypertension. Additional specific recommendations related to UFH treatment can be found in the American Society of Regional Anesthesia and Pain Medicine Practice Guidelines.84

LMWH was developed to create an antithrombotic agent that does not require anticoagulant monitoring in the absence of renal insufficiency. LMWH is produced from UFH by a process that decreases the size of the polysaccharide chains of the heparin molecule, resulting in a smaller molecule with potent anti-activated factor X activity. Both UFH and LWMH require antithrombin to exert their effect on activated factor X (Fig. 16-9). A number of LWMH preparations are available for clinical use: enoxaparin (Lovenox), dalteparin (Fragmin), ardeparin (Normiflo), and tinzaparin (Innohep). Patients with renal insufficiency receiving LMWH require dose adjustment and plasma monitoring of anti-activated factor X activity.85,86 The elimination half-lives of the various preparations of LMWH vary from 3 to 5 hours. Therefore waiting 24 hours after administration is recommended after higher doses (ie, any dose of enoxaparin >1 mg/kg or dalteparin >120 U/kg) of LMWH before an elective surgical procedure or attempting central neuraxial blockade.84 Although the anticoagulant effects of LMWH can be partially reversed by protamine sulfate, its use has been associated with increased bleeding complications perioperatively, despite protamine reversal.87

The activated factor X inhibitor fondaparinux (Arixtra) requires interaction with antithrombin to exert its anticoagulant effect. Once bonded with antithrombin, the fondaparinux-antithrombin complex selectively increases the ability of antithrombin to inhibit activated factor X activity approximately 300-fold. The molecule is smaller than LMWHs and does not elicit formation of or react with the platelet factor 4 (PF4) antibody associated with heparin-induced thrombocytopenia type II. The elimination half-life is 17 to 21 hours. No definite recommendations exist for management of this drug before attempting neuraxial blockade, although it is recommended to use a single-needle atraumatic placement technique and avoid an indwelling catheter if a neuraxial procedure is required. Although no evidence-based recommendations exist, delaying elective surgery for approximately 5 half-lives, or 4 days, is a conservative management strategy for patients treated with this drug.84 Because there is minimal metabolic breakdown and the primary route of elimination is renal, dose adjustment is required for patients with renal insufficiency. Assays of anti-activated plasma factor X levels are not required in the absence of renal insufficiency.85

Direct thrombin inhibitors (DTI) include dabigatran (Pradaxa), lepirudin (Refludan), argatroban, desirudin (Iprivask), and bivalirudin (Angiomax). They exert their effect by interacting with free or clot-bound thrombin to inhibit conversion of fibrinogen to fibrin (Fig. 16-9). DTIs do not rely on the presence of antithrombin to exert their effects. They are given to treat or prevent thrombotic events, including the treatment of patients undergoing percutaneous coronary interventions with or at risk for heparin-induced thrombocytopenia (argatroban) or for PCI (bivalirudin).88 The half-lives of these drugs vary (Table 16-13). Bivalirudin has recently been demonstrated to be a safe and effective anticoagulant for cardiac surgical patients requiring cardiopulmonary bypass.89,90 Although no evidence-based recommendations exist, conservative management suggests that patients on continuous IV infusions of DTIs should have the drug discontinued 5 elimination half-lives before an operation or neuraxial blockade; no antidote exists for DTI-related bleeding.84 DTIs do not elicit formation of or cross-react with the PF4 antibody associated with heparin-induced thrombocytopenia type II and are therefore useful to prevent and treat thrombotic events in patients with this syndrome. Clinical effects of DTIs are commonly followed by either the aPTT or the ACT. Dabigatran is the only FDA-cleared oral DTI available in the United States and has a half-life of approximately 17 hours (considerably longer than the half-lives associated with the IV DTIs) in patients with normal renal function, necessitating discontinuation of this drug approximately 4 to 5 days prior to implementing central neuraxial blockade or a regional anesthetic procedure.84,91

Coumarins, a subclass of oral antithrombotics, exert their anticoagulant effect by inhibiting the hepatic synthesis of vitamin K–dependent coagulation factors. Coumarins inhibit the synthesis of factors II, VII, IX, and X as well as proteins C and S. Assessing the PT or international normalized ratio (INR) can be used to quantify low levels or activities of these specific coagulation proteins. Vitamin K–dependent factors participate in the extrinsic and common coagulation pathways (Fig. 16-9); therefore, the PT or INR is the best test to assess coumarin effects. PT is reported as the INR to avoid inter-laboratory variation of absolute PT results.79 Warfarin is the only coumarin approved by the FDA. It is metabolized by the liver and possesses an elimination half-life of 20 to 60 hours. Anticoagulant effects of warfarin take several days to abate depending on the level of warfarin effect at discontinuation. Administration of vitamin K will decrease the time for the effects of warfarin to abate, but may make it difficult to reinstitute a therapeutic warfarin effect, thereby complicating postoperative anticoagulation therapy. Alternatively, administration of fresh-frozen plasma will acutely reverse the anticoagulant effects of warfarin. However, anticoagulation may recur several hours later due to the longer half-life of warfarin than the coagulation factor levels that it inhibits (ie, warfarin's half-life is 20-60 hours; half-life of factor VII is approximately 6 hours). Administration of prothrombin complex concentrates (eg, Beriplex) may offer another option for rapid and effective reversal of warfarin's effects. Excessive operative hemorrhage and neuraxial anesthesia associated with spinal or epidural hematoma may occur if warfarin is not neutralized.84 Individuals at increased risk for a serious thrombotic event should be anticoagulated with a second intravenous or subcutaneous antithrombotic agent (eg, UFH, LMWH, or a short-acting direct thrombin inhibitor) before warfarin is discontinued or neutralized. Titration of antithrombotic therapy is complex in the presence of more than 1 antithrombotic agent or when a preexisting factor deficiency is present.

Guidelines from the American Society of Regional Anesthesia recommend stopping warfarin 5 days before attempting a central neuraxial blockade or regional anesthesia and verifying that its effects have abated by obtaining an INR value of 1.4 or less.84

Antiplatelet Agents

The 5 subcategories of antiplatelet agents include nonsteroidal anti-inflammatory drugs, glycoprotein IIb/IIIa receptor inhibitors, platelet adhesion inhibitors, platelet adenosine diphosphate (ADP) receptor antagonists, and platelet production-limiting agents. Platelets can be activated by more than 1 stimulus, and therefore not all antiplatelet agents will place surgical patients at an increased risk for bleeding. Indeed, antiplatelet drugs have proven quite beneficial to patients with cardiovascular disease.92,93 The increasing use of herbal medications in the United States has drawn attention to the anticoagulant effects of select agents; limited evidence-based data exist describing the interactions of herbal medications with other medications and their effect on the clinical course of surgical patients.84,94,95

Nonsteroidal Anti-Inflammatory Drugs

Aspirin exerts its antiplatelet effects by irreversibly inhibiting the platelet enzyme cyclooxygenase. This prevents the formation of thromboxane A2 from arachidonic acid for the lifetime of the platelet, which in turn blocks platelet activation and aggregation and the release of thromboxane A2, which propagates further platelet activation. Normal platelets have an average circulating life span of approximately 10 days; thus it may take up to 5 days for half-normal platelet function to return after treatment with aspirin, despite measurement of a normal platelet concentration. Aspirin's half-life is 15 to 20 minutes in the plasma, and it undergoes primarily hepatic metabolism as well as plasma esterase metabolism. Other nonsteroidal anti-inflammatory drugs (NSAIDs), such as ibuprofen, ketorolac, and naproxen, produce reversible platelet cyclooxygenase inhibition. Therefore, the return of platelet function correlates with the drug's half-life, and perioperative management with these antiplatelet agents is easily accomplished when one considers that permitting 5 half-lives to elapse after the last dose will completely eliminate the effects of these NSAIDs. For aspirin a distinct perioperative treatment strategy is needed. After aspirin removal, antiplatelet effects of aspirin can only be overcome by platelet transfusion.81 Doubt exists whether NSAIDs result in significant surgical hemorrhage. If an NSAID is the only inhibitory influence on the coagulation cascade, then later central neuraxial blockade can be safely undertaken.84 In selected patients with severe peripheral arterial vascular disease, it is advisable to continue perioperative aspirin.96

Glycoprotein IIb/IIIa Receptor Inhibitors

The glycoprotein IIb/IIIa (GP IIb/IIIa) receptor inhibitors are available only as IV preparations and are among the most potent inhibitors of platelet function. Currently the following GP IIb/IIIa receptor inhibitors are approved for clinical use in the United States: abciximab (ReoPro), eptifibatide (Integrili), and tirofiban (Aggrastat). Platelets are activated by various stimuli, including thromboxane, ADP, epinephrine, serotonin, and mechanical shear stress. The final common pathway of platelet activation involves platelet aggregation via interaction of GP IIb/IIIa receptors and either fibrinogen or vWF as an intermediate molecule. Complete inhibition of this final step of platelet aggregation could render a patient completely thrombasthenic.97-99

Abciximab is the Fab portion of the chimeric human/murine monoclonal antibody 7E3 that is directed against the platelet GP IIb/IIIa receptor. In therapeutic doses it creates a marked decrease in the ability of platelets to aggregate. In approximately 1% to 2% of patients, it produces significant thrombocytopenia.97 Abciximab also inhibits thrombus formation by suppressing tissue factor-induced thrombin generation. It is FDA approved for use as an adjunct in both percutaneous coronary interventions (PCI) and the treatment of refractory angina. Abciximab has a 30-minute plasma half-life. However, once bound to the platelet GP IIb/IIIa receptor, it remains in the circulation for up to 15 days. The putative mechanism for this is migration from platelet to platelet. The anti-aggregation effect lasts 12 to 48 hours. Platelet transfusion may partially restore platelet function if significant bleeding occurs 12 to 48 hours after infusion.97 Central neuraxial blockade is contraindicated within 24 to 48 hours of abciximab administration.84 Delaying cardiac surgery in the stable patient treated with abciximab for 12 to 24 hours will allow for some return of platelet function.

Tirofiban, a non-peptide chemical antagonist of the platelet GP IIb/IIIa receptor, produces a greater than 90% reversible reduction in platelet aggregation. Tirofiban is FDA approved for the treatment of acute coronary syndrome (ACS) in patients undergoing angioplasty or being medically managed.98 The half-life is reported at 90 to 180 minutes; it undergoes renal excretion largely as unchanged drug. Central neuraxial blockade should be delayed for 4 to 8 hours in patients who are receiving tirofiban.84 It does not appear to confer an increased risk of hemorrhage after emergent or urgent coronary surgery.100 No specific recommendations are available for general surgical patients at this time.

Eptifibatide is a heptapeptide molecule that inhibits the platelet GP IIb/IIIa receptor and is a potent inhibitor of platelet aggregation, with FDA approval for use in PCI and for the treatment of ACS. The drug is cleared from the circulation by the kidneys with a half-life of 2.5 hours. Approximately 71% of the drug is eliminated unmetabolized.99 A 4- to 8-hour delay is recommended after discontinuation of this drug before attempting central neuraxial blockade.84 Data from a large trial suggest that it is safe to perform coronary surgery within 2 hours of the discontinuation of eptifibatide.99

ADP Antagonists

Clopidogrel (Plavix), ticlopidine (Ticlid), and prasugrel (Effient) are FDA-approved ADP receptor antagonists, selectively and irreversibly inhibiting ADP-induced platelet aggregation by blocking the P2Y12 ADP receptor on the platelet's surface. These oral agents lack effects on prostacyclin and thromboxane synthesis and work by preventing the ADP platelet P2Y12 receptor from transmitting a signal to the platelet to activate GP IIb/IIIa receptor complexes.

Clopidogrel is FDA approved to reduce thrombotic events related to a recent myocardial infarction or stroke and established peripheral arterial disease. Additionally, clopidogrel is approved for both medical management and PCI treatment of acute coronary syndrome. Clopidogrel and ticlopidine both undergo extensive hepatic metabolism. Clopidogrel is excreted in both urine and feces. The half-life of clopidogrel is approximately 8 hours.101 The irreversible nature of the inhibitory effect of clopidogrel on platelets requires 5 to 7 days before a sufficient population of functional platelets are restored. Seven full days are recommended after treatment with this drug before attempting central neuraxial blockade.84 Clopidogrel-treated patients exhibit significant bleeding and transfusion requirements after cardiac surgery.102 Evidence-based recommendations do not exist for general surgery patients at this time.

Ticlopidine is FDA approved for coronary stenting in conjunction with aspirin therapy. Ticlopidine is also FDA approved for the prophylaxis of thromboembolic stroke. At least 1 metabolite of ticlopidine is a potent inhibitor of ADP-induced platelet aggregation. This metabolite is excreted primarily in the urine. Ticlopidine has a reported half-life of 12.6 hours, which increases dramatically to 5 days with repeated dosing. Ticlopidine is used less frequently than clopidogrel because of toxicity (eg, neutropenia, agranulocytosis, and thrombotic thrombocytopenic purpura).103 A delay of 14 days is recommended after discontinuation of this drug before attempting central neuraxial blockade.84 Evidence-based recommendations do not exist for general surgery patients at this time.

The recently FDA-approved P2Y12 receptor inhibitor, prasugrel, is metabolized by both intestinal and serum esterases as well as the liver. It is given to reduce thrombotic cardiovascular events (including stent thrombosis) in patients with ACS who are managed with PCI. The effects of prasugrel on platelets are analogous to those of clopidogrel and abate approximately 5 to 9 days after discontinuation. It has a half-life of approximately 7 hours, which is irrelevant because it irreversibly inhibits ADP-mediated platelet aggregation.104 Perioperative recommendations related to bleeding and regional anesthesia are identical to those of clopidogrel.

Platelet Adhesion Inhibitors

Dipyridamole (Persantine, Aggrenox) and cilostazol (Pletal) are drugs classified as platelet adhesion inhibitors. Dipyridamole, available in oral and IV preparations, has a complex mechanism of action that ultimately inhibits platelet aggregation by blocking their activation by ADP, collagen, and platelet-activating factor. Dipyridamole is FDA approved to be used as an adjunct with warfarin for the prophylaxis of thromboembolic complications associated with cardiac valve replacement and for use with thallium in myocardial imaging studies. It is also given to prevent thromboembolic events associated with coronary surgery. Dipyridamole primarily undergoes hepatic metabolism and has a half-life of approximately 9 to 13 hours.105

Dipyridamole administration is associated with bleeding. The risk is greater when dipyridamole is combined with aspirin therapy. A combination of dipyridamole and aspirin, Aggrenox, is commercially available. Aggrenox is FDA approved for the prophylaxis of cerebrovascular accidents in patients with a previous history of stroke or transient ischemic attacks.106 It may be prudent to delay central neuraxial anesthesia for 7 days after Aggrenox therapy to allow the recovery of approximately 50% of normal platelet aggregation. Published literature does not suggest the use of dipyridamole alone to be a direct contraindication to proceeding with elective surgery or neuraxial blockade; however, because dipyridamole works by suppressing platelet aggregation, it may be prudent to avoid elective surgery and neuraxial blockade for 4 days. One reported meta-analysis concludes that among the antiplatelet drugs, dipyridamole was associated with the lowest incidence of bleeding complications.107 In a large trial of antiplatelet drugs used for the prevention of stroke in medical patients, dipyridamole was indistinguishable from placebo in terms of producing bleeding complications.108

Cilostazol increases intracellular cyclic adenosine monophosphate (cAMP) by inhibiting the enzyme phosphodiesterase III. An increase in levels of cAMP results in reversible inhibition of platelet aggregation. It is FDA approved for reduction in the symptoms of intermittent claudication. It is extensively metabolized by the liver into active metabolites that are primarily cleared by the kidneys and has a reported elimination half-life of 11 to 13 hours. Because no evidence-based recommendations exist, a conservative management approach would be to delay elective surgery or central neuraxial blockade for 48 hours after administration.109

Herbal Therapies

Self-medication with herbal therapies has been increasing in the United States.94 Select herbal therapies are now recognized as having deleterious effects on elements of the coagulation cascade, although there is a paucity of research.95 Garlic, ginkgo, ginseng, ginger, feverfew, fish oil, and dong quai have been implicated as inhibiting normal hemostasis; several of these agents can interfere with platelet aggregation.84,94,95 Anesthesiologists should routinely inquire during preoperative evaluation whether patients are taking herbal supplements. Patients taking herbal supplements should be advised to discontinue their use preoperatively. The current lack of literature makes it uncertain whether to recommend delaying elective surgical procedures or neuraxial blockade for patients using these supplements. One may expect that individuals taking these herbal therapies along with the concurrent use of other antiplatelet or antithrombotic agents could exhibit synergistic suppressive effects on their coagulation system.

Hypercoagulable States

Thrombophilias, or hypercoagulable states, place surgical patients at increased risk for experiencing a postoperative thrombotic event. Many of these hypercoagulable states are managed with temporary or lifelong systemic anticoagulation therapy. Environmental influences (eg, trauma, surgery, infection, or administration of a drug) will increase the likelihood of a patient with either a congenital or acquired hypercoagulable disorder to develop a thromboembolic event and thus can impact the perioperative management of these patients.110,111 Several thrombophilias are currently recognized, and anesthesiologists should be familiar with them to aid in the provision of optimal perioperative care. Discontinuation and reinitiation of systemic anticoagulation in the perioperative period is most safely managed in consultation with a hematologist.

Antithrombin Deficiency

Antithrombin (also termed antithrombin III) is a plasma protein produced by the liver that prevents hypercoagulability by binding to and neutralizing thrombin, factors IXa, Xa, XIa, and XIIa (Fig. 16-9). Antithrombin (AT) deficiency is associated with a predilection for thrombosis. AT deficiency is a congenital disorder inherited as an autosomal-dominant trait that is estimated to occur with a prevalence of 1 in 250 to 500 people.112 Individuals with a history of an unexplained thrombotic event should be evaluated to rule out this disorder.

AT deficiency may be due to decreased synthesis (type I) of this protein or due to decreased protein activity despite normal levels (type II). AT levels are between 75% and 120% in normal individuals. AT levels are measured with a specific assay. Disease states associated with reduced AT activity include sepsis, disseminated intravascular coagulation, burns, acute thrombosis, hepatic disease, severe trauma, and nephritis.112

Patients with AT deficiency have a predisposition to develop deep vein thromboses and therefore are often maintained on warfarin. Perioperative management of anticoagulation should be tailored to the invasiveness of the planned intervention or surgical procedure. Surgery induces a prothrombotic state and increases the risk of perioperative thrombosis.113

Anticoagulation with UFH is frequently required perioperatively. Importantly, many AT-deficient patients exhibit a relative resistance to UFH because UFH anticoagulation requires coupling with AT to induce an increase in AT's neutralizing effect on the procoagulant factors IIa, IXa, Xa, XIa, and XIIa.83 ACT or aPTT assays will reveal heparin resistance in patients with an AT deficiency. Heparin resistance in this context is defined as a failure of UFH to produce the expected degree of ACT or aPTT prolongation. Heparin resistance due to AT deficiency is most efficiently treated with plasma-derived, purified AT III concentrate (Thrombate). A bolus of AT III concentrate at 50 IU/kg has been recommended for therapy. If AT III concentrate is not used, heparin resistance can be treated with 2 to 4 units of fresh-frozen plasma with the effect of this therapy assessed by an ACT or aPTT assay.114

Protein C Deficiency

Protein C is a hepatically synthesized vitamin K–dependent protein that, when converted to its active form (activated protein C), exerts an anticoagulant effect through interaction with factors Va and VIIIa (Fig. 16-9). Protein C deficiency may be due to either a quantitative or qualitative defect. It is estimated to occur with a prevalence of 1 in 200 to 500.115

Many patients with protein C deficiency exhibit a prothrombotic tendency and can require systemic anticoagulation. Protein C deficiency is diagnosed with an aPTT-based clotting assay.116 The diagnosis of protein C deficiency in patients who have recently been treated with warfarin is complex. Patients with protein C deficiency requiring a surgical intervention require perioperative modification of their anticoagulation therapy.

Protein S Deficiency

Protein S is synthesized primarily in the liver and functions as a cofactor for activated protein C. Quantitative and qualitative congenital protein S deficiencies have been described. Patients with protein S deficiency can have a prothrombotic tendency and may be maintained on systemic oral anticoagulation as outpatients; therefore, this necessitates perioperative modification of their outpatient regimen.112

Factor V Leiden (Activated Protein C Resistance)

Factor V Leiden, also termed activated protein C resistance, is associated with recurrent venous thromboembolic events. It is caused by a single point mutation in the gene that encodes for factor V. This mutation is common among patients who experience thrombotic events. It is commonly diagnosed by an aPTT-based assay. Patients with the factor V Leiden mutation may be on either temporary or lifetime systemic oral anticoagulation.112

Prothrombin G20210 Mutation

Prothrombin, or factor II, is an essential element of the coagulation cascade. A single base pair substitution mutation results in affected individuals expressing higher plasma prothrombin activity associated with the occurrence of thromboembolic disease in these patients. This disorder is the second most common thrombophilia, with a prevalence of 2% in whites, with considerable geographic variation. It is commonly diagnosed using a polymerase chain reaction–based test.112

Heparin-Induced Thrombocytopenia

Heparin-induced thrombocytopenia (HIT) involves an immune-mediated response to either low-molecular-weight or unfractionated heparin administration that results in pathologic platelet activation causing arterial and/or venous thrombosis with or without associated tissue ischemia and infarction. There are 2 types of HIT. Type I is a non–immune-mediated activation of platelets in heparin-treated patients causing a mild reduction in circulating platelets. HIT type II, or heparin-induced thrombocytopenia with thrombotic syndrome (HITTS), is immune-mediated. It is caused by the interaction of a heparin-dependent immunoglobulin G (IgG) antibody with the molecular complex of platelet factor 4 (PF4) and heparin. The heparin-PF4–bound IgG antibody stimulates platelet activation via the platelet FcγIIa receptor. Once platelets are activated they release internal biochemical mediators that induce platelet aggregation and stimulate other platelets to become activated. The same biochemical mediators that are released from the activated platelets will also stimulate thrombin generation. Heparin-PF4 IgG antibody-induced platelet aggregation along with thrombin contributes to the formation of thrombus. Although venous thrombosis occurs most commonly in HIT type II, arterial thrombosis is also observed, especially at the site of arterial disruption (eg, aortotomy site in cardiac surgery). Thrombi can cause organ ischemia and/or infarction. Loss of limbs and damage to vital organs, such as renal failure or strokes, are recognized complications of this disorder.117

Numerous features of HIT type II are poorly understood. Only a portion of patients exposed to heparin develop the IgG antibody directed against heparin and PF4 (commonly termed the PF4 antibody). Only a small percentage of patients who develop the PF4 antibody will develop thrombocytopenia and a recognized thrombotic event. At least a 50% decrease in platelet count from the pre–heparin administration baseline is preferred to establish the diagnosis of HIT type II. Using the absolute laboratory criterion of fewer than 150,000 platelets/mm3 is not completely appropriate to diagnose HIT type II. Patients who are generating the PF4 antibody on initial exposure to heparin characteristically take at least 5 days to develop sufficient amounts of PF4 antibody to cause thrombocytopenia.117 Patients previously exposed to heparin who have generated PF4 antibody in the past 100 days may develop thrombocytopenia more rapidly. In contrast, the thrombocytopenic response and thrombotic event can occur several days to weeks after the discontinuation of heparin therapy.118

HIT type II may be diagnosed using either an antigen assay (ie, highly sensitive) or a platelet activation assay (ie, highly specific); these tests are indicated in patients with a history of thrombocytopenia temporally related to heparin administration or in patients receiving heparin with previously documented HIT type II. The antigen test uses the enzyme-linked immunosorbent assay (ELISA) to detect the presence of the PF4 antibody and has proven to be highly sensitive.117,119 The activation assays detect PF4 antibodies by detecting their ability to activate platelets in the presence of heparin. The platelet serotonin release assay is an example of an activation assay that detects HIT type II by quantifying the release of 14C-labeled serotonin from platelets activated by heparin in the presence of the PF4 antibody.117

Treatment of HIT type II consists of immediate and complete discontinuation of heparin by any routes. Additionally, immediate systemic anticoagulation is indicated in the setting of thrombosis; the current agents of choice are the direct thrombin inhibitors (DTIs). After systemic levels of a DTI are achieved, oral anticoagulation with a coumarin may be established and the DTI discontinued. Oral coumarin therapy must not begin until effective thrombin suppression has been achieved to avoid precipitating warfarin-induced skin necrosis.120

Given the widespread use of heparin, HIT type II is an important consideration for anesthesiologists. Heparin should be avoided in patients with HIT type II and an elevated PF4 antibody titer in order to prevent precipitation of a thrombotic event. Patients with a remote history of an elevated PF4 antibody titer should not be treated with heparin if this can be avoided (ie, use a DTI). No reversal agents are available for the DTIs; their suppression of thrombin abates when the drug is cleared from the circulation.

Disseminated Intravascular Coagulation

Disseminated intravascular coagulation (DIC) is a deviation from the normal physiologic balance between the processes of thrombus formation and fibrinolysis. Pathologic hemorrhage, pathologic thrombosis, or both can occur in DIC. Numerous disease processes have been identified as causing DIC (Table 16-14). All of these disease states have in common the production of intense, sustained stimulation of the coagulation axis, achieved by inflammatory mediators and resulting in dysregulation of coagulation and/or fibrinolysis. The association of DIC with inflammation explains why this condition can be associated with multiorgan dysfunction syndrome and acute respiratory distress syndrome, 2 disorders that are characterized by insufficient regulation of the inflammatory response. The pathophysiologic mechanisms of DIC are not completely understood. Pathologic thrombosis at the microcirculatory level results in organ malperfusion, ischemia, infarction, and dysfunction as well as pathologic consumption of essential coagulation elements (eg, platelets, fibrinogen, and soluble coagulation proteins). Dysregulation of fibrinolysis can contribute to DIC-associated thrombosis (ie, there is over-suppression of fibrinolysis from excessive plasminogen activator inhibitor-1) or to DIC-associated hemorrhage (ie, with overstimulation of fibrinolysis from high levels of inflammatory mediators including interleukin-1, interleukin-6, interleukin-10, and tissue necrosis factor).121

The diagnosis of DIC is approached from 2 vantage points. First, because DIC is a manifestation of a primary disease process, one must investigate potential clinical causes that may give rise to DIC (Table 16-14). Second, although DIC is often initially suspected as a clinical diagnosis, central laboratory values may help to establish the diagnosis. Some authorities advocate for more readily available central laboratory tests, including the aPTT, PT, TT, fibrinogen, fibrin degradation products (FDP) level, D-dimer level, and complete blood count. Although an elevation of FDPs or D-dimers, prolonged aPTT, PT, and TT as well as a decreased fibrinogen level and platelet concentration are not universally seen in DIC, they can be useful in making the diagnosis with strong clinical evidence.121

The treatment of DIC is most appropriately directed at the underlying cause of the disorder (eg, intra-abdominal abscess drainage and antibiotics in a patient with sepsis). Exhaustive efforts to correct laboratory abnormalities may not be needed in all cases of DIC, especially those in which patients require surgical treatment. Perioperative goals include maintaining the platelet concentration in the range of 25,000 to 50,000/μL and the fibrinogen level greater than 50 mg/dL. However, few studies exist on the most effective management of the patient with DIC. Presently, with the exception of Trousseau syndrome, anticoagulation is not recommended as the treatment of DIC. Administration of antithrombin concentrate, though advocated by some, is also not recommended.121 Hemorrhaging DIC patients should be volume resuscitated as appropriate to restore intravascular volume to a level consistent with organ perfusion. As previously stated, this coagulopathy is not expected to improve until the underlying cause of DIC has been appropriately treated.

Hyperhomocysteinemia

Hyperhomocysteinemia has been established as an independent risk factor for stroke, myocardial infarction, carotid artery disease, and venous thrombosis. Plasma homocysteine levels are partially genetically determined but are also influenced by environmental factors, specifically by dietary intake of folic acid and vitamins B12 and B6. Hyperhomocysteinemia is present in 10% to 20% of patients who develop venous thrombosis. Hyperhomocysteinemia is diagnosed by measuring plasma levels of homocysteine. No specific recommendations exist for the perioperative management of patients with hyperhomocysteinemia, although vitamin supplementation reduces plasma homocysteine levels in the elderly. Reduction in homocysteine levels by vitamin supplementation has not been demonstrated to decrease the risk of vascular disease. It may be prudent to use postoperative venous thrombosis prophylaxis, especially in patient populations considered at risk for developing postoperative thrombosis (eg, orthopedic surgical patients). However, there are no specific recommendations on the best antithrombotic or antiplatelet treatment regimen.112

Antiphospholipid Antibody Syndrome

Antiphospholipid antibodies are directed against proteins that bind to phospholipid. They consist of 2 major subgroups: the lupus anticoagulants and the anticardiolipin antibodies. The lupus anticoagulants (LA) are antibodies that prolong phospholipid-dependent coagulation assays (eg, the aPTT). Anticardiolipin antibodies (aCL) target a molecular congener of cardiolipin (a bovine cardiac protein). The aCL antibodies share a common in vitro binding affinity for cardiolipin. Patients with arterial or venous thrombosis (ie, more common in this disorder) in whom these LA and aCL antibodies are detected are said to have the antiphospholipid antibody syndrome (APS). Primary APS occurs alone, whereas secondary APS occurs in association with other disease states, such as systemic lupus erythematosus (SLE).112

Treatment of APS is complicated by a lack of laboratory standardization and randomized treatment trials. Depending on the assay, antiphospholipid antibodies are reported in up to 10% of healthy people and in 30% to 50% of patients with known SLE. The antibodies are more common in patients with thrombosis, but a direct causal association remains unproven. The clinical relevance of transient or low titers of antiphospholipid antibodies is uncertain.122

The Sapporo criteria define APS as the presence of at least 1 clinical and 1 laboratory criterion. Clinical criteria include confirmed arterial, venous, or small-vessel thrombosis, recurrent fetal loss prior to the 10th week of gestation, or premature birth secondary to placental insufficiency, eclampsia, or preeclampsia. Laboratory criteria include medium or high titers of IgG or IgM aCL antibodies or the presence of LA on 2 or more occasions at least 6 weeks apart.112,122

APS patients with arterial thrombosis tend to have recurrence in an artery, whereas those who experience a venous thrombosis tend to redevelop venous thrombosis. Initial treatment consists of UFH or low-molecular-weight heparin for 5 days and then moderate-intensity warfarin (ie, INR 2.0-3.0) is continued for 6 months or longer or even indefinitely.112,122

Arterial thrombosis in APS most frequently involves the cerebral circulation. Studies have not demonstrated any difference in the risk of thrombotic events in patients treated with warfarin versus those treated with aspirin. Pregnant women with a history of APS and pregnancy loss are treated with UFH (5000 units subcutaneously twice daily) and aspirin (75-81 mg/d). Consensus opinion is that aspirin 81 mg/d may be considered for asymptomatic individuals who are not pregnant. The optimal treatment of patients with antiphospholipid antibodies who have arterial thrombosis not involving the CNS remains uncertain. Many of these patients are treated empirically with long-term warfarin therapy.112,122

Thalassemia

Thalassemia is a congenital hemolytic disorder caused by partial or complete deficiency of hemoglobin α- or β-globin chain synthesis. Homozygous carriers of the β-globin gene defect suffer from severe anemia and require chronic blood transfusion, often resulting in iron overload and progressive organ failure. Concurrent with improved care for patients with thalassemias has been a significant prolongation of their life expectancy; recognition of a chronic hypercoagulable state has accompanied this prolonged life span.123

Cerebral thromboembolic events, deep venous thrombosis (DVT), pulmonary embolism, and recurrent arterial thromboses have been described in different subtypes of thalassemia. Echocardiographic findings of pulmonary hypertension and right ventricular dysfunction in many patients with thalassemias, along with common autopsy findings of thrombotic lesions in the pulmonary arteries, lend support to a major incidence of recurrent pulmonary thromboembolism.123

Thalassemic patients have low levels of proteins C and S, enhanced platelet consumption, and ongoing activation of platelets, monocytes, granulocytes, and endothelium. There is evidence of continuous thrombin generation and enhanced fibrinolysis. Recently it has been suggested that prophylactic antithrombotic therapy should be given to thalassemic patients who are exposed to transient thrombotic risk factors, such as surgery, immobilization, and pregnancy.123

Nephrotic Syndrome

The nephrotic syndrome is a condition characterized by loss of plasma protein in urine, leading to profound hypoalbuminemia. These same urinary losses can also result in deficiencies of natural anticoagulants such as antithrombin and protein S. In addition, increased levels of procoagulants including fibrinogen, factor V, and factor VII as well as increased platelet aggregation can occur. Treatment of the syndrome can involve steroids and cyclosporine; both can increase procoagulant activities.124

The incidence of thromboembolism in adults with the nephrotic syndrome is reported to be as high as 44%. Both arterial and venous thromboses have been reported. Prophylactic anticoagulation is recommended for adults with nephrotic range proteinuria.124 No specific perioperative recommendations exist for patients with nephrotic syndrome.

Paraneoplastic Hypercoagulation

Both arterial and venous thromboses are known complications of cancer, with venous thromboembolism being the most common. Thrombosis is currently the second most common cause of death in cancer. The hemostatic system plays an important role in tumor angiogenesis. Tumor cells produce tissue factor, which directly activates factor X, creating a procoagulant microenvironment125 (Fig. 16-8). The interaction of monocytes and macrophages with malignant cells causes release of tumor necrosis factor, interleukin-1, and interleukin-6. These cytokines produce endothelial damage and sloughing of endothelial cells with exposure of the underlying thrombogenic surface (ie, exposed collagen). The interaction of tumor cells and macrophages also activates platelets, factor XII, and factor X, leading to generation of thrombin and subsequent thrombosis.126 The administration of chemotherapy agents further increases the risk of thromboembolism. The overall risk of risk of thromboembolism varies depending on tumor type, chemotherapy, use of erythropoietin-stimulating agents, presence of central venous catheters, and surgery. Many cancer patients receive long-term anticoagulation therapy with low-molecular-weight heparin.126

Thrombotic Thrombocytopenic Purpura

Thrombotic thrombocytopenia purpura (TTP) is a disorder of platelet aggregation leading to microvascular occlusion. Both familial and acquired forms of this disorder exist; the mechanism of it remains unknown. Idiopathic TTP seems to strike randomly, whereas certain medications such as ticlopidine, clopidogrel, and various chemotherapy agents are associated with TTP. It is characterized by a severe hemolytic anemia and thrombocytopenia along with various neurologic and psychiatric manifestations. Patients may experience hematuria, proteinuria, and ischemia to the retinal, coronary, and abdominal arteries. Coagulation studies are typically normal during the early stages of a TTP episode, but DIC may develop if significant tissue necrosis occurs. Treatment of the disorder consists of the administration of FFP, glucocorticoids, and plasma exchange with fresh-frozen plasma.127

• Avery EG. Anesthetic considerations in the patient with 1) anemia and 2) coagulopathy. In: Longnecker DE, Brown DL, Newman MF, et al, eds. Anesthesiology. New York, NY: McGraw-Hill; 2008: 239-266.
• Doctor A, Stamler JS. Nitric oxide transport in blood: A third gas in the respiratory cycle. Compr Physiol. 2011, 1-28. doi: 10.1002/cphy.c090009
• Francis CW. Antithrombotic agents. In: Kitchens CS, Alving BM, Kessler CM, eds. Consultative Hemostasis and Thrombosis.2nd ed. Philadelphia, PA: Saunders Elsevier; 2007:449-460.
• Hébert PC, Van der Linden P, Biro G, et al. Physiologic aspects of anemia. Crit Care Clin. 2004;20(2):187-212.
• Horlocker TT, Wedel DJ, Rowlingson JC, et al. Regional anesthesia in the patient receiving antithrombotic or thrombolytic therapy: American society of regional anesthesia and pain medicine evidence based guidelines (3rd edition). Reg Anesth Pain Med.2010;35(1):64-101.
• Marik PE, Corwin HL. Efficacy of blood transfusion in the critically ill: a systematic review of the literature. Crit Care Med. 2008;36(9):2667-2674.
• Roberts HR, Escobar MA. Less common congenital disorders of hemostasis. In: Kitchens CS, Alving BM, Kessler CM, eds. Consultative Hemostasis and Thrombosis. 2nd ed. Philadelphia, PA: Saunders Elsevier; 2007:61-79.
• Saidenberg S, Petraszko T, Semple E, Branch DR. Transfusion-related acute lung injury (TRALI): a Canadian blood services research and development symposium. Transfus Med Rev. 2010;24(4):305-324.
• Society of Thoracic Surgeons Blood Conservation Guideline Task Force, Ferraris VA, Ferraris SP, et al. Perioperative blood transfusion and blood conservation in cardiac surgery: The Society of Thoracic Surgeons and The Society of Cardiovascular Anesthesiologists clinical practice guidelines. Ann Thorac Surg. 2007;83:S27-S86.
• Vamvakas EC, Blajchman MA. Transfusion-related mortality: the ongoing risks of allogeneic blood transfusion and the available strategies for their prevention. Blood. 2009;113(15):3406-3417.