Chapter 32: Hematologic Dysfunction in the ICU

Patients admitted to the ICU have frequent hematologic dysfunction as a result of critical illness leading to multiorgan dysfunction and failure. A hematology consultation is often requested for evaluation of hematologic complications in these critically ill patients. The most common reasons for hematologic consultation include evaluation of cytopenias mainly thrombocytopenia, anemia, and less commonly, evaluation of leukocytosis and thrombocytosis. Coagulopathies including severe bleeding and thrombotic disorders are very prevalent in the ICU patients due to their underlying conditions. These patients frequently develop DIC or severe coagulopathy secondary to liver dysfunction or acquired vitamin K deficiency. Bleeding can also be seen as a consequence of renal insufficiency, and the use of antiplatelet agents causing an acquired thrombocytopathy. The administration of anticoagulants in these patients is challenging as the bleeding risk is increased. Hemostasis is frequently disrupted, as these patients often require invasive and/or surgical procedures. Thrombotic complications either venous, arterial, or microvascular are commonly seen as a result of indwelling catheter placement or other invasive procedures, prolonged immobilization, underlying malignancy, autoimmune disorder, or medication related. Frequent exposure to blood products, increases the risk of transfusion reactions including transfusion-related acute lung injury (TRALI) and hemolytic, febrile, and allergic transfusion reactions. In addition dilution coagulopathies are seen in patients who are massively transfused. Prompt interaction between the intensivist and the hematologist is key to optimize the care of these challenging patients.

Overview of Hemostasis

Hemostasis maintains a closed system of vascular integrity and prevents blood loss from injury.¹ The hemostatic response is initiated by injury to endothelial cell surfaces that leads to exposure of tissue factor (TF), collagen, von Willebrand factor (vWF), and fibronectin on the subendothelial matrix. Primary hemostasis consists of platelet adhesion by binding to collagen and vWF on the exposed endothelial surface. Platelets then aggregate via glycoprotein IIb-IIIa receptors, which bind fibrinogen and form platelet thrombi. When activated, platelets release vasoactive, inflammatory, and thrombogenic mediators. For example, ADP binds to the purinergic receptors P2Y2 and P2Y12 to promote aggregation. Thromboxane A2 is also synthesized in the platelets by cyclooxygenase (COX) enzymes and stimulates vasoconstriction and platelet aggregation.

Secondary hemostasis is the process of thrombin generation by coagulation proteins, which are classified into extrinsic, intrinsic, and common pathways (Figure 32–1). The initial step in secondary hemostasis occurs when TF binds to activated factor VII (extrinsic pathway). The intrinsic pathway is composed of factors XII, XI, IX, and VIII, also known as the contact activation pathway. Both pathways activate thrombin (factor IIa) from prothrombin (factor II) through the prothrombinase complex, which is composed of factors Xa, Va, calcium, and phospholipids. Thrombin generation leads to the conversion of fibrinogen to fibrin, which polymerizes and is cross-linked by factor XIIIa, creating a thrombus.

Thrombin activates fibrinogen, platelets, and acts as its own regulator by activating the natural anticoagulants protein C (PC), protein S (PS), and antithrombin (AT) that in turn inactivate coagulation at multiple steps and limit fibrin deposition (see Figure 32–1).

Vitamin K, an essential cofactor for the conversion of glutamic acid residues to gama-carboxyl glutamate (Gla) residues, allows factors II, VII, IX, X, PC, and PS to bind to the surface of plasma membranes and perform their functions. Vitamin K antagonists like warfarin cause multisite blockade of the coagulation cascade by impairing the creation of Gla domains. Fibrinolysis is the process of thrombi dissolution, a necessary step to prevent undesired excess fibrin deposition and pathologic thrombus formation.^2,3 Plasminogen is activated by tissue plasminogen activator (tPA) and urokinase plasminogen activator (uPA) in conjunction with annexin II, which cleave fibrin into fibrin degradation products including d-dimers.

Evaluation of Primary and Secondary Hemostasis

Platelet functional assays and aggregometry measure qualitative defects in platelet hemostasis. These tests include platelet functional assay 100 (PFA-100) and platelet aggregometry (Table 32–1). PFA-100 testing uses special collagen-epinephrine and collagen-ADP cartridges where in whole blood pass through a chamber and aggregate. If one or both of the cartridge closure times are prolonged, it suggests a qualitative platelet defect. Aspirin and nonsteroidal anti-inflammatory drugs (NSAIDs) cause prolongation of the collagen-epinephrine time. Both tests will be prolonged in von Willebrand disease (vWD), congenital or acquired thrombocytopathies seen in renal and hepatic disease. Platelet aggregation and secretion studies either using whole blood or platelet-rich plasma evaluates specific platelet defects. It should be noted that platelet function tests might be abnormal when the platelet count is less than 100,000/L unless the concentrations are adjusted in plasma.

The screening tests of blood coagulation are the prothrombin time (PT) and the activated partial thromboplastin time (aPTT). These tests become prolonged when coagulation factor activities are 20% or less. The PT is used to determine risks of bleeding from defects in the extrinsic and common pathways. The international normalized ratio (INR) is the PT normalized to a pool of known PTs in a population of patients on vitamin K antagonists and adjusted for thromboplastin reagent types. It should be noted that the INR is standardized for bleeding and thrombotic risks for patients on warfarin and does not accurately reflect coagulation dysfunction in liver disease or other coagulopathies.^4,5

The aPTT is a test of the intrinsic and common pathways of coagulation. The aPTT is also used to measure therapeutic levels of intravenous anticoagulants like heparin and direct thrombin inhibitors like argatroban and bivalirudin. When a prolonged PT or aPTT is detected, a variety of diagnoses are implicated and should be related to the patient's clinical presentation and history (Table 32–2). Further evaluation consist on a mixing PT and/or aPTT test that is performed by adding the patient's plasma to a pool of normal plasma in a 1:1 ratio and measuring the PT and/or aPTT immediately and after incubation of 60 to 120 minutes at 37°C. If the clotting time corrects into the normal range upon mixing, this is referred to as an immediate correction. Incubation is performed to detect slow-acting coagulation inhibitors, characteristic of an acquired factor VIII deficiency, which will cause prolongation of the aPTT after initial correction. A failure of the mixing study to correct immediately and after incubation is consistent with the presence of a lupus or lupus-like anticoagulant (LA). A sustained correction is consistent with a coagulation factor deficiency. Falsely positive LA is seen in the presence of heparin, direct thrombin, and Xa inhibitors and may be falsely positive when the patient is on warfarin and the INR is supratherapeutic. Therefore an LA test should not be requested when the patient is receiving these anticoagulants. Bethesda units (BU) are used to measure inhibitor titers to clotting factors and reflect the strength of inhibitors that are detected on mixing studies. One BU is equivalent to the reciprocal dilution of the patient's plasma at which 50% of the specific factor activity is inhibited. For example, in the case of acquired hemophilia A, a BU titer of 5 indicates a dilution of 1:5 where 50% of the factor VIIIa activity was inhibited.

Occasionally, a shortened PT or aPTT is observed in critical illness. A shortened PT is the result of an increase in circulating TF after central nervous system (CNS) injury, stroke, or sickle cell crisis. Similarly, inflammation-induced elevations of factor VIII may result in a shortened aPTT. More specific coagulation testing and their indications are described in Table 32–1.

Inflammation in the ICU

Inflammation occurs in sepsis, systemic inflammatory response syndrome, and other critical illnesses, and causes alterations in both hemostasis and fibrinolysis. Platelets are activated and may become prothrombotic in inflammatory states, white cells are more adhesive to vessel walls by expressing vessel adhesion molecules, and toxic degranulation occurs. Oxidative damage as the result of free radical generation causes decreased red cell membrane flexibility and expression of adhesion molecules on the surface of endothelial cells. Hypoxia is also a prothrombotic trigger. Inflammation may worsen anemia as red blood cells are subjected to cytokines and turbulent blood flow, ultimately leading to hemolysis. Inflammatory cytokines also increase expression of TF on the surface of monocytes and in circulating microparticles. Coagulation factors like fibrinogen, factor VIII, and vWF are acute-phase reactants and their levels may increase secondary to inflammation.

Targeting Inflammation

Initial clinical trials to treat sepsis with drotrecogin-alfa (Xigris), a recombinant activated protein C suggested improvement in outcomes in some critically ill patients. However, reevaluation of this agent in the PROWESS-SHOCK trial showed no benefit of drotrecogin-alfa in patients with septic shock and the drug was withdrawn from the market.^6,7

Disseminated Intravascular Coagulation

DIC is a consumptive coagulopathy that is frequently encountered in critically ill patients. DIC is observed in approximately 50% of patients with sepsis, and is an independent predictor of morbidity and mortality. Other common etiologies of DIC are shown in Table 32–3. DIC occurs as the result of increased circulating procoagulant factors that lead to high levels of thrombin generation and cytokine activation. Systemic activation of thrombin and platelets causes thrombosis in both small and large vessels. Platelets become activated and then aggregate in response to increased thrombin generation causing progressive thrombocytopenia. As the regulatory proteins of thrombin generation and coagulation are progressively overwhelmed, the systemic circulation becomes more thrombogenic. Coagulation factors are consumed in diffuse thromboses, leading to organ failure, ischemia, and tissue damage. Deep venous, arterial, and cerebrovascular thromboses also occur. Fibrinogen levels decrease as DIC continues, resulting in a fibrinogen deficiency and increased bleeding tendency.⁸ In severe cases, gangrene and limb ischemia (purpura fulminans) develops. No single test or clinical finding is able to accurately diagnose DIC. The diagnosis must be based on underlying clinical predisposition. A combination of prolonged PT, prolonged aPTT, and thrombocytopenia are suggestive of DIC, as is a decreased level of fibrinogen.⁹ Approximately 50% of patients with DIC will have prolongation of the PT or aPTT at some point during the course of DIC. The aPTT may initially be shortened as a result of inflammatory increases in FVIII and fibrinogen. Elevated fibrin degradation product (FDP) and d-dimer levels indicate increased thrombin and plasmin generation. However, d-dimer is also elevated in the postsurgical state, after trauma, in the setting of deep vein thrombosis and with liver and renal dysfunction. The platelet count is of particular utility in DIC as thrombocytopenia correlates with thrombin generation.^9,10 Thrombocytopenia may develop rapidly (within 1-4 hours) of the onset of DIC and is an independent predictor of mortality and length of ICU stay.¹⁰ Once the platelet count reaches a nadir, it suggests stabilization of thrombin generation. A platelet count less than 100,000/L is seen in 50% to 60% of patients with DIC, and counts less than 50,000/L are seen in approximately 10% of patients. Several scoring systems for DIC have been developed, and are dependent on clinical presentation and laboratory values. The International Society for Thrombosis and Haemostasis (ISTH) score for overt DIC is shown in Table 32–4, and has a specificity of 91% and a sensitivity of 97%.¹¹ This score is meant to be calculated daily based on serial laboratory measurements and guide clinicians to the improvement or worsening of their patients. The scoring system should be used only when the patient's clinical history is compatible with DIC.

The cornerstone of treating DIC is correction of the underlying cause. Platelets and plasma transfusions should be used for bleeding patients and not solely to correct laboratory values. In the setting of bleeding, plasma should be administered at 15 to 20 mL/kg, and platelets should be transfused to over 50,000/L. Hypofibrinogenemia (fibrinogen concentration 100 mg/L) should be corrected by 10 donor pools (2 units) of cryoprecipitate or 3 g of a fibrinogen concentrate. Antifibrinolytics (tranexamic acid, ε-aminocaproic acid), activated prothrombin complex concentrates (FEIBA) and recombinant factor VIIa (rFVIIa) should be avoided, as they can worsen thrombosis. Anticoagulation should not be withheld in patients with DIC unless the bleeding risk is significant. DIC patients without bleeding should have routine pharmacologic thromboprophylaxis.

Bleeding

Bleeding disorders occur in critically ill patients and their etiology is usually multifactorial. Bleeding is seen as a consequence of thrombocytopenia and/or thrombocytopathies or is related to acquired coagulation factor deficiencies. Thrombocytopenia is commonly drug related, or caused by severe liver disease, or DIC. Thrombocytopathies are secondary to the use of antiplatelet agents such as aspirin, NSAID, and PY212 inhibitors, also seen with liver and renal dysfunction and in patients with congenital thrombocytopathies or vWD. The bleeding pattern may point to a specific etiology. For example, patients with severe thrombocytopenia, thrombocytopathies, and vWF defects present with mucocutaneous bleeding. Conversely patients with severe inherited coagulation factor defects (eg, factors VIII and IX) present with hemarthrosis, muscle, and soft tissue bleeding. Interestingly, the bleeding pattern in patients with acquired factor VIII inhibitors is also mucocutaneous. Some patients may have no known personal or family history of bleeding disorders, but hemostatic challenges that occur in the hospital setting (biopsies, surgeries, placement of intravascular devices) may unmask mild congenital bleeding disorders such as thrombocytopathies, factor XI deficiency, and fibrinolytic defects.

A diagnostic approach to the bleeding patient should include thorough examination of patient's skin, flank, hips, and mucosal surfaces for petechiae, ecchymoses, or hematomas. Inspection of the digits may reveal evidence of ischemia or embolism that reflects a systemic disorder.

In addition careful inspection of chest and endotracheal tubes, urinary catheters, surgical drains, and suction devices should be carefully performed. A high suspicion for local wound complications should be maintained, especially if the patient is within 48 hours of an invasive or surgical procedure. Hemoccult testing is useful in assessment of gastrointestinal bleeding.

The Initial laboratory evaluation should include a complete blood count, PT, aPTT, and fibrinogen. A manual platelet count should be performed by manual examination of the peripheral blood smear if thrombocytopenia is reported in a patient with no signs of bleeding. Pseudothrombocytopenia is a laboratory artifact characterized by platelet aggregation in response to ethylenediaminetetraacetic acid (EDTA), heparin, or citrate, and should be excluded early in the diagnostic evaluation. If large clumps of platelets are observed on the peripheral smear, the platelet count should be repeated using a different anticoagulant tube or by peripheral finger stick to obtain a more accurate platelet count.

Acute management of a bleeding includes discontinuation of all medications that contribute to bleeding, evaluation of a bleeding site, and reversal of hemostatic abnormalities. In addition to local cauterization or compression when a bleeding site is identified, platelets are used for patients with severe thrombocytopenia or thrombocytopathy. Vitamin K is administered for correction of acquired vitamin K deficiency either due to vitamin K antagonists or other acquired etiologies of vitamin K deficiency. Vitamin K is administered at 10 mg intravenously over 15 to 30 minutes and takes effect within 6 to 12 hours. Plasma is used to correct coagulopathy secondary to liver disease and for rapid reversal or vitamin K antagonists. Plasma is infused at doses of 10 to 15 mL/kg given every 6 hours until hemostasis is achieved. Plasma contains approximately 0.7 to 1 unit/mL of clotting factor activity and 1 to 2 mg/unit of fibrinogen. Cryoprecipitate is a protein fraction of plasma, enriched in factor VIII, vWF, fibrinogen, fibronectin, and factor XIII. One unit of cryoprecipitate is obtained from 1 unit of plasma. Cryoprecipitate is indicated to correct hypofibrinogenemia in patients with DIC presenting with bleeding and it is also used in managing dysfibrinogenemia or hypofibrinogenemia in patients with severe liver disease. Cryoprecipitate is administered at 1 unit per 5 kg of body weight, with the goal of increasing fibrinogen to greater than 100 mg/dL.

If excessive fibrinolysis is suspected, antifibrinolytic lysine analogues should be administered. These drugs act as competitive analogues of fibrin and bind plasminogen. Tranexamic acid is given orally at 25 mg/kg or 10 mg/kg IV every 8 hours, ε-aminocaproic acid is given at 50 mg/kg every 6 hours orally or IV 5 g bolus followed by 1 g/h continuous infusion. Tranexamic acid must be dose reduced in patients with renal failure. Both of these agents should be avoided in DIC and in genitourinary bleeding, where the risk for thrombosis is significant. Tranexamic acid has been shown to reduce bleeding and death related to bleeding if administered within 3 hours.

Transfusions of red blood cells should be restricted to actively bleeding patients and the decision to transfuse blood products must be carefully weighed against the risk of transfusion-related fever, circulatory overload, lung injury, and alloimmunization.¹²

Specific Bleeding Diagnoses

Acquired Hemophilia A

Acquired hemophilia A is a rare disorder with an incidence of 1 in 1 million, where patients develop autoantibodies to endogenous factor VIII. This disease is characterized by ecchymoses, soft tissue hematomas, and petechiae, but CNS bleeding can develop. Predisposing factors are poorly understood. In some cases, the identified trigger is pregnancy, new-onset malignancy, and autoimmune disorders and is more commonly seen in the elderly. In nearly 50% of cases, it appears to be idiopathic. Acquired hemophilia A is diagnosed by a prolonged aPTT with initial partial correction by a mixing study, which prolongs after incubation. Treatment should be aimed at controlling bleeding and eradicating the inhibitor. Bleeding is managed with activated prothrombinase complex concentrates (FEIBA at 60-100 IU/kg every 12-24 hours) or rFVIIa (90-100 μg/kg given every 2-3 hours). Simultaneously, treatment to eradicate the inhibitor is initiated with either high-dose prednisone (1 mg/kg) and cyclophosphamide (2 mg/kg daily) or rituximab (375 mg/m²) administered as 4 weekly doses. The inhibitor is eradicated in approximately half of all patients but may take several months, and careful observation for bleeding and complications of immunosuppression is warranted during this time.¹³

Congenital Bleeding Disorders

Congenital bleeding disorders include hemophilia A and B, vWD, and other factor deficiencies. In the presence of a known coagulopathy, factor-specific assays should be sent to determine the baseline activity. Severe hemophilia patients should be maintained on a factor replacement prophylaxis strategy, consisting of factor VIII products administered at doses of 25 to 35 IU/kg every other day (hemophilia A), or factor IX doses of 40 to 60 IU/kg twice a week (hemophilia B). One specific concern with hemophilia B replacement products is the development of anaphylaxis in 5% of patients. Prior to invasive procedures or surgeries, factor VIII or IX activity should be elevated to approximately 100% by the administration of factor-specific products. In mild hemophilia A and vWD, desmopressin acetate (DDAVP) may be administered by IV (0.3 μg/kg) or intranasally 150 μg (1 puff in each nostril) to increase vWF and FVIII activity by 2 to 5 times the baseline within 15 to 30 minutes. Severe vWD patients may require vWF-containing factor replacement (Humate P, Wilate, Alphanate). Hemophilia patients with inhibitors are treated with FEIBA or rFVIIa. New longer acting hemophilia concentrates are now available that can prolong the detectable activity of factors VIII and FIX. (ref 36)

Liver Disease and Coagulation

The liver is the source of synthesis of most coagulation factors. Clues to hepatic dysfunction as the cause of a coagulopathy are a decreased albumin, prolonged PT, or known cirrhosis. Liver coagulopathy may present as both bleeding and clotting disorders. In addition to having reduced procoagulant factors, liver dysfunction results in decreased synthesis of natural anticoagulants namely protein C, S, and antithrombin and impaired fibrinolysis, which can lead to a prothrombotic state. A patient with bleeding should receive plasma and cryoprecipitate, and a patient with thrombosis should be anticoagulated, even in the setting of a prolonged PT.⁵ Aside from liver transplantation, there is no cure for hepatic coagulopathies.

Vitamin K Deficiency

Vitamin K is a fat-soluble vitamin that exists in 2 important forms: vitamin K₁, found in green leafy vegetables, and vitamin K₂, is synthesized by normal gut flora. Antibiotic exposure alters the natural gut flora, decreasing the reserve of bacteria-producing vitamin K. Deficiencies of vitamin K occur in the hospital when the diet is not well balanced or the patient is on broad-spectrum antibiotics for over a week.¹⁴ Other causes are poor oral intake, impaired absorption of fat-soluble vitamins (occurring in pancreatitis and cholestyramine treatment), and the use of nasogastric suction. Vitamin K deficiency leads to gastrointestinal, and postsurgical hemorrhage. Excessive bleeding from needle punctures, and intramuscular hematomas can also be noted. Both the PT and aPTT are prolonged as in DIC, but in vitamin K deficiency plasma activity levels of factors II, VII, IX, and X are less than 50%, while fibrinogen, factor V, and VIII are normal. d-Dimer is also normal in vitamin K deficiency. A 1:1 mixing study should completely correct in this setting. If a vitamin K deficiency is suspected, repletion with 5 to 10 mg of vitamin K should begin by intravenous administration. Correction of the PT should occur 6 to 12 hours after vitamin K administration. Critically ill patients at risk for coagulopathies from vitamin K deficiency should be supplemented by administering 5 mg of vitamin K 2 or 3 times weekly, either orally or intravenously.¹⁴

Anticoagulant Overdose

Overdosing of anticoagulants can result in serious bleeding complications in critically ill patients. Intravenous and subcutaneous UFHs are preferable to oral anticoagulation in the ICU because of their short half-life and reversibility (Table 32–5). If a bleed occurs, heparin must be stopped immediately. Heparin is reversed by administering protamine sulfate 1 mg per 100 units of heparin every 2 to 8 hours, to a maximum dose of 50 mg in 24 hours. If there is an ongoing need for anticoagulation, the heparin should be restarted without bolus at a reduced dose once the bleeding has been controlled. Protamine is less effective at reversing LMWHs and is not effective in reversing fondaparinux.

Warfarin-induced bleeding can be mild or severe. However, supratherapeutic INRs may not need aggressive reversal in the absence of bleeding. Patients with a supratherapeutic INR without bleeding should have their warfarin held and their INR monitored daily until it reaches the therapeutic range. Vitamin K reversal should be given when the INR is 10 or greater in the absence of bleeding. Any bleeding that occurs while on warfarin should be reversed with vitamin K, regardless of the INR. An intracranial bleed should be treated immediately with vitamin K (10 mg IV), plasma (25-35 mL/kg), and 4 factors prothrombin complex concentrate (PCCs) like Kcentra. Kcentra contains factors II, VII, IX, X, PC, and PS, which is, concentrated 25-fold, equivalent to 2 L of plasma. As Kcentra contains heparin, it is contraindicated in patients with HIT. Kcentra is administered over 10 to 15 minutes at doses of 25 to 50 IU/kg, and normalizes the INR within 30 to 60 minutes of administration.¹⁵

The Food and Drug Administration (FDA) approved idarucizumab (Praxbind), a monoclonal Fab fragment for reversal of dabigatran, in 2015 based on the results of the RE-VERSE AD trial (Pollack N Engl J Med 2015, PMID: 26095746). This drug is administered by IV injection at a single dose of 5g. Repeat doses can be considered in consultation with a hematologist and pharmacist. Development of reversal agents for anti-Xa inhibitors are currently in development, including the recombinant protein factor Xa mimic andexanet (Segal DM N Engl J Med 2015; PMID 26559317). In the event of bleeding, besides discontinuation of the offending agent, evaluation for the source of bleeding and supportive therapy including transfusion of red blood cells, activated charcoal (25 g) can be administered orally if ingestion of dabigatran was within 3 hours and within 2 to 3 hours after ingestion of rivaroxaban and apixaban, particularly if intentional overdose. In addition, approximately 60% of dabigatran can be removed by hemodialysis over 2 to 3 hours. Activated prothrombin complex concentrates (FEIBA) 25 units/kg can be considered for dabigatran-related life-threatening bleeds, and rFVIIa has been used as well to control bleeding.¹⁶ Kcentra at a dose of 50 units/kg can be considered for life-threatening bleeds related to rivaroxaban and apixaban. Plasma does not reverse dabigatran, rivaroxaban, or apixaban bleeds, thus is not recommended for reversal. Platelet transfusions are used if the patient was using concomitant antiplatelet agents.

Acquired von Willebrand Disease

Acquired vWD generally manifests as mucocutaneous bleeding. Different mechanisms are responsible for the loss of vWF multimers in acquired vWD, and this disease should be suspected in patients with myeloproliferative neoplasias, plasma cell dyscrasias, hypothyroidism, lymphoproliferative disorders, and solid tumors. Acquired vWD also result from flow obstructions and increased shear that occurs in hypertrophic cardiomyopathy, aortic valve stenosis, and ventricular assist devices. Valve-induced acquired vWD can present with gastrointestinal bleeding, also known as Heyde syndrome. Decreased vWF antigen and loss of high-molecular-weight vWF multimers are suggestive of acquired vWD. Management of bleeding involves using DDAVP or vWF concentrates, but acquired vWD will only resolve after correction of the underlying disorder.¹⁷ Caution must be exercised with DDAVP administration, as patients may develop hypotension, flushing, and fluid retention. Fluid should be restricted to maintenance rate for 24 hours, and electrolytes should to be monitored every 6 to 8 hours. DDAVP should not be administered for more than 3 days because of the development of tachyphylaxis.

Hemodilution

Critically ill patients frequently are exposed to very high volumes of crystalloid and colloid for volume resuscitation and administration of blood products. The resultant volume changes may affect blood count measurements and cause a dilution in hemoglobin, platelets, and coagulation factors. Prior to evaluation for coagulopathy, these tests should be repeated and drawn peripherally, or off lines that are not receiving fluids or medications. Prior studies suggest the amount of crystalloid fluid needed to cause true hemodilution is significant, usually from several liters of normal saline given as bolus dosing.^18,19,20 Maintenance fluids alone are insufficient to cause dilution. A recent and more clinical relevant definition of massive bleeding patient proposed by Savage et al refers to patients transfused with greater than 3 units of RBCs in any 60-minute period (within 24 hours of admission). This new definition includes a clinically relevant rate of transfusion and includes the majority of patients who rapidly exsanguinate.²¹ The Trauma Outcomes Group collected data from 466 massive transfusion patients from 16 level 1 centers from 2005 to 2006 across the United States. This study demonstrated that outcomes were improved with a more balanced ratio of at least 1:1:2 of plasma:platelets:RBC. Subsequently, the Prospective Observational Multicenter Major Trauma Transfusion (PROMMTT) study was performed at 10 level 1 trauma centers where in-house 24/7 research assistants recorded the sequence and timing of all infused fluids in bleeding trauma patients in 2009 to 2010. A balanced use of plasma early in resuscitation was associated with improved early survival. The median time to hemorrhagic death was 2.6 hours, whereas platelets were infused at a median time of 2.7 hours and 30% of patients who died from hemorrhage never received any platelets. An ongoing prospective randomized clinical trial at 12 centers in North America Pragmatic Randomized Optimal Platelet and Plasma Ratios (PROPPR) trial (www.clinicaltrials.gov identifier: NCT01545232) compares a 1:1:1 ratio of plasma:platelets:RBCs with a 1:1:2 ratio in patients predicted to receive a massive transfusion.²² The colloid solution hydroxyethyl starch has also been associated with both a dilutional coagulopathy and an acquired vWD.²³

Using Recombinant Factor VIIa in Critical Illness

The use of rFVIIa has been used in ICU patients to control bleeding. This short-acting activated coagulation factor binds directly to activated platelets and improves systemic hemostasis. It is currently approved for use in hemophilia patients with inhibitors and in patients with congenital factor VII deficiency. The advantages of rFVIIa are rapid onset of hemostasis and minimal volume. However, this drug must be used with caution in patients who do not have congenital coagulopathies, especially in patients over the age of 65, where the risk of arterial thrombotic events was 5.5%.²⁴

Acquired Platelet Dysfunction

Acquired platelet dysfunction may occur as a result of medication use (nonsteroidal anti-inflammatory agents, aspirin, clopidogrel, prasugrel, ticagrelor) or uremia. Uremic platelet dysfunction is multifactorial, resulting from decreased aggregation, displacement from the endothelium, and impaired secretion of granules. Dialysis is used to correct platelet dysfunction. Although desmopressin can shorten the skin-bleeding time in patients with uremia, the use of recombinant erythropoietin stimulating agents has made this abnormality of hemostasis much less frequent than it was previously. The beneficial effect of erythropoietin on hemostasis is based on the increase in red-cell mass, which affects the blood-fluid dynamics, leading to a more intense interaction between circulating platelets and the vessel wall.²⁵ Conjugated estrogens may also be used to correct uremic dysfunction, given at 0.6 mg/kg IV over 30 to 40 minutes for 5 days.

Multiple antiplatelet agents are used in cardiac patients, including aspirin, ibuprofen, ticlopidine, clopidogrel, prasugrel, ticagrelor, dipyridamole, abciximab, tirofiban, eptifibatide, and vorapaxar (Table 32–6). All of these medications create a qualitative platelet dysfunction and can cause bleeding. If bleeding develops, they should be held while the patient receives platelet transfusions to restore the pool of functional platelets.

Thrombocytopenias

Platelets are synthesized from megakaryocytes in the bone marrow by stimulation of the thrombopoietin receptor by thrombopoietin and other cytokines. The megakaryocytes then release platelets into circulation, which are ultimately cleared in the spleen and liver. Splenomegaly or platelet-specific antibody binding may result in increased clearance of platelets and result in thrombocytopenia, defined as a platelet count less than 150,000/L. Thrombocytopenia occurs in 15% to 58% of ICU patients. In medical ICU patients and in septic patients, platelet counts may decrease within 3 to 5 days of admission, dropping to levels 40% to 90% below baseline. Persistent thrombocytopenia at day 14 predicts increased mortality, irrespective of an identified etiology.²⁶ In the absence of thrombotic disease, platelet transfusions may be used to prevent or correct bleeding.²⁷

When the platelet count falls below 50,000/L, patients are at an increased risk of bleeding from procedures. Furthermore, patients may bleed spontaneously when the platelet count drops below 10,000/L. The differential is broad for thrombocytopenia in critical illness, as numerous medications, infections, and concomitant illnesses can be implicated. However, decreases in the platelet count can also be manifestations of severe thrombotic diseases. A diagnostic strategy must take the clinical scenario into consideration. Platelet counts under 100,000/L are associated with increased 30-day mortality and disturbed immune responses to sepsis in critically ill patients.

Initially, the most important diseases to consider and diagnose are the thrombotic microangiopathies (TTP and atypical hemolytic uremic syndrome), HIT, catastrophic antiphospholipid syndrome, and immune thrombocytopenic purpura (ITP). These disorders can be fatal if not immediately diagnosed and treated. After these disorders have been evaluated and ruled out, a more detailed review of causes of thrombocytopenia can be accomplished. Disseminated intravascular coagulation should always be considered as the cause of thrombocytopenia in a patient with sepsis, prolonged PT, elevated d-dimer, and low fibrinogen.

Thrombotic Microangiopathies

Thrombotic microangiopathies (TMA) are characterized by microvascular injury to endothelial cells, which can be provoked by infection, inflammation, drugs, and other factors. These injuries promote endothelial cell injury, platelet aggregation, and hemolysis. Inhibition of the vWF-cleaving protease ADAMTS13 is implicated in the pathology of some types of TMAs. The accumulation of ultra-large vWF multimers causes microvascular thrombi that result in multiple organ failure and dialysis dependence. Rapid recognition and treatment of TMAs is essential, as the mortality can approach 90%. Two important TMAs that must be distinguished acutely are atypical hemolytic uremic syndrome (aHUS) and TTP. Differentiating TTP from aHUS is difficult since the 2 diseases have overlapping manifestations. Higher platelet counts and worse degrees of renal failure are associated with aHUS, whereas platelet counts will almost invariably be low (< 40,000) in TTP and only mild renal failure is seen.²⁴

The initial evaluation of TMAs should include review of a peripheral smear to confirm decreased platelet count and to identify schistocytes. Markers of hemolysis should be requested including lactate dehydrogenase (LDH), haptoglobin, and a direct antiglobulin test. Stool antigens to evaluate for shiga toxins, like O157:H7 and ADAMTS13 activity should be sent prior to treatments. Undetectable ADAMTS13 activity may indicate either a congenital absence of ADAMTS13 caused by mutation of the ADAMTS13 gene (Upshaw-Schulman syndrome), or the presence of an acquired anti-ADAMTS13 IgG inhibitor. Liver disease, vasculopathies, and peripheral artery disease may also have an ADAMTS13 deficiency, and caution should be used in interpreting these results of such patients.²⁸

Plasma exchange restores ADAMTS13 activity in the majority of TTP patients.²⁹ If plasma exchange is not available, simple plasma transfusions should be initiated at doses of 30 mL/kg. If a response to plasma exchange is not seen within approximately 20 exchange sessions or 45 L of plasma, an alternate diagnosis should be considered, including aHUS. Treatment of refractory TTP is less effective and no current standard exists. Second-line therapies include immunosuppression with rituximab, vincristine, mycophenolate, cyclosporine, bortezomib, and splenectomy. Rituximab (375 mg/m²) for 4 weekly doses has been used successfully to induce remission in patients with persistently low ADAMTS13 activity despite plasma exchange.³⁰

aHUS is a TMA that presents like TTP, but ADAMTS13 activity may be normal. This syndrome is caused by complement protein mutations resulting in hyperactivation of the complement system. The anticomplement C5a antibody eculizumab is approved to treat aHUS with an initial dose of 900 mg IV weekly for 4 weeks, followed by maintenance dosing of 1200 mg every 2 weeks. Prior to eculizumab treatment, it is important to ensure the patient is vaccinated against Neisseria meningitidis or covered with antibacterial prophylaxis for the duration of treatment. Clinical and hematologic improvement is usually seen within 1 to 2 weeks.²⁸

Heparin-induced Thrombocytopenia

HIT is a diagnosis not to be missed in a hospitalized patient with new-onset thrombocytopenia. The formation of HIT antibodies occurs when heparin binds to platelet factor 4 and causes a rapid prothrombotic antibody reaction by activating platelet FcyIIaR receptors and increasing thrombin generation. The severity of thrombotic disease in HIT cannot be underestimated. HIT occurs in 1% to 4% of patients on UFH, and fewer than 1% of patients on LMWH, and is more common in postsurgical patients than medical inpatients and in females compared to males.³¹ HIT is a clinicopathologic diagnosis. The first step in evaluation is calculating the 4T score by using the clinical prediction rule (Table 32–7). A low score effectively excludes HIT, and no further workup is needed. If the score is intermediate or high, heparin products, including heparin flushes, and catheters coated with heparin should be discontinued immediately and the patient should be initiated on an alternate nonheparin anticoagulant. Both the heparin antibody ELISA assays and the serotonin release assay (SRA) test should be sent at presentation. If the antibody test is negative but the clinical prediction score remains high, the patient should remain on heparin-free anticoagulant until the SRA is reported. If the SRA is negative, this effectively rules out HIT and heparin can be resumed. If both tests are positive, the patient should remain on an alternate anticoagulant until the platelet count recovers. The patient may be transitioned to warfarin when the platelet count is over 150,000/L. Argatroban, dosed intravenously at 2 ug/kg/min to an aPTT goal of 1.5 to 3 times baseline, is the only anticoagulant currently approved for treatment of HIT in the USA. Dosing needs to be reduced to 25% (0.5 ug/kg/min) of the indicated dose in case of hepatic insufficiency. In a retrospective analysis of 12 ICU patients with multiple organ dysfunction syndrome (MODS) treated with argatroban for suspected or diagnosed HIT, the mean argatroban dose was significantly lower in patients with hepatic insufficiency compared with patients without hepatic impairment (0.10 ± 0.06 μg/kg/min versus 0.31 ± 0.14 μg/kg/min).³²

Immune Thrombocytopenic Purpura

ITP is characterized by isolated thrombocytopenia often occurring in the absence of identifiable and specific triggers. ITP is caused by autoantibody-mediated clearance of platelets. The thrombocytopenia is usually severe, with platelet counts often less than 10,000/L. Physical examination may reveal petechiae on the soft palate and the extremities. The peripheral smear may reveal large platelets, which result from the egress of immature platelets from the bone marrow to compensate for platelet destruction. The most feared clinical complication of ITP is intracranial hemorrhage. Secondary causes of ITP, including HIV, hepatitis C, and autoimmune diseases like systemic lupus erythematosus (SLE) should be identified and treated. The first line of treatment in severe ITP is glucocorticoids (prednisone1 mg/kg daily), intravenous immunoglobulins (IVIg) 1 g/kg given over 1 to 2 days or anti-D for Rh-positive, nonsplenectomized individuals. Treatment options of patients who are unresponsive to or relapse after initial corticosteroid therapies include, rituximab, and the thrombopoietin receptor mimetics romiplostim and eltrombopag. Splenectomy remains a mainstay of therapy in providing sustained remission rates in patients with ITP, although long-term remissions have been attained with Rituximab and other immunosuppressive agents. However, infection and thrombosis remain long-term risks.³³

Medication-related Thrombocytopenia

Drug-induced thrombocytopenia occurs by a variety of mechanisms including direct bone marrow damage and immune-mediated destruction.³⁴ The bleeding rate is approximately 9% secondary to drug-induced thrombocytopenia and is related to the degree of thrombocytopenia. Nonimmune causes are the result of drug toxins on the bone marrow in patients receiving antineoplastic drugs, antivirals, ethanol, and thiazide diuretics that develop slowly over several weeks of exposure. Immune-mediated thrombocytopenia develops within 14 days, and within 1 to 3 days if the patient had been previously exposed to this drug. The causes of drug-induced immune-mediated thrombocytopenia are multiple and are related to the interaction of the drug, the antibody, and the platelet. Penicillin and cephalosporin drugs act as haptens to initiate an immune response. Quinine and related drugs act as binding agents for platelets and antibodies. Tirofiban, eptifibatide, and abciximab bind directly to GP IIb-IIIa. Gold and procainamide induce the production of platelet-specific autoantibodies. The diagnosis of drug-induced thrombocytopenia should start with the timing of thrombocytopenia and duration of exposure to the medication, with careful notation of the platelet trends. This trend will indicate the trajectory of thrombocytopenia. An excellent resource for reports of drug-induced thrombocytopenia and a grading system is available at www.ouhsc.edu/platelets.

It is prudent to suspect a drug-induced thrombocytopenia after exposures of approximately 9 to 10 days (exceptions noted below). Platelet counts may be severe in some cases, resembling ITP. Improvement is seen after 10 to 14 days of discontinuing the offending medication, but can take up to several months. In severe thrombocytopenia with bleeding, stress doses of prednisone (1 mg/kg) and platelet transfusions should be used in conjunction with stopping the offending drug. Bleeding will resolve in 1 to 2 days after discontinuation of the drug, even though improvements in platelet counts may take longer. In the case that a mild thrombocytopenia develops without bleeding, careful risk-benefit assessment should be made as to the indication for the drug compared to the risk for bleeding.

Posttransfusion Purpura

Posttransfusion purpura occurs as a thrombocytopenia that develops 7 to 14 days after blood transfusions, often in postpartum women. These women had been preimmunized during pregnancy against a platelet antigen (HPA-1a). When they receive blood products this increases the production of alloantibodies that binds to and clears their platelets, promoting a platelet transfusion-refractory state. This disorder presents with patients having purpura or petechiae, platelet counts are often less than 10,000/L. Platelet clearance should be measured to determine appropriate response to platelet transfusions as follows:

Corrected Count Increment (CCI): (platelet count posttransfusion – platelet count pretransfusion) × (body surface area)/(number of platelet units transfused × 10¹¹). If the CCI is normal after 1 hour but decreased at 24 hours this suggests a nonimmune mechanism.

If CCIs at 1 hour and 24 hours are less than 7500, the patient is considered platelet refractory and human leukocyte antigen (HLA)-matched platelet transfusions should be considered. Acute bleeding from posttransfusion purpura is treated similarly to ITP, and first-line therapy should be IVIg 1 g/kg for 2 days.

Postsurgery Thrombocytopenia

Platelet counts decline to a nadir within 1 to 4 days after cardiac, vascular, abdominal, or orthopedic surgery as a result of tissue trauma and blood loss, which lead to platelet consumption. Platelet counts should begin to increase after day 4 and reach presurgery levels usually between 5 and 7 days postoperation and may continue to increase for 10 days as a result of reactive thrombocytosis before return to baseline by day 14.

Thrombocytosis

Elevated platelet counts (> 450,000/L) may be seen in the ICU patient. Thrombocytosis should be divided into primary and secondary causes. Elevated platelet counts may occur as a result of myeloproliferative neoplasms including essential thrombocythemia (ET) and polycythemia vera (PV). JAK2 mutations in V617F, exon 12 and exon 13 occur in greater than 95% of PV and 50% of ET patients. Myeloproliferative neoplasms are thrombotic risk factors, especially in patients older than 60 years with prior thrombosis and underlying cardiovascular risk factors. Significant elevation of platelet counts more than 1.5 million may also present with a bleeding diathesis secondary to acquired thrombocytopathy or acquired vWD. Symptomatic patients with elevated platelet counts may benefit from treatment with plateletpheresis and cytoreduction with hydroxyurea, interferon, or anagrelide. Low-dose aspirin 81 mg may be used as primary prevention of thrombosis in some patients without another indication for anticoagulation.³⁵ Secondary thrombocytosis may result from prior splenectomy, postsurgery, and as a result of inflammation. Secondary thrombocytosis does not constitute a significant thrombotic risk.