Chapter 69. Bleeding Disorders

• Assessment of the patient with a bleeding diathesis requires careful history-taking and bedside evaluation and comprehensive baseline coagulation testing.
• Uncomplicated vascular injury related to trauma or surgery should be eliminated as the primary cause of bleeding before the possibility of a coagulopathy is invoked.
• Consideration of coagulation abnormalities should include vascular disorders, platelet problems, impairment of the fibrin generation cascade, and excessive fibrinolytic activity.
• Disseminated intravascular coagulation may present with either thrombotic or bleeding problems. Awareness of the possibility of this diagnosis and prompt, appropriate laboratory confirmation may improve treatment selection and outcome.
• Invasive procedures in critically ill patients with coagulopathies need to be performed with special caution to prevent complications. Some guidelines are presented.

Accurate diagnosis of the mechanism of a bleeding disorder is based on careful history-taking and bedside evaluation of the patient, coupled with appropriate confirmatory laboratory testing.1 Information obtainable from the patient or others about a pre-existing congenital or acquired bleeding diathesis, medications being taken, recent surgery or pregnancy, and underlying disturbed organ function (e.g., hepatic or renal insufficiency) will facilitate the diagnostic process and provide the basis for directing empirical urgent therapy should it be necessary before laboratory testing can be completed.

Careful physical examination may clarify whether observed bleeding is strictly a local problem, i.e., limited to one site and not out of proportion in severity to that expected from an observed injury or fixed lesion, or is due to a systemic bleeding diathesis, which should be suspected if multiple bleeding or bruising sites are present and cannot be accounted for by the extent or severity of known trauma. In some patients with a single mechanism of systemic coagulopathy, it may be possible to distinguish a vascular or platelet disorder, which is characterized by immediate bleeding (prolonged, continued bleeding after onset and dominant petechial/mucosal manifestations), from a problem of fibrin generation (a cascade disorder) or of fibrinolysis, which is typified by delayed bleeding (rebleeding after initial hemostasis and prominent ecchymoses or deep muscle and joint hemorrhage). Complex coagulopathies, such as disseminated intravascular coagulopathy (DIC), may have features of both categories. The presence of findings such as splenomegaly or telangiectases also may be helpful clues.

Several generalizations based on data from the history and physical examination may help in directing the workup and treatment of bleeding problems. The first consideration in the evaluation of a bleeding patient—before a coagulopathy is invoked as a contributing cause—is to exclude uncomplicated vascular injury (related to trauma or surgery) that would be amenable to surgical hemostasis techniques. Second, brisk bleeding is almost never due to spontaneous hemorrhage caused by a coagulopathy alone. Third, clinically significant bleeding often is the result of the coexistence of trauma or a potential bleeding lesion (e.g., a peptic ulcer) with a coagulopathy that may antedate the lesion or may develop while the lesion is present. Fourth, initial and follow-up laboratory screening of critically ill patients for the presence of a potential bleeding diathesis is important in predicting and preventing bleeding complications during interventions such as line placements and surgical procedures.

Systematic laboratory evaluation of the patient with a coagulopathy2 is facilitated by consideration of the factors necessary for normal clot formation and degradation (Table 69-1). An appropriate screening laboratory survey for a hemostatic disorder in the critically ill patient would include a complete blood count including platelet count; review of a peripheral smear for the presence of schistocytes and for the evaluation of platelet number, clumping, size, and appearance; rapid screening of in vitro platelet function3 and measurement of bleeding time (unless the platelet count is less than 50,000/μL); prothrombin time (PT), activated partial thromboplastin time (PTT), thrombin time (TT), and functional fibrinogen level; and the level of fibrin degradation products (FDPs) as well as D-dimer assay. These tests should be done simultaneously rather than sequentially in the critically ill patient.

The initial response to vessel injury with formation of the platelet plug depends on normal vascular and platelet function. Defects in these components are reflected in a prolonged bleeding time. Platelet dysfunction (e.g., aspirin effect, von Willebrand's disease) may be sorted out by in vitro platelet function screening.3 Formation of the definitive fibrin clot requires proper functioning of the clotting factor cascade (Fig. 69-1). Clotting times that are measured in vitro, such as the PT, PTT, and TT, are useful in combination for the detection of significant abnormalities in this pathway. The TT depends on the final step of the cascade (conversion of fibrinogen to fibrin) and, therefore, is sensitive for the detection of fibrinogen abnormalities and inhibitors acting at this level (e.g., heparin). The fibrin plug must be maintained for long enough to permit repair of vascular injury, or delayed or secondary bleeding may occur. Excessive fibrinolysis leading to premature clot instability and lysis is detected by FDP and D-dimer testing. FDPs result from fibrinogenolysis and fibrinolysis, but the D-dimer is a unique product of plasmin degradation of fibrin. Thus elevated D-dimer levels imply that clotting (fibrin formation) has taken place prior to plasmin activity. Screening profiles of a number of common coagulopathies are given in Table 69-2.

Vasculitis4

Patients with vasculitis due to any of a variety of causes (e.g., drugs, infections, neoplasms, or systemic connective tissue diseases) may have an increased risk of bleeding complications owing to vascular fragility resulting from the inflammatory changes in vessel walls. Cutaneous rashes secondary to allergic drug reactions may have a significant vasculitic component. Vasculitis may be suspected by finding so-called palpable purpura—small hemorrhagic areas that are infiltrated and become palpable—in contrast to the typically flat purpura of thrombocytopenia.

Cutaneous vasculitis, such as that associated with drug rashes, may cause a prolonged bleeding time and may compromise the safety of even a small surgical incision through involved areas. A positive diagnosis of vasculitis may be made by a small punch biopsy of a purpuric lesion. Management consists of removing possible offending agents, treating underlying disorders (if possible), minimizing skin trauma, and administering anti-inflammatory agents such as corticosteroids, if not contraindicated by other clinical circumstances.

Vascular Malformations

A variety of vascular malformations may bleed when traumatized or entered inadvertently during closed-space biopsies. Hereditary hemorrhagic telangiectasia (Osler-Weber-Rendu disease) may be associated with cutaneous, mucosal, or deep visceral malformations. Usually there is no laboratory evidence of a coagulopathy. A family history (the disorder has autosomal dominant inheritance) and careful inspection of cutaneous and mucosal sites for blanching “spiders” may suggest the diagnosis. Unfortunately, rare patients have only unsuspected deep visceral malformations, which may bleed. The disorder occasionally coexists with von Willebrand's disease or hemophilia A.

Patients with congenital cavernous hemangiomas may have a coagulopathy related to consumption of platelets and clotting factors in stagnant vascular spaces (Kasabach-Merritt syndrome).5 There may be physical findings and laboratory features of DIC.

Microcirculatory Obstruction

Leukostasis secondary to very elevated myeloblast levels in acute nonlymphocytic leukemias may lead to small-vessel obstruction and resulting purpura.6 Brain, lung, and skin are particularly well recognized sites of clinical importance. Similar vaso-occlusive sequelae may be seen in DIC, thrombotic thrombocytopenic purpura (TTP), heparin-induced thrombocytopenia, and fat embolism after long-bone fractures.

Thrombocytopenias

(Table 69-3) The platelet count is the result of a balance between the rate at which platelets are produced and the rate at which they are removed from the circulation. Bone marrow normally has a six- to eightfold reserve production capacity for the formed elements released into the blood, which can help to compensate for shortening of the normal platelet life span of about 10 days. Although there is no simple, readily available test for the rate of platelet production corresponding to the reticulocyte count as an indicator of new red blood cell (RBC) production, the rate of platelet production can be estimated roughly from evaluation of Wright-stained peripheral blood smears for young (large, basophilic) platelets and from particle counter estimates of mean platelet volume.

Underproduction States

Hypoplastic Marrow

Lack of platelet production, either as an isolated problem or as a factor contributing to thrombocytopenia, is suspected when decreased overall cellularity or a selective decrease in megakaryocyte number and size is found on bone marrow examination. A variety of suppressive factors may cause a generalized decrease in marrow mass or a selective decrease in megakaryocyte number. These include chemicals, both therapeutic drugs and environmental toxins; infectious agents, including bacteria, viruses, fungi, and mycobacteria; immune processes, in which the marrow may be inhibited by either a humoral or a cell-mediated mechanism; radiation injury; and idiopathic hypoplastic or aplastic marrow. Marrow hypoplasia will be evident on examination of a bone core biopsy. Care should be taken not to sample marrow that is in the port of prior radiotherapy. Hypoplasia is expected in such an area and may be misleading if interpreted as representing the status of the full marrow.

Ineffective Marrow

Ineffective marrow states are characterized by normal or increased cellularity but lack of release of precursors into the blood. Measurement of the reticulocyte response to anemia is a convenient way to monitor marrow “effectiveness” in this circumstance. A common cause of ineffective marrow function is folate deficiency causing megaloblastic hematopoiesis. The finding of hypersegmented polymorphonuclear leukocytes or ovalomacrocytes and the lack of a reticulocyte response in conjunction with thrombocytopenia should suggest this mechanism. Increased metabolic requirements, decreased folate intake in the diet, and therapy with certain chemotherapeutic agents predispose an individual to folate deficiency. Of course, vitamin B12 deficiency needs to be considered at the same time because of hematologic features identical to folate deficiency. Ineffective marrow function is also a feature of various myelophthisic conditions in which marrow may be replaced by tumor or granuloma or is impaired functionally by metabolic disturbances such as azotemia or hypothyroidism. A variety of inflammatory conditions may cause increased or decreased platelet counts via mechanisms not clearly defined.

Shortened Platelet Survival

The survival of platelets in the blood is usually shortened in patients who are ill. Fever, bleeding, and sepsis, in general, predispose to shortened platelet survival; some specific circumstances include immune-mediated platelet removal by the reticuloendothelial system, hypersplenism, and consumptive coagulopathy. An otherwise uncomplicated shortened platelet life span can be suspected from the findings of a low circulating platelet count, normal or (usually) increased numbers of megakaryocytes on bone marrow examination, and the presence of large, basophilic platelets on a blood smear (indicative of effective marrow release of platelets).

Immune Thrombocytopenic Purpura (ITP)

Immune thrombocytopenic purpura (ITP) is confirmed by finding antiplatelet antibodies in the patient's serum or on the patient's platelets. When a drug is suspected to be the cause, these studies can be done in the presence and absence of the candidate drug in an effort to help confirm the diagnosis. Since the testing may take some time, often an empirical decision is made to stop administration of one or more suspect drugs, when possible. Immune-mediated thrombocytopenia also may be associated with such conditions as collagen-vascular disease, lymphomas, and viral infections, including HIV-1.

The first step in therapy consists of discontinuation of any potential offending agents (e.g., quinidine, trimethoprim-sulfamethoxazole, diphenylhydantoin, thiazides). In addition, drugs that impair platelet function are contraindicated. It is often difficult to sort out a responsible drug in patients with complex medical conditions who are taking multiple medications, but in our experience, a frequent offender in the ICU setting is intravenous furosemide given repetitively. Corticosteroids may improve platelet survival by loosening the attachment of antibodies to platelets and decreasing the reticuloendothelial clearance of the antibody-coated platelets. A longer-term effect of corticosteroids is suppression of antibody production. Administration of platelets is usually of temporary benefit because their survival is shortened, presumably by mechanisms similar to those affecting the patient's platelets. Another intervention that may help to improve platelet survival is infusion of intravenous γ-globulin, 0.4 to 1 g/kg per day for 3 to 5 days.7 This agent is thought to work by causing reticuloendothelial blockade and, possibly, because of the presence of anti-idiotypic antibodies in the preparation, by reducing the amount of antiplatelet antibody available to attach to platelets. In some patients receiving platelets, pretreatment with intravenous γ-globulin may help to prolong survival of the platelets. A benefit may not be seen for several days, and the resulting increase in platelet number, which may be life-saving and permit surgical procedures, is temporary. Subsequent repeat doses may be effective again, but cost becomes an important consideration when selecting this as an ongoing treatment.

Splenectomy may be an effective treatment for immune thrombocytopenia, but often the ICU patient is not a good candidate for the procedure, and therefore, the emphasis is more on the medical interventions just outlined. Intravenous human anti-Rho(D) IgG also may be effective in treating patients with immune thrombocytopenia who are Rh-positive and still have their spleen.8 The mechanism of benefit is not fully understood, but the rationale is to coat circulating Rh-positive red cells with the anti-D antibody, thereby inducing intrasplenic sequestration of the red cells, resulting in reduced available reticuloendothelial capacity to remove antibody-coated platelets. Careful monitoring of the expected (usually limited) hemolysis is necessary. Some patients may be already significantly anemic and unable to tolerate a significant fall in RBC count. Patients who have received anti-D should receive Rh-negative red cells if transfusion is needed. Other agents, such as azathioprine, vincristine, colchicine, danazol, rituximab, mycophenolate mofetil, and cyclophosphamide, or devices (e.g., staphylococcal protein A column treatment of plasma) may be helpful, but they are usually used in subacute and chronic management situations or for individuals with persistent thrombocytopenia after splenectomy.

Posttransfusion Purpura (PTP)

> Abrupt, often apparently unexplained, delayed appearance of severe thrombocytopenia may occur following blood transfusion in rare predisposed (usually P1A1-negative) individuals. The life span is shortened for both P1A1-positive donor platelets (to which isoantibodies have been produced) and the patient's own platelets (mechanism unclear). Infusions of intravenous γ-globulin (0.4 g/kg per day for 2 to 5 days) have been followed by good responses in patients with this potentially serious problem, regardless of the antibody implicated.9

Thrombotic Thrombocytopenic Purpura (TTP)

In TTP, intravascular clumping of platelets accounts for secondary vaso-occlusion and thrombocytopenia. Recent discoveries10 have supported a proposed pathogenetic mechanism to explain the TTP syndrome (see Chap. 70 for an in-depth discussion of this subject).

Triggers of the acquired syndrome include infections (e.g., HIV-1), various drugs (including the antiplatelet agents ticlopidine and clopidogrel), chemotherapeutic agents (especially mitomycin), and malignancies. The important diagnostic clues are schistocytes on blood smear, a low platelet count (usually with a disproportionately low incidence of hemorrhagic symptoms), a negative direct Coombs' test in the face of laboratory evidence of hemolysis, and (usually) normal findings on coagulation studies. Evidence of vaso-occlusion by platelet hyaline agglutinates may be seen in biopsy specimens from tissues such as skin, marrow, gingiva, and kidney. Treatment with corticosteroids and plasma exchange11 usually results in significant improvement. Also, favorable responses to intravenous vincristine, cyclophosphamide, rituximab,12 and in otherwise refractory cases, splenectomy have been reported. Because exacerbation of the vaso-occlusive process has been noted following platelet transfusions, these should be avoided unless there is serious bleeding or an invasive procedure is mandatory.

Disseminated Intravascular Coagulation (DIC)

When thrombocytopenia is a manifestation of DIC,13 usually other features of the process are present that permit recognition of the syndrome. Accompanying microangiopathic hemolysis, hypofibrinogenemia, elevated levels of FDPs and D-dimers, and prolonged clotting times are cardinal features. In advanced DIC, the bleeding diathesis associated with low platelet counts is compounded by qualitative platelet dysfunction owing to thrombin-induced platelet storage pool depletion, which may make platelet transfusions less effective. Schistocytes are seen less frequently in DIC than in TTP and, paradoxically, may be less evident in more advanced stages when fibrinolysis may be brisk, leading to breakdown of fibrin strands responsible for red cell shearing.

Heparin-Induced Thrombocytopenia (HIT)

Bleeding is the most frequent complication of heparin therapy. An important cause of thrombocytopenia is the immune mechanism related to heparin.14 This is of special interest in the ICU because of the almost ubiquitous exposure of patients to heparin owing to its use in access and monitoring lines and the common insertion of “heparin bonded” or “heparin coated” catheters. In addition, certain patient populations, especially those who have undergone cardiac or orthopedic surgery, appear to be at increased risk of sensitization by heparin.

It is common for patients to have a small, reversible drop in platelet count (usually within the normal range) within 1 to 2 days of the initiation of heparin therapy. This effect is thought to reflect some platelet aggregation caused by the presence of high-molecular-weight heparin species in the commercially available unfractionated heparin. The immune-mediated thrombocytopenia (seen in less than 1% of heparin-treated patients) typically appears at about 5 to 7 days after first exposure, but an anamnestic response due to persistent antibody from previous sensitization to heparin may cause thrombocytopenia to develop earlier in some patients, usually if they are within about 100 days of their initial sensitization. The clinical picture may range from no symptoms to life-threatening thrombotic events, which may be venous or arterial. Hemorrhage is less frequent, despite quite severe thrombocytopenia in some individuals. It is often necessary to make the decision about discontinuing heparin on clinical grounds.

The clinical criteria for diagnosis of HIT include an otherwise unexplained fall in platelet count below 150,000/μL or a fall of 50% or more of platelet count within the normal range. A history of any adverse reaction to heparin in the past or of necrotic changes at subcutaneous heparin injection sites should alert one to risk of serious vascular consequences on re-exposure to the drug. It is often difficult in the ICU setting to sort out mechanisms of thrombocytopenia, but a high index of suspicion is important in avoiding adverse consequences of unrecognized HIT. The “gold standard” confirmatory test (normalization of the platelet count after stopping heparin) also may be difficult to achieve because of other coexisting causes of thrombocytopenia (e.g., other drugs, sepsis).

Current notions of the pathogenesis of HIT suggest that the normally formed platelet factor 4 (PF4)–heparin complex is immunogenic in some individuals. Antibodies developed against the PF4-heparin target antigen form antigen-antibody complexes that can activate platelets after interacting with their Fc receptors. The result is intravascular obstruction by platelet clumps.

Laboratory confirmation of the presence of heparin-dependent antibodies may be helpful but is not always available in a timely fashion for decision making. In addition, false-negative tests are not infrequent. When available, various functional assays, such as heparin-induced platelet aggregation, [14C]serotonin release, and flow cytometric methods, may increase the predictability of clinically significant vaso-occlusive events beyond the detection of sensitization using enzyme-linked immunosorbent assay (ELISA) techniques for antibodies to PF4.

It is important to emphasize that the severe antibody syndrome may occur with any type of heparin (although the incidence appears to be less with low-molecular-weight heparin15) given via either intravenous or subcutaneous route. If the syndrome is suspected clinically, all devices and practices that expose the patient to heparin should be discontinued (including use of heparin locks, flushes of indwelling lines, and indwelling catheters that are heparin-coated).

Patients with HIT are hypercoagulable whether or not a clinically recognized vaso-occlusive event has occurred. As a result, they require immediate alternative anticoagulation. Once sensitized by unfractionated heparin (UFH), patients should not receive low-molecular-weight heparin (LMWH) as an alternative because of the high rate of cross-reactivity of antibodies to UFH with LMWH. The commonly used anticoagulants in this setting are direct thrombin inhibitors that do not cross-react with heparin-induced antibodies. Hirudin and argatroban usually are selected—the choice being made based on the patient's comorbidities (hirudin depends on renal clearance; argatroban depends on hepatic function status). Although each agent has a relatively short half-life, the lack of an antagonist for either increases the risk of bleeding complications. If patients require long-term anticoagulation with oral warfarin, it should not be started for at least 2 weeks because of reported adverse clotting problems when begun earlier in the course of HIT.

Mechanical Surface-Related Thrombocytopenia16

Variable lowering of platelet number is sometimes seen when platelets pass over or through foreign or distorted surfaces. Clinical situations of this type include cardiopulmonary bypass and the presence of intraaortic balloon pumps or pulmonary artery catheters. For cardiopulmonary bypass, at least, platelet-related bleeding during or after bypass is much more likely to be secondary to impaired platelet function (see below) than to severe depression of the platelet count.

Sequestration

Hypersplenism

Thrombocytopenia with or without anemia and neutropenia may be a feature of hypersplenism.17 This state of increased local sequestration, usually with shortened intrasplenic survival, frequently is accompanied by increased spleen size that is detectable on physical examination or imaging studies. Bedside ultrasound is a safe noninvasive method for assessing the spleen in ICU patients.

Hypersplenism may be a chronic, stable condition antedating and then complicating a new problem (e.g., cirrhosis with splenomegaly predating acute deterioration for another reason). In addition, acute onset of hypersplenism with progressive splenic enlargement and sequestration may be a feature of such processes as splenic venous occlusion and sickle cell splenic sequestration crisis in children. It is unusual to encounter a platelet count of less than about 30,000/μL in “compensated” hypersplenism associated with cirrhosis and portal hypertension accompanied by splenomegaly. Bone marrow evaluation is expected to reveal increased megakaryocyte number and overall hypercellularity. The peripheral smear should show decreased platelet numbers but some large basophilic platelets.

In situations of progressive splenomegaly with associated thrombocytopenia (and risk of splenic rupture), splenectomy may be the only available therapy. If the splenic enlargement is secondary to leukemic or lymphomatous involvement, radiation therapy may be of some benefit in shrinking the spleen, but the effect may be temporary, and there is the risk of paradoxical thrombocytopenia or neutropenia rather than count improvement.

Thrombocytosis/Thrombocythemia

Elevated platelet counts are seen commonly in critically ill patients because inflammation, bleeding, surgery, hemolysis, severe injury, and neoplasia are among the major causes of reactive thrombocytosis. Usually, the count is elevated to levels less than 1 × 106/μL, and no adverse effect is expected from such a secondary increase in qualitatively normal platelets.

In contrast, patients with elevated platelet counts secondary to myeloproliferative disorders18 are at an increased risk of bleeding or thrombotic sequelae, especially when the platelet count exceeds 1 × 106/μL. The diagnosis of an underlying myeloproliferative disorder may have been established prior to the patient's presentation with a complication, or it may be suspected from associated clinical features such as splenomegaly, abnormal platelet morphology on smear, and bone marrow examination revealing panmyelosis (increase in all cell lines) as well as qualitative abnormalities in megakaryocytes, which are also markedly increased in number.

If the patient is asymptomatic, a prolonged bleeding time may portend hemorrhagic problems. No definitive laboratory tests predict impending thrombotic complications in such patients, but it appears that the risk may be increased in older persons with fixed vascular disease. It is especially important to avoid splenectomy, either intentional or inadvertent, at the time of abdominal surgery for other reasons because marked platelet elevations may cause fatal vaso-occlusive events.

Platelet levels may be reduced quickly and effectively in symptomatic thrombocythemic individuals by plateletpheresis. Treatment with anagrelide or hydroxyurea can be used to maintain normal platelet counts (the usual goal is to keep the platelet count less than 500,000/μL).18 These drugs also can help to prevent secondary rises in platelet count following surgical procedures and other factors that may precipitate reactive thrombocytosis.

Thrombocytopathy

Qualitative platelet abnormalities are encountered frequently in the ICU setting.19 Often they contribute to other, more serious bleeding diatheses, but on occasion they present a significant problem in their own right. Exposure to external agents, such as ethanol, aspirin, and nonsteroidal anti-inflammatory drugs (NSAIDs), prior to admission is common. Other drug-induced causes of thrombocytopathy in the hospital include exposure to the parenteral NSAID ketorolac (TORADOL) and glycoprotein (GP) IIb/IIIa inhibitors (e.g., abciximab, ticlopidine, clopidogrel). The resulting prolongation of bleeding time varies in duration and severity. Once the offending agent has been removed, the bleeding problem can be overridden promptly by administration of normal platelets or, in the case of aspirin or ticlopidine, by intravenous infusion of 1-diamino-8-D-arginine vasopressin (DDAVP), if necessary. Often, unless there is significant bleeding or invasive procedures are needed, it is sufficient to wait several days and let the patient's own production of new platelets unexposed to the offending agent correct the problem.

Uremia

A more difficult problem is the thrombocytopathy associated with uremia. A variety of interventions may be helpful temporarily. Correction of the prolonged bleeding time may be needed prior to an intended surgical intervention (e.g., renal biopsy for diagnosis). The key step is appropriate dialysis; additional benefit sometimes is noted with administration of corticosteroids, infusions of cryoprecipitate20 (usually 10 cryopacks every 12 hours), intravenous DDAVP21 (0.3 μg/kg), platelet transfusions, and intravenous conjugated estrogens.22 Intravenous DDAVP, if effective in reducing the bleeding time, may be repeated at 6- to 12-hour intervals, but tachyphylaxis is expected after 1 to 2 days of this therapy. It is also important to seek and correct other coexisting coagulopathies, such as vitamin K deficiency.

Myeloproliferative Disorders

Patients with myeloproliferative disorders may have platelet dysfunction, especially when platelet counts are elevated. Usually, normalization of the platelet count results in return of the bleeding time to normal, but individuals with long-standing polycythemia vera or myeloid metaplasia, for example, may still have abnormal bleeding, especially with surgical procedures, even when all counts have been normalized.

Drugs

Aspirin

The classic drug inhibitor of platelet function is aspirin. It causes irreversible inhibition of cyclo-oxygenase, thereby leading to impaired prostaglandin metabolism. The net effect of aspirin is the result of inhibition of production of the platelet aggregant thromboxane A2 versus decreased synthesis of prostacyclin—the potent endothelium-derived inhibitor of platelet aggregation. In most individuals on Western diets that contain limited amounts of the omega fatty acid precursors of prostacyclin synthesis, the net result is a prolonged bleeding time. When aspirin is discontinued, the exposed platelets remain impaired, but newly synthesized platelets with normal function soon begin to contribute to normalization of the bleeding time. Usually, the effect of aspirin on bleeding time has dissipated significantly by 3 days after discontinuation.

Nonsteroidal Anti-Inflammatory Drugs (NSAIDs)

NSAIDs cause reversible inhibition of cyclo-oxygenase and, therefore, a shorter period of bleeding risk than aspirin after discontinuation.

Antibiotics

Antibiotics of the penicillin family, epitomized by carbenicillin, may cause a complex coagulopathy. In addition to disturbance of the synthesis of vitamin K by the intestinal flora, a thrombocytopathy characterized by a prolonged bleeding time is recognized. These phenomena usually are not of major clinical significance, but they may add to other coagulation defects and affect the interpretation of coagulation laboratory evaluations. When bleeding from these mechanisms is serious, platelet transfusion after discontinuation of the drug should compensate for the impaired function of platelets exposed to the drug.

Dysproteinemias

Abnormal immunoglobulin concentrations, as seen in myeloma and Waldenstrom's macroglobulinemia, interfere with platelet function (causing prolonged bleeding times) presumably by physically “coating” platelet surfaces and thus limiting the access of clotting factors to the phospholipid surfaces of platelets, which are needed to promote key reactions in the clotting cascade. In addition, prolongation of the TT is often noted.

Cardiopulmonary Bypass16

Acquired thrombocytopathy is regularly produced by contact of platelets with membrane surfaces in bypass circuits. In some patients this effect may add to other causes of increased bleeding risk after bypass (such as hypofibrinogenemia and inadequate heparin neutralization). Transfusion of donor platelets (if clinically urgent) is expected to normalize the bleeding time once bypass is discontinued.

Thrombin-Induced Storage Pool Defect

Thrombocytopathy may contribute to the bleeding problems in DIC. During active DIC (e.g., in acute promyelocytic leukemia), an acquired platelet function defect with prolonged bleeding time may appear owing to release of granule contents by excess thrombin. This is likely to be especially troublesome when patients are thrombocytopenic as well. Control of the DIC is the key to successful correction of this problem. Infusion of donor platelets is of benefit for only limited periods because these platelets are also exposed to excess thrombin while the DIC is active.

Sample Collection and Interpretation

Accurate values for clotting times, which form the basis for assessment of the integrity of the coagulation cascade, depend on collection of the correct quantity of blood relative to the amount of anticoagulant present in the sample tube; in addition, extraneous contamination by heparin must be avoided. It is important not to overfill collection tubes; rather, the vacuum should be allowed to determine the correct amount of specimen per tube, and the tube then should be inverted promptly to ensure adequate mixing with anticoagulants. Rapid delivery to the testing location decreases problems with cascade activation in the tube (which leads to prolonged clotting times). When drawing from heparinized lines, a double draw (discarding at least 10 mL initially and using the second aliquot) should prevent heparin contamination. Unexpectedly long PTTs that are not corrected by mixing the sample with an equal volume of pooled normal plasma but that are corrected by incubating the sample with heparin neutralizer powder (triethylaminoethyl cellulose) should indicate the presence of heparin contamination.

Consideration of the results of PT and PTT in combination can help to identify the site(s) of cascade deficiency (Table 69-4). Prolonged clotting times should be restudied after the patient's plasma is mixed with an equal volume of pooled normal plasma to differentiate between absence of the factors in question and inhibition of their activity by circulating inhibitors. If the clotting time is corrected by this mixing test, a deficiency is present; if the clotting time of the mix remains prolonged, an inhibitor (such as heparin or one directed against a specific clotting factor) is present.

Isolated Factor Deficiencies

Isolated severe factor deficiencies may antedate and complicate the need for intensive care, as in hemophilia A or B, or may be acquired as a consequence of the critical illness, as in the case of hypofibrinogenemia secondary to DIC.

Hypofibrinogenemia

Low functional fibrinogen levels in ICU patients most commonly result from decreased hepatic synthesis of fibrinogen or increased removal due to the action of thrombin and plasmin during DIC. Bleeding risk is increased by the anticoagulant effect of nonclottable fibrin split products in high titer. Spontaneous clinical hemorrhage from hypofibrinogenemia is not expected above a concentration of 100 mg/dL in plasma and usually is not a serious risk until levelsg of 50 mg/dL are reached. Fibrinogen may be replaced with fresh frozen plasma (FFP) or, more efficiently in terms of volume, with cryoprecipitate; each cryopack is expected to raise the fibrinogen level approximately 4 to 10 mg/dL in an adult. Although the plasma half-life of fibrinogen is approximately 4 days under normal circumstances, serial plasma fibrinogen level determinations will dictate the frequency of infusions (as often as every 6 to 12 hours, especially in acutely ill patients with a shortened fibrinogen half-life).

Hemophilia a (Factor VIII Deficiency)

Critical illness in a patient with hemophilia A requires careful correction of the pre-existing coagulopathy, especially when the situation is complicated by such new problems as ITP, DIC, sepsis, and hepatic or renal dysfunction. Appearance of an inhibitor to factor VIII and development of a progressive pseudotumor may further increase the complexity of management. Preparation of the severe hemophiliac for surgery consists of checking to exclude an inhibitor and then replacing with factor VIII concentrate to achieve the desired percent correction (100% equals 1 unit of factor VIII per milliliter of the patient's plasma). For major surgery, a starting level at or above 100% is desirable. Additional doses are given intraoperatively. The doses are dictated by the amount of bleeding and stat factor VIII levels. The shortened half-life of factor VIII during bleeding leads to increased requirements for replacement (both higher doses and shorter intervals between doses). To give a sample calculation: Suppose that a 70-kg patient with a hematocrit of 40% has a blood volume of 4900 mL (70 mL/kg) and a plasma volume of 2940 mL (60% of the blood volume). If the starting factor VIII level is less than 1%, approximately 3000 units of factor VIII concentrate will be needed to achieve correction to the 100% level. It is important to assess the adequacy of correction by monitoring the factor VIII level before invasive procedures and to continue support for 10 to 14 days after a major surgical procedure or hemorrhage to prevent serious delayed bleeding. Maintenance postoperative factor VIII support usually is given every 8 to 12 hours; the goal is to prevent the between-dose nadir in factor VIII plasma levels from falling below 40% to 50%. Several recombinant factor VIII products are available.

Similar considerations apply to patients with hemophilia B (Christmas disease), except that replacement is with FFP, prothrombin complex, or recombinant factor IX in severe deficiency states.

Factor VII Deficiency

Factor VII deficiency is a much less common inherited deficiency than the preceding. It is a non–X-linked disorder with a variable severity, which in some patients presents as a severe hemophilia-like condition. Replacement with FFP or prothrombin complex is appropriate. The relatively short half-life of factor VII (4 to 7 hours) may require dosing at 6- to 8-hour intervals during major hemostatic challenges, but less frequent maintenance doses may be sufficient beyond the fourth day following some orthopedic procedures.23

Isolated factor VII deficiency is caused most often by vitamin K deficiency or liver disease or occurs during the initiation of warfarin anticoagulation. When factor VII deficiency cannot be corrected with vitamin K therapy, factor replacement with FFP or, when indicated, prothrombin complex can be used. Recombinant activated factor VII (rVIIa) also is available.

Factor XI Deficiency

Factor XI deficiency is usually inherited, although in advanced liver disease a deficiency of factor XI may coexist with low levels of the other factors synthesized in liver parenchymal cells. FFP is the only currently available replacement product. For severely deficient patients with a bleeding tendency, preoperative plasma exchange may be used to establish safe hemostatic levels that can then be maintained by follow-up plasma infusions.

Factors II, V, and X

These deficiencies may appear in isolation or coexist with others. Isolated factor X deficiency with significant bleeding may be seen in amyloidosis. FFP (or prothrombin complex for factors II and X) replacement therapy is indicated, if needed, but it may not readily correct an amyloid-associated factor X deficiency because of rapid clearance of the factor, which attaches to the amyloid fibrils. Splenectomy may be helpful in such instances.

Factor XII

Factor XII deficiency causes a prolonged PTT without a clinically significant bleeding diathesis. Some workers have suggested that, paradoxically, patients who are severely deficient in factor XII may be hypercoagulable. It is important to identify this factor deficiency both because needed invasive procedures can be done in its presence when it is an isolated abnormality and to avoid unnecessary correction of the PTT with FFP.

Factor XIII

Deficiency of factor XIII24 is usually lifelong and may be associated with significant bleeding problems of the delayed type. In homozygotes there is an increased risk of fatal intracranial bleeding; spontaneous abortion also has been noted. Rarely, an acquired deficiency is seen in patients who develop an inhibitor while taking a drug such as isoniazid. The diagnosis cannot be made from the usual clotting or bleeding time tests. Rather, this deficiency must be suspected and tested for directly. Clot dissolution in 5 M urea during overnight incubation indicates deficiency (<1% activity). The deficiency is readily corrected with FFP, cryoprecipitate, or a placental concentrate. Only 1% to 2% of the normal factor XIII activity is adequate for hemostasis and intact clot stability. The long half-life of this factor (8 days) permits long dosing intervals and relatively easy prophylaxis when a hemostatic challenge is anticipated.

von Willebrand's Factor (VWF)25

Deficiency of VWF results in a bleeding disorder more akin to that associated with platelet dysfunction than to that of the hemophilias. Cardinal findings include a prolonged bleeding time and a moderately prolonged PTT with depression (usually mild) of the factor VIII coagulant (VIII:C) level. Decreased levels of ristocetin cofactor and von Willebrand's antigen also are found. Rapid screening for von Willebrand's disease using an in vitro platelet function analyzer can be helpful in diagnosis.3 The most common variety of this disorder (type I) is characterized by a defect in secretion of normal VWF from vascular endothelial cells. Less commonly, qualitative defects in the multimeric structure of the factor (type II) or a severe deficiency in its synthesis (recessively inherited type III) is seen. Type I patients may respond to infusion of DDAVP (0.3 μg/kg IV) with a rise in VWF and other released endothelial products, some of which may have anticoagulant effects (such as prostacyclin [PGI2] and tissue plasminogen activator). Usually there is a net hemostatic benefit, which is seen as an improved bleeding time over a several-hour period. Repeat doses may be given, but tachyphylaxis after 1 to 2 days precludes effective long-term therapy with DDAVP or reliance on it alone for major surgical procedures.

The Duke bleeding time was considered by some to be the best indicator of bleeding risk prior to invasive procedures, but critical review of the bleeding time26 raises questions as to the predictability of bleeding from the result of this test in individual cases. A more extensive workup, including both functional studies (ristocetin cofactor, PTT, VIII: C, ristocetin-induced platelet aggregation) and antigenic studies (VIII-related antigen and multimer analysis), is useful in characterizing the type and potential severity of the defect. In acute situations when the diagnosis is known, measuring the VIII-related antigen level, if available, is a helpful way to monitor the patient's status and adequacy of replacement. The platelet function analyzer also may be used for rapid confirmation of improvement after treatment. The usual goal of therapy is normalization of the VWF level during active bleeding and before anticipated surgery or another invasive procedure. For severe deficiency and prolonged treatment, virucidal-treated factor VIII concentrates, which contain high levels of VWF (e.g., HUMATE p), are indicated. Monoclonal antibody–purified factor VIII concentrates are not useful because VWF is depleted from these products. DDAVP may be contraindicated in type IIB and platelet-type (or pseudo-) von Willebrand's disease—two conditions in which platelet aggregation and thrombocytopenia may ensue.

Inhibitors27

Coagulation factor inhibitors may create life-threatening situations, especially in otherwise critically ill patients. The potency of the inhibitor can be determined by titration of the effect of plasma dilutions on the result of mixing the patient's plasma with pooled normal plasma (see above).

Factor VIII Inhibitors

The most commonly encountered coagulation factor inhibitors are directed against factor VIII. They are seen most frequently in hemophiliac patients receiving supplementary factor replacement, but they also can occur sporadically in elderly individuals, in the postpartum period, and in persons with an autoimmune diathesis or lymphoma.28 Factor VIII inhibition usually is measured in Bethesda units (BU); levels greater than 10 BU indicate more potent inhibitors. Testing to exclude a factor VIII inhibitor should be done before any significant invasive procedure is undertaken in a hemophiliac in order to make it possible to predict the response to factor replacement. It is also important to confirm that the expected response to factor administration has occurred before proceeding with an invasive procedure. For lesser titers, infusion of factor VIII concentrate often can override the inhibitor in the short term, although in some patients this approach will elicit a rise in the inhibitor titer. Other modes of therapy for high-titer inhibitors include bypassing the factor VIII–dependent step with prothrombin complex concentrate (PCC) or activated PCC, use of porcine factor VIII concentrates, and infusions of rVIIa. Combined immunosuppressant programs that include steroids, cytotoxic agents, plasma processing over columns that remove immunoglobulin G (IgG) and IgG-containing immune complexes,29 and high-dose intravenous immunoglobulin are approaches to the subacute and chronic management of hemophiliacs with inhibitors. Recent favorable experience with rituximab has been noted.30

Inhibitors found in nonhemophiliacs tend to be more responsive to simpler immunosuppressive treatments, such as treatment with azathioprine and prednisone. There is a significant risk of bleeding in these individuals when the inhibitor is not controlled. If an underlying tumor or autoimmune disorder is present, its management may help to reduce the level of inhibitor. Surgical procedures in patients with inhibitors obviously are fraught with high risk, and meticulous hemostatic correction perioperatively is necessary. Correction needs to be maintained for at least 10 days after major procedures (as in uncomplicated hemophilia) to prevent delayed bleeding.

Lupus Anticoagulants

The presence of the so-called lupus anticoagulant31 is important to recognize because, although it is usually associated with a prolonged PTT, it may imply a hypercoagulable state rather than a bleeding diathesis. Rare individuals with lupus anticoagulants have a bleeding diathesis. Prolongation of the PT in addition to the PTT may be an indicator of such a case. This phospholipid inhibitor was first identified in patients with de novo systemic lupus erythematosus (SLE), but it is known to occur also in drug-induced SLE syndromes, in individuals with other autoimmune diseases, and in otherwise normal persons. The typical laboratory result is a prolonged PTT and/or Russell's viper venom time (RVVT) with evidence of an inhibitor and interference on several of the coagulation factor assays. An abnormally steep rise in the PT with serial dilution of thromboplastin in vitro is a characteristic finding (the tissue thromboplastin inhibition [TTI] test). This latter phenomenon may be mimicked by the presence of heparin and warfarin anticoagulation; therefore, it is preferable to perform the test, if possible, before anticoagulation is begun. Absorption of the patient's plasma with platelet phospholipid in vitro may remove the inhibitor and lead to normalization of the PTT, TTI, and RVVT test results, thus enhancing the specificity of diagnosis.

Most patients with the lupus anticoagulant do not have coagulation problems, but approximately 25% have clinically important hypercoagulability. It is unclear whether empirical prophylactic long-term anticoagulation with warfarin is useful. Perioperative prophylactic dose heparinization is recommended to offset added surgical hypercoagulable risks. If thrombotic problems—typically venous (deep venous thrombosis, pulmonary emboli)—do occur, chronic anticoagulation with warfarin is recommended. Results of the PT are sometimes unpredictable in patients with lupus anticoagulants. Chromogenic factor X levels may be more reliable for monitoring the degree of warfarin effect. Increased risk for arterial clotting events also is present.

The potency of lupus anticoagulants may diminish spontaneously as a result of the discontinuation of an offending drug (e.g., procainamide, phenothiazines) or in response to immunosuppressive therapy (e.g., with prednisone).

Factor V Inhibitors32

Some patients exposed to topical bovine thrombin used as a surgical sealant may develop an inhibitor to factor V (a contaminant of the bovine product). On occasion, there can be significant lowering of the human factor V level because of cross-reactivity. The antibodies eventually disappear in a period of weeks, but treatment with plasma and platelets (thought to be site of factor V protected from the inhibitor) may be needed in individuals who bleed significantly.

Combined Factor Deficiency States

Vitamin K Deficiency33

The combination of poor dietary intake and antibiotic therapy that is so common among ICU patients predisposes them to vitamin K deficiency. Impaired absorption of fat-soluble vitamins, as in biliary obstruction, may contribute as well. Prolongation of the PT initially and then of both the PT and the PTT is characteristic. Documentation of selected reduction of the levels of factors II, IX, and X is rarely needed to make this diagnosis; usually the initial decline in factor VII in the absence of other explanations (such as liver disease or warfarin anticoagulation) is sufficient. Vitamin K supplementation may be given orally, subcutaneously, intramuscularly, or intravenously. The oral and subcutaneous routes are preferred, if feasible, to avoid a risk of hematoma or, rarely, anaphylaxis after intravenous dosing.

Dysfibrinogenemia

A qualitative abnormality of fibrinogen may or may not be associated with clinically significant hypofibrinogenemia.34 The diagnosis is made by finding a disproportionately low functional level of fibrinogen relative to the antigenic fibrinogen level. In addition to the variety of inherited fibrinogen abnormalities associated with bleeding, clotting, and wound dehiscence problems that have been recognized, an acquired decrease in fibrinogen function is seen commonly in the presence of parenchymal liver disease35 and hepatomas.

If it becomes clinically necessary to correct the hypofibrinogenemia, infusion of cryoprecipitate is the method of choice (see above). A functional level of 100 mg/dL or greater is the usual goal of therapy.

Definitions and Laboratory Testing Methods

Clot strength and stability are the keys to prevention of rebleeding and consequent development of hemorrhagic infarction after initial hemostasis. The covalent cross-linking of fibrin mediated by factor XIII is important for clot strength, and carefully regulated activity of the fibrinolytic system allows eventual clot degradation once the vessel is strong enough to handle re-established flow. The details of regulation of the fibrinolytic process, including its intensity and timing, are incompletely understood.

Fibrinolytic activity is mediated by plasmin, which is derived from the inactive plasma precursor plasminogen. Plasminogen can be activated by various endogenous kinases, including tissue plasminogen activator and urokinase, as well as by exogenous agents such as streptokinase. The activity of plasmin is regulated by a series of inhibitors, the most important of which is α2-antiplasmin. The activity of the fibrinolytic system is assessed indirectly in vitro by measurement of concentrations of substrate (fibrinogen) and plasmin-generated breakdown products of fibrinogen and fibrin (FDPs, D-dimer). Lesser degrees of fibrinolysis may be detected by measuring the level of plasmin-antiplasmin complexes.

Primary Fibrinolysis

It is helpful to distinguish so-called primary fibrinolysis, in which plasmin is generated from plasminogen by an activator in the plasma, from secondary fibrinolysis, in which plasmin is generated within a formed clot by an activator that is also present. Primary fibrinolysis (fibrinogenolysis) is seen in states of tissue injury such as liver disease or therapeutic administration of activators such as urokinase and streptokinase. Secondary fibrinolysis is the normal mechanism of clot degradation initiated by tissue plasminogen activator. The latter process is accelerated pathologically in DIC.

Secondary Fibrinolysis

The in vitro tests of clot lysis, such as the euglobulin lysis time, are sensitive to generation of circulating plasmin and therefore give shortened times in primary fibrinolysis but normal times in secondary fibrinolytic states (e.g., DIC). Levels of fibrinogen and FDPs do not distinguish the two mechanisms. D-dimer is a unique degradation product of cross-linked fibrin that is not seen among the products of plasmin digestion of fibrinogen. As such, it is helpful in distinguishing fibrinogenolysis from fibrinolysis. This test is especially helpful in evaluating states of increased clot formation such as DIC. However, it must be emphasized that elevated D-dimer levels do not indicate the mechanism of clot generation. For example, high levels of D-dimer may be present during the dissolution of a hematoma, and it would be wrong to assume that DIC necessarily was the cause of the original clot.

Acute Disseminated Intravascular Coagulation13

The most important complex coagulopathy in intensive care medicine is acute DIC. DIC is essentially a state of increased propensity for clot formation; it may be triggered by a variety of stimuli related to such diverse underlying disorders as sepsis, tissue injury, and neoplasm. There may be clinical and laboratory evidence of hypercoagulability, but in acute cases, consumptive coagulopathy with hemorrhagic manifestations may predominate. The full-blown state is characterized by thrombocytopenia, prolonged clotting times, depressed circulating levels of cascade factors (especially factor VIII), and increased fibrinolysis manifested by high FDP and D-dimer levels as well as hypofibrinogenemia. There may be associated hemolysis with evidence of microangiopathic changes on blood smear (schistocytes) in some patients. Organ dysfunction secondary to vaso-occlusion or hemorrhage may be evident. DIC should be suspected when such a complex coagulopathy appears, especially if the clinical setting would predispose the patient to it.

The key step in management is treatment of the underlying condition that is responsible for the hypercoagulable state (e.g., antibiotic treatment for infection). Transfusion of blood products, including RBCs, platelets, FFP, and cryoprecipitate, if needed for the correction of severe hypofibrinogenemia, may protect the patient from hemorrhagic complications while the mechanism triggering the DIC is being eliminated. Although, in theory, blood product support may temporarily “feed the fire,” often patients benefit from such treatment without requiring heparin therapy with its attendant potential complications.

In cases of purpura fulminans and massive thromboembolism, where the pace of the process is catastrophic, a trial of heparin therapy is appropriate (or drotrecogin in severe sepsis), along with blood product support, especially in patients who have potentially treatable underlying disorders that can be expected to respond reasonably promptly with appropriate specific therapy. The doses of heparin required are variable because of heparin resistance in DIC (owing to low levels of antithrombin III and elevated levels of PF4). The necessary dose may vary initially from 5 to 10 units/kg/h, given by continuous intravenous infusion up to full therapeutic levels. The criteria usually used to determine whether heparin is helpful include clinical observation of lessened bleeding and a rise in fibrinogen levels and platelet counts.

There are obvious risks to the use of heparin because of uncertainty as to the optimal dose and because the addition of an anticoagulant drug makes it difficult to interpret the results of clotting tests. Heparin therapy usually is not indicated for patients with bleeding in critical areas, such as intracranial locations. In addition, isolated antifibrinolytic therapy with agents such as ϵ-aminocaproic acid is contraindicated because of the risk of extensive thrombotic complications. For more chronic, low-grade DIC, subcutaneous heparin (e.g., 5000 units every 8 to 12 hours) or low-molecular-weight heparin may be helpful in preventing clinical thrombotic complications.

Laboratory parameters for following DIC include clotting times, platelet count, and fibrinogen level (assuming they are not disturbed simultaneously by an underlying process such as leukemia or liver failure). FDP levels are good markers for the detection of fibrinolysis but tend to lag as indicators of improvement.

Massive Transfusion

Patients who have had one to two blood volumes replaced with stored RBCs within 24 hours are subject to the development of a complex coagulopathy, the most important feature of which is thrombocytopenia owing to both dilutional and consumptive mechanisms.36,37 In addition, there may be impaired platelet function and lowered levels of plasma factors VIII and V. Successful management of this coagulopathy depends on recognition of the developing pattern and administration of sufficient platelets and FFP to offset the effects of dilution. More serious disturbances of hemostasis appear if DIC develops as a consequence of the underlying shock and/or tissue injury (see above). There has been recent interest in use of rVIIa to help control hemorrhage and thereby reduce the need for massive transfusion.38,39

Antibiotic Therapy

Use of multiple or broad-spectrum antibiotics predisposes the patient to combined coagulation defects. Vitamin K deficiency is common. Impaired platelet function manifested by prolonged bleeding time has been observed with carbenicillin 40 and related antibiotics.

Liver Disease41

Because of the key role of the liver in clotting factor and thrombopoietin synthesis, coagulopathies are an important part of liver dysfunction. Not only are clotting times prolonged, but if there is associated hypersplenism, thrombocytopenia also may be present. Underproduction of thrombopoietin and immune thrombocytopenia secondary to hepatits C, alcohol, and folate deficiency are also potential factors contributing to lowered platelet counts. Platelet dysfunction (usually mild) also may be present. In addition, primary fibrinolysis secondary to liver cell injury may occur, creating a picture that may mimic DIC. It is not unusual to see DIC superimposed on liver failure causing a severe complex coagulopathy. The pattern of fibrinolysis and the factor VIII:C level may be helpful for distinguishing the coagulopathy of advanced liver failure from DIC. Primary fibrinolysis (short euglobulin lysis time) and elevated factor VIII:C levels are seen in liver disease; findings of secondary-type fibrinolysis (normal euglobulin lysis time) and low factor VIII:C levels (as well as more elevated levels of D-dimer) favor DIC. Placement of shunts to drain ascites into the venous circulation may be associated with accelerated DIC and bleeding risk.

Therapy for the coagulopathy of liver disease consists of replacement of clotting factors and platelets as well as correction of vitamin K deficiency. If volume considerations permit, FFP is used first. Often it is necessary to use prothrombin complex for adequate correction prior to invasive or surgical procedures. The risks of administration of this product include enhanced coagulation and, in some instances, DIC owing to the presence of activated clotting factors in the concentrate. Prothrombin complex concentrates replace only the vitamin K–dependent factors, not all the clotting factors made in the liver. Other approaches to life-threatening hemorrhage in patients with liver failure have included use of fibrinolysis inhibitors and, more recently, rVIIa.39

Renal Disease41

Uremic patients are likely to have a bleeding diathesis, which most commonly is a thrombocytopathy characterized by a prolonged bleeding time. The disease underlying the renal failure also may, of course, predispose the patient to a coagulopathy (e.g., immune thrombocytopenia due to SLE, factor X deficiency secondary to amyloid, low antithrombin III and factor IX levels in nephrotic syndrome of various causes). The thrombocytopathy becomes especially important if surgical interventions such as renal biopsy are needed. Adequate acute or subacute hemostasis usually can be achieved by using one or more of the treatments discussed earlier, but chronic protection from bleeding is more difficult to achieve. Hemostasis may be improved in uremia after treatment with recombinant human erythropoietin.42

Dysproteinemias43

Patients with multiple myeloma or Waldenstrom's macroglobulinemia may have significant bleeding problems in addition to the thrombocytopenia that may accompany advanced disease or be a consequence of treatment. Both impaired platelet function (prolonged bleeding time) and prolonged TT secondary to interference with the conversion of fibrinogen to fibrin may be seen. The mechanism of these effects is not fully understood but is thought to involve the physical “coating” of the platelets by the paraproteins, which reduces the exposure of the phospholipid template that is needed for key steps of the cascade. Steric hindrance of the process of fibrin formation also may play a role. These adverse effects on coagulation tend to improve with a decline in paraprotein levels (owing to chemotherapy-induced decreased production and/or to removal by plasmapheresis).

General Considerations

The following guidelines should be used in the management of patients with coagulopathies:

1. Invasive procedures should be performed only when absolutely necessary in patients who have coagulopathies or can be expected to develop them as a result of their disease or its therapy.

2. Intervention(s) should be as limited as possible, and procedures that can be performed under direct vision are preferred. Direct vision allows the operator to be sure that primary hemostasis has been achieved and facilitates follow-up for detection of delayed bleeding.

3. The clinician should be certain that the preliminary coagulation status is fully delineated so that potential bleeding risks can be anticipated accurately and the correct therapeutic agents can be available.

4. If a preparatory treatment (e.g., infusion of factor VIII concentrate) or a change in medication (e.g., discontinuation of an anticoagulant drug) is indicated, the clinician should check that the expected beneficial effect actually has occurred before the procedure is begun.

5. The timing of procedures for which supportive products will be needed should be coordinated with the providers (e.g., the blood bank or pharmacy) so that an adequate supply is available for the anticipated duration of support. One should be liberal in estimating needs so that unexpected complications can be covered.

6. The patient should be monitored carefully for both immediate and delayed bleeding to detect status changes early. Other health care personnel should be alerted to the potential for bleeding and the appropriate therapy based on previous evaluation.

7. In a patient who has been evaluated and treated for a coagulopathy, it is important to be alert to the possible effects on coagulation of changes in either medical status (e.g., the onset of sepsis) or medications. Reassessment may be necessary if the patient shows a decreasing response to therapy that was effective initially.

Specific Coagulopathies

Vasculitis

Patients with vasculitic rashes may have very long bleeding times that prohibit skin incision. Discontinuation of an offending drug and trial of corticosteroids may be useful. Administration of platelets is not indicated unless thrombocytopenia or thrombocytopathy coexists.

Thrombocytopenia

For procedures that will be performed under direct vision, such as limited biopsies or peripheral line insertions, platelet counts of 50,000 to 80,000/μL usually are adequate, depending on the extent of the planned intervention. For major surgical procedures, insertion of lines in large central vessels, or closed-space needle biopsies, initial counts in the range of 80,000 to 100,000/μL are preferred (assuming that platelet function assessed by the bleeding time is normal). Lumbar punctures usually are performed safely at platelet counts of greater than 50,000/μL.

Thrombocytopathy

Thrombocytopathic bleeding can be very difficult to manage, especially if the environment into which new platelets are transfused is deleterious to their function. Thus, in uremia, for example, multiple transfusions of normal platelets may be ineffective. Other methods of therapy discussed in earlier sections of this chapter may be helpful in this situation. The bleeding time may not be corrected significantly in some patients. The predictive value of the skin incision bleeding time in individual patients has been questioned, but it seems prudent to attempt to shorten a prolonged bleeding time, when possible, before an invasive procedure is performed.

Drug-induced thrombocytopathies may be easier to manage if the drug has a reasonably short half-life or if its inhibitory effect on platelet function is reversible (as with NSAIDs, for example). After the drug has been discontinued and its effect has dissipated, new platelets produced by the patient's marrow and (if needed for an acute intervention) transfused normal platelets will correct the prolonged bleeding time. In urgent situations, DDAVP infusions may be used to override, at least temporarily, the thrombocytopathic effects of aspirin and ticlopidine.44

Thrombocythemia

Patients with myeloproliferative disorders and elevated platelet counts (usually in excess of 1 × 106/μL but sometimes in the range of 0.5 × 106 to 1 × 106/μL) may have excessive bleeding, either spontaneously or following trauma or surgery. It is generally advisable to lower the platelet count into the normal range, but even when this goal is achieved, some patients have considerable bleeding difficulties. The bleeding time may be a helpful predictor. It is wise, when possible, to avoid or minimize invasive procedures in these individuals. Similar precautions are indicated in individuals with a bleeding diathesis due to a myelodysplastic disorder.

Fibrin Generation Cascade Defects

A PT and PTT that are at least normal should be insisted on, if they are achievable, for procedures in which excess bleeding could seriously compromise the outcome (e.g., intracranial neurosurgery) or for closed-space procedures in which mechanical hemostasis is difficult to achieve or in which the extent of bleeding may not be obvious until it is significant. More precise correction of factor levels than that produced by achieving normal clotting times is indicated for such procedures in patients with known factor deficiency states, such as hemophilia A.

For procedures done under direct vision or for line placements, there may be a greater tolerance for impaired clotting times. The clotting times should not exceed 1.2 to 1.3 times the baseline value under these circumstances. In contrast to platelet-related problems, bleeding in patients with cascade deficiencies may be delayed in onset and, therefore, unexpected by uninformed medical and nursing staff. The risk of bleeding with liver biopsy, for example, may be reduced by using transjugular or laparoscopic techniques.

Fibrinolytic States

Hyperfibrinolytic states leading to short clot survival, hypofibrinogenemia, and elevated levels of FDPs may be associated with serious bleeding during and after procedures. Intracranial bleeding, thrombotic events, and recent major surgical procedures are contraindications to the initiation of fibrinolytic therapy because of the risk of clot breakdown. If possible, invasive procedures should not be performed while the patient is in a fibrinolytic state. If bleeding occurs during fibrinolytic therapy, the usual approach is to discontinue the drug, provide blood product support (including cryoprecipitate as a source of fibrinogen, if needed), and wait until the fibrinolytic state subsides.

Anticoagulant Drug Effects

Unfractionated Heparin

The heparinized patient who needs to undergo an invasive procedure usually can be managed by discontinuation of heparin approximately 6 hours beforehand. In urgent circumstances, neutralization with protamine sulfate can hasten preparation.

Low-Molecular-Weight Heparin (LMWH)

Invasive procedures, e.g., epidural anesthesia, have been complicated by bleeding in patients who stopped LMWH therapy shortly beforehand. Because of lack of a predictably effective antagonist, it is important to know the anticoagulated status of these patients before procedures are initiated. LMWH levels can be helpful in assessing residual presence of these agents and possible bleeding risk.

Warfarin

After discontinuation or tapering of warfarin therapy prior to an intended procedure, it may take 2 days or more for the PT to normalize. In emergent situations, use of FFP or prothrombin complex (if volume overload is a problem) can provide immediate correction by supplying the missing factors. Oral or intravenous vitamin K (1 to 5 mg) has its onset of effect approximately 8 to 12 hours after administration. The use of vitamin K in higher doses may lead to a subsequent state of warfarin resistance. The use of prothrombin complex may be associated with induction of a hypercoagulable state. This agent should be reserved, therefore, for urgent indications. During the period following tapering of warfarin, temporary heparinization can allow more flexible planning of interventions; heparin should be discontinued 6 hours before an invasive procedure. In patients unable to receive heparin or LMWH, warfarin can be continued throughout the perioperative period, with FFP used to override its anticoagulant effect, until it is safe to re-establish anticoagulation.

Direct Thrombin Inhibitors

Although their half-lives are short, hirudin and argatroban effects need to be monitored closely by following the PTT level once they have been discontinued to be certain that there is no lingering anticoagulation prior to an invasive procedure. Lack of an effective antagonist for either of these drugs makes rapid reversal of their effects problematic.

Drotrecogin Alfa (Xigris)45

Drotrecogin alfa is a recombinant activated protein C product with antithrombotic and profibrinolytic properties used in the management of patients with severe sepsis. Its major side effect is risk of bleeding, especially on day 1 of the 96-hour infusion period. Forty-three percent of serious bleeding events were procedure-related.46 Coexisting thrombocytopenia of 30,000/μL or less was a frequent concomitant risk factor. It is recommended that infusion of the drug be discontinued 2 hours prior to an anticipated procedure and that the platelet count be maintained above 30,000/μL.

The drug variably affects the PTT and assays of intrinsic pathway factors but has little effect on the PT or assays of factors that influence the PT.

If otherwise safe, it is recommended that a 12-hour period elapse after a significant procedure before cautiously restarting the infusion.