Thrombocytosis

Primary thrombocytosis is noted not only in essential thrombocythemia but also in polycythemia rubra vera, agnogenic myeloid metaplasia, and chronic myelogenous leukemia.⁶⁵ Related hemorrhage is usually from the skin or mucous membranes, and thrombosis involving the arterial and/or venous systems can also occur. Although an elevated whole-blood viscosity, intrinsic defects in platelets, and elevated platelet counts may all predispose to bleeding, it is difficult to assess risks in an asymptomatic patient. Therapy should be utilized for the symptomatic patient (eg, phlebotomy to correct polycythemia), which primarily consists of cytoreductive therapy of the underlying disorder.

Secondary or reactive thrombocytosis can also occur in other cancers but is characterized by only mild elevations of platelet counts without associated changes in platelet morphology and function.⁶⁶ This can sometimes serve as a diagnostic clue and is not associated with Trousseau syndrome of migratory thrombophlebitis. Indeed, persons with thrombocytosis over 400,000/mm³ should be evaluated for chronic or acute inflammatory disorders, acute hemorrhage, anemia, iron deficiency, or neoplasm. After splenectomy, the platelet count rises in the first week to levels of 1 × 10⁶/mm³ or more and then falls slowly to normal over 3 months.

Thrombocytopenia

Thrombocytopenia in cancer patients is usually attributable to treatment with chemotherapy and radiotherapy. Impaired production of platelets due to a decrease or absence of megakaryocytes is therefore the most common cause of thrombocytopenia in patients with cancer (Table 149-4). However, thrombocytopenia may also be due to splenic sequestration in patients who have splenomegaly as part of their primary neoplastic process. In this setting, increased numbers of megakaryocytes are evident unless extensive marrow infiltration is present. Immunemediated thrombocytopenia may also occur related to anti-human leukocyte antigen (anti-HLA) or antiplatelet-specific alloantibodies. Interestingly, combination chemotherapy and pulsed high-dose dexamethasone therapy have been shown to be beneficial in some patients in whom immune thrombocytopenia is refractory to corticosteroids and splenectomy.^67,68 Finally, thrombocytopenia may be related to diffuse intravascular coagulation (DIC), especially in patients with acute myelocytic leukemias, lymphomas, and carcinoma of lung, breast, gastrointestinal, or urologic origin. DIC most commonly complicates acute promyelocytic leukemia due to the presence of both thromboplastic material and fibrinolytic proteases in the promyelocytic subcellular components.⁶⁹

Table 149-4

Causes of Thrombocytopenia.

Abnormalities in Platelet Function

Platelet function can be abnormal in several chronic myeloproliferative disorders. Although most bleeding in patients with acute myelogenous leukemia is related to thrombocytopenia, intrinsic abnormalities in platelet function have been described including decreased platelet procoagulant activity and decreased aggregation and serotonin release responses to ADP, epinephrine, or collagen.⁷⁰ These defects may reflect the fact that megakaryocytes have originated from a leukemic stem cell. Platelet transfusions, coupled with treatment of the underlying disease, remain the mainstay of therapy.

Platelet dysfunction is evident in a fraction of patients with IgA myeloma or Waldenström macroglobulinemia, multiple myeloma, and monoclonal gammopathy of undetermined significance.⁷¹ In addition to thrombocytopenia, the following factors may also predispose to bleeding: hyperviscosity, acquired factor × deficiency in the setting of amyloidosis, a circulating heparin-like anticoagulant, fibrinolysis, and interference by myeloma protein with fibrin polymerization and with the function of other coagulation proteins. Cytoreductive therapy is primary, with plasmapheresis reserved for acute bleeding.

Platelet Transfusion Support

In 1910, fresh whole blood was first transfused to thrombocytopenic patients, resulting in a significant rise in the platelet count, hemostasis, and improvement of the bleeding time.⁷² In the 1950s, platelets were first used for the treatment of thrombocytopenia related to combination chemotherapeutic treatments of leukemias.⁷³ Data from the National Cancer Institute in the early 1960s clearly demonstrated that leukemia patients died of hemorrhage during induction of remission with chemotherapy and established the quantitative relationship between platelet count and hemorrhage.^44,74 It was shown that platelet therapy could modify the course of hemorrhage in both pediatric and adult settings, the only difference being the doses required. In the 1970s, studies in children and adults confirmed the efficacy of platelet transfusion to prevent rather than control hemorrhage.^75,76

Single- and Multiple-Donor Platelets

One unit of platelet concentrate is obtained from one unit of whole blood by centrifugation, and contains approximately 5.5 × 10¹⁰ platelets/unit. Concentrates from multiple (6 to 8) donors are pooled to produce a single component for transfusion. Individual platelet concentrates can be stored for up to 5 days prior to pooling and transfusion. Apheresis technology has permitted harvesting the equivalent of several platelet concentrates from a single donor during a single donation.^77,78 A single-donor platelet unit typically contains at least 3 × 10¹¹ platelets/unit. The major advantage of multiple-donor platelet concentrates is their availability, since they are derived from conventional whole blood donations. However, some studies suggest that alloimmunization can occur early in patients receiving multiple-donor platelets and that limiting the number of donors per transfusion may postpone the development of refractoriness to platelet transfusions in thrombocytopenic patients.^79,80 A second potential advantage of single-donor platelets stems from the decreased risk of infection when exposed to fewer donors. Indeed, the Retrovirus Epidemiology Donor Study demonstrated that the prevalence of any viral infection was 50% higher in whole blood donors as compared to apheresis donors.⁸¹ The use of single-donor platelets is increasing; they are generally are considered to be the platelet product of choice for patients being treated for malignancy.³⁹ Single-donor platelet collections are utilized to provide HLA-matched donors for alloimmunized recipients who have not responded to platelets from random donors.

Indications for Therapeutic and Prophylactic Platelet Transfusion

A minority of patients with cancer require platelet transfusions; however, platelets are more commonly transfused to patients with cancer than to patients with any other category of disease. The majority of platelet transfusions are given prophylactically to prevent bleeding, as opposed to therapeutically, to treat active bleeding.⁸² The appropriate indications for transfusion of platelets have been the subject of a National Institutes of Health Consensus Development Conference as well as a more recent American Society of Clinical Oncology clinical practice guideline .^83,83a At present, 6 to 8 units of random-donor platelet concentrates, or the equivalent platelet dose in the form of a single-donor apheresis platelet product, are often routinely transfused to cancer patients with platelet counts less than 10,000 to 20,000/mm³ to prevent hemorrhage. Gmür and colleagues demonstrated that the threshold for prophylactic platelet transfusion can safely be set at 5 × 10⁹/L in patients with acute leukemia without fever or bleeding manifestations and at 10 × 10⁹/L in patients with such signs.⁸⁴ More recently, they have shown this policy to be safe in the outpatient management of patients with aplastic anemia. Two important randomized trials have shown that the risk of major bleeding was similar whether 10 × 10⁹/L or 20 × 10⁹/L was used as the platelet-transfusion threshold in patients with acute leukemia, and that the lower threshold reduced platelet use.^85,86 The implications of thrombocytopenia as a risk factor for hemorrhage and therefore the timing and dose of prophylactic platelets may vary in different clinical settings. Moreover, the use of higher doses of platelets per transfusion may extend the interval until additional transfusions are needed.⁸⁷

The risk of bleeding at a given platelet count may vary in distinct clinical settings. For example, patients with thrombocytopenia due to acute myelocytic leukemia were reported to have increased bleeding at < l0,000/mm³ platelets, in contrast to patients with acute lymphoblastic leukemia, who had similar risk of hemorrhage at < 20,000/mm³ platelets.⁸⁸ Young platelets are more efficient at controlling hemorrhage, so the need for platelet transfusion will be greater if the count is falling after chemotherapy compared to a similar level during a rise from a nadir.⁸⁹ Patients with chronic thrombocytopenia due to decreased platelet production (ie, myelodysplastic disorders) may require transfusions, in contrast to patients with accelerated destruction but active production of platelets (ie, idiopathic thrombocytopenic purpura), who may not require routine platelet transfusions. Moreover, patients with chronic thrombocytopenia may tolerate lower absolute platelet counts without transfusion. In patients with abnormalities of platelet function, it is not the absolute platelet count but rather the number of functional platelets that is important for the prevention of bleeding. The bleeding time may aid in defining the risk of bleeding. Thus, it is difficult to define an absolute platelet threshold for transfusion for all patients, and both the timing and the dose of prophylactic platelet transfusion must therefore be determined on a clinical basis.^90–94

Clinical and Laboratory Assessment of the Effectiveness of Platelet Transfusion

The effectiveness of platelet transfusion can be assessed by laboratory parameters, the corrected increment (CCI) in platelet count 1 hour or 10 count to 15 minutes after transfusion and the bleeding time, as well as by the observed clinical outcome after transfusion.^95–98 The corrected platelet increment is defined as the increment in platelet counts from pre- to posttransfusion corrected for the number of units transfused and for the body surface area of the recipient. When low numbers of platelets are transfused, regression analysis of posttransfusion increments may provide a more accurate assessment than CCI.⁹⁹ Measurement of the bleeding time after transfusion serves as a measure of the number of functional platelets, particularly in patients known to have dysfunctional platelets. Techniques such as radiolabeling platelets can be utilized to diagnose and identify the site of accelerated platelet destruction, but the most important monitor of the effectiveness of platelet transfusion is critical clinical assessment for the presence and extent of hemorrhage.

Alloimmunization

Platelets bear HLA-A and -B but lack HLA-C and -DR Ags, and there is a high correlation between the development of lymphocytotoxic anti-HLA antibodies in the recipient and refractoriness to random-donor platelets.¹¹² Anti-HLA antibodies are most easily detectable using the patient's serum and a panel of lymphocytes representing known HLA specificities. The incidence and the timing of production of anti-HLA antibodies after platelet transfusion remain controversial and may vary with the recipient population. Most studies document alloimmunization in 50% to 90% of multitransfused patients.¹¹³ Some studies demonstrate that the rate of alloimmunization increases with the number of transfusions, whereas other reports find no relationship between the number of platelet transfusions given and the rate of alloimmunization.^114,115 Moreover, a fraction of patients with cancer never become sensitized.^116,117 Nonetheless, it is crucial to test for anti-HLA antibodies whenever recipients become refractory to random-donor platelet transfusion, since response to random-donor platelet transfusion is poor in the sensitized host, and HLA-matched or family-member platelets may be useful in this setting.^118–121

Yankee and colleagues first demonstrated that platelets obtained from HLA-identical siblings or from unrelated donors matched at the HLA-A and -B loci (grade A or B matches) could result in satisfactory posttransfusion increments in alloimmunized recipients who were refractory to random-donor platelet transfusions.^119,120 Subsequently, Duquesnoy and colleagues found that donors whose HLA antigens were the same (B match) or crossreactive with the patient's Ags (BX match) were equivalent.¹²²

The evaluation of donors for the same or crossreactive Ags became even more complex. For example, platelets from donors lacking HLA-A2 who bear one of two (grade C match) or three of four (grade D match) Ags not present in the recipient may have favorable posttransfusion outcomes in alloimmunized recipients; in contrast, platelets from HLA-A2+ donors who were HLA-A and -B matches were unsuccessful.¹²² Weak anti-HLA antibodies, which can cause platelet destruction in vivo, may not be detected by standard assays, and, alternatively, excellent platelet transfusion recoveries have been observed despite a positive lymphocytotoxicity cross-match.^112,121 Additional crossmatching techniques may be required in those 20% of sensitized patients who remain refractory even to HLA-matched platelets.

The recognition of the refractoriness associated with the development of anti-HLA antibodies led to attempts to either avoid or delay alloimmunization by modifying the platelets to be transfused. Since HLAs are expressed on leukocytes, investigators have attempted to (1) remove white blood cells from platelets or treat platelets with ultraviolet irradiation (UV) to abrogate the leukocyte antigen-presenting function; (2) use single rather than multiple donor platelets to minimize exposure to HLA ; and (3) transfuse only HLA-matched or leukocyte-depleted HLA-matched platelets. Leukocyte-reduced platelets have been prepared either by additional centrifugation or by use of filters, both of which deplete white cells by 2 to 3 logs, with varying associated losses of platelets. In particular, filtered platelets with no more than 10⁶ contaminating leukocytes reduced the likelihood of formation of anti-HLA antibodies; in contrast, transfusions of similar numbers of non-leukocyte-reduced platelets did sensitize recipients.^40,123 However, some reports suggest that neither bedside filtration nor UVB irradiation is effective in the prevention of alloimmunization to platelets.^124,125 In an animal model, prestorage, but not poststorage, leukodepletion did abrogate alloimmunization.¹²⁶

Use of single-donor platelets or of ultraviolet irradiation of platelets has been evaluated in small studies to prevent the development of alloimmunization.^79,80,127 It is important to note that the Trial to Reduce Alloimmunization to Platelets (TRAP) showed that the incidence of anti-HLA and platelet-specific antibodies alone, as well as the incidence of antibodies associated with platelet refractoriness, were reduced in leukemic recipients who received either filtered or UV-treated pooled random-donor concentrates or filtered single-donor platelets compared to similar patients who received nonfiltered pooled random-donor concentrates (Table 149-5).¹²⁸ Only 13% of patients developed platelet refractoriness associated with lymphocytotoxic antibodies, suggesting that the majority of unresponsiveness to transfusion is related to other factors. A meta-analysis of seven randomized controlled trials published between 1983 and 1995 supports the TRAP findings and suggests that these findings can be extended to other than leukemic patients and applied to the prevention of platelet transfusion refractoriness.^123,129

Table 149-5

Trial to Reduce Alloimmunization to Platelets in Patients Undergoing Myeloablative Therapy for Acute Myelocytic Leukemia.

When sensitized recipients remain refractory to HLA-matched platelets, attempts have been made to carry out additional cross-matching to identify more compatible platelet donors.^{109,130–132} These include leukoagglutination,⁵¹chromium-lysis, immunofluorescence, enzyme-linked immunosorbent assay (ELISA), assays of¹³¹iodine-labeled platelet-associated IgG, and platelet aggregometry. They have enjoyed variable success determining which platelets will be effective in refractory patients, and no one assay or combination of methods has been universally accepted to predict response to transfusion.

Potential mechanisms to explain unexpectedly suboptimal posttransfusion recoveries in patients receiving HLA-matched ABO-compatible platelet transfusions include unrecognized HLA specificities, circulating immune complexes, and antibodies to platelet-specific Ags. Plasma exchange or intravenous immunoglobulin therapy has been used prior to platelet transfusion in those patients who remain refractory to ABO compatible HLA-matched platelets, with mixed benefit.^133,134

The ABO blood group determinants, presumably absorbed from the plasma, are present on platelets with a structure similar to that on erythrocytes.¹³⁵ Major Ags of the Rh, Duffy, Kidd, Kell, and Lutheran systems, in contrast, are not expressed on the surface of human platelets.¹³⁶ Clinical studies have shown that platelet transfusions may induce the formation of isohemagglutinins because of a small number of contaminating erythrocytes, suggesting that ABO compatibility between donor and recipient is important.¹³⁷ However, in vitro studies have demonstrated that exposure of group A and AB platelets to the appropriate anti-A and anti-AB alloantibodies and complement does not cause ultrastructural damage to the platelets or induce platelet aggregation, suggesting that an ABO mismatch between donor and recipient is not an absolute contraindication to platelet transfusion.¹³⁸

The need for ABO compatibility between platelet donor and recipient has been directly examined in several studies. A randomized trial of ABO-compatible versus ABO-incompatible platelets demonstrated that the corrected increments after the first transfusion were equivalent, but higher increments were noted after subsequent transfusion of ABO-compatible platelets in a subset of patients.¹³⁹ Rh-incompatible platelets may not be immunogenic in certain patients with cancer.¹⁴⁰ Therefore, ABO matching of donor and recipient is unimportant for the majority of platelet transfusions. The fact that transfusions that are not group specific are considered safe and clinically effective greatly expands the availability of platelets for transfusion. However, in specific clinical settings, donor and recipient ABO compatibility may be significant. For example, if HLA-matched but ABO-mismatched platelet transfusion does not result in a satisfactory increment after transfusion, HLA- and ABO-matched platelets may be of benefit.¹⁴¹ Heal and colleagues have postulated that soluble plasma HLA-A and -B and ABO Ags contribute to the destruction of donor and sometimes recipient platelets by an immune complex or another “innocent bystander” mechanism.¹⁴² Finally, hemolytic reactions caused by very high titers of isohemagglutinins present within the plasma in the transfused platelet concentrates, have very rarely been noted , but this is seldom a reason to provide only ABO-compatible platelets.¹⁴³

Platelet Alloantigens

There is now a variety of alloantigens implicated in alloimmune thrombocytopenia that have important biologic functions: platelet activation-dependent fibrinogen binding (gpIIb:IIIa complex), platelet-von Willebrand factor interactions (gpIb:IX complex), and platelet/collagen interactions (gpIa:IIa complex).^144,145 Antibodies directed at these Ags can cause immune-mediated destruction of platelets in the presence or absence of anti-HLA antibodies. Sensitization to these Ags is rarely the cause of refractory thrombocytopenia in patients with cancer.^109,128

One clinical sequela of antibodies to platelet alloantigens is posttransfusion purpura, resulting from antibody directed at the PL^A1 Ag. In posttransfusion purpura, profound thrombocytopenia develops approximately 1 week after transfusion, primarily in women who have been immunized either by earlier pregnancies or, less frequently, by a previous transfusion.^146,147 Almost all patients have been PL^A1 negative with anti-PL^A1 antibody in their plasma at the time of thrombocytopenia. Partial exchange transfusion, plasmapheresis, and high-dose IV immunoglobulin have all been used to accelerate recovery from posttransfusion purpura.¹⁴⁷ The abrupt onset of thrombocytopenia after transfusion can be related to passive alloimmune thrombocytopenia to HPA-1a and HPA-5b, and transplantationassociated alloimmune thrombocytopenia can result from residual host cells producing antibodies against donor platelet Ags (HPA-1).¹⁴⁵

Alternatives to Platelet Transfusion

Studies to date have not shown reproducible effects of G-CSF or GM-CSF on recovery of platelets. However, numerous molecules have shown thrombopoietic effects in vitro and in vivo.^6,11,13 Examples of these include IL-1, IL-3, IL-6, IL-11, leukemia inhibitory factor (LIF), and c-kit ligand (KL). Additionally, with the tools of recombinant DNA, there is the ability to “cut and splice” novel recombinant hybrid molecules that may possess thrombopoietic activity, such as the PIXY 32l molecule, which is a recombinant hybrid of the coding regions from the human GM-CSF and IL-3 genes. Although these growth factors do elevate platelet counts in patients, studies in humans to date suggest that fever and constitutional symptoms may be dose limiting for IL-1, IL-3, and IL-6. Interleukin-11 is now approved by the US Food and Drug Administration (FDA) as a platelet growth factor.¹² A most exciting development is the recent isolation of thrombopoietin, the c-mpl ligand.^11,13 This promotes both megakaryocyte progenitor expansion and megakaryocyte differentiation, and its injection in nanogram quantities raises platelet levels of mice to three or four times the normal. Moreover, mice that are deficient in the c-mpl receptor have few megakaryocytes and are profoundly thrombocytopenic but have normal numbers of other hematopoietic cell types.¹⁴⁸ Thrombopoietin may be useful to enhance collection of PBSCs and for ex vivo expansion of stem cells, but its future development to enhance platelet recovery is unclear, since its use in normal donors induced antibodies and related thrombocytopenia.

Attempts have also been made to develop alternatives to platelet transfusion.¹⁴⁹ Klein and colleagues have administered lyophilized whole platelets to treat thrombocytopenic bleeding in pediatric patients and observed clinical improvement in 50% of patients, suggesting that morphologic integrity of platelets might not be essential for their in vivo function.¹⁵⁰ Platelets preserved in gelatin also retain hemostatic efficacy.¹⁵¹ New technologies for cryopreservation of platelets that achieve satisfactory platelet recovery after transfusion are being described.¹⁵² Aldehyde-fixed, dried, and rehydrated platelets adhere to exposed subendothelium and may therefore remain hemostatically active.¹⁵³ By crosslinking arginine-glycine-aspartic acid (RGD) tripeptides to erythrocytes, Coller and colleagues have produced “thromboerythrocytes” that interact with gpIIb:IIIa receptors on activated platelets.¹⁵⁴ Synthetic liposomes, composed of phospholipids and platelet glycoprotein complexes, can shorten prolonged bleeding times in thrombocytopenic animals.¹⁵⁵ Platelet membrane particles can normalize prolonged bleeding times in animal models and will soon undergo clinical testing.^156,157 Finally, synthetic phospholipids promote procoagulant activity on damaged vessels, suggesting their potential utility as platelet substitutes.^158,159 It is likely that cytokines and/or new products will be developed in the future to substitute for platelet transfusion therapy. The availability of new technology will then allow more effective strategies to both prevent thrombocytopenic hemorrhage and avoid the potential immunohematologic complications of platelet transfusion.