Thrombopoiesis, the generation of platelets from megakaryocytes in the bone marrow, is complex and incompletely understood. Megakaryocytes are large end-stage cells from which platelets bud. The earliest recognized committed progenitor is the burst-forming unit (BFU)-Meg.¹ Fig. 33.1 shows megakaryocyte development from stem cell stage through to platelet production. BFU-Megs develop into colony-forming unit (CFU)-Megs in the presence of growth factors thrombopoietin (TPO), interleukin-3 (IL-3) and IL-11. Megakaryocyte nuclei are large polyploid structures with chromosome contents between diploid (2N) to 64N. Such polyploid status is achieved through a process termed nuclear endoduplication: that is, successive doubling of chromosome content in the absence of cell division. Platelets are produced from megakaryocytes that are 8N or greater.² A single megakaryocyte can generate around 3000 platelets of which 20–30% are pooled in the spleen. In health the peripheral blood platelet count is 150–400 × 10⁹/l but this fluctuates, for example, following heavy exercise, ‘stress’, and around the menstrual cycle. This transient rise in platelet count may be caused by mobilization of platelets pooled in the spleen. There are also racial differences in the ‘normal’ platelet count and some Mediterranean populations have platelet counts as low as 80 × 10⁹/l in health. Platelets are produced at a rate of 35 000–44 000 per microliter per day³ and have a lifespan of 9–10 days.

Fig. 33.1 Hematopoietic lineages, showing megakaryocyte development and platelet production.

(From Provan D, Gribben JG 2000 (eds), Molecular Haematology. Blackwell Science, Oxford, with permission.)

Platelet structure and function

Normal platelet function requires the presence of key membrane proteins and two major types of cytoplasmic granules. Because of their limited metabolic activity, and the presence of polymorphic glycoproteins on their exterior surface, platelets are vulnerable to attack by many agents including drugs, toxins, viruses and the immune system. In addition, drugs or diseases that interfere with platelet function do so for the lifetime of the platelet and it may take several days for the effects of any interfering drugs to diminish once the offending agent is stopped.

Platelet membrane constituents

The platelet plasma membrane contains a variety of polymorphic glycoprotein molecules which interact with ligands such as coagulation factors, vessel wall components and other molecules in order to generate the primary hemostatic plug.

The integrin family of proteins

Integrins are key platelet membrane proteins and have been characterized on a large number of leukocytes and many other cells. For example, the fibronectin receptor on mammalian fibroblasts is one of the best characterized matrix receptor proteins ⁴ (Fig. 33.2). This receptor, in common with all other integrins, is a heterodimer consisting of a non-covalently associated complex of two distinct high-molecular-weight polypeptides, α and β. The receptor functions as a transmembrane linker which mediates the interaction between the intracellular actin cytoskeleton and fibronectin in the extracellular matrix. Like all integrins, the fibronectin receptor recognizes so-called RGD (Arg-Gly-Asp) sequences in matrix components. In platelets, integrins recognize and bind a variety of proteins in order to form a hemostatic plug through a complex mechanism of platelet adhesion, shape change and activation of the clotting pathway.

Fig. 33.2 Fibronectin receptor on mammalian fibroblasts: the platelet integrins share homology with the fibronectin receptor.

Platelet integrins and related proteins

Platelets contain five integrin α subunits and two β subunits producing: α_IIbβ₃, α_νβ₃, α₂β₁, α₅β₁ and α₆β₁.^5,^6,⁷ These are shown in Fig. 33.3. Further detail is provided by Table 33.1. These proteins are essential for normal platelet function, and are often the target of immunological attack in disorders such as idiopathic thrombocytopenic purpura (ITP).

Fig. 33.3 Cartoon of platelet membrane, showing platelet glycoproteins.

(Courtesy of Peg Berrity, Science Photo Library.)

Table 33.1 Platelet integrins and related proteins

Platelet alloantigens

These can be platelet-specific or shared with other cells. Important shared antigens include HLA class I and ABH (blood group A and B) antigens. Platelet-specific antigens fall into five well-defined human platelet antigen (HPA) groups (Table 33.2; see also Chapter 37): HPA-1, HPA-2, HPA-3, HPA-4 and HPA-5, each of which has an α and β allele. Each platelet allotype represents a single amino acid substitution in the platelet glycoprotein molecule. Because some platelet glycoproteins carry epitopes that play a major role in platelet function, platelet alloantibodies may not only cause thrombocytopenia but also affect primary hemostasis.

Table 33.2 Human platelet alloantigen system.

Cytoplasmic platelet constituents

Platelets contain two principal types of granule: dense bodies and α granules (Fig. 33.4). Dense bodies contain ADP, ATP, 5-HT, calcium and pyrophosphate. The α granules contain more than 300 releasable proteins including adhesion molecules, chemokines, cytokines, coagulation factors, fibrinolytic regulators, growth factors and pro- and anti-angiogenic factors such as PF4, β-thrombospondin, PDGF, vWF, fibrinogen, factor V and fibronectin. The contents of these granules are integral components of the platelet’s biological activities.

Fig. 33.4 Schematic representation of the platelet showing dense granules and alpha granules, with details of the contents of each.

Biological function of platelets

The primary role of the platelet is the prevention of blood loss from damaged tissues and vessels, i.e. primary hemostasis. This is achieved through platelet activation, adhesion, shape change and aggregation. Platelets may also play a role in the maintenance of vascular integrity by the constitutive release of cytokines and growth factors from their granules that bind to endothelial cell surface receptors resulting in intracellular signaling that stabilizes the molecular complexes that form the junctions between adjacent endothelial cells.⁸

Platelet adhesion (Fig. 33.5A)

This process involves interaction between the platelet integrins and the subendothelium. The main integrin involved in platelet adhesion is GPIIb/IIIa which can be activated via a number of signals including exposure to extracellular matrix components such as collagen, in addition to other key activators such as adenosine diphosphate (ADP). Known ligands for platelet integrins include molecules such as fibrinogen, fibronectin and von Willebrand factor (vWF).

Fig. 33.5 (A, B) (A) Platelet adhesion to subendothelium: the platelet membrane contains a variety of glycoprotein receptors to adhesive proteins. VN, vitronectin; Lm laminin receptor; ABP, actin-binding protein; a-A, a-actin; vWF, von Willebrand factor; TSP, thrombospondin; FN, fibronectin. (B) Platelet aggregation: the terminal event in platelet aggregation is thought to be the binding of fibrinogen to GPIIb/IIIa receptors on adjacent platelets. The fibrinogen bridge may be stabilized by thrombospondin bound to its receptor (GPIV).

(From Mackie IJ. The biology of hemostasis and thrombosis. In: Ledingham JGG, Warrell DA, editors. Concise Oxford Textbook of Medicine. Oxford: Oxford University Press; 2000, with permission.¹⁴²)

Platelet activation

If there is damage to the endothelial lining of blood vessels, platelets become exposed to subendothelial structures. Platelets release a variety of attractants and other chemicals and recruit other platelets which are attracted to the site of injury. Important platelet activators include ADP, thromboxane A₂ and thrombin. The activated platelets provide a strong procoagulant surface on which the main clotting cascade is amplified, with the final product being the generation of thrombin and the production of a fibrin clot, composed of platelets enmeshed within a fibrin network.

Platelet shape change

After adhesion to the subendothelium, platelets undergo a major shape change, from a discoid shape to one which is irregular, with projections (Fig. 33.6). This process is initially reversible but ultimately becomes irreversible.

Fig. 33.6 Platelet shape change. Scanning electron micrograph of platelet shape transformation during adhesion to the arterial wall. Assemblies of actin filaments create spike-like attachment points and arrange themselves in an expansive webwork that mantles the neighboring space.

(From Roth GJ. Developing relationships: arterial platelet adhesion, glycoprotein Ib, and leucine-rich glycoproteins. Blood. 1991;77:5–19, with permission.¹⁴³)

Platelet aggregation (Fig. 33.5B)

There are a number of compounds that can induce platelet aggregation, including ADP, collagen, thrombin, adrenaline, vasopressin and others. Aggregation occurs following the shift of GPIIb/IIIa, the dominant integrin on the platelet surface, from a low affinity to a high affinity state and is mediated through the binding of fibrinogen or von Willebrand factor to adjacent platelets.

Quantitative platelet abnormalities: thrombocytopenia

Thrombocytopenia, a reduction in platelet count, may be caused by:

• impaired production

• increased destruction

• altered distribution

• a combination of these.

Pseudothrombocytopenia

This describes patients in whom the peripheral blood platelet count is found to be spuriously low, and is caused by a variety of mechanisms. Platelet clumping caused by the anticoagulant ethylenediaminetetraacetic acid (EDTA) is the commonest cause, with an incidence of around 0.1%. EDTA-induced pseudothrombocytopenia is easily excluded by examination of a blood film which will confirm the presence of numerous large platelet clumps.^9,¹⁰ Other causes of pseudothrombocytopenia include the presence of giant platelets, which are not counted as platelets by modern automated counters, and platelet satellitism (platelets attach themselves to monocytes or granulocytes) (Fig. 33.7).

Fig. 33.7 Platelet satellitism. The figure shows a neutrophil surrounded by adherent platelets.

(From Lewis SM, Bain B, Bates I, Dacie J 2000 (eds), Practical Haematology, 9th edn. Churchill Livingstone, Edinburgh, with permission.)

Pooling of platelets in the spleen

This accounts for the thrombocytopenias seen in patients with hepatic cirrhosis and portal hypertension. The term hypersplenism is used for patients in whom there is thrombocytopenia through excessive splenic pooling of platelets.

Mechanism involved

In health, the spleen may pool up to one third of the total platelet mass, and in disease states this may rise to 90%.¹¹ Although the peripheral blood platelet count may only be a fraction of the normal range, the patient generally has an overall normal platelet mass since production is entirely normal but the low peripheral counts simply reflect a larger than normal mass of platelets pooled in the spleen.

Thrombocytopenia due to failure of platelet production

Acquired amegakaryocytic thrombocytopenia

This describes severe thrombocytopenia caused by a selective reduction in megakaryocytes in an otherwise normal bone marrow, and is the result of damage to the megakaryocyte stem cell. The disorder is analogous to acquired pure red-cell aplasia (PRCA). Acquired amegakaryocytic thrombocytopenia may be caused by drugs, toxins and connective tissue disorders.¹²^–¹⁴

Drug-induced megakaryocyte hypoplasia

All drugs that are myelosuppressive will inhibit megakaryocyte stem cells. Agents that are cell cycle phase-specific have the most profound effects on megakaryocytes (e.g. methotrexate and cytosine arabinoside). Other drugs may selectively ‘poison’ megakaryocytes, and induce profound hypoplasia, for example thiazide diuretics, ethanol and estrogens.

Bone marrow failure syndromes

The bone marrow failure (BMF) disorders are those in which there is a failure of bone marrow precursor (stem) cells in contrast to those disorders such as myelodysplastic syndromes in which normal or increased numbers of abnormal cells are produced, or those in which the survival of the cells is reduced. The BMF syndromes are a diverse groups of disorders with a common endpoint in which there is loss of hematopoietic stem cells. BMF can be acquired or inherited (see Table 33.3).

Table 33.3 Acquired bone marrow failure syndromes

Aplastic anemia

This is discussed in Chapter 13.

Thrombocytopenia due to increased platelet destruction

Causes may be immunologic or non-immunologic.

Non-immunologic causes of thrombocytopenia

Disseminated intravascular coagulation (DIC)

DIC is characterized by excessive activation of the coagulation cascade (see also Chapter 35). In most cases DIC is an acute event, but chronic DIC is well described, although clinically less important. The main problem faced in patients with DIC is bleeding, which may be mild but is often severe with generalized oozing from venepuncture sites, central lines and other indwelling cannulae, gastrointestinal and genitourinary tracts. Microthrombi are found in 5–10% of cases, often affecting digits, with resulting peripheral gangrene.

Pathogenesis

DIC is triggered by the release or exposure of tissue thromboplastins which contain a high concentration of phospholipids following trauma, surgery, mismatched blood transfusion and a variety of other triggers (Table 33.4). In addition to systemic activation of coagulation (leading to fibrin clot formation, organ failure and consumption of platelets and coagulation factors that may cause bleeding), there is dysregulation of natural anticoagulant systems and fibrinolysis.

Table 33.4 Triggers for disseminated intravascular coagulation (DIC)

Trauma	Including surgical
Dissemination of cancer cells	Malignancy, following administration of chemotherapy
Massive hemolysis	Post mismatched blood transfusion
Venoms	e.g. snake venoms
Endothelial injury	Gram-negative sepsis
Infections
Burns
Septicemia

TTP is a disseminated microangiopathy, first described by Moschowitz in 1924.¹⁷ It is uncommon, occurring primarily in adults and with an annual incidence in the US of 4–11 cases per million people. HUS has clinical and laboratory features that often overlap with those of TTP. However, HUS is more common in children, the renal abnormalities are more marked than in TTP (Table 33.5; see also Chapter 10) and the underlying pathogenetic mechanisms of the two conditions differ.

Table 33.5 Classification of thrombocytopenic purpura and hemolytic uremic syndrome.

Idiopathic TTP/HUS	Classic adult TTP and childhood non-verotoxin-associated HUS-TTP
Secondary TTP-HUS	Pregnancy-related	TTP, postpartum HUS
	Verotoxin-induced	Escherichia coli and Shigella dysenteriae I
		Childhood HUS
		Epidemic adult TTP-HUS
	Malignant disease	Especially metastatic carcinomas
	Drug-induced	Chemotherapy agents, e.g. mitomycin C, cisplatin, and other drugs
		Immunosuppressive agents, e.g. cyclosporin, quinine, ticlopidine
	Post-marrow/stem cell transplantation	Especially in conjunction with total body irradiation or high-dose (intensive) chemotherapy

HUS, hemolytic uremic syndrome; TTP, thrombocytopenic purpura.

(Modified from George JN, El-Harake M 1995. Thrombocytopenia due to enhanced platelet destruction by non-immunologic mechanisms. In: Beutler E, Lichtman MA, Coller BS, Kipps TJ (eds), Williams Hematology, 5th edn. McGraw-Hill, New York, 1290–1315)

Clinical features of TTP

TTP is characterized by the pentad:

• thrombocytopenia

• microangiopathic hemolysis

• neurologic symptoms and signs (typically fluctuating)

• renal function abnormalities

• fever.

TTP can affect any organ system but it is the involvement of the hematopoietic, renal and central nervous systems that leads to the typical clinical features.

The previously high mortality rate has been reduced through effective treatment and good supportive care. The disorder is sporadic, but appears to be more common in people of black origin and in obese individuals. It may develop in pregnancy and in this setting must be differentiated from pre-eclampsia, eclampsia and the HELLP (hemolysis, elevated liver enzymes and low platelets) syndrome.

TTP typically has a sudden onset, with fever and neurologic symptoms including paralysis, coma, fits and psychiatric disturbance. There is usually purpura and the clinical picture is one of a fluctuating course. The causes of TTP can be subclassified into: 1) congenital; 2) idiopathic; and 3) non-idiopathic.¹⁶

Laboratory investigations

A full blood count and blood film will show anemia (hemoglobin ~ 8–9 g/dl) with polychromasia and other evidence of hemolysis. The film will usually show red-cell fragments and thrombocytopenia (Fig. 33.8). There is usually evidence of hemoglobinemia reflecting the presence of intravascular hemolysis. Lactate dehydrogenase will be elevated. Clotting tests are generally normal although occasionally there may be features of DIC. The direct Coombs test is negative. Renal failure is uncommon but elevated serum creatinine is seen in about a third of cases.

Fig. 33.8 Peripheral blood film from a patient with thrombotic thrombocytopenic purpura (TTP). Note the many fragmented red cells, characteristic of this disorder.

(Figure kindly provided by Dr John Amess, Barts & The London NHS Trust.)

Pathogenesis

Failure to cleave large von Willebrand factor (vWF) multimers is thought to be crucial in the pathogenesis of TTP. Von Willebrand factor is synthesized in vascular endothelial cells and megakaryocytes and assembles into multimers linked by disulfide bonds which are stored in the Weibel–Palade bodies of endothelial cells and α-granules of platelets. A small proportion of these multimers is constitutively secreted by endothelial cells into the circulation, stabilizing factor VIII and mediating both platelet adhesion to sites of vascular injury and platelet aggregation in conditions of high shear. Endothelial cell stimulation causes unusually large vWF multimers to be secreted and these attach to the endothelial cell surface.¹⁶

A key recent discovery to understanding the pathogenesis of TTP has been the recognition of the role of a deficiency of a von Willebrand factor-cleaving protease, termed ADAMTS 13 (an acronym for a disintegrin and metalloprotease with thrombospondin-1-like domains). ADAMTS 13 cleaves the large vWF multimers secreted by endothelial cells to generate the range of multimer sizes that normally circulate in the blood. This function appears to be critical in preventing thrombosis in the microvasculature because hereditary or acquired deficiency of ADAMTS 13, resulting in an inability to cleave newly secreted von Willebrand factor multimers, leads to increased binding of vWF to platelets and to the disseminated platelet thrombi that are characteristic of TTP.

More than 50 mutations of the ADAMTS 13 gene have been described in patients with the rare congenital or familial form of TTP. About half of these patients present during childhood but the remainder present in adulthood when the acute event may be triggered by conditions such as infection or pregnancy.

In contrast, in the more common idiopathic TTP, ADAMTS 13 deficiency is caused by autoantibodies that inhibit its activity or enhance its clearance from the circulation. Mostly the antibodies are IgG.

There are a number of causes of non-idiopathic thrombotic microangiopathy. These include cancer, pregnancy (see below), certain drugs (quinine, mitomycin, ciclosporin, ticlopidine and clopidogrel) and hemopoietic progenitor cell transplantation.

The characteristic histological features include hyaline thrombi in terminal arterioles and capillaries. The thrombi are composed of platelets and are rich in von Willebrand factor but contain little fibrinogen or fibrin. Immunoglobulins and complement are also often present consistent with the autoimmune origin of TTP.

Diagnosis

There is no specific diagnostic test for TTP and in practice for management purposes the diagnosis is based on laboratory confirmation of thrombocytopenia and microangiopathic hemolytic anemia without an apparent alternative cause and with or without the characteristic clinical features described above.

Management

Plasma exchange with cryosupernatant or FFP is now regarded as the treatment of choice for TTP and should be instituted as soon as possible after diagnosis and continued on a daily basis with replacement of 1.0 to 1.5 times the predicted plasma volume of the patient. Plasma exchange should continue until there is normalization of the neurological state, the platelet count has been >150 × 10⁹/l for at least 2 days, the LDH is normal and the hemoglobin rising.¹⁸ The disease is highly variable in its course and it is difficult to predict the clinical outcome at the outset. Patients require intensive care nursing, and may require ventilation. Hemodialysis may be required along with plasma exchange. A short course of high dose corticosteroids has also been recommended along with aspirin when the platelet count has recovered to 50 × 10⁹/l.¹⁸ Platelet transfusions should be avoided except in cases of severe hemorrhage. Remission is achieved when the platelet count remains normal for 30 days after discontinuation of plasma exchange.¹⁹ Patients with severe deficiency of ADAMTS 13 appear to have an increased risk of relapse, most within a year. Rituximab has been used to try to reduce the risk of relapse.^20,²¹

Hemolytic uremic syndrome

The disorder was first described in 1955 by Gasser,²² who reported on five patients (all infants) with acute renal failure, who died of renal cortical necrosis. HUS is classified into D+HUS and D-HUS.¹⁶ D+HUS accounts for >90% of cases and is caused by enteric infection with Shiga toxin-producing bacteria, most commonly Escherichia coli O157:H7 or occasionally Shigella dysenteriae I.²² It can occur at all ages but is seen typically in young children.²²^–²⁴ HUS usually develops 4–6 days after the onset of diarrhea. There is a seasonal incidence, with the highest number of new cases reported in the summer months.^25,²⁶

D-HUS is uncommon and clinically heterogeneous. These individuals do not have an antecedent enteric infection with a Shiga toxin-producing organism and have a relatively poor prognosis with a mortality rate of about 25%.¹⁶

The histopathological features of HUS are different from those of TTP. The kidneys are preferentially involved and the thrombi are composed primarily of fibrin with few platelets and little vWF.¹⁶ Endothelial damage is pronounced in childhood D+HUS. Marked destruction of the renal cortex may occur and glomerular thrombosis is characteristic with an appearance suggesting capillary congestion rather than ischemia.¹⁶ D-HUS has been little studied but the renal lesions appear to differ with more pronounced mesangial involvement and less glomerular thrombosis than in D+HUS.¹⁶

Pathogenesis. Most D+HUS results from E. coli O157 infections caused by food-borne outbreaks from cattle. Unwashed fruit is another source of some outbreaks. Subunits of Shiga toxin produced by the bacteria bind to microvascular endothelial cells particularly in the kidney and on monocytes and platelets. The toxin is subsequently internalized and inhibits protein synthesis leading to cell death. Cytokine release is stimulated further increasing Shiga toxin binding. Resulting endothelial cell damage has prothrombotic consequences that lead to the histopathological changes described above.¹⁶

D-HUS in contrast is associated with complement dysregulation and mutations in one of the complement regulatory proteins (factor H, factor I, membrane cofactor protein (MCP) and factor B) have been described in about 50% of these patients.¹⁶ In some patients an autoantibody against factor H has been recognized and in about 5% of cases mutations that impair thrombomodulin function have been identified.²⁷

Clinical features. Children typically present with diarrhea (often bloody), vomiting and abdominal pain. Because of the large volume of fluid loss, patients are often oliguric or anuric at presentation. Fever, hypertension and fits are also features.

Laboratory findings. There is evidence of acute renal failure and features of microangiopathic hemolysis.²⁸ The blood count will generally show an elevated WBC, most usually neutrophilia; the prognosis is worst for patients with total neutrophil counts exceeding 20 × 10⁹/l.²⁵

Management. Around half of the children affected will require hemodialysis for between 1 and 2 weeks. Unlike in adult TTP, plasma exchange using FFP replacement is not often required in children.

Outlook. Unlike adult TTP, childhood D+HUS is rarely fatal and the mortality is around 3–5%. The outlook is less favorable in D-HUS. In cases with end-stage renal failure that have undergone renal transplantation, the outcome appears to correlate with the cause of the complement dysregulation. Recurrence of HUS has not been seen in transplanted kidneys in patients with MCP mutations whereas the condition did recur following renal transplantation in patients with factor H or factor I mutations, perhaps because these latter two factors are made in the liver. Combined liver and kidney transplants may be successful in some of these cases.¹⁶

Pre-eclampsia and HELLP syndrome

In addition to TTP and HUS there are a number of causes of thrombotic microangiopathy that can occur during pregnancy, in particular pre-eclampsia/eclampsia and the HELLP syndrome. Differentiation between these different conditions can be problematic. Pre-eclampsia occurs after 20 weeks of pregnancy and is associated with hypertension, edema, sodium retention, proteinuria and DIC. It may progress to eclampsia which is characterized by convulsions.²⁹ Previously, pre-eclampsia/eclampsia and HELLP have been considered distinct disorders but it seems more likely that they represent a spectrum of a pathologic process.

HELLP syndrome

This disorder, characterized by hemolysis, elevated liver enzymes and low platelet count, occurs in 0.5–0.9% of all pregnancies,^30,³¹ and is associated with a mortality rate of 1–4%.³² It occurs in up to 10% of cases of severe pre-eclampsia. Perinatal mortality is high and may reach 10–20%.³²

Clinical features. HELLP is associated with generalized weakness, nausea and vomiting, right upper quadrant pain, headache and visual upset.³³ In some cases HELLP may occur following delivery.³⁶

Pathophysiology. There may be abnormalities of the placental vessels with resulting placental ischemia. This is associated with the systemic release of a thromboxanes, angiotensin, tumor necrosis factor-α (TNF-α) and other procoagulant proteins.^31,³³ DIC may result leading to thrombi which threaten major end-organs including placenta, renal, hepatic and central nervous systems. The thrombi lead to endothelial damage and there is microangiopathic hemolytic anemia (MAHA) and failure of the organs affected by the thrombotic process. Liver failure and occasionally liver rupture may result.³⁴

Management. Prompt delivery, and control of the hypertension are required, in addition to correcting the factor deficiencies caused by the DIC. Corticosteroids have been used as well as plasma exchange and plasmapheresis.³⁵ Maternal deaths, where these occur, are usually due to uncontrolled DIC. Cases in which thrombotic microangiopathy persists post-partum should be considered for plasma exchange.¹⁸

Thrombocytopenia caused by massive blood transfusion

Massive transfusion whereby a patient’s total blood volume is replaced over a short time period (24 h or less) may lead to thrombocytopenia, and although the degree of thrombocytopenia is largely related to the amount of blood transfused, the mechanism is not purely dilutional. Patients receiving 15 units of red cells within a 24-hour period develop mild thrombocytopenia, with platelet counts between 47 and 100 × 10⁹/l, whereas patients transfused with 20 units of red cells over the same period develop more pronounced thrombocytopenia (25–61 × 10⁹/l).^36,³⁷

Management

Severe thrombocytopenia can be prevented by administering platelet concentrates prophylactically to patients receiving more than 20 units of red cells within a 24-hour time period.

Liver disease

Thrombocytopenia secondary to alcohol

Thrombocytopenia occurring in alcoholic patients may be caused by a variety of mechanisms including cirrhosis, splenomegaly and folic acid deficiency. However, thrombocytopenia may be found in the absence of any or all of these pathologies, and is probably due to the direct toxic effects of alcohol on the bone marrow itself, since ethanol is a poison and can suppress the production of platelets by the marrow.^38,³⁹

Thrombocytopenia in hepatocellular failure

In addition to the other coagulopathies induced by liver failure, thrombocytopenia is often present and generally reflects the hypersplenism that occurs in portal hypertension. In fulminant hepatic failure, there are abnormalities of both platelet structure and function. On laboratory testing there may be evidence of mild DIC, although this is not generally of major clinical importance.⁴⁰ In addition, patients with chronic liver disease often have elevated levels of fibrin degradation products which interfere with platelet aggregation.

Thrombocytopenia caused by infection

There are numerous infections caused by a wide variety of pathogenic bacteria, fungi, viruses and protozoa that result in thrombocytopenia in humans. The mechanism underlying the thrombocytopenia is variable. In many cases there is suppression of marrow function, and this is particularly the case with viral infections and accounts for most cases of mild thrombocytopenia. Implicated viruses include mumps,⁴¹ varicella,⁴² EBV,⁴³ rubella⁴⁴ and many others.

Mechanism of thrombocytopenia

There is a variety of mechanisms involved in thrombocytopenia induced by infection. Following measles infection in children, there is a reduction in marrow megakaryocytes and by day 3 of the infection many of the megakaryocytes have vacuoles within the nucleus and cytoplasm.⁴⁵

Thrombocytopenia induced by HIV infection

HIV infection probably induces thrombocytopenia through a variety of mechanisms, including immune-mediated thrombocytopenia with reduced platelet life span,⁴⁶^–⁴⁸ and through direct infection of megakaryocytes themselves. This may occur at any stage in the course of HIV infection, and 40% of HIV positive individuals develop thrombocytopenia at some point in the illness.⁴⁹^–⁵¹ Unlike classical idiopathic thrombocytopenic purpura (ITP), males and females are affected equally and thrombocytopenia may be the presenting feature in 10% of HIV positive people.⁵²

Thrombocytopenia due to hematinic deficiencies

Thrombocytopenia may be a feature of vitamin B₁₂ or folate deficiency, and mild thrombocytopenia is found in 20% of patients with megaloblastic anemia caused by vitamin B₁₂ deficiency.⁵³

Immunologic causes of thrombocytopenia

Here we have restricted our discussion to the more important immune-mediated causes of thrombocytopenia, namely neonatal alloimmune thrombocytopenia (NAIT), post-transfusion purpura (PTP), ITP, drug-induced thrombocytopenia and heparin-induced thrombocytopenia (HIT) (Table 33.6). Since the targets involved in several of these disorders are human platelet alloantigens we outline briefly their salient features.

Table 33.6 Disorders associated with immune-mediated thrombocytopenia.

Autoimmune	Idiopathic (ITP) Secondary immune Autoimmune disorders, e.g. SLE, rheumatoid, thyroid disease Lymphoproliferative disorders, e.g. CLL, NHL Cancer, e.g. solid tumors Miscellaneous, e.g. post-BMT, chemotherapy Viral infection e.g. HIV, measles, mumps, rubella, EBV, varicella Drug-induced
Alloimmune	Post-transfusion purpura Neonatal alloimmune thrombocytopenia

BMT, bone-marrow transfer; CLL, chronic lymphocytic leukemia; EBV, Epstein–Barr virus; HIV, human immunodeficiency virus; ITP, idiopathic thrombocytopenic purpura, SLE, systemic lupus erythematosus; NHL, Non-Hodgkin lymphoma.

(From Chong BH 1998, Diagnosis treatment and pathophysiology of autoimmune thrombocytopenias. Critical Reviews in Oncology Hematology 20(3):271–296)

Molecular basis of HPA antigens

Platelets have a variety of cell surface antigens, some of which are shared by other cells, such as HLA class I, and blood groups A and B. Others are platelet-specific and are not found on any other type of cell. To date there are 19 alloantigen systems described on platelets, all of which map to membrane proteins (Table 33.2). Eleven of the 19 are carried on GPIIb/IIIa (α_IIbβ₃ integrin heterodimer), three are on GPIb/IX/V, two on GPIa/IIa (α₂β₁), and one on each of GPIV, GPV and CD109. The molecular basis for most of these is now elucidated (see Chapter 37).

Antibody response to non-self HPA antigens

The difference between self and non-self HPAs is determined by a single amino acid substitution, and for this reason HPAs are not particularly immunogenic, in comparison with, for example, Rh(D) and HLA class I antigens. The most clinically important HPAs are HPA-1a and HPA-5a, since antibodies to these antigens account for 95% of cases of NAIT. Alloantibodies to the other HPAs also occur but at much lower frequency.

Alloimmunization against platelet-specific antigens is associated with three major clinical syndromes: neonatal alloimmune thrombocytopenia, post-transfusion purpura and refractoriness to platelet transfusions. Only the first two are discussed here.

Alloantibody-mediated thrombocytopenia

Neonatal alloimmune thrombocytopenia

Antibodies to HPA occur in 1 in 365 pregnancies and cause severe thrombocytopenia in 1 in 1100 neonates at term, accounting for around 20% of cases of thrombocytopenia in neonates. NAIT occurs when there is feto-maternal incompatibility for HPA. The condition was first described by van Loghem et al in 1959.⁵⁴

During pregnancy, or following a blood transfusion, the mother becomes sensitized and produces alloantibodies against HPA. Maternal IgG anti-HPA crosses the placenta resulting in premature destruction of fetal platelets.

Antigens involved in NAIT

In Caucasian females NAIT is most commonly due to anti-HPA-1a (98% of the population are HPA-1a⁺). Other implicated antigens include HPA-1b,⁵⁵ 101 HPA-2a,⁵⁶ HPA-3a,⁵⁷ HPA-3b,⁵⁸ HPA-4a,⁵⁹ HPA-4b,⁶⁰ HPA-5b⁶¹ and HPA-5a.⁶²

Diagnosis of NAIT

In contrast to maternal ITP, the mother in NAIT has a normal platelet count and is completely well. The neonate is generally asymptomatic at birth, apart from a low platelet count. In cases where neonates are symptomatic the platelet count is generally <30 × 10⁹/l. If no treatment is given, the baby’s platelet count may remain low for up to 2 weeks, and occasionally longer.

Maternal serum is tested against the father’s platelets in addition to normal control platelets. Occasionally, the maternal antibody titer is low at delivery and difficulty to detect using the above techniques. However, there are techniques available for the detection of low-titer maternal antibodies.⁶³

Clinical features

Unlike hemolytic disease of the newborn where the firstborn child is unaffected, in NAIT the firstborn child may be affected. Recurrence in subsequent pregnancies is common, if there is feto-maternal incompatibility. The newborn infant is usually normal at birth but most will have some degree of bleeding due to the marked thrombocytopenia.

Management

The NAIT may be very mild and require no therapy apart from close monitoring of the neonate’s platelet count. If treatment is required there are a variety of options available including:

1. Corticosteroids, although it is uncertain whether corticosteroids reduce the period of thrombocytopenia.⁶⁴

2. Platelet transfusions, which are useful if there is serious bleeding. These can be random donor platelets but more ideally should be antigen-negative (HPA-1a and HPA-5a negative). The mother’s platelets may be used, but these should be washed to remove antibody which is present in the mother’s plasma.⁶⁵

3. Intravenous immunoglobulin.⁶⁶

4. Exchange transfusion, which removes antibody and reduces the period of thrombocytopenia.⁶⁷

Future pregnancies

This depends on the father’s genotype for the particular HPA. For example, if he is homozygous then each pregnancy will be affected. In cases where the father is heterozygous, 50% of pregnancies will be affected and the fetus can be typed using DNA obtained from chorionic villus sampling.⁶⁸

Post-transfusion purpura (PTP)

This rare disorder occurs 7–10 days following a red-cell transfusion in recipients who possess alloantibodies against platelet antigens of the donor. In addition to destroying the incoming platelets, the alloantibody also mediates destruction of the recipient’s own platelets (i.e. lacking the target antigen).⁶⁹

Pathogenetic basis

Platelet alloantibodies may be present in the recipient against any one of the six major platelet antigens: anti-HPA-1a, anti-HPA-1b,⁷⁰ anti-HPA3a,⁷¹ anti-bak-b,⁷² anti-HPA-4a or anti-HPA-5b.⁷³ In most cases the antibody has specificity for HPA-1a (2% of the population is HPA-1a–).⁷⁴

Clinical features

PTP affects multiparous women, although it has been reported in males,⁷⁰ between the ages of 16 and 80 years. Patients have usually been exposed to platelet antigens through either pregnancy or transfusion or both. The platelet antigen most commonly involved is HPA-1a.

Laboratory features

Patients usually have a platelet count less than 10 × 10⁹/l and the bone marrow will show normal or increased numbers of megakaryocytes. Diagnosis of PTP requires the demonstration of the presence of platelet-specific alloantibodies in the serum of the affected patient. Most patients with PTP are HPA-1a⁻ and the presence of HPA-1a platelets in the transfused blood boosts the primary response.^70,^72,^75,⁷⁶ The antiplatelet antibodies produced are mainly of IgG1 and IgG3 class. Why patients destroy their own platelets which are HPA-1a⁻ is unclear but is believed to involve IgG3. Possibly HPA-1a⁺ platelets release HPA-1a which combines with the newly-formed anti-HPA-1a. This complex is absorbed on to the surface of the patient’s HPA-1a⁻ platelets.

Management

Corticosteroids may help the purpura but do not appear to be effective in increasing the platelet count. Intravenous immunoglobulin (IVIg) is the mainstay of treatment.⁶⁹ Occasionally, plasma exchange may be required. If platelet are required these should be HPA-1a⁻. Fatal intracranial hemorrhage occurs in 10% but most patients recover within 1–6 weeks.⁶⁹

Autoantibody-mediated thrombocytopenia

This includes acute and chronic ITP, in addition to thrombocytopenia secondary to other autoimmune disorders, lymphoproliferative diseases or drugs.

Immune thrombocytopenic purpura (ITP)

In ITP platelets are opsonized with antiplatelet autoantibodies and removed prematurely by the reticuloendothelial system (RES) leading to a reduced peripheral blood platelet count. In addition, in many patients platelet production is reduced. ITP is therefore a disorder of platelet destruction and relative failure of production.⁷⁷ The etiology of ITP is obscure and the clinical course is variable and unpredictable. ITP has an incidence of 5.8–6.6 new cases per million population per year in the US,⁷⁸ with a similar incidence in the UK. Childhood ITP is generally seasonal and typically follows a trivial viral infection or vaccination, and in most cases is transient, requiring no treatment with spontaneous recovery in 80% of cases. In the adult (generally chronic) form there is usually no obvious antecedent illness and most patients have chronic thrombocytopenia; spontaneous recovery is uncommon.⁷⁹ In most cases of adult ITP the platelet glycoprotein (GP) antigen targets are GPIIb/IIIa and GPIb/IX.⁸⁰

The terminology used in ITP has been updated following a consensus meeting of international experts. The term ‘acute ITP’ is no longer in use, and chronic ITP, which used to define ITP persisting longer than 6 months, now refers to ITP lasting beyond 12 months ⁸¹ (Fig. 33.9).

Fig. 33.9 This schematic shows the new terminology used for the phases of ITP. The disease begins as newly diagnosed ITP which may last beyond 3 months after which it is called persistent ITP. If ITP lasts longer than 12 months it is known as chronic ITP.

Self-limiting ITP is the most common form of ITP found in children, with an annual incidence of between 3 and 8 per 100 000 per year.^82,⁸³

Pathophysiology

It is believed that this type of ITP is most likely due to an inappropriate immune response to an environmental trigger; the nature of this trigger is not yet identified.^84,⁸⁵ The disorder may represent an abnormality of antigen-presenting cells, with an increase in the numbers of CD4⁺ and CD8⁺ cells. The platelets are rapidly destroyed by the immune complexes that bind to the Fc receptors on the platelets, or due to autoantibodies that bind to the antigenic site on the platelets. Platelets that are coated with antibody or immune complexes are rapidly cleared by the reticuloendothelial system.

Clinical features

The children affected by ITP are, in general, well. The disorder is commonest in children between the ages of 2 and 5 years and, unlike adult chronic ITP, there is no sex predominance. A viral illness predates the development of ITP in most cases of childhood ITP. Physical signs of thrombocytopenia usually take the form of bruising or petechial hemorrhage. Splenomegaly, hepatomegaly and lymphadenopathy are not features of newly diagnosed ITP. A recent UK study reported that in some 76% of children with ITP the disorder was mild with bruising and occasional epistaxis.⁸² Around 4% had no bleeding symptoms and only 3% had severe bleeding from the gastrointestinal tract, nose or vagina requiring hospitalization. In around 15% of children the disease persisted longer than 6 months and fell into the ‘chronic ITP’ group. The chronic form was found to be commoner in older children and in females.

Laboratory investigation

The blood count will show an isolated thrombocytopenia but should otherwise be normal. The platelet count is often less than 10 × 10⁹/l.⁸² Anemia may be present in cases in which there has been significant bleeding.

Bone marrow aspirates

The need to carry out a bone marrow aspirate is debatable, but when this is carried out there tends to be normal or increased numbers of megakaryocytes confirming that the thrombocytopenia is due to peripheral destruction rather than a failure of production. However, if there is any suspicion regarding the presence of underlying disease accounting for the thrombocytopenia or the child has features not typical of ITP then performing a bone marrow aspirate may be indicated.

Thrombopoietin (TPO) levels

These are not measured routinely, but in cases where TPO has been assayed in ITP the levels tend not to be elevated.⁸⁶^–⁸⁸

Management

Most children with ITP need no therapy. There is no set threshold for medical intervention and what constitutes a ‘safe’ platelet count is not known.⁸⁹ From studies of the natural history of patients with newly diagnosed ITP, we know that patients with this disorder have far fewer bleeding problems than those patients with comparable platelet counts caused by other diseases, such as acute leukemia or aplastic anemia. This, in part, reflects the fact that platelet function in ITP is extremely good, with a large proportion of reticulated (young) platelets in the peripheral blood.⁹⁰ If therapy is required to elevate the platelet count then the options comprise oral corticosteroids, intravenous immunoglobulin and splenectomy.

Chronic immune thrombocytopenic purpura (ITP)

Clinical features of ITP

This is the most common form of ITP in adults. Patients may be asymptomatic or may have purpura, bruising or mucosal bleeding including gum bleeding, retinal hemorrhage, epistaxis, melena or menorrhagia (Fig. 33.10). The degree of bleeding is largely dependent on the platelet count, and patients with platelet counts below 10 × 10⁹/l are at greatest risk of bleeding. Splenomegaly is not a feature of ITP and if present, tends to suggest a diagnosis other than ITP.

Fig. 33.10 Petechial hemorrhages on the shin of a patient with severe chronic ITP.

Pathophysiology of ITP

ITP is an autoimmune disease characterized by increased platelet destruction due to the presence of antiplatelet antibodies. This results in increased platelet clearance by the RES. Several investigators have demonstrated specific autoantibodies against platelet membrane antigens, thus confirming the autoimmune nature of the disorder.^91,⁹² The cause is unknown but it appears likely that in a genetically predisposed individual, a trigger such as infection leads to loss of self-tolerance⁹³ (Fig. 33.11).

Fig. 33.11 Possible mechanism involved in the development of autoimmune disease such as ITP. Self-reactive lymphocytes are generated during B-cell development. These are normally eliminated but some may ‘leak’ into the periphery. This, by itself, is probably insufficient to lead to autoimmunity, and requires a specific genetic predisposition in addition to the ‘wrong environment’ (e.g. viral infection) to produce the autoimmune phenotype.

(From Mackay IR. Tolerance and autoimmunity. British Medical Journal. 2000;321:93–96, with permission.⁹³)

Glycoprotein (GP)-specific autoantibodies may be important in the pathogenesis of chronic ITP;⁹⁴ from available data GPIIb/IIIa appear to play a major role in the development of chronic ITP in 30–40% of cases.^62,⁹⁵ Fig. 33.12 illustrates the structure of GPIIb/IIIa schematically. Previous investigators have looked for autoantigenic epitopes on the GPIIb/IIIa molecule using competitive binding between human autoantibodies and mouse monoclonal antibodies (MoAbs).^96,⁹⁷

Fig. 33.12 Schematic representation of GPIIb/IIIa.

(From Provan D, Gribben JG (eds) 2000, Molecular Haematology. Blackwell Science, Oxford, with permission.)

Implicated epitopes

Kekomaki et al. have shown that the 33kDa chymotryptic core fragment of IIIa is a frequent target in chronic ITP.⁹⁸ Fujisawa and colleagues have used synthetic peptides corresponding to IIIa sequences and have shown that in five of 13 sera from patients with chronic ITP binding was to residues 721–744 or 742–762, corresponding to the carboxy terminal of IIIa.⁹⁷

GP-specific human MoAbs have been developed as important tools in the search for GP autoepitopes in chronic ITP.^94,⁹⁹ Some investigators have localized certain autoantigenic epitopes to regions of IIb or IIIa but blocking experiments using murine MoAbs have produced contradictory data in terms of homogeneity of the IIb/IIIa antigenic repertoire.⁹⁶ Only a few cryptic epitopes on IIb/IIIa have been recognized using GP-specific human MoAbs.

Antibody class

The autoantibodies involved in ITP are generally IgG, but IgA and IgM autoantibodies have also been reported.⁶²

Diagnosis

Despite advances in serologic and other techniques the diagnosis of ITP remains largely clinical, and one of exclusion. Secondary causes include systemic lupus erythematosus (SLE), lymphoproliferative disease, HIV infection and others. Standard investigative tests include full blood count which will confirm the presence of isolated thrombocytopenia, blood film to ensure there are no red cell fragments or other diseases such as leukemia or parasitic infections, and an autoimmune profile, to exclude a secondary cause for the thrombocytopenia. A bone marrow aspirate is often carried out in adults, but not usually in children, and will usually show normal or increased numbers of megakaryocytes in an otherwise normal marrow (Fig. 33.13).

Fig. 33.13 Bone-marrow aspirate in patient with ITP showing increased numbers of megakaryocytes, some of which are young with fewer nuclei than mature megakaryocytes.

(Figure kindly provided by Dr John Amess, Barts & The London NHS Trust.)

Standard first-line therapy

Therapy is seldom necessary for patients whose platelet counts exceed 20–30 × 10⁹/l and in whom there are few spontaneous bleeding episodes ⁴⁷ unless they are undergoing any procedure likely to induce blood loss.¹⁰⁰ Standard treatments, including oral prednisolone,³² IVIg,^79,^101,¹⁰² and splenectomy, will elevate the platelet count sufficiently in the majority of adults. However, some 20–25% of adults with ITP are refractory to first-line therapy.

Chronic ITP failing to respond to therapy

This defines those patients who fail to respond to first-line treatment or require unacceptably high doses of corticosteroids to maintain a safe platelet count. A number of agents have been used as second-line therapy for ITP including high-dose steroids, high-dose IVIg, intravenous anti-D, vinca alkaloids, danazol, azathioprine, combination chemotherapy and dapsone. An excellent summary is provided by McMillan.⁷⁸

Experimental therapies

For those who fail to respond to standard first- and second-line therapy and who require treatment the options are limited and include: interferon-α,⁷⁹ cyclosporin A, CAMPATH 1H, and protein A columns.¹⁰³

Thrombopoietin receptor agonists (TPO-mimetics)

Traditionally anti-platelet autoantibodies accelerating platelet clearance from the peripheral circulation have been recognized as the primary pathophysiologic mechanism in chronic immune thrombocytopenia (ITP). Recently, increasing evidence supports the coexistence of insufficient megakaryopoiesis. Inadequate low thrombopoietin (TPO) levels are associated with insufficient proliferation and differentiation of megakaryocytes, decreased proplatelet formation and subsequent platelet release.¹⁰⁴ The successful isolation and cloning of thrombopoietin (TPO) in the mid-1990s and identification of its key role in platelet production was a major breakthrough, rapidly followed by the development of the recombinant thrombopoietins, recombinant human TPO and a pegylated truncated product, PEG-rHuMGDF. Both agents increased platelet counts but development was halted because of the development of antibodies that cross-reacted with native TPO, resulting in prolonged treatment-refractory thrombocytopenia. Second generation thrombopoietin receptor agonists were developed with no sequence homology to native TPO and these have now been extensively used with no significant side-effects.¹⁰⁵ Two agents are currently available and have been used in early clinical studies and are licensed in some countries at the time of writing. Romiplostim (AMG 531, Nplate®; Amgen, Thousand Oaks, CA, USA) is a novel recombinant thrombopoiesis-stimulating Fc-peptide fusion protein (‘peptibody’) and eltrombopag, a non-peptide, synthetic TPO-receptor agonist (GSK, London, UK; marketed as Promacta in the US and Revolade in the UK). These two novel activators of thrombopoietin receptors have been used in several phase III studies and both agents demonstrate increase of platelet counts in about 80% of chronic ITP patients within 2–3 weeks. These agents substantially broaden the therapeutic options for patients with chronic ITP although long-term results are still pending.^106,¹⁰⁷

Secondary immune thrombocytopenia

Immune-mediated thrombocytopenia, similar to ITP, may occur in patients with other underlying autoimmune diseases such as systemic lupus erythematosus (SLE). Table 33.7 summarizes the main causes of immune-mediated thrombocytopenia caused by autoantibodies.

Table 33.7 Causes of thrombocytopenia due to autoantibodies

Idiopathic	ITP (newly diagnosed, peristent and chronic)
Secondary to other autoimmune or inflammatory disorders	SLE Other autoimmune disorders Lymphoproliferative disease, e.g. CLL Post-BMT
Associated with viral infections	e.g. HIV
Drug-induced	e.g. quinine, quinidine, heparin, gold salts

BMT, bone-marrow transfer; CLL, chronic lymphocytic leukemia; HIV, human immunodeficiency virus; ITP, idiopathic thrombocytopenic purpura; SLE, systemic lupus erythematosus.

Drug-induced thrombocytopenia

Drugs may induce thrombocytopenia through a variety of mechanisms, both immune and non-immune. Only drug-induced immune thrombocytopenia is discussed here; drug-induced qualitative abnormalities are discussed later in the chapter. In drug-induced immune-mediated thrombocytopenia, drug-dependent antibodies most commonly recognize epitopes of glycoproteins Ib-IX and IIb-IIIa.¹⁰⁸ Fig. 33.14 shows the pathways involved in prostaglandin metabolism highlighting the sites of action of aspirin and non-steroidal anti-inflammatory drugs (NSAIDs).

Fig. 33.14 Arachidonic acid metabolism and the generation of thromboxane A₂ and PGI₂. This is inhibited by aspirin and other NSAIDs leading to excessive bleeding.

Clinical features

Thrombocytopenia is an uncommon drug-related side-effect. The predominant features are of bleeding or bruising. The number of drugs suspected of causing thrombocytopenia is quite large but strong evidence is only available for a few of these including: quinine,^109,¹¹⁰ quinidine,¹¹¹ sulfonamides,¹¹² trimethoprim, abciximab, and gold salts.¹¹³ The incidence of thrombocytopenia in the cases of abciximab is relatively common at 0.5–1% after first exposure rising to 10–14% after second exposure.¹¹⁴ The incidence is also high with gold at about 1%.¹¹⁴

The onset of thrombocytopenia is usually between 5 and 14 days after first exposure to the drug in cases of primary exposure, but in only a few hours if it is a secondary exposure.

Patients present with petechial hemorrhages, bruising and in some cases bleeding from either the gastrointestinal or the genitourinary tract. There may be associated hemolysis or neutropenia.

Investigations

Serologic tests may confirm the presence of drug-dependent antibodies. However, there are many patients in whom immune-mediated platelet destruction is believed to be occurring who have no detectable drug-dependent antibodies,⁷³ and the reverse is also true, whereby patients have drug-dependent antibodies in their plasma but with no accompanying thrombocytopenia.¹¹⁵ Overall, therefore, the diagnosis of drug-induced thrombocytopenia through the action of drug-dependent antibodies remains largely clinical.

Management

The drug should be stopped and in most cases complete recovery will result in a few days. In cases of life-threatening bleeding therapeutic options include: plasma exchange, platelet transfusions, IVIg or corticosteroids.

Predictable (non-immune) thrombocytopenia

There are many drugs that will induce thrombocytopenia in a predictable fashion. These include anticancer (chemotherapy) drugs, which inhibit growth of stem cells, resulting in a reduction in all three cell lines (i.e. induce pancytopenia).

Drug-induced immune thrombocytopenia

Despite many reports of agents capable of inducing immune-mediated platelet destruction, there is good evidence for only a handful of these, namely heparin (see below), quin(id)ine, gold and trimethoprim-sulfamethoxazole. Heparin-induced thrombocytopenia represents a serious drug-induced platelet disorder and is discussed in detail below.

Heparin-induced thrombocytopenia (HIT)

Heparin-induced thrombocytopenia (HIT) is a common and serious complication of heparin therapy. It has an immunologic basis that is not yet fully understood and, in contrast to other drug-induced causes of thrombocytopenia, is associated with a thrombotic rather than a bleeding tendency. This condition was previously called type II HIT to distinguish it from the mild and clinically benign type I HIT which is sometimes observed early in the course of heparin administration, is non-immune mediated and self-limiting, and is not associated with a bleeding or clotting tendency (Fig. 33.15).

Fig. 33.15 Heparin-induced thrombocytopenia. Generation of heparin-platelet factor 4 complexes (H-PF4) in heparin-treated patients. Development of heparin-dependent antibodies which focus the immune response onto platelets and endothelial cells. Cell activation and cell–cell interactions, a phenomenon which is strongly enhanced at pathologic sites. Formation of a white clot, blood activation and clotting, and release of circulating aggregates and procoagulant microparticles in the blood circulation. PMN, polymorphonuclear cells (neutrophils).

(From Amiral J, Meyer D. Heparin-induced thrombocytopenia: diagnostic tests and biological mechanisms. Baillière’s Clinical Haematology. 1998;11:447–460, with permission.¹⁴⁴)

Pathogenesis

PF4 is a chemokine produced by megakaryocytes and released from platelet α-granules upon platelet activation. Released PF4 binds cell-surface glycosaminoglycans and heparin. Formation of the PF4-heparin complex causes a conformational change in the protein leading to exposure of cryptic epitopes to which the antibodies in HIT are usually generated. The anti-PF4/heparin antibodies, which are of IgG isotype, bind to PF4 through their F(ab) domains. Serial PF4 molecules align on the platelet surface and several IgG molecules bind, leading to the formation of large immune complexes that cross-link the platelet FcγIIa receptors, and resulting in platelet aggregation and an associated platelet procoagulant response.¹¹⁶ HIT antibodies also bind to endothelial cells and monocytes resulting in additional procoagulant effects through increased expression of tissue factor and formation of platelet-monocyte aggregates leading to thrombin generation. Therefore in HIT the clinical manifestations are the result of both FcγIIa-dependent platelet activation and activation of the coagulation system through tissue factor expression and platelet activation.

Clinical features

The risk of HIT is greater with 5 or more days of exposure to unfractionated heparin (1–3%) than to low-molecular-weight heparin (0–0.8%).¹¹⁷ The risk is also greater in certain patient groups such as surgical patients than in others, e.g. obstetrics. In HIT the platelet count usually falls by >50% and this typically occurs 5–14 days after the start of therapeutic or prophylactic heparin administration (the first day of heparin administration is day 0). The platelet count generally reaches a nadir in the range 20–100 × 10⁹/l and it is very unusual for it to drop below 10–15 × 10⁹/l in contrast to some other drug-induced thrombocytopenias. HIT can also be of rapid onset, developing within 24 hours in patients who have had heparin exposure within the previous 3 months. Occasionally delayed-onset HIT can occur beginning several days to weeks after heparin exposure and caused by antibodies that activate platelets independently of heparin. Although hemorrhage is uncommon, because of the enhanced thrombin generation as described above, there is a high rate of thrombotic events affecting major vessels. In patients treated with heparin for ischemic vascular disease, arterial thrombosis is common, and in those receiving heparin for thrombotic diseases such as deep vein thrombosis (DVT), the associated thrombus is usually venous. Even in the presence of heparin there may be considerable extension of the DVT, which may prove fatal. The mortality rate is high at 30%. Other manifestations of HIT occasionally occur, for example skin lesions (erythema with or without central necrosis) at sites of heparin injections or acute systemic reactions after intravenous heparin administration. DIC may also be seen in some patients. Thus the development of any of these features, new or recurrent thrombosis, or the development of thrombocytopenia during or following heparin administration should always prompt consideration of HIT.

Diagnosis

The differential diagnosis of HIT is wide and includes sepsis, multi-organ failure, malignancy, the antiphospholipid syndrome and other drugs that cause thrombocytopenia. HIT is a clinicopathological syndrome and the diagnosis is based on a combination of clinical and laboratory features. The laboratory tests are of two types, measurement of antibody binding to PF4/heparin complexes or detection of heparin-dependent activation by patient serum.¹¹⁶ The former is performed by enzyme-immunoassay (EIA) and is highly sensitive but with a specificity that ranges from 50% to 90%. The strength of the assay correlates with the probability of HIT – the diagnosis is likely in those that are strongly positive (>1.0 optical density (OD)) but unlikely in those that are weakly positive (0.4–1.0 OD).¹¹⁶ In the functional assays, aggregation of donor platelets by patient serum or plasma plus heparin is assessed. The optimal assays of heparin-dependent platelet activation use washed platelets. The sensitivity of the functional assays for HIT is generally lower than the EIAs but the specificity is higher and can reach 95–99% with assays that use washed platelets. The specificity of both types of assay is further enhanced by the finding that a positive result at low heparin concentration is inhibited at high heparin concentration due to disruption of PF4/heparin complexes.

Combining clinical assessment and appropriate laboratory testing is essential in the diagnosis of HIT.¹¹⁶ The first stage is estimation of the pretest probability, for example, using the so-called 4Ts scoring system which takes account of: 1) degree of thrombocytopenia; 2) timing of thrombocytopenia; 3) presence of thrombosis or other clinical manifestations of HIT; 4) other potential causes of thrombocytopenia. In patients with a low pretest probability no laboratory testing is required and heparin can be continued. If the EIA is negative HIT is very unlikely and heparin can be continued. If the EIA is weakly positive at low heparin concentration and reactivity is not inhibited at high concentration, HIT is unlikely and heparin can be continued. A strongly positive IgG EIA indicates an increased risk of platelet activating antibodies and a platelet activation assay should be performed. If this is positive, HIT is extremely likely. Finally the clinical situation should be reassessed to support or exclude the diagnosis.

Management

In HIT the heparin should be stopped immediately and a non-heparin anticoagulant should be substituted. Stopping heparin alone is insufficient as the risk of thrombosis approaches 50% even in those who have isolated thrombocytopenia and are clinically asymptomatic at the time of diagnosis. Cross-reactivity between unfractionated and low-molecular-weight heparin approaches 100% and therefore the latter should not be used when HIT occurs in patients receiving the former. Since initiation of vitamin K antagonists may worsen the thrombosis associated with HIT, these should be stopped and vitamin K administered. The main agents that are used instead of heparin are two direct thrombin inhibitors, lepirudin and argatroban, and the heparinoid danaparoid. Other agents that are sometimes used, though not approved in this setting, are the direct thrombin inhibitor bivalirudin, and the synthetic pentasaccharide and factor Xa inhibitor fondaparinux. The main disadvantage of these agents is that they all carry a significant risk of bleeding and none has an antidote. They each have advantages and disadvantages in different clinical settings and expert advice should be sought when considering their use. If oral anticoagulants are required they should be initiated at low doses after the platelet count has recovered to >150 × 10⁹/l and overlapped with one of the agents mentioned above for a minimum of 5 days and until the INR has been in the therapeutic range for 48 hours. Platelet transfusions are contraindicated in HIT.

Although exposure to heparin is usually avoided following the diagnosis of HIT, the antibodies do not usually persist beyond 100 days and re-exposure to heparin after this time does not generally lead to recurrence of HIT.

Other drugs causing immune-mediated platelet destruction

There are many drugs implicated in immune-mediated thrombocytopenia. The evidence is strongest for quinine, quinidine, heparin and gold salts,¹¹³ as already discussed. Other implicated agents, though with fewer reports to date, include: α-methyldopa,¹¹⁸ diclofenac,¹¹⁹ rifampin,¹²⁰ carbamazepine¹²¹ and sulfonamides. Readers are referred to George et al. for an excellent review of the topic.¹²²

Acquired functional abnormalities of platelets

Bleeding problems may arise through either inadequate numbers of platelets or functional abnormalities of the platelets themselves. This section discusses disorders in which there are abnormalities of platelet function along with their pathogenetic basis (Table 33.8).

Table 33.8 Drugs and disorders interfering with platelet function

Systemic disorders	Uremia Cardiac by-pass surgery
Hematologic disorders	Myeloproliferative diseases Leukemia Myelodysplasia Paraproteinemias, including multiple myeloma
Drugs	Aspirin Antibiotics Anticoagulants Others

Uremia in renal failure

Bleeding may be a feature of either acute or chronic renal failure,^123,¹²⁴ with spontaneous bleeding into the skin, mucous membranes including the gastrointestinal or genitourinary tracts, central nervous system and other sites (see also Chapter 35).

Pathogenesis

Platelet function, in the presence of uremia, is abnormal,^123,¹²⁵^–¹²⁷ and a variety of laboratory studies have shown that all aspects of platelet activity are affected, including platelet adhesion, aggregation and procoagulant activity.¹²⁸^–¹³⁰

The normal process of platelet adhesion has been shown to involve contact of platelets to endothelial structures. This is dependent on the binding of vWF to GP lb/IX.¹²⁴ In the presence of uremia there may be a qualitative or quantitative abnormality of vWF or GPlb/IX itself.

Management

The most important factor to consider is whether the patient is actively bleeding, rather than abnormalities detected using bleeding time, and other coagulation assays. Hemodialysis itself will help correct the bleeding induced by the uremia.¹²⁷ Desmopressin (DDAVP) which promotes the release of vWF from vascular endothelial cells may reduce the bleeding time in some uremic patients.¹²⁶

Myeloproliferative disorders

The myeloproliferative diseases (MPDs) are neoplastic hematologic stem cell disorders and include essential thrombocythemia (ET) (Fig. 33.16), polycythemia rubra vera (PRV), idiopathic myelofibrosis (IMF) and chronic myeloid leukemia (CML). These disorders are associated with both bleeding and thrombosis.

Fig. 33.16 Trephine biopsy in patient with essential thrombocythemia, one of the myeloproliferative disorders, showing greatly increased numbers of megakaryocytes with clustering.

(Figure kindly provided by Dr John Amess, Barts & The London NHS Trust.)

Pathophysiology

In PRV there is a rise in whole blood viscosity through elevation of the hematocrit which may contribute to thrombosis.^133,¹³⁴ Abnormalities in platelet function have been reported in MPDs, and the bleeding time is prolonged in a minority of patients. However, bleeding may occur even if the template bleeding time is normal.¹³³

Platelet abnormalities in the myeloproliferative disorders

Platelets may be larger than normal, although in some cases they are smaller. Their survival may be reduced, especially in essential thrombocythemia. Platelet aggregation is abnormal with the standard aggregants including ADP and collagen, though this is not a feature of thrombocytosis when the underlying cause is reactive, thereby excluding thrombocytosis per se as a cause of the abnormal aggregation.¹³⁴ It may, in fact, be a secondary consequence of the conversion of arachidonic acid to prostaglandin endoperoxides or lipooxygenase products,¹³⁵ or a decrease in platelet responsiveness to thromboxane A₂.¹³⁶

Management

For PRV, a reduction in hematocrit, aiming for a level less than 0.45 (45%) is an approach adopted by most hematologists.¹³⁷ Ongoing clinical trials may help determine the optimal level of hematocrit. The bone marrow may be suppressed effectively using hydroxyurea or busulfan with an overall reduction in platelet count in patients with essential thrombocythemia.^138,¹³⁹ However, even though the peripheral platelet count is lowered, this does not necessarily correct the associated platelet abnormalities. The management of the patient who is actively bleeding is more complex.

Platelet abnormalities in leukemia and myelodysplastic syndromes

Bleeding is common in both types of disease, and is often due to thrombocytopenia. However, functional abnormalities of platelets are also recognized. In acute myeloid leukemia (AML) for example, platelets may be larger than usual, with abnormalities of shape and the appearance of their granules (Fig. 33.17).

Fig. 33.17 Bone-marrow aspirate (high power) of AML M4 (acute myelomonocytic leukemia), showing large myeloblasts.

(Figure kindly provided by Dr John Amess, Barts & The London NHS Trust.)

Management

Patients with functional platelet abnormalities generally respond to platelet transfusion. Correction of the underlying malignancy with definitive therapy will also improve the acquired bleeding diathesis. Similar abnormalities are also seen in MDS and treatment generally consists of regular platelet transfusions ^140,¹⁴¹ (Fig. 33.18).

Fig. 33.18 Bone-marrow aspirate in multiple myeloma showing large numbers of plasma cells. The paraprotein may cause platelet dysfunction through a variety of mechanisms including hyperviscosity, development of uremia and other mechanisms.

(Figure kindly provided by Dr John Amess, Barts & The London NHS Trust.)

Drugs that interfere with platelet function

Drugs may cause bleeding through the induction of profound thrombocytopenia. In addition, there are numerous agents that may induce bleeding through the interference with the normal function of platelets. Thus, although the platelet count may be entirely normal, the platelets are rendered functionally defective. Aspirin and non-steroidal anti-inflammatory agents (NSAIDs) are the most common cause of acquired platelet dysfunction (Table 33.9). Their effects are mediated through irreversibly inhibiting cyclooxygenase activity in the platelet resulting in impairment of the granule release reaction and defective aggregation. Aspirin, in particular, acetylates the serine residue at position 530 of prostaglandin synthase, the enzyme responsible for converting arachidonate to prostaglandin cyclic endoperoxides, and thereby inhibits the synthesis of prostacyclin and thromboxane A₂ (Fig. 33.19). These effects are seen in both the platelet and endothelium, but the effect of aspirin on platelet function is detectable for several days after the drug is stopped since there is a lag phase before new platelets lacking the drug enter the circulation. Endothelial cells, on the other hand, are able to generate prostaglandin synthase much more rapidly.

Table 33.9 Drugs interfering with platelet function

NSAIDs	Aspirin Diclofenac Mefenamic acid Others	Cyclooxygenase inhibitors
Antibiotics	Penicillins Cephalosporins Nitrofurantoin	In high doses, particularly in ill patients, many antibiotics may interfere with platelet aggregation
Anticoagulants	Heparin Epsilon aminocaproic acid
Drugs that increase platelet cAMP	Dipyridamole Iloprost	Dipyridamole is a phosphodiesterase inhibitor
Cardiovascular system drugs	Diltiazem Isosorbide dinitrate Nifedipine Propranolol
Psychotropics	Tricyclic antidepressants such as imipramine and amitriptyline Phenothiazines, e.g. chlorpromazine, promethazine
Anesthetics	Local and general anesthetics (e.g. halothane)
Anticancer drugs	Chemotherapeutic agents such as mithramycin, BCNU and daunorubicin
Anticoagulants	Heparin and coumadin
Miscellaneous	Dextrans Ticlopidine Lipid-lowering drugs, e.g. clofibrate Quinidine Ethanol

NSAIDs, non-steroid anti-inflammatory drugs.

(Modified from Rao AK, Carvalho ACA 1994. In: Colman RW et al (eds), In Hemostasis and Thrombosis: Principles and Practice, 3rd edn. JB Lippincott, Philadelphia.)