High-risk individuals include those over the age of 60 years, those who have had a previous ET-related thrombotic or haemorrhagic event, patients who require therapy for systemic hypertension or diabetes or those with a platelet count greater than 1,500  ×  10⁹/L. It should be noted, however, that microvascular symptoms are not generally regarded as thrombotic events for the purpose of risk assessment, unless they are severe, or fail to respond to platelet anti-aggregating agents. It is now accepted that high-risk patients should be treated with platelet lowering agents, ideally to a target value of 400  ×  10⁹/L, together with aspirin. Such a conclusion is supported by an Italian prospective study that randomised 114 such cases to receive either hydroxycarbamide or no cytoreductive therapy. After a median follow-up of 27 months, patients on hydroxycarbamide suffered significantly fewer thrombotic events [8].

Which therapy should be offered as first-line? This question has been answered by the Medical Research Council’s Primary Thrombocythemia-1 (MRC PT-1) trial, the largest prospective randomised trial ever undertaken in the field of myeloproliferative neoplasms [9]. In this study, 809 high-risk patients with ET were randomised to receive hydroxycarbamide plus aspirin, or anagrelide plus aspirin. Compared to hydroxycarbamide plus aspirin, patients in the anagrelide arm suffered higher rates of arterial thrombosis, major ­haemorrhage and myelofibrotic transformation, but intriguingly a reduced rate of venous thromboembolism. These data support the view that hydroxycarbamide and aspirin should be prescribed as first-line therapy for the majority of high-risk ET cases. Importantly, a comparison of arterial thrombotic rates in the PT-1 anagrelide arm (8 % at 2 years) with the untreated arm of the Cortelazzo study (26 % at 2 years) suggests that anagrelide provides partial protection from this complication. The PT1 study, therefore, supports anagrelide’s licensed indication as second-line therapy for patients refractory, or intolerant, of hydroxycarbamide (Table 12.4).

Patients under the age of 40 years, with no high-risk features, are regarded as low risk and should be offered anti-platelet drugs only, since such cases have a thrombotic risk that is only slightly greater than, or equal to, that in the normal population. It will be important, however, to gain better quality information about this patient group, ideally in large clinical studies, such as the ongoing low-risk arm of the UK PT1 study. The management of intermediate risk cases, defined as those aged 40–60 years with no high-risk features, remains unclear and routine cytoreductive therapy should be avoided until the results of an international randomised study (PT-1) are known.

Microvascular occlusion in ET, resulting in symptoms such as painful digits, that can lead to gangrene, transient ischaemic attacks, migraine and erythromyelalgia is improved by aspirin therapy [10]. However, the only randomised controlled trial investigating the role of aspirin in MPN has been in the related disorder polycythemia vera (PV). The so-called ECLAP study (European Collaboration on Low-dose Aspirin in PV) nevertheless supports the safety and utility of routinely prescribed aspirin, as treated patients had a significantly reduced risk of thrombotic complications, including non-fatal myocardial infarction and stroke, death from cardiovascular causes, pulmonary embolism and major venous thrombosis [11]. A maintenance daily dose of 75 mg is ­usually sufficient and reduces the risk of peptic ulceration associated with larger doses. Treatment should be avoided in patients with a history of bleeding, or in those with high platelet counts, i.e. 1,000–1,500  ×  10⁹/L, especially in the context of a bone marrow trephine [12]. For the occasional patient with ongoing thrombotic events, despite adequate platelet control, consideration should be given to the combination of aspirin and clopidogrel, while clopidogrel alone may be helpful in patients with aspirin-related side effects.

Hydroxycarbamide, or hydroxyurea, is an anti-metabolite that slows DNA synthesis and repair by inhibiting ribonucleotide diphosphate reductase activity, an enzyme that catalyses the reduction of ribonucleotides. The end result is a block in cell cycle at the G1/S phase and cell death. Side effects are uncommon, mostly mild and reversible, and include neutropenia and anaemia, gastrointestinal tract symptoms (anorexia, nausea, vomiting and diarrhoea) and fever. Rare complications include pneumonitis, azoospermia and liver dysfunction. Prolonged use is associated with a range of dermatological complications, including actinic keratosis, squamous cell carcinoma and nail changes such as onychodystrophy and melanonychia. The evidence for teratogenicity is weak, although it is recommended that patients should avoid conception while on therapy [13].

Anagrelide is an orally active imidazoquinazoline compound, with potent platelet-reducing activity as a result of inhibiting megakaryocyte size, ploidy and maturation, and consequently is not thought to be leukaemogenic. Its efficacy is well established in ET patients, including both treatment-naïve cases and those refractory to other cytoreductive agents, being effective in over 90 % of cases. The response is rapid, with most patients reaching the treatment target within a few weeks. The results of the largest randomised study, comparing hydroxycarbamide and anagrelide (MRC PT-1 Study), indicated that despite an equivalent platelet count, anagrelide-treated patients experience higher rates of arterial thrombosis, major haemorrhage and myelofibrotic transformation, while venous thrombosis was greater in the hydroxycarbamide arm [9]. In contrast, the recently published ANAHYDRET study suggests that anagrelide is not inferior to hydroxycarbamide [14]. However, it should be noted that the number of patients and end-point events were considerably less than that in the PT-1 study and, as a result, the ANAHYDRET study may not have been powered to detect a clinical difference. Overall, these two randomised studies support the use of anagrelide as second-line therapy in high-risk ET.

The most frequent side effects of anagrelide relate to its inhibition of cyclic nucleotide phosphoesterase III, an enzyme expressed in heart tissue and vascular smooth ­muscle, and include headache, palpitations and tachycardia. Such problems are usually mild and subside over time, with lower incidences reported during long-term therapy compared with the initial treatment phase. Side effects can be minimised by starting at a low dose (0.5  −  1.0 mg/day) and increasing gradually, while avoiding obvious precipitating factors, such as stress and anxiety, alcohol, caffeine and smoking. Benign palpitations that do not subside can be treated with low-dose beta-blockers (e.g. metoprolol or atenolol), especially if exercise or stress related. Aggravation of cardiac insufficiency is a rare but important side effect, and special caution is advised in patients with previous cardiac failure. Anagrelide may also be associated with a mild fall in haemoglobin levels as a result of its vasodilatory properties.

The attraction of interferon, a presumed non-leukaemic and non-teratogenic agent with the potential to eradicate an abnormal clone, has led to the use of recombinant interferon-alpha (IFN-alpha), particularly in younger patients and during pregnancy. Early studies in ET showed that IFN-alpha can result in a reduction in platelet count to less than 600  ×  10⁹/L within 3 months, using an average of three million IU per day [15]. However, the use of conventional IFN has been limited by toxicity, leading to treatment discontinuation in approximately a third of patients. Side effects include flu-like symptoms, fatigue, musculoskeletal pain, depression and autoimmune disorders. Recently, pegylated interferon has been shown to be better tolerated, with only 8 % of patients with PV having to stop therapy at 1 year [16], a rate comparable with that reported for hydroxycarbamide. Furthermore, pegylated interferon alfa-2a (PEG-IFN-alpha-2a) has been shown to reduce the mutant allele burden in JAK2 V617F-positive ET patients, with 6 % of cases achieving a complete molecular remission [17]. Additional studies are required to confirm these interesting findings and to determine whether molecular responses are associated with improved clinical outcomes.

Suspicion of the potential long-term side effects of all alkylating agents, as well as the risk of prolonged bone marrow suppression, has limited the use of busulphan as a routine treatment of ET. Furthermore, its use has been associated with pulmonary toxicity, especially broncho-pulmonary dysplasia and pulmonary fibrosis, while a study reporting an increased risk of second malignancies after the sequential use of busulphan and hydroxycarbamide is a further cause for concern [18]. Although it has been successfully and conveniently used in small intermittent doses by some centres [19], its use should be confined to elderly patients. A relatively non-toxic but effective regimen consists of busulphan 2–4 mg/day for 1–2 weeks administered every 4–6 weeks until the platelet count is below 400  ×  10⁹/L, with further doses being prescribed once the count rises above this level.

Radioactive phosphorus (32-P) has been used for the treatment of MPN for nearly 80 years. The isotope has a half-life of 14.3 days, is a pure beta emitter and has a maximum range in vivo of only 8 mm. It is an effective agent at controlling blood counts, has few acute side effects and rarely results in haematological complications [20]. A single intravenous dose of between 2.5 and 3.0 mCi/m2 is often sufficient to provide adequate control for up to 6–36 months, although occasionally a second smaller dose is required after 3–4 months. There is conclusive evidence, however, that it is leukaemogenic and as a result, its use should be restricted for patients over 70 years of age.

The diagnostic criteria for PMF have recently been defined by the WHO and rely heavily on the features of the bone marrow trephine (Table 12.5). The differential diagnosis encompasses the many causes of bone marrow fibrosis (Table 12.6). Whereas increased fibrosis may be found in PV, ET and chronic myeloid leukaemia, myelofibrotic transformation of these conditions can occur with features resembling PMF.

Current evidence indicates that PMF is a true clonal disorder of the haematopoietic stem cell. Such a conclusion was first reached by the analysis of glucose-6-phosphate dehydrogenase (G-6PD) isoenzymes and subsequently by studies using X-linked restriction fragment length polymorphisms (RLFPs), which confirmed PMF to be characterized by monoclonal haematopoiesis. This conclusion is supported by the karyotype-based demonstration of clonality of erythroid, granulocyte–monocyte and granulocyte–monocyte–erythroid progenitors, but not fibroblasts [24]. Recurrent cytogenetic abnormalities occur in approximately 50 % of PMF cases at diagnosis, with the most common lesions, ­representing over 80 % of abnormal cases, being del(13)(q12;q22), del(20)(q11;q13), trisomy 8, trisomy 9, del(12)(p11;p13), monosomy or long arm deletion of chromosome 7 and partial trisomy 1q [25, 26]. No cytogenetic lesion appears specific for PMF, with the possible exception of der(6)t(1;6)(q23–25;p21–22), which has recently been proposed as a marker for PMF [27]. Interestingly, the gene coding the FK-506-binding protein 5 (FKBP51) is located in the involved region (6p21) and the protein has been shown to be over-expressed in CD34+ cell-derived megakaryocytes from PMF patients.

A major advance in the understanding of the disease’s pathogenesis was the identification of the somatic mutation JAK2 V617F, which occurs in approximately 50 % of cases. JAK2 V617F, a G to T substitution that generates valine instead of phenylalanine at position 617 within the JAK2 pseudokinase domain, leads to activation of multiple downstream signalling pathways, enabling cytokine-independent haematopoietic cell growth [28]. Somatic mutations in the thrombopoietin receptor, MPL, were subsequently documented in 5 % of JAK2 V617F-negative cases, confirming the importance of constitutive activation of the JAK-STAT pathway in the disease’s pathogenesis [29]. Although the discovery of JAK2 and MPL mutations has provided important insights, there remain a number of unanswered questions: for example how do the mutations lead to constitutive activation of the protein, what mutations are involved in those cases lacking a known genetic abnormality and what role does gene dosage play?

The pathogenesis of excessive bone marrow fibrosis that is a feature of PMF is believed to be a reaction to the underlying clonal haematopoietic tissue. The prevailing hypothesis involves an aberrant release of cytokines from the clonal megakaryocytes and monocytes that include transforming growth factor-β, basic fibroblast growth factor, epidermal growth factor, platelet-derived growth factor, vascular endothelial growth factor and calmodulin, all of which induce a secondary fibroblast proliferation and dysregulation of extracellular matrix formation [24]. The latter is composed of an excessive deposition of interstitial and basement membrane glycoproteins, including collagens type I, III, IV, V and VI, fibronectin, vitronectin, laminin and tenascin, as well as a marked increase in angiogenesis. Furthermore, the pathological ingestion of neutrophils by megakaryocytes (emperipolesis), induced by altered P-selectin expression, may contribute to cytokine release [30], while neutrophil-derived enzymes facilitate the release of CD34+ stem cells into the peripheral blood, a characteristic feature of PMF [31]. Interestingly, PMF CD34+ cells generate a 24-fold greater number of megakaryocytes (MKs) than normal CD34+ cells. Such derived MKs exhibit a delayed pattern of apoptosis and an over-expression of the anti-apoptotic BCL-xL protein, findings that lend support to the current hypothesis of the pathogenesis of PMF [32].

PMF characteristically occurs after middle age, with a median age at presentation of 60 years. About a quarter of patients are asymptomatic at diagnosis and are identified following routine examination. The most common symptoms are the consequence of anaemia, namely fatigue, weakness, palpitations and dyspnoea. Palpable splenomegaly is characteristic and when massive can manifest in a range of complaints, particularly abdominal discomfort and early satiety. Splenic infarction, resulting from the organ outgrowing its blood supply, may produce transient discomfort, although rarely can result in severe abdominal pain simulating an abdominal emergency. Hepatomegaly occurs in approximately 70 % of cases and portal hypertension may result from increased hepatic blood flow or intra-hepatic obstruction. Constitutional symptoms may dominate the clinical picture, including low-grade fever, night sweats and weight loss, and their presence carries a poor prognosis [33]. Patients may also complain of bone pain, especially in the lower extremities. Bleeding may complicate the clinical course and, although often mild manifesting as petechiae and ecchymoses, can be life threatening due to massive gastrointestinal haemorrhage. The haemorrhagic diathesis may result from a combination of thrombocytopenia, acquired platelet dysfunction and low-grade disseminated intravascular coagulation.