A guideline for the management of specific situations in polycythaemia vera and secondary erythrocytosis

The previous BSH guideline for the management of erythrocytosis was published in 2005 (McMullin et al, 2005) and amended in 2007 (McMullin et al, 2007). Here, we re‐evaluate the literature formulate guidance on the management of specific situations encountered in polycythaemia vera (PV) and the management of the other types of secondary erythrocytosis. Recommendations for the diagnostic pathway of investigation of an erythrocytosis, risk stratification and management of PV are in the accompanying guideline (McMullin et al, 2018). We review evidence and outline guidance on management of acute thrombotic events and secondary prevention of thrombosis in PV. The unusual thrombotic events, splanchnic vein and cerebral vein thromboses are discussed and haemorrhage. The specific situations of surgery and pregnancy and guidance on management of pruritus are included. The evidence for the management of other causes of erythrocytosis, including idiopathic erythrocytosis, congenital erythrocytosis, hypoxic pulmonary disease and post‐transplant erythrocytosis, is reviewed and recommendations made.


Introduction
The previous BSH guideline for the management of erythrocytosis was published in 2005(McMullin et al, 2005 and amended in (McMullin et al, 2007. Here, we re-evaluate the literature formulate guidance on the management of specific situations encountered in polycythaemia vera (PV) and the management of the other types of secondary erythrocytosis. Recommendations for the diagnostic pathway of investigation of an erythrocytosis, risk stratification and management of PV are in the accompanying guideline (McMullin et al, 2018). We review evidence and outline guidance on management of acute thrombotic events and secondary prevention of thrombosis in PV. The unusual thrombotic events, splanchnic vein and cerebral vein thromboses are discussed and haemorrhage. The specific situations of surgery and pregnancy and guidance on management of pruritus are included. The evidence for the management of other causes of erythrocytosis, including idiopathic erythrocytosis, congenital erythrocytosis, hypoxic pulmonary disease and posttransplant erythrocytosis, is reviewed and recommendations made.

Management of specific situations in polycythaemia vera
Thrombosis Thrombotic events are the major cause of morbidity and mortality in PV, and their prevention is the main objective of treatment. About one-third of patients present with a thrombotic event (Elliott & Tefferi, 2005;Hultcrantz et al, 2015;Kaifie et al, 2016) and younger patients diagnosed with PV have an increased risk of early death from cardiovascular disease over the general population, accounting for 45% of all deaths in PV (Marchioli et al, 2005;Hultcrantz et al, 2015).
Arterial thrombosis involving large arteries and peripheral vascular disease are more common than venous thromboembolism in PV. A prospective study (Marchioli et al, 2005) found that three quarters of PV patients who experience a thrombotic event had arterial and one quarter venous thrombosis.
Concerning the management of thrombosis: cardiovascular risk factors should be assessed and controlled, particularly hyperlipidaemia and hypertension, smoking and diabetes mellitus. The long-term cardiovascular risk should be formally assessed at baseline and annually using a validated score, such as that determined by the computerised scoring tool, QRISK, although the impact of myeloproliferative neoplasms (MPN) on this score is not defined (National Institute for Health and Care Excellence [NICE], 2014). This assessment is usually conducted in primary health care, but the results should be used in primary and secondary care to optimise the management of vascular risk factors.
Risk stratification for thrombosis identifies age ≥65 years and previous thrombosis as high-risk factors for further thrombosis. The role of thrombocytosis as a risk factor in PV is less clear than in essential thrombocythaemia (ET) (Di Nisio et al, 2007). The mechanism behind thrombosis is complex and factors include functional cellular changes, such as increased leucocyte-platelet aggregation and endothelial activation, and the overexpression of activation and adhesion molecules by the neoplastic cells (Falanga & Marchetti, 2012). Some of these factors may explain the increased incidence of thrombosis at unusual sites, such as splanchnic vein thrombosis (SVT) and cerebral venous thrombosis (CVT).
Factor V Leiden (F5 R506Q) and prothrombin gene mutation (F2 G20210A) have been associated with thrombosis in PV and ET (Trifa et al, 2014). However, routine testing for inherited or acquired thrombophilia is not recommended as it has is no impact on treatment and standard guidelines for thrombophilia testing should be applied for PV patients.
Both primary and secondary prevention should include control of cardiovascular risk factors in accordance with current recommendations (NICE, 2014). Of interest, there is some evidence that angiotensin-converting enzyme inhibitors (ACEi), as anti-hypertensive agents, may have an additional role in controlling the haematocrit (Hct) . In accordance with evidence for the management of PV, low-dose aspirin should be offered to all patients (Landolfi et al, 2004). Low-dose aspirin, which describes various doses (e.g. 75, 81 or 100 mg) used in different countries, is the only dose to have been assessed in recent clinical trials. The risk of major bleeding, particularly gastro-intestinal, increases with age, and proton pump inhibitors should be given to patients over 75 years of age as well as to those a with high bleeding risk. The role of the ADP-receptor antagonist, clopidogrel, in the long-term prevention of MPNrelated thrombosis has not been investigated although it is widely used as an alternative in those who cannot tolerate aspirin. The Hct should be maintained below 0Á45 in accordance with evidence from the Cytoreductive therapy in PV (CYTO-PV) study (Marchioli et al, 2013).

Management of acute thrombotic events
Acute thrombotic events should be managed according to current guidelines. Cytoreductive therapy should be instituted to ensure control of blood counts to the therapeutic target. Venesection to reduce the Hct to <0Á45 should be undertaken if it is not controlled at the time of acute thrombosis. For arterial thrombosis, low dose aspirin should be offered to all patients, and specialist advice should be sought.
For venous thromboembolism (VTE), anticoagulation should be commenced. Low molecular weight heparin (LMWH) remains the first choice in the acute setting, followed by vitamin K antagonists (VKA) as per guidelines (NICE, 2012). Direct oral anticoagulants (DOACs) are increasingly used in the non-MPN population for prophylaxis and VTE therapy. However, there is limited data for their specific use in MPN, where they appear to be safe and efficacious .

Secondary prevention of thrombosis
Indefinite anticoagulation for VTE is recommended because of the presence of continuing risk (NICE 2012;Watson et al, 2015;Kearon et al, 2016). In MPN, the role of continuing Guideline anticoagulation, particularly in VTE, has been assessed by several studies. Around a third of patients who stop anticoagulation after initial treatment suffer a recurrence (Barosi et al, 2013;Kaifie et al, 2016). Two retrospective studies (Hern andez-Boluda et al, 2015;De Stefano et al, 2016a) showed that cessation of VKA after 3 months of treatment for a first VTE in MPN patients increased the risk of recurrent thrombosis compared with continued treatment. The later study also suggested a role for VKA in reducing the risk of arterial thrombotic recurrence and the possible benefit of adding aspirin to VKA to prevent recurrent arterial thrombosis. We are unable to make a specific recommendation about the addition of aspirin to VKA when a secondary thrombosis occurs in a patient already on VKA, as there is not yet evidence to support this. However, trials of this combination are in progress in other conditions and evidence may emerge.
For unprovoked VTE, where there is no physical precipitating factor, we therefore recommend indefinite anticoagulation to reduce the incidence of recurrence. In a significant number of patients with recurrent thrombosis , cell counts were sub-optimally controlled. Cytoreduction should therefore be reviewed to ensure optimal counts are achieved. For recurrent arterial events, specialist advice should be sought from cardiovascular, stroke or vascular physicians.

Splanchnic vein thrombosis
Unusual sites of venous thrombosis are more common in patients with MPN, especially in younger patients. The splanchnic circulation is particularly vulnerable in patients with JAK2 V617F mutation. SVT comprises thrombosis of hepatic, portal, superior mesenteric and splenic veins singly or in combination. The incidence of SVT during the course of PV ranges from 0Á8 to 7Á8% (Sekhar et al, 2013). Overall, portal-mesenteric axis venous thrombosis is commoner than hepatic vein thrombosis or Budd Chiari Syndrome (BCS), however BCS is more prevalent in PV (Smalberg et al, 2012). The mortality of MPN-SVT is high, at 20-40% at 10 years, due to the index occurrence or recurrent thrombosis, bleeding or leukaemic transformation (Hoekstra et al, 2011;Ageno et al, 2015). The presence of hypersplenism, SVT-related haemodilution and iron deficiency due to blood loss and/or masked polycythaemia can artificially lower the blood counts and a normal Hct may be found despite a high red cell mass (Kiladjian et al, 2008). Therefore, it is important to investigate all patients presenting with SVT without attributable local pathology, such as malignancy, sepsis or pancreatitis, for the presence of JAK2 V617F mutation, which is present in 80-90% of patients with MPN-SVT (Kiladjian et al, 2008), and if this is negative, testing should be done for a CALR mutation, which is present in about 2Á5% (Sekhar et al, 2016).
Although PV is commoner in males, PV-related SVT is commoner in young females. The reason for the predilection of splanchnic venous circulation to thrombosis in MPN is not clear. Many series report co-existing inherited or acquired thrombophilia in 25-30% patients, especially protein S deficiency which has been associated with unusual site thrombosis in MPN patients . In patients with SVT or cerebral vein thrombosis (CVT), testing for inherited and acquired thrombophilia should be considered if it would influence management decisions, such as intensity and duration of anticoagulation in the face of higher risks of bleeding (Tait et al, 2012).
Anticoagulation treatment in the acute phase should be undertaken with LMWH unless there are compelling contraindications (Tait et al, 2012). In addition, the treatment of acute phase SVT includes interventions to recanalize the blocked veins, such as thrombolysis and insertion of stents. Optimal management results in long-term survival of 85% at 10 years in BCS. Patients requiring liver transplantation have a high mortality (Potthoff et al, 2015). The management of PV-related SVT therefore needs close coordination between hepatologists, interventional radiologists and haematologists to achieve an accurate diagnosis, anti-coagulate effectively, undertake interventional procedures and optimise cytoreduction in the acute phase.
Recurrence risk is high and prevention is important because of significant morbidity and mortality associated with recurrence. Overall 25-30% of MPN-SVT patients suffer a recurrence by 10 years (Hoekstra et al, 2011;Ageno et al, 2015). Although recurrence is more likely in the splanchnic circulation (Amitrano et al, 2007), it can occur in other venous or arterial locations. Anticoagulant practice in MPN-SVT is heterogeneous (Ellis et al, 2014). Current VTE guidelines recommend continuing anticoagulation in the presence of continuing active risk (Kearon et al, 2012;Tait et al, 2012). On this basis, PVrelated SVT is an indication for long-term anticoagulation. Several prospective and retrospective registry studies conclude that long-term anticoagulation reduces recurrence risk at splanchnic and other sites with an acceptable level of bleeding complications (Amitrano et al, 2007;Ageno et al, 2015;De Stefano et al, 2016b). Recurrence in the arterial circulation has been observed in patients not on anti-platelet agents (Hoekstra et al, 2011;De Stefano et al, 2016b). The strategy of using anti-platelet agents in addition to VKA is followed in selected patients with BCS with stents in-situ but registry data do not suggest any benefit from this combination for prevention of recurrence of thrombosis outside of this group. Concerning anticoagulation with DOACs, 20% of patients with non-cirrhotic SVT suffered thrombosis or haemorrhage at 2 years (De Gottardi et al, 2017). These agents have not been studied in MPN-related SVT where LMWH followed by warfarin is the standard of care.
The impact of cytoreduction in MPN with SVT has not been systematically studied. Cytoreduction was used in 40-70% patients across registries and did not uniformly influence recurrence risk. However, abnormally high blood counts due to inadequate cytoreduction were present in over half the patients with recurrent thrombosis (De Stefano et al, 2016b) indicating the need for rigorous clinical monitoring of patients on cytoreduction. Venesection alone is not appropriate to treat PV and cytoreduction should be undertaken in the presence of SVT. Patients with PV-related SVT are younger and interferon may be suitable as a first line agent, however data on interferon in this group of patients are not yet available. There are no data to guide therapeutic targets of cytoreductive treatment. In particular, whether more aggressive cytoreduction offsets the residual risk of recurrent thrombosis is not known. One strategy is to use standard targets for treatment of PV with an aim to achieving normal Hct, white blood cell (WBC), neutrophil and platelet counts (Barbui et al, 2011). An alternative strategy is to use lower targets, as defined by experts in the ongoing trial of pegylated interferon 2 alpha in MPN including MPN-SVT aiming for a Hct ≤0Á42, WBC count between 2 and 8 9 10 9 /l, and platelet count between 100 and 200 9 10 9 /l (Hoffman, 2012). As this study is not yet complete the evidence for the use of these targets is weak. Nevertheless, the fact that these patients do develop thrombosis with normal counts suggests functional thrombogenicity, which may be altered by using diseasemodifying agents to achieve these lower targets. Small doses of cytoreductive agents are often sufficient to achieve therapeutic targets and care must be taken to avoid significant thrombocytopenia in the face of continuing anticoagulation. Treatment with ruxolitinib was found to reduce the volume of spleen in a third of MPN-SVT patients after 2 years .
Complications of treatment include bleeding, especially gastrointestinal haemorrhage and, in particular, variceal haemorrhage (Ageno et al, 2015;De Stefano et al, 2016b). A quarter of major bleeds are intracranial. Other than thrombocytopenia, cytoreduction does not have any impact on bleeding risk. Bleeding risk can be minimized by careful management of oesophageal varices and concurrent use of proton pump inhibitors.

Cerebral vein thrombosis
Cerebral vein thrombosis occurs in about 1% of MPN patients, primarily in those with ET bearing the JAK2 V617F mutation (Martinelli et al, 2014). In selected patients with unexplained CVT, testing for JAK2 V617F can identify the aetiology and assist management. Recurrence rates are higher in MPN-CVT, especially in spontaneous CVT, and a third of patients suffer recurrence despite anticoagulation and cytoreduction (Miranda et al, 2010;Martinelli et al, 2014). Recurrence is systemic, affecting venous and arterial circulation and often involves splanchnic veins. Patients with PV-related CVT should be treated with long-term anticoagulation and cytoreduction. The role of aspirin and JAK inhibitors have not been studied in this group of patients.

Haemorrhage
Haemorrhage is both a less frequent and generally less severe clinical complication of PV than thrombosis. The reported incidence varies greatly between studies (Elliott & Tefferi, 2005;Marchioli et al, 2005;Kander et al, 2015;Kaifie et al, 2016). The principal sites affected are skin, mucous membranes and gastrointestinal tract. Patients with SVT are at a particularly high risk of bleeding from gastro-oesophageal varices.
Factors contributing to bleeding in patients with PV include extreme thrombocytosis (platelet counts >1500 9 10 9 /l) (Finazzi et al, 1996) and associated acquired von Willebrand syndrome (aVWS). This has been reported to affect 12% of PV patients, usually those with high platelet counts, and does not always correlate with bleeding (Kander et al, 2015;Mital et al, 2015). Other platelet function defects are reported in PV but they are not predictive of bleeding. However, the risk of bleeding is likely to be higher with combined use of anti-platelet therapies, such as dual anti-platelet therapy or anti-platelet therapy and VKA (Hallas et al, 2006;Kaifie et al, 2016). Other studies have reported splenomegaly (Kaifie et al, 2016) and a high leucocyte count (Chou et al, 2013;Lim et al, 2015) as predictors of haemorrhage.
Blood counts should therefore be optimised. Other measures to consider include adjustment of any concomitant anti-platelet and/or anticoagulant therapy. Clinically significant bleeding may, paradoxicall,y require platelet transfusion (Terasako & Sasai, 1998) and a role for tranexamic acid has been suggested (Spivak, 2002). The utility of recombinant activated factor VII (rFVIIa) is unknown in MPN patients with uncontrolled life-threatening bleeding and perhaps worthy of further study.

Peri-operative management
Polycythaemia vera patients undergoing surgery are, paradoxically, at risk of both haemorrhage and thrombosis. Even in treated PV patients with well controlled counts, there is an increased risk of haemorrhage, demonstrated by their increased transfusion requirements (Weingarten et al, 2015) compared to non-PV patients . VTE incidence was increased 5-fold despite prophylaxis. It is therefore important that PV patients are assessed pre-operatively and abnormal counts optimised, balancing the acute need for surgical procedure against the risk of bleeding and thrombosis.
Polycythaemia vera patients who bleed may have significant qualitative platelet abnormalities. Patients with a prior history of bleeding should be tested for platelet function and for aVWS preoperatively. Anti-thrombotic prophylaxis should be administered according to standard postoperative guidelines. There is a little evidence that use of cytotoxic agents impairs wound healing: their use in the 4-6 weeks after surgery needs to be considered based on thrombotic and haemorrhagic risk and could be avoided if venesection alone is sufficient.

Recommendations: surgery
• Pre-operative planning should involve a haematologist to optimise count control and individualise peri-operative plan. (GRADE 1B) • Use standard protocols for managing antithrombotic prophylaxis. (GRADE 1B) • If the patient has a bleeding history, screen for coagulation tests, aVWS and platelet function tests pre-operatively. (GRADE 2C)

Pregnancy
Polycythaemia vera is uncommon in females of reproductive age, occurring in less than 0Á3 per 100 000 (Srour et al, 2016). There are only small case series describing the management of pregnancy in patients with PV (Harrison, 2005;Robinson et al, 2005;Griesshammer et al, 2008). Larger case series are published for pregnancy management in patients with ET or cohorts of patients with all MPNs (Alimam et al, 2016;Skeith et al, 2017). The guidance detailed here refers to these case series, practices extrapolated from the management of pregnant patients with ET and personal practice. Pregnancy is a prothrombotic state with increased risk of thromboembolism in patients with PV. Consequently, there is a significant risk of obstetric complications, such as fetal loss throughout all trimesters, intra-uterine growth retardation, prematurity, maternal thromboembolism and haemorrhage. Previously significant fetal loss and maternal morbidity was seen, but a recent prospective study of pregnancy outcomes in MPNs showed better outcomes. There were no maternal deaths or thrombotic events (Alimam et al, 2016). This improvement in pregnancy outcomes is probably partly due to a more protocol-based management and a multidisciplinary approach.
Pregnancies in patients with PV should be managed under the joint care of an obstetrician experienced in the care of high-risk patients and a haematologist experienced in MPNs. Management should start with preconceptual planning for both male and female patients. The Hct should be kept within a gestational-appropriate range and potentially teratogenic medications (such as hydroxycarbamide) should be stopped a minimum of 3 months prior to planned conception in males and females. There is insufficient data regarding the safety of anagrelide during pregnancy and it should therefore be withdrawn at least 3 months prior to conception in females. Interferon should be considered for those patients who require cytoreductive therapy, though patients should be counselled regarding the potential for interferon to reduce fertility. All patients should receive low dose aspirin (unless there are patient-specific contraindications) throughout pregnancy and the postpartum period.
Patients with high risk pregnancies should be identified (Table I). These are patients with previous arterial or venous thrombosis or haemorrhage attributed to PV, previous pregnancy complications (>3 first trimester losses, >1 s or third trimester loss, birth weight <5th centile for gestation, intrauterine death or stillbirth, pre-eclampsia), extreme thrombocytosis before or during pregnancy, diabetes mellitus or hypertension requiring pharmacological treatment.
Those patients with a prior history of thrombosis and/or fetal morbidity should commence cytoreduction therapy with interferon and prophylactic LMWH in addition to aspirin. The dose and frequency of LMWH should be adjusted according to renal function, body weight and previous history. For those patients of normal body weight, no renal impairment and with a history of previous venous thrombosis or fetal morbidity, standard dose thromboprophylaxis (e.g. enoxaparin 40 mg/day) should be commenced as soon as pregnancy is confirmed. This should be increased to intermediate dose thromboprophylaxis (e.g. twice daily enoxaparin 40 mg) at 16-20 weeks. Patients with a previous history of arterial thrombosis should be commenced on intermediate dose prophylaxis (e.g. twice daily enoxaparin 40 mg) throughout pregnancy (Harrison, 2005). Interferon should be initiated for those patients with a history of PV-associated haemorrhage, extreme thrombocytosis (>1500 9 10 9 /l) before or during pregnancy and/or diabetes mellitus or hypertension requiring pharmacological treatment. All other patients (standard risk pregnancy) should receive low dose aspirin throughout pregnancy with the addition of once daily prophylactic dose LMWH for 6 weeks postpartum.
During pregnancy, the Hct should be maintained within the normal range for gestation with venesection (Table II). Interferon can be considered if venesection fails to adequately control the Hct or is not tolerated. Patients should be reminded to avoid iron supplementation in the absence of proven iron depletion. Where supplementation is offered, this should be at a low dose with regular monitoring. The full blood count, blood pressure and urinalysis should be checked every 4 weeks until 24 weeks and then every 2 weeks. Fetal growth should be assessed by regular ultrasound. Uterine Dopplers should be performed at 20 weeks' gestation to assess placental function by measuring the mean pulsatile index. If the index is >1Á4, the frequency of growth scans should be increased and consideration given to escalating treatment to include LMWH and interferon.
Dehydration should be avoided in all patients during labour and the third stage of labour should be actively managed. Thromboembolic deterrent (TED) stockings are advisable during labour and if immobile during the postpartum period. LMWH and aspirin should be stopped prior to elective caesarean section or once spontaneous labour has begun, as per local guidelines.
Low molecular weight heparin should be restarted at the earliest opportunity post-partum provided there is no significant bleeding. The dose of LMWH should be once daily and should be continued for 6 weeks post-partum. Aspirin should be continued throughout the post-partum period. Breastfeeding is safe with both aspirin and LMWH, and permissible with interferon. The Hct should be monitored in the post-partum period and maintained at less than 0Á45. Attention should be paid to the platelet count, as a rebound thrombocytosis may develop, necessitating the introduction of cytoreductive therapy.
There is no evidence regarding the management of PV during fertility treatment. Ovarian hyperstimulation is associated with an increased risk of thrombosis and therefore cytoreductive therapy (with interferon) and thromboprophylaxis with LMWH should be considered.   It can occur spontaneously or be precipitated by water or changes in temperature and can have a significant negative impact on quality of life, affecting sleep, participation in social activities and bathing (Siegel et al, 2013). The intensity of pruritus varies but can be severe causing emotional depression, anxiety and even suicide ideation. The pathogenesis of pruritus in PV is not clear. Mechanisms involving basophils (Pieri et al, 2009) and mast cells (Ishii et al, 2009) have been postulated. It can also be associated with the iron deficiency that commonly results from frequent venesection therapy. Pruritus has been found to be correlated with granulocyte JAK2 V617F homozygosity and allele burden Gangat et al, 2008). The mechanism for this association is unclear though it may relate to cytokine levels.

Recommendations: pregnancy management
The management of pruritus is challenging. Symptoms can improve or resolve with control of blood counts, using venesection alone or with cytoreductive therapy (Le Gall- Ianotto et al, 2017). Improvement in pruritus has been reported with hydroxycarbamide (Sharon et al, 1986), interferon (although this drug can also worsen pruritus) (Lengfelder et al, 2000), busulfan and 32 P (Jackson et al, 1987). Trials of ruxolitinib in PV have shown this agent to be effective in controlling pruritus in patients who have failed to respond to hydroxycarbamide (Vannucchi, 2015;Mesa et al, 2017;Passamonti et al, 2017). In those who are severely iron deficient as a result of venesection, pruritus may improve with iron replacement, if this is possible, but iron replacement must be undertaken with extreme caution and close supervision of blood counts.
However, pruritus can persist in patients with PV once their blood counts have normalised (Gangat et al, 2008). Antihistamines (both H1 and H2 receptor antagonists) are commonly prescribed though demonstrate mixed results (Weick et al, 1982;Steinman & Greaves, 1985) and combination therapy at high doses may be required to be effective. Other adjunctive therapy that have been shown to be of benefit include selective serotonin reuptake inhibitors (SSRIs) (Tefferi & Fonseca, 2002) and anticonvulsants (such as gabapentin and pregabalin). Tricyclic antidepressants (Weisshaar et al, 2012), thalidomide and naltrexone have also been found to be beneficial (Phan et al, 2010). Sometimes creams containing menthol can also provide some relief (Patel et al, 2007). From a practical point of view, it is advisable to cut nails to reduce skin damage. Joint management with a dermatologist should be considered for refractory symptoms and for those patients requiring narrow-band-ultraviolet (UVB) or ultraviolet A (UVA) (Rivard & Lim, 2005). The addition of oral psoralen to UV light (PUVA) is only occasionally used. In patients who do not require cytoreductive therapy, the above options should be tried prior to initiating cytoreductive therapy just to control pruritus.

Idiopathic erythrocytosis
Idiopathic erythrocytosis (IE) is a diagnosis of exclusion. It is an absolute erythrocytosis of no identifiable cause that is more frequent in males (Randi et al, 2016). Every effort should be made to exclude identifiable primary and secondary causes of erythrocytosis (including PV and congenital erythrocytosis) before a diagnosis of IE is made. Erythropoietin (EPO) levels are unhelpful as they are below normal in a third of patients, suggesting a primary erythrocytosis, and elevated in two-thirds, suggesting a secondary erythrocytosis. Recent studies using sequencing analysis have identified a number of patients who were previously diagnosed with IE as having mutations in either erythrocytosis-associated genes or novel genes (Bento et al, 2014;Camps et al, 2016), suggesting that at least some of these cases have a genetic basis.
Although previous data suggested that the risk of thrombosis and/or haemorrhage was considerably elevated in IE, these observations were based on old evidence that predated the identification of the JAK2 V617F mutation and therefore some patients with PV are likely to have been included in the study cohorts. More contemporary data suggests that the risk of thrombosis/bleeding is low in IE (Bertozzi et al, 2017). There are no clinical studies evaluating the use of aspirin or venesection in IE and therefore evidence-based recommendations cannot be made. Cytoreductive therapy is not indicated in patients with IE. The thrombotic risk factors should be evaluated in each case and, in selected cases, the Hct can be controlled by venesection. There is no evidence on which to determine the target Hct as relevance of data from PV patients for IE is unclear. A pragmatic approach is required for these patients with rigorous control of vascular risk factors, such as diabetes mellitus, hypertension and smoking and use of aspirin in cases where this would be otherwise clinically indicated for primary or secondary prevention. It would be reasonable to venesect patients with an arbitrary target Hct of <0Á55, although a lower target Hct of <0Á45 may be appropriate for a patient with a history of thrombosis considered to be related to the erythrocytosis.

Congenital erythrocytosis
Germ-line defects can result in a congenital erythrocytosis, which is usually diagnosed in children or young adults, often in those with a family history of erythrocytosis. A wide range of defects have been described leading to primary and secondary erythrocytosis including high oxygen affinity haemoglobins (Bento et al, 2014;Camps et al, 2016). These conditions are rare and clinical events and outcomes are poorly described. In Chuvash polycythaemia, where affected individuals have a homozygous mutation in the VHL gene increased thrombotic complications have been reported (Sergueeva et al, 2017). Major thrombotic events have also been reported to occur in young individuals with other types of congenital erythrocytosis (Bento et al, 2014). There are also reports of increased pulmonary artery pressure in both Chuvash polycythaemia and other congenital erythrocytosis (Bushuev et al, 2006;Smith et al, 2006). Recently, somatic gain-of-function mutations in EPAS1 (HIF2A) have been associated with neuroendocrine tumours in the presence of erythrocytosis (Zhuang et al, 2012) and, therefore, it may be advised to screen for these events in individuals with known mutations.
There is little hard evidence to advise on management of congenital erythrocytosis. Low-dose aspirin is of benefit in myeloproliferative neoplasms in the prevention of thromboembolic events (Landolfi et al, 2004) and, by extrapolation, may be of benefit in congenital erythrocytosis. Venesection can be used to reduce the Hct but it must be considered that the raised Hct is in keeping with the physiological consequences of the mutation and in Chuvash polycythaemia there are suggestions that venesection may be detrimental. However, in those with symptoms for which the raised Hct may be contributory or if a previous thrombotic episode has occurred, or in individuals who are asymptomatic but who have affected family members with the same genetic lesion and a thrombotic episode, venesection to a target of 0Á52 is suggested. This advice particularly pertains to those with high oxygen affinity haemoglobins (Mangin, 2017).
In Chuvash polycythaemia the VHL protein is a SOCS1coperative regulator of JAK2 and JAK2-targeted therapy may be of benefit in this disorder (Russell et al, 2011). Recently there have been reports of clinical improvement in patients with Chuvash polycythaemia on ruxolitinib (Zhou et al, 2016) and ruxolitinib may therefore be considered for the management of this specific cause of congenital erythrocytosis.

Hypoxic pulmonary disease
Erythrocytosis can be associated with advanced chronic obstructive pulmonary disease (COPD) and with obstructive sleep apnoea syndrome (OSA). In both conditions, erythrocytosis is relatively rare. In COPD, the incidence of erythrocytosis, usually defined as Hct > 0Á55, ranges from 6 to 8% (Chambellan et al, 2005;Cote et al, 2007). Incidences of 1Á7-8% are reported in OSA (Gangaraju, 2016;Nguyen & Holty, 2017). The prognostic significance of erythrocytosis in hypoxic pulmonary disease (HPD) is not known but the development of an erythrocytosis in patients with HPD is associated with an increased risk of the development of cor pulmonale and poor median survival of 2-3 years (Criner, 2000). However, in a retrospective analysis of severe COPD patients on long term oxygen treatment, a Hct ≥ 0Á55 was associated with better prognosis compared to COPD patients with anaemia (Chambellan et al, 2005;Cote et al, 2007).
The risk of PE in this setting is also unclear (Nadeem et al, 2013) as an increased Hct in the general population is also associated with increased VTE risk (Braekkan et al, 2010). A recent study suggests that patients with COPD at low risk of VTE had increased incidence of pulmonary embolism if they had concurrent erythrocytosis (Guo et al, 2016).
Additional risk factors affecting circulatory compromise and tissue oxygen delivery include carbon monoxide in smokers, extent of hypercapnia, renal blood flow, acid-base balance (pH), capacity of the bone marrow to respond to erythropoietic drive, position on the oxygen dissociation curve and changes in the peripheral vascular circulation. Furthermore, for an individual patient, co-existent age-related vascular disease may also affect thrombotic risk.
The evidence base for recommendations for the management of erythrocytosis in HPD remains limited. However, treatment of the underlying hypoxia reduces the Hct. Long term oxygen therapy improves survival in patients with COPD and severe hypoxaemia (PaO 2 below 7Á3 kPa or <8Á0 kPa with nocturnal hypoventilation) (Crockett et al, 2001).
All patients with erythrocytosis consequent upon HPD should therefore be evaluated by a respiratory physician for consideration of long-term oxygen therapy or alternative methods of improving oxygenation. If they are smokers they should be strongly advised to stop. In addition to supplemental oxygen, nocturnal oxygenation may also be improved by the use of non-invasive ventilation. Therefore, failure to achieve adequate oxygenation in HPD should not be accepted without review by a specialist respiratory physician. Of benefit regarding limited venesection in patients with HPD, reducing the Hct to 0Á50-0Á52 led to an improvement in exercise tolerance but a further staged reduction of Hct to 0Á45 did not give additional benefit as discussed in the previous guideline (McMullin et al, 2005). However, a Hct lower than 0Á45 is detrimental to these patients, and associated with poorer outcome (Chambellan et al, 2005). It is of interest that ACEi (Leshem-Rubinow et al, 2012) have the ability to reduce the Hct in this setting.
In the case of obstructive sleep apnoea, erythrocytosis is associated with nocturnal oxygen desaturation and such patients should be referred for appropriate investigation. Long term non-invasive continuous positive airway pressure (CPAP) has been shown to reduce erythrocytosis (Krieger, 1992).

Recommendations: hypoxic pulmonary disease
• Patients with hypoxic pulmonary disease who develop erythrocytosis should be evaluated by a respiratory physician for consideration of long-term oxygen therapy or alternative therapy. (GRADE 1A) • Patients who are symptomatic as a result of hyperviscosity or have a Hct >0Á56 should be considered for venesection to reduce this to 0Á50-0Á52. (GRADE 2C)

Post-transplant erythrocytosis
The definition of post-transplant erythrocytosis (PTE) varies across publications, which impacts on the estimate of incidence. A Hct over 0Á51 lasting more than a month after renal transplantation is a commonly used definition (Vlahakos et al, 2003); however, gender-specific cut-offs (males, Hct > 0Á52; females >0. 50), use of haemoglobin concentration in the definition (>170 g/l) and variable duration (>1 month, 3 months, 6 months) have also been used. The incidence of post-renal transplant erythrocytosis is between 5% and 20% and is decreasing (Kiberd, 2009). PTE also occurs in approximately 1% of patients after haematopoetic stem cell transplantation (HSCT) (Ahmed et al, 2011;Atilla et al, 2016) and after combined pancreatic-renal transplant, with a reported incidence of 16% (Guerra et al, 2010). Post-transplant erythrocytosis following renal transplantation is a benign, often self-limiting condition which typically develops within the first year following transplantation, although it can occur at 2-4 years (Charfeddine et al, 2008;Kiberd, 2009). Spontaneous resolution can occur over a period of 1-4 years (Einollahi et al, 2005) and treatment with ACEi or angiotensin receptor blockers (ARB) successfully reduces the Hct in the majority (Charfeddine et al, 2008). Indeed, the increasing use of these drugs for hypertension at earlier stages after transplantation has been deemed responsible for the falling incidence of PTE (Kiberd, 2009). Risk factors for developing erythrocytosis include male gender, retained native kidney, hypertension, renal artery stenosis, a short period of pre-transplant dialysis, reduced need for pretransplant use of EPO or being transfused. Studies also list good allograft function, diabetes, rejection-free course and use of ciclosporin adversely affecting the incidence and use of mycophenolate mofetil favourably affecting the incidence (Kiberd, 2009). Symptoms due to high viscosity, comprising dizziness, malaise, headache, lethargy and plethora, are reported. No thrombosis or mortality has been reported in recent studies (Kiberd, 2009). Venesection has been used by some in patients with persistently high Hct (>0Á57-0Á60) and helps reduce the Hct but does not reduce the incidence of thrombosis (Charfeddine et al, 2008). Aspirin use has been variable and its effect is not clear.
Mechanisms of post-renal transplant erythrocytosis have been reviewed (Vlahakos et al, 2003). Recipient-related, rather than donor-related, features underlie the development of PTE. The erythrocytosis is EPO-responsive, with high EPO levels at onset reducing to lower levels after suppression of the renin-angiotensin-system (RAS) using ACEi, ARBs or native kidney nephrectomy; recurrence after discontinuation of these drugs is accompanied by increased EPO levels (Vlahakos et al, 2003). Retention of native kidneys is thought to lead to a 'hyper-erythropoietinaemia'. Unlike suppression of the RAS, venesection increases EPO levels. The renin-angiotensin, androgen and adrenergic systems increase erythropoiesis and sensitivity of erythroid progenitors to EPO. The likely common pathway is via insulinlike growth factor-1 (IGF1) which increases erythropoiesis. ACEi reduce IGF1 levels thus reducing erythropoiesis. Pancreatic plus renal transplant increases IGF1 levels, and use of systemic venous drainage (as opposed to portal venous drainage) was associated with a higher incidence and earlier onset of PTE (Guerra et al, 2010). These patients also required more frequent venesection to maintain a stable Hct. Treatment with ACEi or ARB reduces Hct to a nadir level at 3 months with stable results for several years. Discontinuation usually leads to recurrence of post-transplant erythrocytosis, but 20-30% of patients maintain a normal Hct after cessation. The use of aggressive venesection leads to iron deficiency.
Haematopoetic stem cell transplantation-related erythrocytosis has been observed in about 1% of patients transplanted for aplastic anaemia but also in other conditions, including myelodysplastic syndromes and acute leukaemia (Ahmed et al, 2011;Atilla et al, 2016). Patients who developed PTE did not have total body irradiation and were mostly men. Erythrocytosis was asymptomatic, with no thrombosis or mortality reported. All were treated with venesection on an ongoing basis. PTE after HSCT starts, on average, about 11 months after the transplant and can be self-limiting, lasting approximately 6 months.
Persistent post-transplant erythrocytosis unresponsive to ACEi, ARB or venesection requires further investigation to define underlying causes, including JAK2 V617F mutation analysis.

Cyanotic congenital heart disease
Patients with congenital cyanotic heart disease are best managed by specialist cardiology clinics. Decisions about venesection for these patients should not be made in a haematology clinic.

Conclusion
Using the available evidence and best practice we have suggested practical guidance for the management of specific situations and complications of polycythaemia vera.

Declaration of Interests
The BSH paid the expenses incurred during the writing of this guidance. All authors have made a declaration of interests to the BSH and Task Force Chairs which may be reviewed on request.