2018/11/27 by Mary Frances McMullin, Claire Harrison, Claire N. Harrison +14 · 25 citations
Medicine · Biochemistry, Genetics and Molecular Biology · #Myeloproliferative Neoplasms: Diagnosis and Treatment #Kruppel-like factors research #Renal Diseases and Glomerulopathies
paper · doi:10.1111/bjh.15648
This guideline was compiled according to the British Society for Haematology (BSH) process at b-s-h.org.uk. The Grading of Recommendation Assessment, Development and Evaluation (GRADE) nomenclature was used to evaluate levels of evidence and to assess the strength of the recommendations. The GRADE criteria can be found at www.gradeworkinggroup.org. The literature review was conducted on 2 March 2017. Databases searched include MEDLINE(OVID), Embase (OVID) and CENTRAL(The Cochrane library) using the search terms (and relevant MESH terms): polycythaemia vera, erythrocytosis, familial, high oxygen affinity haemoglobin, defects of oxygen sensing pathway, diagnosis, investigation, molecular, mutation, JAK2, MPL, CALR, bone marrow, red cell mass, erythropoietin, risk, management, treatment, cytoreduction, venesection, hydroxyurea, interferon, busulfan, pipobroman, radioactive phosphorus, aspirin, anagrelide, ruxolitinib, thrombosis, haemorrhage, pregnancy, pruritus, surgery and management. The search covered the period from 2005, the date of last version of the guideline (McMullin et al, 2005), to February week 3 2017. Exclusions included articles not in English, studies not in humans, single case reports and case series of under 5 cases. A total of 6062 articles were identified which, with exclusions and duplications, resulted in 1215 articles which were reviewed. Review of the manuscript was performed by the BSH Guidelines Committee General Haematology Task Force, the BSH Guidelines Committee and the General Haematology Sounding Board of BSH. It was also placed on the members section of the BSH website for comment. A patient representative from MPN-Voice ( www.mpnvoice.org.uk) participated in the guideline writing meeting. The guideline has been reviewed by MPN-Voice; this organisation does not necessarily approve or endorse the contents. The previous guideline was published in 2005 (McMullin et al, 2005) with an amendment in 2007 (McMullin et al, 2007) to update the diagnostic criteria following the discovery of the JAK2 mutation in patients with polycythaemia vera (PV). Since that time, there has been a considerable amount of research in the area concerning diagnostics, risk stratification, new agents and reinvestigation of existing agents. It was therefore decided to evaluate the literature to formulate guidance on the diagnostic pathway for erythrocytosis, risk stratification of PV, management of PV (including specific situations) and the management of secondary erythrocytosis. Here we provide evidence-based guidance on diagnosis, risk stratification and management of PV. Our review of the evidence led us to some differences in diagnostic criteria and risk stratification than have been proposed by other international organisations. We discuss the reasons for this. An accompanying guideline looks at management of specific situations in PV and management of secondary erythrocytosis (McMullin et al, 2018). Patients with a persistently raised venous haematocrit (Hct) (males, >0·52; females, >0·48) should be investigated. As suggested in our previous guideline (McMullin et al, 2005) and confirmed in recent literature, Hct has been consistently shown to perform better in identifying patients with a raised red cell mass (RCM) than haemoglobin concentration (Alvarez-Larrán et al, 2012; Ancochea et al, 2014). Patients should be investigated according to the proposed algorithm (Fig 1). Investigation requires knowledge of diagnostic criteria for both PV and potential secondary causes of erythrocytosis (Tables 1 and 2). A detailed history, examination and stage 1 investigations (listed below) should identify a potential cause in the majority of patients, although a proportion will require more extensive testing and, in some cases, a cause may not be found (idiopathic erythrocytosis). The potential for dual pathology should also be considered. A detailed clinical history and examination are essential and, in the absence of a molecular marker of disease, will determine further investigations and management. Particular attention should be paid to the drug history (prescribed and recreational), smoking, alcohol consumption and body habitus. Systematic questioning should elicit symptoms related to other potential secondary causes of erythrocytosis (see Table 2). A proportion of patients, who have a clear secondary cause for their erythrocytosis, may not need any further investigations. The full blood count analysis will not only confirm a raised Hct but will also identify neutrophilia and thrombocytosis, which are common in JAK2 V617F-positive PV and part of the criteria for JAK2-negative PV (Table 1). As smokers have a significantly higher neutrophil count than non-smokers (Whitehead et al, 1995), neutrophilia is defined as >12·5 × 109/l in this patient group. A blood film should be reviewed in all patients to look for any atypical features. In those with confirmed PV, abnormalities, such as circulating blasts, leucoerythroblastic features and monocytosis, would be indications for bone marrow assessment. A number of renal and hepatic diseases can cause erythrocytosis. Serum calcium levels should also be determined to exclude a parathyroid adenoma/carcinoma, which rarely causes secondary erythrocytosis. Identifying tissue hypoxia, a cause of secondary erythrocytosis, can be achieved most simply by using pulse oximetry in the clinic. An SaO2 of <92% has been shown to be associated with an absolute erythrocytosis (Berlin, 1975). Clinicians should however be aware of three situations of hypoxic erythrocytosis where this testing is unreliable and will give a normal result. These are: carbon monoxide poisoning, high oxygen affinity haemoglobins and sleep apnoea syndrome. Those with suspected high oxygen affinity haemoglobins should undergo genetic testing as described below. In those with suspected sleep apnoea (heavy snoring with daytime somnolence or increased body mass index >30 kg/m2), referral should be made to a respiratory or sleep physician. Carboxyhaemoglobin (COHb) levels are significantly higher in smokers compared with non-smokers and cigarette consumption has been shown to be directly related to COHb levels (Castleden & Cole, 1975). Testing can therefore be performed at baseline where smoking is suspected. Low serum ferritin levels are common in PV patients and iron deficiency can mask the presentation of PV, giving a misleadingly low Hct because iron deficiency limits erythropoiesis and hypochromic microcytosis develops. Erythrocyte production is controlled by the hormone erythropoietin (EPO). Measurement of serum EPO can provide information on potential causes of erythrocytosis and help stratify further testing (see Fig 1). EPO levels are commonly high in hypoxic conditions or when erythrocytosis is secondary to exogenous administration or endogenous overproduction. In contrast EPO levels are typically low in PV, although their diagnostic utility in this setting is limited in the era of JAK2 mutation testing (Ancochea et al, 2014). The identification of JAK2 mutations in almost all PV patients has revolutionised the diagnosis of PV. The JAK2 V617F mutation can be found in over 95% of PV patients (James et al, 2005) and an exon 12 mutation in most remaining patients (Scott et al, 2007). Testing for JAK2 V617F in peripheral blood is sensitive and bone marrow samples are not required to identify this (Takahashi et al, 2013). Testing for JAK2 V617F is advised as a stage 1 investigation and should confirm the diagnosis the vast majority of PV patients. Separate guidance is available for assays used for detection of JAK2 mutations (Bench et al, 2013). Further investigations are warranted in those patients with a persistent, significant erythrocytosis if JAK2 V617F studies are negative and a secondary cause is not immediately apparent (See Fig 1). Secondary causes must be considered because PV is rare in the absence of a JAK2 V617F mutation. Patients with Hct >0·60 (males) or >0·56 (females) can be assumed to have an absolute erythrocytosis, but in others RCM studies can be helpful to confirm an absolute erythrocytosis. An RCM more than 25% above the mean predicted value is diagnostic of an absolute erythrocytosis (Pearson et al, 1995). Those with a raised Hct but an RCM within the normal range have an apparent erythrocytosis. A relative erythrocytosis, found in states of dehydration, can be confirmed when the RCM is within the normal range and plasma volume is below normal. Patients with a relative or apparent erythrocytosis require no further investigation. It is noted however, that due to the many drawbacks of this test including cost and labour, access to RCM studies is variable nationally. Radiological splenomegaly is a minor criterion for JAK2 V617F-negative PV (Table 1) and ultrasound is the simplest method for detection. Abdominal ultrasound can also exclude secondary causes of erythrocytosis, particularly renal and hepatic pathology, including hepatocellular carcinoma. Further testing can be stratified according to the EPO level measured during stage 1 investigations. Compared with JAK2 V617F, patients with exon 12 mutated-PV tend to be younger, with higher haemoglobin concentrations, lower white blood cell (WBC) and platelet counts, and an isolated increase in erythropoiesis without granulocytic or megakaryocytic morphological abnormalities (Scott et al, 2007; Passamonti et al, 2011). In contrast to JAK2 V617F testing, a discrepancy between exon 12 mutant allele burden in bone marrow and peripheral blood has occasionally been described (Kjær et al, 2012). Bone marrow histology may be helpful in distinguishing PV from secondary erythrocytosis (Thiele et al, 2005). Bone marrow aspiration in PV typically reveals markedly increased erythropoiesis with moderate to marked increase in granulopoiesis and megakaryopoiesis; widely variable megakaryocyte size, including large forms with hyperlobated nuclei; and absent iron stores. The bone marrow trephine biopsy sections show hypercellularity, trilineage expansion of haemopoiesis (rarely preferentially erythroid) and normoblastic erythropoiesis. Granulocytic maturation may be left-shifted and disorderly; megakaryocytes show increased variation in size, often with a predominance of large forms with uneven or reduced nuclear lobulation, and megakaryocyte clusters are common. Reticulin is increased in a minority of patients [up to World Health Organization (WHO) grade 1 in most cases]. Presence of an acquired genetic abnormality is a major criterion for JAK2-negative PV and the presence of an abnormal karyotype can therefore support this diagnosis. An acquired SH2B3 (LNK) mutation would also support the diagnosis. More recently, mutations in a number of other genes, most commonly TET2 and DNMT3A, have been reported in PV (Delic et al, 2016) These gene mutations are not, however, disease-specific and have also been reported in healthy individuals (Genovese et al, 2014; Jaiswal et al, 2014), limiting their application in the diagnostic setting, particularly when found in isolation. A raised EPO level should lead to a thorough search for secondary causes of erythrocytosis (Table 2), which may require additional supplementary investigations (Fig 1). Further imaging (e.g. head and neck computed tomography) is indicated in this setting if a cause for the high EPO has not been identified. Imaging should aim to exclude rare tumours, such as a cerebellar haemangioblastoma, pheochromocytoma, meningioma or a parathyroid tumour, all of which can rarely cause erythrocytosis. A number of germline mutations in genes involved in oxygen sensing, erythropoiesis and oxygen transport have now been implicated in patients with otherwise unexplained erythrocytosis. These include mutations in the EPO receptor and genes in the oxygen sensing pathway (VHL, EGLN1, EPAS1), high oxygen affinity haemoglobinopathies caused by mutations in the globin genes HBA1, HBA2, HBB, and 2,3-bisphosphoglycerate deficiency as a result of BPGM mutations (Bento et al, 2014). These patients can present with low, normal or high EPO levels; high affinity haemoglobins and 2,3-BPG deficiency also cause a left shift of the oxygen dissociation curve. In the past, this group of mutations was detected by Sanger sequencing of individual genes, in an order directed by EPO levels and P50 analysis, but this approach is labour-intensive and time-consuming (Bento et al, 2014). More recently, next generation sequencing-based targeted panels have been developed to assess established and novel genes implicated in erythrocytosis (Camps et al, 2016), negating the need for P50 testing. A targeted panel should be performed in patients in whom congenital erythrocytosis is suspected, particularly young patients or those with a family history. Polycythaemia vera (PV) presents at a median age of 60 years with a slight male predominance. Patients can present with arterial or venous vascular occlusive events, microvascular disturbances or, occasionally, haemorrhage. Splenic pain and/or enlargement, pruritus, gout and constitutional symptoms, such as fatigue, may be present. Alternatively, asymptomatic patients may be identified incidentally following a full blood count. All patients who are newly diagnosed with PV should be discussed in a multidisciplinary team setting. The recommended diagnostic criteria for JAK2-positive and the very rare JAK2-negative PV are given in full in Table 1. Although the WHO classification considers histology to be useful in distinguishing PV from other myeloproliferative neoplasms (MPNs) (Arber et al, 2016), several studies have reported high rates of non-consensus or failure to reach a histological diagnosis in patients with PV (Koopmans et al, 2011; Madelung et al, 2013) (Alvarez-Larrán et al, 2014a) Given the uncertain utility of bone marrow histology in the diagnosis of uncomplicated PV, it is not mandatory in all patients, but should be considered if there are atypical features, such as marked splenomegaly or a history of splanchnic vein thrombosis, where it is necessary to establish if there is an occult MPN. The degree of baseline fibrosis can also be ascertained, which, as discussed below, may have a prognostic role. Bone marrow biopsy may nonetheless be useful in those patients likely to have a long disease history, as a baseline sample for comparison in the event of suspected disease transformation. Abnormal karyotype and other molecular abnormalities (e.g. TET2 mutations) have been reported in PV and some may have prognostic value (Delic et al, 2016; Cerquozzi et al, 2017) but these tests are not routinely required at diagnosis. In patients with a JAK2 V617F mutation, haemoglobin concentration and/or Hct are currently used as a surrogate for RCM to distinguish between PV and essential thrombocythaemia (ET) (Arber et al, 2016), but Hct has the better accuracy in predicting RCM (Alvarez-Larrán et al, 2012; Ancochea et al, 2014). Concerns have however been raised that distinguishing PV from ET based on blood count thresholds alone may fail to identify a subgroup of patients with "masked" PV, who may be better managed as PV rather than ET. The definition of masked PV has been inconsistent across studies. When using a raised RCM to define an erythrocytosis, studies have shown that a Hct threshold of 0·52 in males will fail to identify approximately 20% of male patients with a raised RCM, whilst the threshold of 0·48 in women is more sensitive (Alvarez-Larrán et al, 2012). These ''masked'' PV patients were reported to have similar outcomes to those with "overt" PV when managed equivalently (Alvarez-Larrán et al, 2016). By contrast, other studies have defined masked PV as those patients who did not meet the haemoglobin-based thresholds for PV but did meet the other WHO criteria, mainly bone marrow histology and JAK2 status. Patients meeting this definition had poorer outcomes in terms of myelofibrotic or leukaemic transformation and survival, but no difference in rate of thrombosis. Hct thresholds of 0·48/0·49 (females/males) were subsequently proposed to discriminate ET from PV in the WHO 2016 revision (Arber et al, 2016). The proposed BSH Hct-based thresholds have good specificity but will miss a minority of patients with a raised RCM. By lowering the Hct threshold it may be possible to identify patients with histology more typical of PV who may have certain adverse outcomes, but these findings have not yet been reproduced independently. It is unknown whether management of any of these patients using a strict Hct target benefits their vascular risk. However, in patients with a JAK2 V617F mutation and borderline Hct levels (especially males with Hct 0·48–0·52), the possibility of true erythrocytosis should be considered, especially if the patient is at high risk of vascular events. Options in this group include performing an RCM study to clarify the diagnosis or, pragmatically, managing the patient with a Hct target as for PV. It should also be noted that Hct is a poor surrogate of RCM in patients who have had splanchnic vein thrombosis (Lamy et al, 1997), and these high-risk patients are best managed with standard blood count targets regardless of blood count parameters at diagnosis. Hct is also reduced by pregnancy and gestation-specific ranges should be used when considering the distinction between PV and ET in a patient presenting during pregnancy. Quantitative assessments of the JAK2 V617F allele burden in peripheral blood granulocytes have shown that this parameter tends to be higher in PV than ET. A higher mutant allele burden correlates with certain clinical features at presentation, including higher haemoglobin levels, higher WBC counts, lower platelet counts, lower mean cell volume (MCV), lower serum ferritin and EPO, more splenomegaly and more pruritus (Dupont et al, 2007; Tefferi et al, 2007; Vannucchi et al, 2007; Passamonti et al, 2010). However, there is no validated threshold at which JAK2 V617F allele burden can confirm or refute a diagnosis of PV and this investigation is not recommended routinely. A low level JAK2 V617F mutation (allele burden <1–3%) should be interpreted in the context of clinical, haematological and other laboratory findings (Bench et al, 2013). If the result is reproducible and does not represent a false positive, this finding may provide support for a diagnosis of a PV in a patient with otherwise unexplained, significant erythrocytosis (Perricone et al, 2017). However, the JAK2 V617F mutation has been identified in normal individuals, often at a low allele burden, with a frequency that increases with age (Genovese et al, 2014; Jaiswal et al, 2014). Caution is therefore warranted and comprehensive investigations should exclude an alternative secondary or congenital cause of erythrocytosis. The test should preferably be repeated within 3–6 months, and clinical assessment for other features of an MPN, e.g. splenomegaly, bone marrow histological features, and screening for an additional mutation in JAK2 exon 12 may be helpful. The principal aims of risk stratification in PV are a) to select patients at higher risk of thrombosis for consideration of and to provide the most information to patients on the and of a diagnosis of PV. diagnosis, in the study to the on in Polycythaemia age years and a history of thrombosis were found to be the most of et al, 2005). A baseline WBC count of × 109/l is a significant of thrombosis, particularly an increased risk of et al, however, prognostic including have not been risk arterial also to risk in PV et al, 2017) et al, 2007) et al, 2007). is more in patients with Hct et al, 2013) and when the WBC count × 109/l et al, A between at diagnosis or and risk has not been established in PV et al, but × is associated with increased risk of due to acquired disease and should be considered an for of & The of degree of and history of venous thrombosis on long are In the age years was associated with and age years was associated with increased of et al, 2005). In the and years was associated with and et al, 2013). disease has been associated with increased risk of myelofibrotic transformation et al, 2005). with WBC count × 109/l is associated with an A history of venous thrombosis also on The these three parameters WBC count and to risk for et al, but this prognostic has not been validated in studies. other clinical and laboratory have been reported to and/or risk of disease transformation. The presence of splenomegaly in PV patients has been associated with and increased risk of transformation to both and et al, 2011). The presence of an abnormal karyotype and et al, 2013). A raised level and the presence of fibrosis at diagnosis a higher rate of transformation to but not to (Alvarez-Larrán et al, et al, 2012). analysis that a JAK2 mutant allele burden of is also associated with increased risk of not of or et al, but the clinical utility of this is not yet Although JAK2 disease has a clinical to JAK2 PV haemoglobin lower WBC there to be no difference in et al, 2011). gene sequencing is a approximately of PV patients have mutations of of and and these patients have a reduced rate of in analysis et al, 2016). Patients with PV may present with thrombosis or PV symptoms, such as microvascular pruritus may be and fatigue, may also be presenting features which can significantly on of et al, 2017). However, patients may be asymptomatic at The of are to and therefore (Table is related to and the principal aim of is to this risk. assessment and management of risk such as and smoking, is in burden is also a target for is evidence that patients with controlled PV, as determined by splenomegaly and have a significantly higher measured by the et al, 2016). for may the need for and or of The target for Hct in PV was based on from assessment of of vascular at Hct levels and it was that a target below should be (Pearson & This target has now been validated in a clinical by the in PV who the of Hct to compared with a more target range of et al, 2013). Patients with a Hct target of had a significantly lower rate of and major thrombosis than those with a target of et al, 2013). It was noted that the median WBC count was significantly lower in the low Hct which may have been related to variation in the of between the The of this parameter on the difference in between the has been (McMullin et al, 2013). The by recommended criteria for PV. however, evidence that of these to outcomes from the Hct These are a of to assess with across clinical but are not as useful in clinical et al, is considerable published evidence of an between increased WBC count and thrombosis risk in PV et al, et al, et al, et al, et al, 2017). In contrast, study did not such an et al, 2010). An analysis of of patients the study that in patients with WBC × 109/l there was increased of thrombosis in comparison with those with WBC × related to an increase in et al, 2007). In a study of PV to determine whether blood the rate and survival, age and level at diagnosis were found to be risk for vascular When the vascular of the patients with a had WBC compared with 20% of patients without a et al, study which a lower risk in those managed to Hct a lower WBC which may have to the lower rate of et al, 2013) (McMullin et al, 2013). is no evidence from to determine whether targeted at count on and therefore no to target WBC as a can be no evidence for or lower thrombosis risk was in patients or according to criteria in an analysis of PV patients with a better was when there was a WBC and platelet (Alvarez-Larrán et al, 2012). is evidence at of platelet there is a risk for and which may in those with high is currently no to allele burden the clinical setting. many studies have used allele burden to assess of treatment, there is currently no clear clinical of this as a burden over may with to et al, but there is no evidence that this and no evidence that lowering allele burden is no that of bone marrow or fibrosis grade is of value but this should be if there is suspected from blood or that of blood at for patient should be used to and a Hct of et al, 2013). In low risk patients this is to target is to this target an alternative approach using a may need to be considered. levels of have been reported to have an with higher thrombosis risk in patients on in those patients 3 or more (Alvarez-Larrán et al, 2017). study has defined a difference in Hct A target Hct in males and is not deficiency may result from this is concentration fatigue, and and other symptoms may a administration must be with and with and of blood symptoms may an alternative such as The value of in patients with PV was in the In this those to had significantly vascular at 3 years compared to was a in the risk of the