Guidelines for the diagnosis, investigation and management of polycythaemia/erythrocytosis
Year: 2005
Type: article
Abstract: Traditionally, polycythaemia has been used to identify a group of varied disorders with an increase in circulating red cells that are typified by a persistently raised haematocrit (Hct). Since only the red cell lineage is involved, the term erythrocytosis has more validity and will be used throughout this article. Polycythaemia will be retained in relation to the clonal disorder, polycythaemia vera (PV), in which three cell lineages are involved. The purpose of this guideline is to provide a rational approach to the diagnosis, investigation and management of patients with an erythrocytosis. This will include recommendations on the management of PV, apparent and relative erythrocytosis, idiopathic erythrocytosis and the secondary erythrocytoses because of high oxygen affinity haemoglobin, hypoxia because of chronic lung disease, congenital cyanotic heart disease and postrenal transplantion. The guideline group was selected to include UK-based medical experts. The drafting group met (real or virtual) on four occasions and communicated by e-mail. Each member of the group was allocated responsibility for the preparation of a selected component of the first draft. Medline, CANCERLIT and EMBASE were systematically searched for publications in English from 1966 to June 2004. Relevant literature in group members own collections and older references generated from initial papers were also examined. The Cochrane controlled trials register and the Cochrane optimal search strategy for randomised controlled trials was searched but no additional material was identified. Randomised trials and series of patients and single case reports were considered if appropriate. Meeting abstracts were not included in the systematic search strategy. The group leader synthesised the draft components, which were subsequently revised by consensus. No recommendations are included for which full consensus was not achieved. The guideline was reviewed by sounding boards, and British Committee for Standards in Haematology (BSCH) and comments were incorporated where appropriate. Criteria used to quote levels and grades of evidence are outlined in Table I. Patients with a persistently raised venous haematocrit (Hct) (>0·52 males, >0·48 females for >2 months) should, in general, be investigated by measurement of their red cell mass (RCM). A number of physiological factors have been shown to influence the Hct value, although in practice the use of only minimal or no venous occlusion when taking the blood sample is the most important. In addition, Coulter ‘S’ and Coulter ‘S Plus’ analysers are known to underestimate Hct [approximately 7% at reduced mean cell haemoglobin (MCH) values]. As a result, a correction factor should be applied (Guthrie & Pearson, 1982). However, males and females with Hct values above 0·60 and 0·56 respectively can be assumed to have an absolute erythrocytosis and do not require confirmatory studies (Pearson, 1991). The World Health Organisation (WHO) criteria (Pierre et al, 2001), which accept a haemoglobin of greater than 18·5 in males and 16·5 in females, has not been verified. PV can be masked in patients that present with iron deficient anaemia and it may be necessary to administer iron to correct this. If iron is administered, it should be done extremely judiciously with at least weekly monitoring of Hct as this parameter can rise very rapidly and may precipitate thromboembolic events. Recommended methods for the measurement of RCM have been drawn up by the Radionucleotide Panel of the International Committee for Standardisation in Haematology (ICSH) (ICSH, 1980). These methods have now been widely adopted and provide within-method accuracy in the order of 2–3%. Traditionally, results and normal ranges have been expressed in terms of ml/kg total body weight. Indeed, the widely quoted Polycythemia Vera Study Group (PVSG) criteria include RCM values expressed in ml/kg (Berlin, 1975). However, this approach can lack precision in obese individuals as, since adipose tissue is relatively avascular, it may result in high predicted normal values and low measured values. In order to overcome this limitation, the ICSH prediction formulae, based on surface area for RCM, showed that the scatter of results from 98% of males and 99% of females fall between ±25% of the mean value at any given surface area (Pearson et al, 1995). Therefore, using these limits as the reference range, the diagnosis of absolute erythrocytosis is made when an individual's measured RCM is more than 25% above their mean predicted value. Individuals with a raised venous Hct but whose RCM falls within the reference range have an apparent erythrocytosis. ‘Relative erythrocytosis’ as a pathological state, where the RCM is in the normal reference range and the plasma volume is below the reference range, generally only exists in true pathological states of dehydration (Brown et al, 1971). RCM should be expressed in relation to surface area as recommended by the ICSH. The term ‘relative erythrocytosis’ should be reserved for states of dehydration. Apparent erythrocytosis should be used for those individuals who have a raised venous haematocrit but with a red cell mass within the reference range. The classification of absolute erythrocytoses is shown in Table II. Once an absolute erythrocytosis has been confirmed it is desirable to identify the underlying aetiology, although this may not be possible either initially or after prolonged investigation. Nevertheless, the starting point is knowledge of the underlying causes of a secondary erythrocytosis (Table II) and the diagnostic criteria of PV (Table III). Frequently, it can be difficult to prove conclusively that an erythrocytosis is either secondary or primary and dual pathologies resulting in an erythrocytosis should be considered, especially in the elderly. Two stages of diagnostic tests are listed in Table IV (see below for comment). The number and order of tests and intensity of investigation depends on the clinical features after assessment. However, stage 1 investigations, including serum erythropoietin (EPO) levels and blood gas measurements, should be undertaken in all patients and can facilitate the diagnosis in the majority of cases. Selected tests in stage 2 are undertaken following an evaluation of the initial results. A neutrophilia is present in approximately two-thirds and thrombocytosis in 50% of PV cases (Berlin, 1975) and, as a result, provide useful minor criteria for the diagnosis of PV (Pearson & Messinezy, 1996). It should be noted that smokers have significantly higher neutrophil counts than non-smokers (Whitehead et al, 1995) and it has been suggested that, in smokers, the upper limit of normal for neutrophil count should be taken as 12·5 × 109/l. Low serum ferritin levels are more commonly seen in PV than secondary erythrocytosis. Indeed, the absence of iron stores is a frequent finding in PV marrow histology. Although an elevated B12 level is characteristic of PV, resulting from transcobalamin release from an increased granulocytic mass, it is not an essential investigation. Erythrocytosis may be associated with both renal and hepatic disease. Serum calcium levels should be determined to exclude the very rare secondary erythrocytosis caused by parathyroid adenomas or carcinoma. The measurement of arterial oxygen saturation (SaO2), a sensitive indicator of tissue hypoxia, is most easily achieved with the use of a pulse oximeter. However, there are three situations causing hypoxic erythrocytosis in which the SaO2 can be misleading: carbon monoxide poisoning, the presence of high oxygen affinity haemoglobins and sleep apnoea syndrome. Most instruments provide carbon monoxyhaemoglobin (COHb) measurements and this value should be subtracted to give an accurate SaO2 result. Smokers generally have higher COHb levels, although a secondary erythrocytosis because of smoking alone is uncommon (Smith & Landaw, 1978). High oxygen affinity haemoglobins, as well as congenitally low 2,3-bisphosphoglycerate levels, will give rise to a normal SaO2, despite tissue hypoxia and measurement of the p50 is important to exclude these rare conditions. A SaO2 below 92% has been taken to indicate a causal relationship with an absolute erythrocytosis (Berlin, 1975). A normal daytime SaO2 can also be seen in the sleep apnoea syndrome and supine hypoventilation because of premature airway closure. It is important therefore to consider these conditions and to enquire about symptoms relating to nocturnal oxygen desaturation, for example snoring, nocturnal restlessness and daytime somnolence. Nocturnal reduction in SaO2, usually but by no means exclusively seen in obesity (Pearson & Treacher, 1990), may be found in 10–20% of patients who would otherwise have been classified as having idiopathic erythrocytosis (Moore-Gillon et al, 1986). Review by a chest or sleep physician (to assess the need for a respiratory sleep study) is recommended in patients with erythrocytosis of all types who are known to snore heavily and either have unwanted daytime somnolence [defined as an Epworth Score >10/24 (http://www.stanford.edu/dement/epworth.html)] (Johns, 1993) or who are significantly overweight (body mass index >30 kg/m2). A chest X-ray is also recommended to exclude lung pathology. Since erythrocyte production is controlled by EPO, a serum EPO level can provide information as to whether the erythrocytosis is hormonally mediated or autonomous. In patients with erythrocytosis secondary to hypoxia, serum EPO levels are typically raised. In contrast, EPO levels in patients with PV are characteristically reduced and remain low in the majority of cases even following adequate venesection (Messinezy et al, 2002a). EPO values below the reference range, however, can be seen in idiopathic erythrocytosis, a fact that lowers the specificity of low EPO levels for PV. In addition, a normal serum EPO level excludes neither hypoxia nor PV as the cause of erythrocytosis. Nevertheless, with the availability of specific, sensitive and reproducible EPO assays a low serum EPO levels can now be used as a minor criterion in the diagnosis of PV (Messinezy et al, 2002b). Abdominal ultrasound is an essential investigation in all patients with a proven absolute erythrocytosis to exclude underlying renal and hepatic pathology. In the absence of liver disease, a palpable spleen is a reliable sign of PV and, as such, has been adopted as a major criterion for its diagnosis. Splenomegaly can be found in two-thirds of PV cases by various imaging techniques, although ultrasound is the simplest (Messinezy et al, 1997). However, in view of the significant inter-observer error of scanning detection of non-palpable splenomegaly, it has been proposed that this finding be taken as a minor criterion (Table III). In females, pelvic ultrasound would detect leiomyomas, which have occasionally been found to be a cause of secondary erythrocytosis. Bone marrow aspirate and trephine biopsy are not required to meet the diagnostic criteria for PV (Table III). However, these investigations provide useful information including confirmation of the diagnosis of PV, differentiation from secondary erythrocytosis and other myeloproliferative disorders (MPDs), and assessment of the degree of fibrosis. They also provide a baseline which can be compared with subsequent bone marrow examinations to assess disease progression or response to therapy. The finding of a chromosomal abnormality establishes clonality. In PV, aspirated bone marrow is expected to have dense particles and cellular trails. There is usually marked erythroid hyperplasia with moderate to marked hyperplasia of granulopoiesis and megakaryopoiesis. As well as increased cells of the neutrophil lineage, eosinophils and basophils, but not monocytes, are increased. Wide variations in megakaryocyte size, including larger variants with hyperlobated nuclei are characteristic. Iron stores are typically absent. The bone marrow trephine core is hypercellular for the patient's age. There is trilineage involvement, not selectively and only rarely preferentially erythroid (Pierre et al, 2001). Erythroid maturation is maintained and is normoblastic, although erythropoietic nests may be abnormally located abutting trabeculae. Granulocyte maturation is also maintained. It may be left-shifted and distributed in a disorderly fashion, involving loss of the usual preferential localisation of promyelocytes and myelocytes at trabecular margins. There is increased variation in megakaryocyte cell size. Large variants are often predominant and may have uneven or reduced nuclear lobulation. Clusters of megakaryocytes are common and typically pleomorphic, i.e. contain mixed large and small cells. There is usually a mild to moderate increase in stromal reticulin (grade 2–3) (Imbert et al, 2001). Histological features supporting a diagnosis of secondary erythrocytosis rather than PV include the presence of stromal inflammatory features, such as increased plasma cells, increased haemosiderin in stromal macrophages and evidence of abundant background apoptotic activity (Thiele et al, 2001a). Bone marrow histology is a B criterion in the WHO classification and is helpful in distinguishing PV from reactive conditions with secondary erythrocytosis. However, there is histological overlap with other chronic myeloproliferative disorders (CMPD), which currently limits the usefulness of histology in sub-classification (Pierre et al, 2001). (In doubtful diagnostic situations, haematologists or pathologists may find it helpful to refer the sample to a haematopathologist with expertise in trephine biopsy interpretation.) Cytogenetic abnormalities are found in 10–20% of patients with PV. The abnormalities trisomy of chromosomes 8 and 9, del (20q), del (13q) and del (1p) are the most commonly found. An abnormal karyotype, a clonality marker, is a major diagnostic criterion. Patients who progress to myelodysplastic syndrome or acute leukaemia almost always have a karyotypic abnormality (Pierre et al, 2001). The culture of the mononuclear non-adherent fraction of peripheral blood cells or bone marrow cells of patients with PV in serum-containing medium without the addition of EPO leads to the growth of BFU-E – so-called ‘endogenous erythroid colonies’. This feature, which is not usually found in normal individuals or patients with secondary erythrocytoses, has been shown to be a good marker for PV (Partanen et al, 1989). Despite these observations, culture techniques are not standardised and are expensive. It is for these reasons that the finding of endogenous erythroid colonies is best regarded as a minor diagnostic criteria. It is important to examine the p50 in patients with unexplained erythrocytosis to exclude a high affinity haemoglobin. There are a large number of beta-chain haemoglobin variants that have increased oxygen affinity and a resultant left-shifted oxygen dissociation curve. (For this test, contact Special Haematology, St Thomas', London on 020 7188 3421 or Department of Haematology, City Hospital, Birmingham, UK, on 0121 507 4577 to discuss. A 5 ml EDTA sample will be required from the patient and a normal control, bled at the same time.) Patients with an unexplained erythrocytosis and low serum EPO levels should be considered for investigation of an EPO receptor mutation (Percy et al, 1998). The Chuvash form of erythrocytosis, an autosomal recessive disorder common to a large number of families in central Russia, has been shown to result from mutations in the VHL gene. Recently, a small number of patients of Asian and Western European ancestry with erythrocytosis have also been reported to have VHL mutations (Percy et al, 2003). These patients have inappropriately normal or high EPO levels for their Hct. (Investigation of EPOR and VHL gene mutations can be analysed by sending an EDTA sample by post to Dr M Percy, Department of Haematology, C Floor, Belfast City Hospital, Lisburn Road, Belfast, UK, 028 9032 9241, ext. 3325.) A novel cell surface receptor Polycythaemia rubra vera-1 (PRV-1) has been found to be overexpressed in granulocytes from PV patients. Decreased c-MPL protein expression has been found in platelets from patients with PV. These tests are interesting but not currently useful diagnostic tests (Kralovics et al, 2003). Polycythaemia vera presents at a median age of 60 years. The incidence is the same in males and females. Reports of the annual incidence of PV vary widely, from 0·2/106/year in Japan (Kurita, 1974) to 28/106/year in Goteborg, Sweden (Kutti & Ridell, 2001). Patients often present with either arterial or venous vascular occlusive events (Barabas et al, 1973). Coronary and cerebral events are prominent and microvascular disturbances can also occur (Gruppo Italiano Studio Policitemia, 1995). Occasionally, they can present with haemorrhage particularly involving the skin and gastrointestinal tract. Splenic pain, pruritus, gout and constitutional symptoms, such as fatigue, may also be presenting features. The disease may progress over time in various ways. Thrombosis and haemorrhage continue to occur. Splenomegaly may develop and increase. A few patients will proceed to develop massive splenomegaly and some myelofibrosis but there is variation in the terminology used. Older terms such as ‘spent phase’, ‘burnt out PV’, and ‘postpolycythaemic myeloid metaplasia’ are poorly elucidated. Modern definitions have relied on the demonstration of significantly increased bone marrow reticulin staining (grade III–IV on a four-point scale) together with a clinical syndrome of one or more of progressive splenomegaly, anaemia, leucoerythroblastosis and constitutional symptoms (Thiele et al, 2001b). With this definition, the rates of myelofibrotic transformation range from <5% (Berk et al, 1995; Tatarsky & Sharon, 1997), up to 15% (Najean & Rain, 1997a) at years. leukaemia is of the of PV, in patients and in those with venesection only (Berk et al, 1986). PV a as shown by an & where the median in patients was with the cause of The of of PV are therefore the of and the of transformation to acute leukaemia and which may occur including haemorrhage and and These were used when the evidence and management There have been randomised clinical trials of for PV (Table Most trials have used the for the diagnosis of PV as the that the are of the trials are without Most of these in studies of any disease with a patient and significant It also difficult to patients on allocated when information No an of the They have also been reported in ways. The of incidence of leukaemia is particularly as in some cases the incidence of acute leukaemia of a group of patients on a is reported and in the on an to is These are not However, there are about the and of various that can be from of the clinical The venesection as the first for PV. In with and was significantly in the venesection and associated with of leukaemia and An increased of was seen in the venesection but this was the first when the Hct was (Berk et al, 1995). In it was reduced to There was a large degree of between with of patients randomised to venesection having to by in the of patients in the (Najean et al, the of using venesection as for PV is was a that compared with venesection high of and et al, 1986). The for this was to use to the increased of that was initially in the of The haemorrhage and was significantly increased in the venesection and and the was therefore In the majority of cases a high count was found at the time of haemorrhage but the count was controlled by in the other The European Organisation for on of a to was in to the Hct between and was significantly in the with the major for the an increase in vascular There were no between the for other such as myelofibrosis or The Polycythaemia Study Group randomised trials in The first was a of alone with known as in patients over the age of (Najean & Rain, of patients between the was not significantly No were for vascular or progression to The of leukaemia was significantly greater for the and with the apparent after 5 and the to up to the In addition, the of was also greater for the and with a The from the was a of with in patients the age of (Najean & Rain, was in the at compared with an for the and There were no between the in vascular or rates of leukaemia or was significantly increased in the and to occur The (Najean et al, also very in their patients They patients in on 60 in and in on in and in with the patients randomised to were with or The are small that any of is of very However, the patients in the myelofibrosis but very Therefore, to is not The of myelofibrosis was higher if by The European on in Polycythemia vera the of in PV et al, and an (Gruppo Italiano Studio Policitemia, 1997). Patients were randomised between and significantly reduced the of the of thromboembolic or from The of major or minor was also significantly There was no significant increase in The results of this the about the and of that were raised by the et al, and evidence for the use of in the management of PV. secondary to the raised Hct is a cause of in PV. There is a and between the Hct and rates of (Pearson & 1978). However, platelets may also be of the evidence for the of the count is by studies in essential where the count significantly reduced the incidence of vascular events et al, 1995). In the there was no between a raised count and events (Berk et al, 1986). However, reduced the incidence of in PV et al, a for platelets in thromboembolic events. vary in the of myelofibrotic although one randomised a greater incidence for compared with (Najean & Rain, In this of the count was associated with progression to This that the count will thromboembolic events and influence the of transformation to An Hct has been associated with increased and in patients & and therefore the Hct should be controlled for can be used to the Hct in PV. In a patients with PV who were with venesection and the incidence of arterial and venous thromboembolic events increased as the Hct increased. The cerebral blood was significantly below normal in PV patients with a raised Hct and by when the Hct was than et al, that the Hct is reduced to below by There is currently no evidence to a level of Hct in males and females. has been used to the RCM et al, 1997). be used it would have significant but it be used in the very patient with an the Hct should be maintained at than by The volume should be with the patient's and B level The was included in the randomised (Berk et al, as listed in Table It was associated with a higher of acute The of acute leukaemia was higher with higher of is not now recommended in the of PV. The was used in the randomised and found to be to (see Table although both a low incidence of acute leukaemia at a median of 8 years. A single (Messinezy et al, patients with venesection and low The median was years. leukaemia and myelofibrosis were significantly increased above the normal but there were some with These studies that low is in PV since is an it should be reserved for the elderly. is a of to which and and of It has been used in but not in the It has been used in a number of series of patients and one randomised (Najean & Rain, of acute leukaemia from to have been in one series where a of was et al, and in which some patients were also with other of myelofibrosis from to have been by a by the which has a in the of It has been used in a and a number of case The was a that included and patients et al, was in the patients and in the This showed that was but recommended reduction because of The patients in this were compared with randomised to venesection in et al, with the Hct maintained at than There was including the after the of and no in the of With prolonged there was no in the incidence of myelofibrosis between the et al, 1997). Therefore, is using the There has been on the of In a series of reports on the use
Source: British Journal of Haematology
Language: English
Subfield: Hematology
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W2151757571