Aspirin responsive erythromelalgia and migraine-like microvascular cerebral transient ischemic attacks (MIA) are the presenting features of early stage essential thrombocythemia (ET) and polycythemia vera (PV) [1-4]. Prefibrotic stages of ET and PV are diagnosed by the combined use of clinical, laboratory and bone marrow histopathology features. The 1980 RCP criteria of ET were determined by careful prospective documentation of peripheral blood and bone marrow smears and bone marrow biopsy material. In 1980 we modified the PVSG criteria for PV by including bone marrow histology as a pathognomonic clue to myeloproliferative disease (Table 1) [1,2]. We replaced O₂-saturation of >92% by bone marrow biopsy, changed splenomegaly by bone marrow histology as a major criterion (A3), and used splenomegaly as a minor criterion (Table 1B) [1-10]. We skipped raised B12 (>900 ng/L) or raised B12 binding capacity (>2200 ng/L) as completely irrelevant for early stage PV.

Table 1: The 1980 Rotterdam Clinical and Pathological (RCP) criteria for Essential T) and Polycythemia Vera (PV) [1,2].

Platelets in excess of 400 × 10⁹/L, and an increase of clustered enlarged megakaryocytes in a bone marrow biopsy material was found to be pathognomonic diagnostic for ET and PV and excluded reactive thrombocytosis (Figure 1). The 1980 RCP criteria for PV, bone marrow histopathology and erythrocyte count above 6 × 10¹²/L are diagnostic for PV and do exclude all variants of primary and secondary erythrocytosis (Table 1) [4-6]. The combined use of clinical and bone marrow features on top of JAK2 and MPL mutation screening significantly improved the 2006 European Clinical Molecular and Pathological (2006 ECMP) [5,6] and the 2008 WHO classification of the JAK2^V617 and JAK2 wild type MPNs ET, PV and myelofibrosis (MF) [11,12]. The WHO-ECMP criteria [4-6] for prefibrotic normocellular ET, hypercellular prodromal PV and PV combined the use blood cell counts, bone marrow histology and JAK2^V617F mutation screening to clearly distinguish JAK2^V617F mutated trilinear MPN entity (Tables 2 and 3) from JAK2 wild type MPN [6]. In the present prospective clinical research study of newly diagnosed ET and PV at presentation we evaluated the 1980 RCP and the WHO. ECMP features of JAK2^V617F ET and PV patients.

Figure 1: 1980 RCP ET (upper) , ET/PV (lower left) , and PV (lower right) bone marrow features in ET and PV patients. Pleiomorphic megakaryocytes in ET (upper panels) have less hyperlobulated nuclei as compared to PV (left bottum). The LAP score is increased in ET and PV complicated by erythromelalgia and the clustered pleiomorphic megakaryocytes in prefibrotic ET ET/PV and PV patients complicated by erythromelalgia are identical. Observations between 1975-1985 [1,2].

Table 2: The 2015 WHO-European Clinical Molecular and Pathobiological (2015 WHO-ECMP) criteria for the classification of prefibrotic JAK2^V617F mutated essential throbocythemia (ET) [5,6,28-30].

Table 3a: The 2015 WHO-ECMP criteria for the diagnosis of prodromal and overt polycythemia vera (PV) and primary or secondary erythrocytoses [5,6,28-30].

Table 3b: WHO bone marrow histology and CMP criteria for staging of early, overt, and advanced PV. A0, A2, B1 establish early PV (mimicking ET) prodromal PV CMP stage 0 [6,35].

Hemoglobin, hematocrit, erythrocytes, leukocytes, platelets, iron status and chemical parameters were routinely performed. Erythrocyte volume (red cell mass RCM) was measured using Cr51 (natriumchromate) labeled autologous erythrocyte and plasma volume was measures with J131-human serum albumin. The leukocyte-alkaline- phosphatase (LAP) stain was performed with Na-naphtyl and Fast Garnet G.B.C Salt. Iron stain of bone marrow smears were performed with Prussian blue reagents. Bone marrow biopsies from the iliac crest were stained with hematoxylin and eosin for histopathology evaluation. All bone marrow biopsies were evaluated by expert hematopathologists for morphology, grading of cellularity and scoring of silver stained reticulin fibers according to PVSG (Ellis et al. [13], Table 1C) and WHO recommendations [14]. On top of JAK2^V617F mutation screening using the PCR test according to Baxter et al. [15], and quantitative JAK2 mutation allele burden, the diagnostic work-up of ET and PV patients included serum erythropoietin (EPO) levels, bone marrow aspirate for morphology and endogenous erythroid colony formation (EEC), red cell mass (RCM) measurement according to Pearson et al. [16] and spleen size on echogram. The minimum criterion for the diagnosis of ET or thrombocythemia associated with PV is 400 × 10⁹/ L^1-10. Collagen staining of Masson was used for objective detection of collagenisation of reticulin fibers to clearly distinguish between early stage reticulin fibrosis (RF) from advanced reticulin/collagen fibrosis (RCF) for grading of myelofibrosis (MF, Table 1C) according to WHO recommendations [12,14].

In the period of January 1975 to December 31, 1980, we prospectively studied 30 consecutive early prefibrotic stage MPD patients (mean age 56.7 years, range 33-86), who presented with erythromelalgic thrombotic thrombocythemia, 14 ET and 16 PV patients. The clinical features of erythromelalgia and transient neurologic and ocular ischemic manifestations caused by platelet-mediated arteriolar inflammation and thrombosis in thrombocythemia in 26 MPN (ET and PV cases A in Table 4) have been reported in great detail [1-3]. Eleven of 14 ET patients had platelet counts between 400 and 1000 × 10⁹/L. Spleen size on scan was slightly increased in 5 of 14 ET and in 13 of 16 PV patients. Leukocyte count counts was increased (>10 × 10⁹/L) in 5 out of 14 ET and in 14 of 16 PV patients meeting the WHO criteria. LAP score was increased (>100) in 12 out of 14 ET and in all PV patients. Increase of clustered large pleomorphic megakaryocytes in bone marrow smears and biopsies was diagnostic for MPD in all 14 ET and 16 PV patients (Table 4). A normocellular bone marrow picture (cellularity <60%) with increase of clustered pleomorphic megakaryocytes and no increase of cellularity was seen in 7 of 14 ET and in one of 16 PV patients. A moderate increase of cellularity (1 ± 60-80%) in the bone marrow due to increased erythropoiesis leading to a WHO-CMP defined prodromal PV picture was seen in 3 ET and 4 PV patients. A typical trilinear PV bone marrow picture with pronounced increase of cellularity (2 ± 80-100%) due to predominantly increased megakaryo-erythropoiesis or megakaryocytic erythrocytic and granulocytic proliferation (panmyelosis according to Dameshek [9,10]) was seen in 2 of 14 ET and in 11 of 16 PV. These results indicate that bone marrow histopathology on its own is not reliable to differentiate between ET and PV but appeared to be a powerful tool to differentiate ET and PV from all variants of primary or secondary erythrocytosis and reactive thrombocytosis with a sensitivity and specificity near to 100% [5,6]. The morphology of large pleomorphic megakarocytes was not different in ET and PV. Pleomorphism of megakaryocytes became more pronounced in the hypercellular (80-100%) bone marrows with increased reticulin fibrosis (RCF) in advanced PV. The clinical manifestations and laboratory features in 50 consecutive cases of thrombocythemia, (30 ET and 20 PV) are described in great detail [16]. The peripheral blood findings in 30 ET (primary thrombocythemia; PT) and 20 PV seen between 1975 and 1985 are shown in Figure 1 [17]. Hemoglobin levels are normal in ET and elevated in PV. Leukocytes were normal or elevated in ET and PV. LAP score was much more elevated in PV than in ET (Figure 2). Out of 30 ET patients LAP score was increased in 24 and normal in 6. Platelet counts were between 400 and 1000 × 10⁹/L in the majority of the 30 ET patients. Serum uric acid levels were normal in ET and usually elevated in PV (Figure 2) [17].

Table 4: Peripheral blood and bone marrow data in 14 ET and 16 PV patients observed between 1975 and 1981, University Hospital, Dijkzigt, Rotterdam [1]. A=complicated by erythromelalgia and/or migraine-like atypical transient ischemic attacks. B=asymptomatic ET or PV. ET:L Essential Thrombocythemia. PV: Polycythemia Vera. PVT=PV after phlebotomy.

Figure 2: Laboratory findings at time of presentation in symptomatic with thrombocythemia subdivided in 30 PT (=ET) and 20 PV patients seen between 1975 and 1985 [17]. Hb levels are normal in PT and elevated in PV [16]. Leukocytes were normal or elevated in PT and PV. LAP score was much more elevated in PV than in PT with normal values in 6 PT patients. Platelet counts were were between 400 and 1000 × 10/L in the majority of the 30 PT patients. Serum uric acid levels werenormal in PT and usually elevated in PV.

Between 1997 and 2014 we prospectively studied 10 early stages of JAK2^V617F mutated MPN patients (6 ET and 4 PV), who presented with erythromelalgia or migraine-like microvascular cerebral ischemic attacks (MIA). The clinical diagnoses without the use of bone marrow histopathology (The French Approach according to Villeval, James, Pisani Casadevall and Vainchenker 2006 [18]) were ET in 6 and PV in 4 (Table 5A). The 6 ET were heterozygous for the JAK2^V617F mutation. A typical example of hetrozygous JAK2^V617F mutated ET with asymptomatic stable disease for more than 20 years is shown in Figure 3. Three PV patients were homozygous for the JAK2^V617F mutation (case 8-10, Table 5). All ET patients had an erythrocyte count far below 6 × 10¹²/L and all PV had erythrocyte counts above 6 × 10¹²/L (Table 5A). Increase of erythrocytes counts above 6 × 10¹²/L for the diagnosis of PV appears to be independent from the iron deficient status and persists in PV in a clinical remission obtained by repeated venesection (Figures 3 and 4).

Table 5a: Clinical and laboratory data, and clinical diagnosis of ET or PV JAK2^V617F *: + is heterozygous, ++ is homozygous . NT=not tested, PV patients with documented increased RCM had erythrocytes above 6.0 × 10¹²/L at diagnosis and at time of transition of prodromal PV into classic PV.

Figure 3: Heterozygous JAK2V617F positive normocellular ET with stable platelet counts during 20 years follow-up 1995-2015. Figure

Figure 4: Clinical course of acute onset Idiopathic Erythrocythemic (IE)=PV stage zero  in complete remission by phlebotomy according to Dameshek 1950 [10] on top of low dose aspirin according to Michiels et al 1985 [2].

The diagnosis of the 10 JAK2^V617F positive MPN patients based on bone marrow histology alone, as blindly judged by pathologists, was consistent with ET in 3 (Table 5B) and PV in 7 cases (Table 4B). The 3 ET patients with PV bone marrow histology (Figure 2) had very low serum EPO levels and EEC, but erythrocyte counts far below 6 × 10¹²/L consistent with the diagnosis of prodromal PV (Table 5B). These three ET patients with prodromal PV bone marrow histology developed a slow onset PV after long-term follow-up of 8, 9 and more than 10 years (Figure 3 and Table 5).

Figure 5: Erythrocyte counts x10¹²/L vertical axis versus red cell mass (RCM) horizontal axis. At erythrocyte values below 5.8 × 10¹²/L the red cell mass (RCM) values are between 25 and 30 ml/kg in ET. At erythrocyte values above 5.8 × 10¹²/L all values of RCM area bove 30/kg in all PV cases indicating that the erythrocyte cut-off level of 5.8 × 10¹²/L dicriminates between ET and PV (see details in Table 6) The numbers in the blue line are erythrocyte counts x10¹²/l.

Table 5b: Bone marrow histology features diagnosis according to 2008 WHO and WHO-CMP criteria in 10 patients with JAK2V617F positive MPN at diagnosis and during long-term follow-up. M/E=myeloid erythroid ratio, US: unstainable. uc: unclassifiable. Pro-PV=prodromal PV [16]..

The diagnoses according to WHO-CMP criteria for diagnosis and staging of ET and PV patients (Table 5) were normocellular ET in 2, prodromal PV in 3 cases, hypercellular ET with a megakaryocytic granulocytic myeloproliferation on bone marrow histopathology (EMGM or masked PV) in 1, and acute onset PV in 4 patients (Table 4B). Bone marrow histopathology alone cannot distinguish JAK2^V617F mutated ET, prodromal PV and PV.

Bone marrow cellularity of the 10 JAK2^V617F mutated MPN patients (6 ET and 4 PV) ranged from range 60% to 90% (Table 5B). There was an increased erythropoiesis in 8/10 and increased granulopoiesis in 4/10 patients with JAK2^V617F mutated thrombocythemia (Table 5B).

RCM, red cell counts and bone marrow histology in ET and PV

The 2008 WHO cut-of levels for the diagnosis of PV are: Hb>18.5 g/dl and Ht>0.60 in men and Hb>16.5 and Ht>0.56 in women for the diagnosis of PV with the compelling need to measure red cell mass (RCM) to distinguish JAK2^V617F mutated ET from PV in border line cases with hemoglobin (Hb) and hematocrit (Ht) in the upper level of normal [4-6,11,12]. In the two prospective Rotterdam studies we assessed the RCP and used the WHO-ECMP criteria related to RCM, Hb, Ht and erythrocyte counts in 10 ET and 16 PV patients in whom RCM, peripheral blood and bone marrow data were available (Table 6). The correlation curves between erythrocyte count, Hb or Ht versus RCM showed the best correlation between erythrocyte counts and RCM (Figure 5). At RCM above 30 ml/kg the erythrocytes are above 5.8 × 10¹²/L in all 19 WHO-ECMP defined PV patients (Table 6 and Figure 5). At erythrocyte counts above 5.8 × 10¹²/L the hematocrit values range from 0.46 to 0.72 in WHO-ECMP defined PV (Figure 5). At erythrocyte counts below 5.8 × 10¹²/L the hematocrit values range from 0.40 to 0.45 in WHO-ECMP defined ET who had normal RCM (Figure 5 and Table 6). At erythrocytes above 5.8 x10¹²/L in PV patients the Hb values ranged from 15.0 to 20.9 and are below 2008 WHO criteria in 3 females and 2 males (Table 6 in blue), who had incraesed RCM. At erythrocytes above 5.8 × 10¹²/L in PV patients the Ht values ranged from 0.46 to 0.72 and are below 2008 WHO criteria but had increased RCM in 7 females and 1 male (Table 6 in blue). Seven ET patients had normal RCM at erythrocyte counts between 4.4 to 5.3 × 10¹²/L of whom 4 had WHO normocellular (<60%) ET and 3 had hypercellular (60-80%) prodromal PV bone marrow histology. The morphology of clustered medium to large megakaryocytes in bone marrow smears and biopsies were not different in ET and PV patients. Increase of erythrocytes counts above 5.8 × 10¹²/L for the diagnosis of PV appears to be independent from the iron deficient status and persists in PV in a clinical remission obtained by repeated venesection (Figure 4) thereby confirming the observations of Dameshek [9,10].

Table 6: The relation between RCM, erythrocyte count and bone marrow histology findings at time of diagnosis in 10 ET and 16 PV. Three ET cases could be diagnosed as PV according to the 2008 ECMP criteria. PV cases not meeting the crude 2008 WHO cut-of levels (hemoglobin (Hb) and hematocrit (Ht) for PV: Hb >18.5 g/dl and Ht>0.60 in men and Hb>16.5 and Ht >0.56 in women) for the diagnosis of PV [11,12] are indicated in blue.All PV cases had increased red cell mass (RCM) and increased
erythrocytes above 6 × 10¹²/L (red) meeting the 1980 RCP and the ECMP criteria for the diagnosis of classical polycythemia vera (PV).

In the Basel cohort of 100 MPN patients the JAK2 status and allele burden was determined from peripheral blood samples of all included patients. Isolation of granulocytes, T lymphocytes, platelets, buccal mucosa, RNA, and DNA, as well as cDNA synthesis were performed as described (Kralovics et al.) [19,20]. The clinical, laboratory and patholgical features of 60 MPN patients subdivided in WHO defined 24 ET and 46 PV are reported in 2009 and shown in Table 7 [20]. We could correlate erythrocyte count and serum EPO levels to hemoglobin (Hb), hematocrit (Ht), platelet count and JAK2^V617F mutation load in these 60 evaluable MPN patients including 24 ET and in 46 PV when the 2015 WHO-ECMP criteria (Tables 2 and 3) are applied to each individual case. The JAK2^V617F mutation load in 24 ET patients was zero in 10 (JAK2 wild type ET) and positive in 14 ranging from 3 to 20%, from 20 to 42% and above 50% in 6, 5 and 2 cases respectively. The JAK2^V617F mutation load in 36 evaluable PV patients ranged from 3 to 20%, from 20 to 50% and above 50% in 5, 12 and 19 PV cases respectively. Increased erythrocyte counts above the normal level of normal (5.5 × 10¹²/L) correlated with increased red cell mass measurement in 8 PV patients and the 6 ET patients had normal values for erythrocyte count and RCM (real life observations, Table 8). In 8 JAK2^V617F mutated PV increased red cell mass (RCM) was associated with erythrocyte counts above 5.5 × 10¹²/L in 7 cases except one with prodromal PV (Table 7 and Figure 6). Increased RCM and erythrocytes above 5.5 × 10¹²/L was related to hemoglobin from 14.6 to 18.9 g/L, to hematocrit from 0.46 to 0.57, and platelet count between 122 to 1158 × 10⁹/L. In 6 ET patients, normal red cell mass (RCM) was related to erythrocyte counts of 4.6 to 5.4 × 10¹²/L, to Hb from 14.0 to 16.1 g/L, to Ht from 0.39 to 0.47, and platelets from 575 to758 × 10⁹/L in 5 and 1579 × 10⁹/L in 1. These findings in the Basel cohort of ET and PV patients confirm that increase of erythrocytes counts above 5.5 × 10⁹/L in males and females is related to increase RCM. The increased erythrocyte counts above the normal value of 5.5 × 10¹²/L is independent from the iron deficient status and persists in PV in a hematological remission by phlebotomy alone (Figure 4).

Figure 6: Erythrocyte counts x10¹²/L vertical axis versus red cell mass (RCM) horizontal axis. At erythrocyte values below 5.5 × 10¹²/L the red cell mass (RCM) values are between 25 and 30 ml/kg in ET. At erythrocyte values above 5.5 × 10¹²/L all values of RCM area bove 30/kg in all PV cases indicating that the erythrocyte cut-off level of 5.5 × 10¹²/L discriminates between ET and PV (see details in Table 7). The numbers in the blue line are erythrocyte counts x10¹²/l.

Table 7: The Basel data in 13 cases with documented MPN in bone marrow biopsy and 5 case of erythrocytosis: direct comparison and erythrocytes, serum EPO and RCM.

Table 8: WHO-CMP criteria for PV including increased erythrocyte counts.

Four studies showed that WHO defined elevated hemoglobin concentration cannot be used as a surrogate marker for absolute erythrocytosis in PV patients indicating the need that RCM is mandatory for patients who do not meet the WHO defined, rather crude hemoglobin and hematocrit value [21-24]. In a series of 77 consecutive patients (31 males and 46 females) with PV in the study of Johansson et al. only 35% of male and 63% of female PV patients had Hb values above 18.5 and 16.5 g/dL respectively21. Laboratory features at diagnosis of 266 PV and 381 ET patients diagnosed according to Pearson et al. [16] from a nation-wide survey of 647 patients with chronic myeloproliferative disease in Japan are summarized in Table 8 [24]. In this study of 266 PV patients with increased red cell mass, erythrocytes counts were in the range of 5.6 to 7.6 × 10¹²/L (mean 6.61 × 10¹²/L). All PV patients had increased erythrocytes above the normal value of 5.5 × 10¹²/L in 100%, whereas hemoglobin was above 18 gm/dL in 50%, hematocrit above 0.55 in 46%, and decreased serum EPO (<3.3 u/mL) in 94% [24]. The corresponding values in 381 PVSG defined ET patients had completely normal values for 4.74 + 0.79 × 10¹²/L (range 3.9-5.3 × 10¹²/L), hemoglobin and hematocrit. The platelet counts in PVSG defined ET and PV of 531 ± 332 × 10⁹/L and 1063 ± 434 × 10⁹/L are significantly different. A PV bone marrow histology with clustered pleiomorphic large megakaryocytes and increased of erythropietic cellularity (60-80%) has been observed by Thiele in cases of so-called initial (latent) PV with thrombocythemia at platelet counts between 600 × 10⁹/L and 1260 × 10⁹/L mimicking WHO ET [25]. The laboratory data of 23 cases diagnosed as initial (latent) PV did not meet the PVSG and WHO defined levels of hemoglobin and hematocrit required for diagnosis of PV [25], but did meet the WHO-CMP criteria for PV including increased erythrocyte counts above 5.8 × 10¹²/L in men and above 5.6 × 10¹²/L in female (Table 8) [25].

The present study demonstrates that erythrocyte count at a cutoff level of 5.8 × 10/12L in males and 5.6 × 10/12L in females differentiates ET and prodromal PV from classical PV (Tables 6, 7, Figures 5 and 6) obviating the need to measure RCM in JAK2^V617F and exon 12 mutated patients. It is the degree of erythrocythosis on top of characteristic bone marrow histology, increased LAP score and decreased serum EPO levels that distinguishes WHO JAK2^V617F mutated classical PV from ET and prodromal PV and in particular also from the JAK2 wild type MPN phenotypes carrying the MPL or calcireticulin (CALR) mutation (Figure 7) [26-30]. The reduction in iron reserve in PV leads to an insufficient amount of iron for the synthesis of haemoglobin in the developing red cells, and as a result that bone marrow iron stain is negative in PV [10,26]. JAK2 wild type ET lacks features of PV at the clinical laboratory and bone marrow level [27-30]. As iron deficiency develops in PV on treatment with phlebotomy, the mature red cells produced are smaller than normal and occupy less room in the circulation, which is associated with the relief of hypervolemic symptoms [10,26]. The haemoglobin and hematocrit levels remain low for periods of months to years in PV patients in complete haematological remission by phlebotomy alone, but the erythrocyte count persist to remain above 5.8 × 10¹²/L (Figures 5 and 6) [10,26]. As the mean corpuscular volume of red cells becomes reduced to levels below 70 cubic micron due to the chronic iron deficiency state, the discrepancy between the high red cell count far above 6 × 10¹²/L and low hemoglobin level becomes increasingly more striking [10,26].

Figure 7: WHO-CMP Classification of five distinct clonal Myeloproliferative Neoplasms (MPNs) and Transitional states according to Michiels et al. [6,28-30].

Piche et al. described the bone marrow histopathology findings in 59 JAK2^V617F positive ET and 44 JAK2 wild ET cases [27]. These original observations confirm our findings in the present study that ET patients with JAK2^V617F mutation indeed have PV-like morphological bone marrow changes of pleomorphic large megakaryocytes similar to our findings in WHO-CMP defined ET, prodromal PV patients and PV patients [6,28-30]. At time of first presentation symptomatic JAK2^V617F positive ET and prodromal PV usually have platelet count between 400 and 1000 × 10⁹/L, low serum EPO, increased LAP score, and slight to moderate increased bone marrow cellularity due to increased erythropoiesis. Increase of bone marrow erythropoiesis, granulopoiesis and serum LDH levels and spleen size are more pronounced in JAK2^V617F positive ET in particular at higher JAK2^V617F mutation load but clusters of large and giant megakaryocyte with ‘staghorn’ nuclei are rare [27].

In conclusion, a typical MPN bone marrow histology, erythrocytes above 5.8 × 10¹²/L in males and 5.6 × 10¹²/l in females (normal cut-off value is 5.5 × 10¹²/L in females) separates overt and masked PV from ET and prodromal PV obviating the need of RCM measurement. The superiority of 2015 WHO-CMP [28-30] above the 2008 WHO [11,12] criteria is that the 2015 WHO-CMP clearly distinguish within the JAK2^V617F mutated MPN normocellular ET (WHO-ET), hypercelluar ET due to increased erythropoiesis (prodromal PV) and ET with hypercellular megakaryocytic-granulocytic myeloproliferation (EMGM or masked PV) from overt and advanced PV with splenomegaly and bone marrow reticulin fibrosis (Table 3 and Figure 7). With the advent of molecular screening for JAK2, MPL and CALR since 2013 at least four main types of clonal myeloproliferative neoplasm (MPN) can be distinguished (Figure 7) [28-30]. First, JAK2^V617F-positive ET, prodromal PV and masked PV, slow onset PV and rapid onset PV. Second, JAK2 wild type ET and myelofibrosis (MF) carrying the MPL515. Third, JAK2 wild calreticulin (CALR) mutated ET and MF. Fourth, a small proportion of ET and MF patients are JAK2, MPL/CALR wild type.