The treatment of essential thromobocytosis revisited

To the editor:

In “How I treat essential thrombocythemia,” Beer et al¹  unfortunately chose not to use all the evidence available to them. Essential thrombocytosis (ET) and its companion myeloproliferative disorders (MPDs), polycythemia vera (PV), and primary myelofibrosis (PMF) exhibit substantial phenotypic mimicry but have different natural histories, complication rates, and treatment requirements. Thus, diagnostic accuracy is vital for therapy to be safe and effective. Much ink has been spilled over the utility of marrow histology in distinguishing ET from PMF, but marrow histology, although distinctive, is not always diagnostic for any of the MPDs.²  Furthermore, isolated thrombocytosis is not a disease per se but, like myelofibrosis, it is only a disease manifestation that must be assessed in the context of both the patient and time. It takes approximately 8 years for isolated thrombocytosis to evolve into a PMF phenotype,³  and this is also true for its evolution to PV. Viewed from this perspective, the authors cannot be sure of the correct diagnosis of the patient illustrated in their Figure 2. However, considering that clonal dominance is least common in ET as is JAK2 V617F homozygosity, whereas both are most common in PMF,⁴  a high circulating CD34⁺ cell concentration or a high JAK2 V617F allelic burden, when present, should separate ET from PMF, but this was not discussed. With respect to PV, the distinction is even easier, yet Beer et al err by assuming that JAK2 V617–positive thrombocytosis patients with “normal iron stores/MCV” cannot have PV, which is incorrect,²  and then they err again: conflating the venous hematocrit with the total body hematocrit derived by isotope dilution, they remarkably assert, “However, in the presence of a normal hematocrit and normal iron stores, the clinical significance of a raised red cell mass is unclear, and so we do not measure the red cell mass in our ET patients.”^1 p1475 This conceptual error, that the hematocrit is an accurate reflection of the red cell mass in the MPD, has been repeatedly discredited⁵  and has no place in modern medical practice. Indeed, because 64% of JAK2 V617–positive ET patients had an elevated red cell mass but a normal hematocrit,⁶  and because venous thrombosis is more common in PV than in ET, this conceptual error has important therapeutic implications. The authors also misinterpret the PT-1 study: hydroxyurea was only more effective than anagrelide in preventing transient ischemic attacks because it is a nitric oxide donor, but not arterial thrombosis, and was markedly less effective for preventing venous thrombosis; others have made similar observations. Finally, given that life span is normal in ET, whereas hydroxyurea is at best palliative, it is unclear how the authors can dismiss an established body of evidence identifying hydroxyurea as leukemogenic in the MPD (reviewed in Spivak⁷ ) when discussing its use.