Response: essential thrombocythemia: seeing the wood for the trees

We thank Dr Spivak and Dr Silver for their interest in our review.¹  We completely agree that diagnostic accuracy is important and that isolated thrombocytosis is not a disease entity. However, we disagree with the remainder of their comments, which fall into 3 categories: the relationship of ET to PMF, the relationship of ET to PV, and the appropriate use of hydroxyurea therapy.

With regard to the relationship between ET and PMF, Drs Spivak and Silver suggest that, in a patient presenting with isolated thrombocytosis, CD34 count, JAK2 V617F homozygosity, and mutant allele burden should be used routinely to distinguish these disorders. We feel this would be inappropriate because these tests are neither standardized nor widely available, show considerable overlap between MPN subtypes, and are of unproven prognostic utility. In our article, we outline a pragmatic approach to diagnosis that is widely applicable, while acknowledging that future improvements in molecular diagnosis will likely further illuminate the many gray areas. In Figure 2 we describe a patient who presented with an isolated thrombocytosis and markedly increased bone marrow reticulin, but no other features of PMF. Importantly, such patients are common (15%-20% of ET patients), cannot be classified as having either ET or PMF according to WHO criteria,²  and, because they have a relatively good prognosis, are best managed as ET. We also cite several lines of evidence supporting the concept that PMF represents presentation in accelerated phase of a previously undiagnosed MPN, usually ET.

Concerning the relationship of ET to PV, Drs Spivak and Silver are simply wrong when they state that we assume “JAK2 V617F–positive thrombocytosis patients with ‘normal iron stores/MCV’ cannot have PV.” Instead, we clearly explain that “there are inherent problems in the use of continuous variables, such as hemoglobin, hematocrit, or red cell mass, to make this binary distinction [ie, between ET and PV], because the group of patients with borderline values will inevitably include both disorders.”^1 p1475 They are wrong again when they state that we conflate “the venous hematocrit with the total body hematocrit derived by isotope dilution.” Instead, we clearly state, “We recognize that, unless markedly elevated, hematocrit does not accurately predict a raised red cell mass.”^1 p1475 When considering how to manage a patient with a normal hematocrit but a raised red cell mass, it is important to remember that clinical trials of PV have never included this subgroup of patients. It is therefore correct for us to assert that “in the presence of a normal hematocrit and normal iron stores, the clinical significance of a raised red cell mass is unclear.”^1 p1475 Drs Spivak and Silver do not reveal whether they propose measurement of red cell mass in all patients diagnosed with ET, or indeed how an abnormal result should influence therapy. In the absence of informative clinical studies we take a pragmatic approach and base a diagnosis of PV on the presence of a JAK2 mutation and a raised hematocrit (with or without supporting features such a low serum erythropoietin), an approach consistent with both WHO and BCSH guidelines,^3,4  and we do not measure red cell mass in our ET patients.

In the context of hydroxyurea, Drs Spivak and Silver claim that we have misinterpreted the results of the PT-1 trial. This is incorrect. We state that “Hydroxycarbamide (also known as hydroxyurea) is the only cytoreductive agent proven to reduce thrombotic events in a randomized controlled trial”^1 p1477 and feel it would be inappropriate to ignore carefully documented transient ischemic attacks given the considerable evidence that they are harbingers of completed strokes.⁵  Last, Drs Spivak and Silver claim we dismiss evidence that hydroxyurea is leukemogenic. Once again this is wrong. After weighing up the strength of the published evidence and citing multiple papers supporting both sides of this debate, we conclude “At this time it is unclear whether single agent hydroxycarbamide is leukemogenic; however, any increased risk is likely to be small and should be balanced against the reduction in thrombotic complications.”^1 p1478