We present here a cohort of MPN patients harbouring the V617F mutation in the JAK2 gene. We have compared clinical characteristics and VAF between patients with PV, ET and PMF (Tables 1 and 2), examined the rate of thrombotic events in each MPN subtype (Table 3), examined the association of VAF with risk of experiencing thrombotic events (Figs. 1 and 2) and compared the clinical characteristics of patients that experienced or did not experience thrombotic events prior to or post diagnosis (Table 4 and 5).

We have shown that significant differences in the clinical characteristics are present between patients with PV, ET and PMF. Namely, the mean haemoglobin concentration and haematocrit are significantly higher in patients with PV, whilst platelet counts are significantly higher in patients with ET. These results are in line with current diagnostic criteria for the classification of MPNs. Diagnosis of PV requires an elevated haemoglobin concentration or elevated haematocrit as a major criterion, whilst diagnosis of ET requires an elevated platelet count [10]. Whilst the mean age at diagnosis is similar for all three of the MPN sub-groups, there was a significant skew in gender in our ET cohort with 76.3% being female. This is consistent with previous published studies indicating a predominance of females in the ET population [11, 12]. The percentage of patients with splenomegaly is significantly higher in the PMF group compared to both the ET and PV group. As expected, the mean JAK2V617F VAF was found to be significantly higher in the PV and PMF group than in the ET group.

Within our cohort, we observed significantly lower JAK2V617F VAF in ET versus PV and PMF (Table 2). This is consistent with previously published data showing a mean VAF below 20% for ET and above 40% for PV [13]. Interestingly, in vivo studies have demonstrated that low level JAK2V617F expression in transgenic mice results in an ET-like phenotype with marked thrombocytosis, whilst high JAK2V617F expression results in erythrocytosis as seen in PV [14]. This suggests that JAK2V617F VAF may play a direct role in MPN pathogenesis. The prognostic impact of JAK2V617F VAF in MPN has been documented in a number of previous studies showing increased risk of thrombotic events, higher rate of progression to secondary myelofibrosis and reduced overall survival in patients with higher JAK2V617F VAF [6,7,8,9]. As such, JAK2V617F genotype has been included in the International Prognostic Score of thrombosis in ET (IPSET-thrombosis) [15]. Indeed, vascular thrombotic complications are a major cause of morbidity and mortality in MPNs [16].

In our cohort, we observed a significantly higher JAK2V617F VAF in patients that had a thrombotic event versus those that did not (Fig. 1). This was due to a significant increase in the JAK2V617F VAF in patients that had thrombotic events post diagnosis. No associations were observed between thrombotic events prior to diagnosis and JAK2V617F VAF in this study (Fig. 2). The majority of previous studies have not separated thrombotic events into prior to or post diagnosis; however, one study showed that a VAF above 75% is associated with increased risk of thrombosis post diagnosis [17]. The underlying mechanism by which JAK2V617F VAF modulates post diagnosis thrombotic risk specifically is not clear. However, post diagnosis other factors known to contribute to thrombotic risk including raised haematocrit and total WCC are more closely controlled potentially increasing the proportional contribution of JAK2V617F VAF. These data highlight the importance of addressing clonal expansion of JAK2V617F with an aim to reducing the clinical burden of thrombosis in MPN patients post diagnosis. Both interferon and the JAK2 inhibitor ruxolitinib have been shown to reduce JAK2V617F VAF in the RESPONSE and PROUD-PV trials [18, 19]. Future trials addressing clonal expansion of JAK2V617F may provide further opportunities to improve outcomes, prognostication and monitoring in patients with MPNs.

We observed a higher rate of thrombotic events in patients with PV when compared with ET and PMF (Table 3). This result is consistent with previous studies and is likely a result of higher JAK2V617F VAF, WCCs and erythrocytes in PV, all of which have been correlated with increased thrombotic risk [20, 21]. Interestingly, raised platelet counts have not been independently associated with risk of thrombosis; on the contrary, platelets counts above 1500 are associated with increased risk of bleeding as a result of the clearance of von Willebrand factor (VWF) multimers [22, 23].

We recognise limitations in this retrospective study, such as the data regarding thrombotic events. This introduces a degree of uncertainty as thrombotic events could not be independently verified and may be impacted by the lack of clinical record or patients’ inability to precisely recollect their past medical history. Previous studies have shown a correlation between JAK2V617F VAF and venous but not arterial thrombosis. We observed no associations between venous or arterial thrombosis and JAK2V617F VAF when taken independently (data not shown). Our low patient number with venous thrombosis may limit our ability to identify this association. Moreover, the cohort only included a small number of patients diagnosed with PMF. As such, caution must be taken when drawing conclusions from this study data regarding JAK2V617F VAF and its prognostic impact in PMF.