Gender Differences In the Prevalence of Thrombosis In the JAK2 V617F-Positive Myeloproliferative Disorders

Thrombotic events are prevalent and morbid conditions in essential thrombocytosis (ET), polycythemia vera (PV), and primary myelofibrosis (PMF), but the exact mechanisms aside from hyperviscosity and dysfunctional platelet behavior are largely undefined. Since ET, PV and PMF are chronic, latent disorders of middle and older age, we examined the influence of traditional thrombotic risk factors on the tendency toward thrombotic events. To this end, we retrospectively analyzed a cohort of 270 JAK2 V617F-positive patients from the Johns Hopkins Center for Myeloproliferative Disorders.

Thrombotic events within 5 years preceding or at any time following an official diagnosis of ET (n=59), PV (n=166), or MF (post-ETMF, post-PVMF, and PMF) (n=45) were captured by retrospective review of the medical record; the events were characterized as microvascular or macrovascular, arterial or venous. Traditional thrombotic risk factors were recorded, including hypertension, diabetes, dyslipidemia, and history of smoking) and since there were gender differences in diagnostic class (ET and PV were more common in women), age at diagnosis, smoking history, dyslipidemia, and leukocyte count, these variables were incorporated into a multivariable logistic regression model.

Of the entire MPD cohort, 67 patients (25%) had 79 microvascular or macrovascular thrombotic events, including 14 (24%) with ET (18 events), 45 (27%) with PV (51 events), and 8 (18%) with MF (10 events). The thrombosis prevalence did not differ statistically by disease class. However, thrombosis prevalence did differ by gender: women had a higher prevalence of thrombosis compared to men (29% vs. 18%; p=0.035) and the odds ratio remained significant after multivariable adjustment (OR 1.98, p=0.04). There were no gender differences with respect to treatment (aspirin or cytoreductive agents), nor were there differences due to age (median 56 years) or disease duration at the time of thrombosis. The frequency of JAK2 V617F homozygosity also did not differ by gender. In men, 9 of 13 (69%) thrombotic events occurred at diagnosis or in the preceding 5 years; however, in women, thrombotic events during follow-up were as likely since only 20 of 45 (44%) thrombotic events occurred at diagnosis or in the preceding 5 years. Nine patients experienced multiple thrombotic events, of these, 6 were women; in 8 cases, these thrombotic events involved different vascular beds, suggesting a general proclivity for thrombosis. Overall, a dependence between sex and the type of thrombosis could not be identified, possibly due to inadequate power (p=0.36). However, there was an indication of gender differences in some thrombotic subgroups: for example, of 20 abdominal venous thromboses (AVT, including hepatic vein, portal vein, mesenteric vein and splenic vein thrombosis), 16 (80%) occurred in women (p=0.07). AVT was the most common event in PV, occurring in 23% of cases, and of these, 11 of 12 were reported in women. Further, in 7 of these women, an AVT was the presenting manifestation of their MPD. AVT was also the most common type of thrombosis in MF, occurring in 50% of cases; of these, 4 of 5 also occurred in women. In contrast to PV, rather than manifesting at diagnosis, 3 of the 5 events (60%) complicated splenectomy. Neurologic events also appeared to be more frequent in women: 6 of 7 CVA's (86%) occurred in women (p=0.14), and of 11 TIA's, 8 (73%) occurred in women (p=0.4).

In this retrospective analysis of 270 cases, thrombotic events were equally prevalent in ET, PV, and MF, though importantly more common in women in this patient cohort. There was also a trend suggesting differences in thrombosis type between women and men, since TIA, CVA, and especially, AVT appeared to be more common in women. Indeed, in women, AVT was often the presenting feature leading to a diagnosis of PV. This study indicates unique disease and gender specific interactions in the generation of thrombosis in the MPD, and suggests that women are paradoxically at higher risk for thrombosis in the MPD. Though a mechanism is not yet defined, this analysis contributes to a growing literature emphasizing gender differences in all three MPD, which already includes differences in the JAK2 V617F allele burden and disease class. Further, this analysis supports the important impact of individual and host variation on the clinical manifestations of the MPD.

No relevant conflicts of interest to declare.