Thrombosis is the major cause of morbidity and mortality in polycythemia vera (PV) and essential thrombocythemia (ET). Age ≥60 years (y) and/or history of thrombosis labels patients (pts) as high-risk for thrombosis. Yet, thrombosis frequently occurs prior to the diagnosis of PV/ET. In a multicenter study of the East German Study Group (HINC-207; OSHO #091), the interaction between age and time occurrence of the first thrombosis as risk factors for thrombosis after diagnosis was studied.


After IRB approvals, JAK2 mutated adults with PV or ET were prospectively enrolled in 9 centers and centrally stratified in a one to two ratio (group A: pts with a history of thrombosis; group B: pts without thrombosis) with a pre-planned minimum of 60:120 pts. Based on a longitudinal and cross-sectional design, clinical and laboratory data at diagnosis, last follow-up, and thrombosis (for group A) were collected. Thrombosis prior to diagnosis was labeled as A1 and thrombosis after diagnosis as A2. Thrombosis risk factors were grouped into age-, previous thrombosis-, thrombosis prior to PV/ET-, cardiovascular (CV)-, thrombophilia-, and disease- (JAK2 allele burden, Hct, and WBC) related. Additionally, therapies [aspirin (ASS), anticoagulation, phlebotomy, and cytoreduction] and data from a study-own patient questionnaire were included. All pts signed informed consent. The primary endpoint was the phenotypic diversity in JAK2-mutated ET and PV pts with or without thrombosis.


From April to Dec, 2019, 246 pts were recruited. Data on 237 pts (median age 62y; 59% females, 58% PV) are available. At diagnosis, pts in group A (n=71, median age 59.5y) tended to be younger than those in group B (n=166, median age 63y) (p=0.07). Yet, 70.4% thrombotic events (venous: median age 46.5y; arterial: median age 57y) occurred in A1 and correlated with younger age (p=0.03). Only 3 pts developed a second event after diagnosis. These were counted in A2 (n=24, median age at thrombosis: 61y). Overall, thrombosis occurred either prior to or within the first 3y after diagnosis in 63/71 (89%) pts. Age>60y could not be identified as a risk factor for thrombosis or type of thrombosis at any time point. The 5 y probability of no thrombotic event after diagnosis in pts >60y was 90.4% vs. 89.2% for pts <60y (p=0.8) and that of a thrombotic event >3y after diagnosis in pts >60y was 3.7% vs. 4.9% for pts <60y (p=0.7). Similarly, A1 did not correlate with A2 (p=0.3). With 1691 patient-years for the entire cohort, the incidence of thrombosis after PV/ET diagnosis was 0.7 for arterial and 0.6 for venous events per 100 patient-years. Smoking was more prevalent in pts >60y (p=0.003) and was not associated with thrombosis. Irrespective of age, hypertension (65%, p=0.03), hyperlipidemia (19%, p=0.008), and diabetes (16.4%, p=0.05) were frequent and correlated with A2 while atrial fibrillation (p=0.03) and inherited thrombophilia risk factors (p<0.00) with A1. JAK2 allele burden (median 19%) and Hct >45% (median 45%) at diagnosis correlated strongly with age >60y (p=0.005) but not with A, A1, or A2, although Hct >45% at diagnosis correlated with A2 in PV (p=0.001). Surprisingly, a Hct >45% at thrombosis was more frequently present in A1 (55%) vs A2 (30%) (p<0.00). Median WBC at diagnosis was higher in B compared to A (p=0.004), strongly associated with age >60y (p<0.00) but not with A2. WBC >15% at thrombosis did not correlate with A. Age rather than thrombosis was the trigger for cytoreduction [82% hydroxyurea (HU) in B pts >60y vs 53% in A pts <60y] (p<0.00). In PV, ASS did not correlate with thrombosis (25% of pts in B did not receive ASS). Cytoreduction, interval between diagnosis and cytoreduction, nor the duration of exposure correlated with thrombosis.


The majority of thrombotic events occurred prior to or within the first 3 years after the diagnosis of JAK2 mutated PV/ET and were associated with CV-risk factors rather than older age. Phenotypic features such as Hct >45%, high WBC, and JAK2 allele burden were associated with age >60y and less with thrombosis. Their value as surrogate markers for therapeutic interventions to reduce thrombosis needs to be critically evaluated in larger series. Whether adequate PV/ET- or CV-risk- treatments account for the low rate of CV events after diagnosis (despite a higher incidence of CV-risk factors) compared to the general population could not be answered due to study design and needs to be addressed prospectively.


Al-Ali:Novartis: Honoraria, Research Funding; Celgene: Honoraria, Research Funding.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution