PF688 JAK2 ALLELIC RATIO IMPACTS ON VASCULAR EVENT IN MYELOFIBROSIS BY INCREASING THE RISK OF THROMBOSIS. A SINGLE CENTER EXPERIENCE ON 150 PATIENTS

Background Vascular complications are a recognized cause of morbidity and mortality in Myelofibrosis (MF). However, the mechanism underlying both thrombosis and bleeding as well as which risk factors can predict them remain poorly understood. Aims To evaluate the clinical characteristics, incidence and prognostic factors for vascular events (VE) in a cohort of MF patients (pts). Methods In 150 pts consecutively diagnosed with MF from 2000 to 2018 at the Hematology Unit of ASST Spedali Civili in Brescia major vascular complications (AMI, stroke, TIA, pulmonary embolism (PE) and vascular thromboses, CNS or GI bleeding, haemorrhagic events requiring active treatment) were retrospectively analyzed. Parameters tested as potentially related to VE were: age, sex, mutational status, previous thromboses, cardiovascular risk factors, hematologic parameters at diagnosis, IPSS, cytoreductive therapy and splenectomy. A competing-risk analysis was carried out to identify possible risk factors associated with VE. Results Clinical characteristics of the cohort are detailed in table 1, according to the occurrence of VE. Overall, 68 pts (45%) received antiplatelet agents after MF diagnosis, in 17 cases due to a previous history of thrombosis. After a median follow-up of 45.5 months (5-488) from diagnosis, 34 VE (15 bleedings and 19 thrombosis) were recorded in 30 pts. The cumulative incidence of vascular complications was of 27% at 10 years, accounting for 3.55 vascular events per 100 person-years. Notably, all 4 pts who started an antiplatelet therapy after thrombosis have developed a hemorrhagic event. Thromboses were venous in 11/19 pts (6 splanchnic, 2 deep venous thromboses, 2 PE), whereas arterial in 8 cases (3 AMI, 2 strokes, 1 TIA, 2 peripheral arterial occlusion disease). Bleedings occurred in 40% pts at GI tract, 30% in CNS and 10% in renal district, while in 2 pts after splenectomy. At last follow-up, 51 pts had died, in 5 cases due to VE (2 AMI and 3 hemorrhage). An MF treatment was ongoing in 26/30 pts at the time of VE. Among parameters analysed, a Jak2 allelic ratio higher than 75% was associated with VE in MF (HR2.6; 95%CI 1.1-6.5; p 0.04). Its influence on vascular complications reflects its impact on thrombosis (p 0.005), while a high Jak2 allelic ratio seems not to be related to bleeding (p 0.1). At univariate analysis also no antiplatelet agents treatment correlated with VE (p 0.02). However, only the JAK2 allele burden maintained the statistical significance in predicting vascular complications by multivariable analysis (HR4.3; 95%CI 1.8-9.6,p 0.001). No one of the other parameters investigated impacted on the occurrence of both thromboses and bleeding. By considering thrombotic events only, the risk was significantly higher also for patients with a previous history of thrombosis (p 0.05) but again, by multivariate analysis, only a Jak2 allele burden higher than 75% proved significantly related to thrombosis (HR 6; CI95% 1.9-18.4;p 0.002; Fig1). None of the clinical and molecular parameters analyzed could predict bleeding, except for having less than 100.000 platelets per microliter at diagnosis at univariate analysis (p 0.03). Conclusion In our study, a higher Jak2 allele ratio at diagnosis was the only predictive factor for VE, in particular thrombosis, confirming what described in Essential Thrombocythemia and Polycythemia Vera. Instead, no one of the possible risk factors analyzed could be associated to bleeding events. Studies on a larger population are needed to confirm these data, which would suggest to adopt a more strict anti-thrombotic surveillance in this subset of pts.