Thrombotic Events in Primary Myelofibrosis: Prevalence and Prognostic Factors in 207 Consecutive Patients.

Thromboembolic disease (TE) is a major cause of morbidity and mortality in polycythemia vera (PV) and essential thrombocythemia (ET). There is limited information on the prevalence of TE in primary myelofibrosis (PMF) and associated prognostic factors.

Study patients were recruited form the Mayo Clinic database for PMF. Diagnosis was according to the 2001 WHO criteria. Clinical and laboratory information was collected at time of diagnosis and correlated with the occurrence of TE in general and arterial (AT) or venous (VT) thrombosis in particular. Leukocytosis was defined as a leukocyte count of > 10 × 10⁹/L. Qualitative JAK2V617F analysis was done according to previously published PCR-based methods. In addition, a subset of the study patients underwent quantitative PCR for JAK2V617F and rs12343867 SNP genotyping (performed by a commercially available Taqman assay). All molecular studies were performed on stored bone marrow. Standard statistical methods were used to test significance of associations between thrombosis and other variables.

207 patients with PMF were studied (median age 62, range 28-87; 132 males). At presentation, the median (range) of hemoglobin (Hb), leukocytes (WBC) and platelets (plt) were 11 g/dl (6.4-15.4), 8.6 × 10⁹/L (0.8-156.7) and 273 × 10⁹/L (13-1916), respectively. Dupriez prognostic scores at diagnosis were low, intermediate and high in 57%, 33% and 10% of the patients, respectively.

130 (63%) patients were JAK2V617F positive. Among 73 patients in whom quantitative JAK2V617F analysis was performed, the mutation was not detected in 30 (41%) patients and the median mutant allele burden in the 43 mutation-positive patients was 28% (range 1-85). The same 73 patients had rs12343867 SNP genotyping performed that revealed C/C genotype in 14 (19%) patients, C/T in 34 (47%) and T/T in 25 (34%).

At or before diagnosis, a total of 29 patients (14 %) had TE: 23 (11%) AT and 9 (4%) VT. The AT included acute coronary syndromes (ACS; n=19), transient ischemic attack/ cerebrovascular accidents (TIA/CVA; n=6) and peripheral arterial thrombosis (PAT; n=1). The VT included abdominal vein thrombosis (AVT; n=4), deep vein thrombosis/pulmonary embolism (DVT/PE; n=6) and dural sinus thrombosis (n=1). Several patients had more than one event and some both AT and VT. During follow-up, 22 (11%) patients experienced TE: 17 (8%) VT and 8 (4%) AT. The post-diagnosis VT included AVT (n=8), DVT/PE (n=13) and both DVT/PE and AVT (n=4). Post-diagnosis AT included CVA (n=4), TIA (n=1), ACS (n=2) and PAT (n=1).

AT at or before diagnosis was associated with advanced age (p=0.003) but not with sex, leukocytosis, hemoglobin level, platelet count, JAK2 mutational status, JAK2 mutant allele burden or rs12343867 SNP genotype. VT at or before diagnosis also did not correlate with all these variables or age. These variables were also evaluated regarding their relationship with post-diagnosis AT, which showed significant associations with hemoglobin level (p=0.006), platelet count (p=0.04) and high JAK2V617F allele burden (0.03). The corresponding significant variables for post-diagnosis VT were leukocytosis (p=0.02) and high JAK2V617F allele burden (p=0.01). Of note, AT or VT at or before diagnosis did not predict post-diagnosis events. During multivariable analysis, only higher hemoglobin level predicted post-diagnosis AT.

The current study provides an accurate description of both arterial and venous thrombotic events in PMF. The association between higher hemoglobin and leukocyte counts at diagnosis and post-diagnosis occurrence, respectively, of arterial and venous thrombosis is similar to observations in patients with PV or ET and suggests a potential benefit from early institution of cytoreductive therapy for the appropriate patient groups.

No relevant conflicts of interest to declare.