The Platelet COUNT at Diagnosis of Essential Thrombocythemia Is a Prognostic Factor for Thrombosis-Free Survival: Retrospective Analysis on 1201 Patients

The protective effect of higher platelet count at diagnosis of Essential Thrombocythemia (ET) was reported in some papers (Carobbio A. 2011, Palandri F. 2012, Montanaro M., 2014). As at our knowledge, there is no study specifically addressing this point; in this retrospective analysis we have examined 1201 ET patients (pts) followed in 11 Hematological centers of our region from 1/1978 to 12/2010. The diagnosis of ET was made with PVSG, WHO 2001 and WHO 2008 criteria, respectively, according to the period of 1^st observation. The main features of our cohort were as follows: median age 62,9 yrs (19-96), male/female 435/766 (36.2%/63.8%), median WBC count 8,8 x 10⁶/L (1.2-57.7), median Hb level 14.0 g/dl (6.0-20.5), median platelet count 813 x 10⁶/L (457-3582), JAK-2^V617F mutation in 498/834 performed pts (59,7 %) with a median allele burden of 19.6% (0.2-99.9%), spleen enlargement in 226 pts (18.7%), previous thrombosis in 17.9% of pts (arterial 14.1%, venous 3.8%). The median follow-up of the entire cohort was 7.75 yrs. Thrombosis-free survival curves were plotted according to Kaplan-Meier method and independent risk factors were identified with the Cox proportional-hazards method. At the multivariate analysis, negative prognostic factors for TFS resulted: previous thrombotic events (p= 0.012), age ≥60 yrs (p= 0.008) and spleen enlargement (p= 0.039): on the contrary, platelet count ≥ 944.10⁹/L resulted a protective factor for TFS [p= 0.031 with an HR 0,57 (C.I. 95% 0,35-0,95)]. Receiver operating characteristic (ROC) analyses based on thrombotic events during follow-up were used to identify the baseline platelet count of 944 x 10⁹/L as the best threshold for predicting thrombotic events. Thrombotic events according to this cutoff were 40/384 (10.3%) in pts with platelet count ≥ 944 x 10⁹/L and 109/817 (13.3%) in pts with platelet count < 944 x 10⁹/L. The sites of thrombosis are reported in the table. A comparison of the main features in these two populations showed that pts with PLT count < 944 x 10⁹/L were older (median age 60.4 yrs vs 57.1 yrs, p= 0.016), had a lower median WBC count (8.8 x 10⁹/L vs 10.6 x 10⁹/L, p< 0.0001), an higher median Hb level (14.1 g/dL vs 13.6 g/dL, p< 0.0001) and an higher rate of JAK-2^V617F mutation (67.2% vs 41.6%, p< 0.0001); no differences were observed between the two groups as to thrombotic events before diagnosis, spleen enlargement and cardiovascular risk factor (p=NS). As to the treatment, both groups resulted equally treated with anti-aggregant agents (84,6% vs 87,4%, p= 0,76) while in pts with platelet count <944 x 10⁹/L the oral anticoagulants (7.1% vs. 3.1%, p= 0.01) were more often used. Pts with higher platelet count were more frequently treated with cyto-reductive drugs (90,4 % vs 76,4 %, p< 0.0001). No significant difference resulted for Hydroxyurea (70,8 % vs 64,3%, p= 0,34) and Interferon ( 11,7% vs 6,9%, p= 0,07); on the contrary, more pts with higher platelet count were treated with anagrelide (10.7% vs 5.0%, p= 0.001) and alkylating agents (8.9% vs 5.1%, p= 0.03). In conclusion, our retrospective analysis confirmed the protective role for thrombosis of an higher platelet count at diagnosis. Pts with platelet count ≥ 944 x 10⁹/L were more frequently treated with cyto-reductive drugs and this could possibly explain the better TFS, even if the platelet count closer to the occurrence of a thrombotic event resulted near the normal values in both groups. On the other hand, the higher rate of JAK-2^V617F mutation in the group of pts with a baseline lower platelet count could be responsible of this counterintuitive finding: it is worth of note, however, that in our series the JAK-2^V617F mutation did not result a significant factor for TFS.