Refining thrombotic risk stratification in essential thrombocythemia: Impact of european leukemia net clinical signs and symptoms in patients treated with hydroxyurea

Abstract Introduction Essential thrombocythemia (ET) is a chronic myeloproliferative neoplasm characterized by an increased risk of thrombotic complications, which are a major cause of morbidity and mortality. Traditionally, ET patients (pts) are stratified into low-risk (LR; absence of risk factors) and high-risk (HR; age ≥60 yrs and/or history of thrombosis) categories for thrombosis. More recently, the revised IPSET-thrombosis (rIPSET) model has incorporated the presence of JAK2V617F mutations, resulting in 4 risk groups: very low risk (no adverse factors), LR (JAK2V617F mutation only), intermediate risk (IR) (age ≥60), and HR (history of thrombosis or both age ≥60 and JAK2V617F). While the prognostic significance of European LeukemiaNet (ELN) clinical signs and symptoms (CSSs) has been validated in polycythemia vera (PV) pts (Palandri et al., Leukemia, 2025), their potential role in refining thrombotic risk assessment in ET remains unexplored. Aims This study aimed to: (1) determine the prevalence of ELN-defined CSSs in a real-world cohort of ET pts treated with first-line hydroxyurea (HU) and (2) evaluate the impact of CSSs on thrombotic risk during HU therapy, within conventional and rIPSET risk categories. Methods A retrospective analysis was performed on 399 WHO2022-defined ET pts managed at IRCCS AOU S. Orsola-Malpighi (Bologna, Italy) between 2010 and 2024. ELN CSSs included progressive splenomegaly, persistent or progressive leukocytosis, extreme thrombocytosis, severe pruritus, and high hematocrit (HCT), defined as HCT >45% but below the WHO2022 thresholds for PV and in the presence of normal erythropoietin levels. Relevant cardiovascular risk factors (CVRFs) were identified as the simultaneous presence of active smoking and hypertension, based on logistic regression identifying these as the most significant contributors to thrombosis risk (p=0.02 and p=0.03, respectively). Thrombosis-free survival (TFS) was calculated from the start of HU therapy to the first thrombotic event or last contact. Hazard ratios (HRs) were estimated via Cox regression with frailty terms. Results Out of 399 pts, 71.4% were aged ≥60 yrs at HU start, 65.7% carried the JAK2V617F mutation, and 21.3% had a previous thrombotic event. According to conventional stratification, 79 pts (19.8%) were classified as LR and 320 (80.2%) as HR. Using the rIPSET model, 19.8% were LR, 15.8% IR, and 64.4% HR. The median HU starting dose was 500 mg, and 93.2% of patients received antithrombotic prophylaxis. ELN-defined CSSs were observed in 208 pts (52.1%): extreme thrombocytosis in 14.5%, persistent or progressive leukocytosis in 4.8%, progressive splenomegaly in 0.3%, high Hct in 35.8%, relevant CVRFs in 5.3%, and severe pruritus in 5.0%. Notably, CSSs were more frequent in conventional LR compared to HR pts (70.9% vs 48.1%, p<0.001), and in rIPSET LR (69%) compared to IR (46%) and HR (47.9%) pts. All LR pts without CSSs started HU due to persistent microvascular symptoms. During HU therapy, 50 thrombotic events (29 arterial, 21 venous) occurred in 38 pts, yielding an overall incidence rate ratio (IRR) of 1.97% patient-yrs (p-y) (95%CI: 1.46–2.60). Thrombosis IRR was significantly higher in pts with CSSs (2.74 vs 0.93% p-y without CSSs; p=0.001). Within both conventional and rIPSET risk categories, CSSs were associated with increased thrombotic risk. In conventional HR pts, the IRR was 3.02 with CSSs vs 1.11 without (p=0.003), with 10-year TFS of 81.4% vs 90.3% (p=0.002). In LR pts, all thrombotic events occurred in those with CSSs (IRR: 2.01% p-y), with 10-year TFS of 83.9% vs 100% for LR without CSSs. Similarly, in rIPSET HR pts, IRR was 2.99% p-y with CSSs vs 1.17% without (p=0.011); in IR, 2.89% vs 1.07% (p=0.041); and in LR, 2.80% p-y with CSSs, with no events in CSS-negative pts. Multivariate analysis (including CSSs, age, JAK2 mutation, prior thrombosis) showed that ≥1 CSS (HR 3.01, 95%CI: 1.50–6.05, p=0.002) and prior thrombosis (HR 1.76, 95%CI: 1.04–3.37, p=0.038) remained independently associated with thrombotic risk. Among individual CSSs, relevant CVRFs (HR 4.0, p<0.001) and high Hct (HR 1.7, p=0.05) were independent predictors. Conclusions ELN-defined CSSs identify ET pts at higher risk of thrombosis during HU therapy, across conventional or rIPSET risk stratification. Incorporating CSSs into clinical assessment may enhance risk prediction and support more individualized management strategies in ET, paralleling advances seen in PV.