Revised ELN Criteria in Polycythemia Vera Identify an Increased Risk Phenotype for Thrombotic Events Beyond Conventional Risk Stratification. a Multicenter Cooperative Study

Introduction

The European LeukemiaNet (ELN) criteria for initiating cytoreductive therapy aim to manage this risk, primarily using hydroxyurea (HU) (Marchetti M et al, Lancet Haematol 2022). However, the impact of ELN criteria for therapy start (CTS) on thrombotic risk, particularly among low-risk (LR, age <60 yrs & no previous thromboses) and high-risk (HR: HR-AGE, age >60 yrs; HR-THRO: previous thrombosis regardless of age) patients (pts), needs further exploration.

Aims

This study evaluates the incidence of thrombosis among PV pts undergoing HU treatment across different risk categories and the impact of CTS on thrombotic risk.

Methods

The PV-ARC study (NCT06134102) is a multicenter, retrospective study of 1162 WHO2022 PV pts.

Among these, 739 pts treated with HU were evaluated for CTS, which were revised to apply real-world practice as follows: 1) persistent/progressive leucocytosis: 100% increase if white blood cells (WBC) <10 x10⁹/L or 50% increase if WBC>10 or WBC>15 at diagnosis and HU start; 2) extreme persistent thrombocytosis: platelet (PLT) >1000 x10 ⁹/L at diagnosis and HU start; 3) progressive splenomegaly: >5 cm below costal margin (BCM) from diagnosis; 4) inadequate hematocrit (HCT) control: 5) >6 phlebotomies (PHL)/yr or HCT >53% at diagnosis and HU start or PHL intolerance; 6) uncontrolled cardiovascular risk factors; 7) severe itching (score ≥5/10).

Incidence rate ratios (IRR) were calculated per 100 patient-years (%p-y). Thrombosis-free survival (TFS) was assessed using Kaplan-Meier analysis from HU start, and multivariate Cox regression analysis was used to identify factors independently associated with thrombotic risk.

Results

Among the 739 HU-treated pts, 137 (18.5%) were LR and 602 (81.5%) HR (HR-AGE: 70.4%; HR-THRO: 29.6%). Revised CTS were identified in 445 (60.3%) pts: 95 (69.3%) LR, 242 (57.1%) HR-AGE, and 109 (61.2%) HR-THRO, with a significant difference between LR and HR pts. More than one CTS was present in 152 (34.1%) pts, mainly in LR pts (43.2% LR vs. 31.6% HR, p=0.02).

Additional reasons to start HU (spleen palpable at 2-5 cm BCM, mild leucocytosis/thrombocytosis, microvascular disturbances, PHL requirement <6/yr) were present in 100% LR and in 71.3% HR pts without CTS.

Median HU starting dose was 0.5 g/die (34.2% received ≥1 g/d). Over time, maximum HU dose was ≥1 g/d more frequently in LR (70.3%), compared to HR-THRO pts (48.7%, p=0.05) and HR-AGE (41.9%, p=0.005) pts.

Antiplatelets and/or anticoagulants were used in 94.5% of pts, comparably across risk categories.

The IRR of thrombosis during HU was 1.7 %p-y. It was similar in LR and HR-AGE pts (1.1 vs 1.3 %p-y, p=0.68) but significantly higher (3.0 %p-y) in HR-THRO pts (p=0.006 vs LR and p=0.002 vs HR-AGE). The IRR of arterial thrombosis was significantly higher in HR-THRO pts (1.1 %p-y) compared to LR (0.4 %p-y, p=0.05) and comparable to HR-AGE (0.6 %p-y, p=0.10). The IRR of venous thrombosis was significantly higher in HR-THRO pts (1.3 %p-y) compared to both LR (0.5 %p-y, p=0.05) and HR-AGE (0.6 %p-y, p=0.02).

CTS were associated with a significantly increased IRR of thrombosis in the total cohort (2.2 vs 0.7 %p-y, p<0.001) and across all risk categories: LR pts (1.6 vs 0 %p-y, p=0.05), HR-AGE pts (2.0 vs 0.5 %p-y, p=0.001), and HR-THRO pts (4.0 vs 1.7 %p-y, p=0.04).

During HU treatment, TFS at 5 years was 88.7% in pts with CTS compared to 96.1% in those without CTS (p<0.001). Across all the risk categories, best TFS at 5 yrs was observed in LR and HR-AGE (LR with no CTS, 100%; HR-AGE with no CTS: 97.8%). LR/HR-AGE pts with CTS and HR-THRO pts with no CTS had comparable TFS (86.4%, 91.4% and 88.1%, respectively). HR-THRO pts with CTS had the worse TFS (79.5%).

Multivariate Cox analysis considering CTS, age>60 yrs and previous thrombosis, confirmed that CTS (HR: 3.1, p<0.001) and previous thrombosis (HR: 3.2, p<0.001) were independent predictors of thrombotic risk.

Conclusion

This study demonstrates that ELN criteria for therapy start effectively identify PV pts at increased thrombotic risk, regardless of their conventional risk category. These findings highlight the need for incorporating CTS into existing prognostic models to improve risk stratification and therapeutic decision-making in PV. Additionally, the data suggest that most LR pts require HU due to CTS, emphasizing the necessity for tailored management strategies and further research into the long-term impact of HU therapy in younger pts.