Thromboembolic Risk Remains High Across Polycythemia Vera Groups

Thromboembolic events (TE) remain a substantial risk across all clinical risk groups of patients living with polycythemia vera (PV). During the post-index observation period of a recent large retrospective study that included 20,089 patients, a quarter of them experienced at least 1 TE, according to the findings recently published in Clinical Lymphoma, Myeloma and Leukemia.¹

The authors used de-identified electronic health records (EHRs) and linked claims from the Optum Market Clarity database for years 2007 to 2019, covering more than 105 million patients in the US and Puerto Rico. To be included, each adult patient (mean [SD] age, 63 [13.7] years) had to have at least 2 PV diagnosis codes separated by at least 60 days and meet data minimums.

How Common Were Thromboembolic Events?

The final study cohort was 20,089 patients, 57.9% of whom were male patients and 88.6% Caucasian. At study entry, 16.2% had event-based high-risk PV, defined by a prior TE at any age; 49.4% had age-based high-risk PV, defined as age 60 or older without a prior TE; and 34.4% were classified as low-risk, meaning they were younger than 60 years and had no history of TEs. The median (IQR) overall post-index period was 4.4 years (2.4-5.9), with the event-based high- and low-risk groups having identical mean (SD) post-index periods (4.7 years [2.4]) and the age-based high-risk group having 4.3 years (2.5). Median baseline hematocrit was 49%, with 58% of this group having a baseline hematocrit of at least 48%.

Although patients with a history of TEs had the greatest risk (50.2%), 13.3% of those classified as low-risk also experienced a post-index TE, suggesting that conventional risk stratification may not capture the full thrombotic burden in PV. Overall, TEs were seen in 25.1% of the study cohort. The most common venous events were deep vein thrombosis or deep thrombophlebitis, reported in 8.1% of patients, and pulmonary embolism, reported in 4.5%. For arterial events, stroke occurred in 7.1% and myocardial infarction/acute coronary syndrome in 6.4%.

A history of thrombosis was the strongest predictor of subsequent TE (HR, 2.60; 95% CI, 2.38-2.85), the authors explained, with other significant predictors including white blood cell (WBC) counts (ref, < 11 x 10³ µL) of at least 25 x 10³ µL (HR, 1.89; 95% CI, 1.50-2.39), pre-index use of cytoreductive therapy (HR, 1.80; 95% CI, 1.63-1.99), WBC of 15 to less than 25 x 10³ µL (HR, 1.45; 95% CI, 1.27-1.66), and age of at least 60 years (HR, 1.44; 95% CI, 1.32-1.58). Each percentage-point increase in post-index hematocrit, WBC of 11 to less than 15 x 10³ µL, chronic kidney disease, syncope, and splenomegaly were also associated with increased TE risk.

The authors cautioned that the association between cytoreductive therapy and subsequent risk may reflect the fact that patients with more advanced or aggressive disease were more likely to receive cytoreductive therapy “to help achieve and maintain hematocrit below 45%.” They also emphasized that these data echo previous research that cytoreductive therapies do not eliminate the risk of TEs in high-risk patients; they simply reduce it in patients considered high risk.^2-4

The Clinical Implications of These Findings

Current risk categories are largely based on age and previous thrombosis, but the results suggest that other factors—particularly leukocytosis and hematocrit—may contribute meaningfully to thrombotic risk.¹ The authors suggest that current risk stratification may underestimate risk for some patients considered low risk. Overall mortality was 21% in the follow-up period but was substantially higher among event-based high-risk patients and age-based high-risk patients than among low-risk patients (31.5% vs 27.8% vs 6.3%, respectively).

The authors conclude that thromboembolic risk reduction should remain a priority across all PV risk groups. They emphasize the need to better identify patients at elevated risk and more effective therapeutic management to reduce thrombotic complications.

Still, there are limitations to their findings, including the retrospective nature of EHR and claims data, possible coding and documentation errors, incomplete records, and the absence of screening for TEs. JAK2 mutation data were also missing or indeterminate for most patients, preventing analysis of JAK2 status and allele burden as risk factors.

References

1. Kuykendall AT, Molina A, Reaven N, et al. Risk of thromboembolic events in patients with polycythemia vera: a real-world observational study.Clin Lymphoma Myeloma Leuk. 2026;26(8):549-561.e23. doi:10.1016/j.clml.2026.05.006
2. Verstovsek S, Pemmaraju N, Reaven NL, et al. Real-world treatments and thrombotic events in polycythemia vera patients in the USA. Ann Hematol. 2023;102(3):571-581. doi:10.1007/s00277-023-05089-6
3. De Stefano V, Za T, Rossi E, et al. Recurrent thrombosis in patients with polycythemia vera and essential thrombocythemia: incidence, risk factors, and effect of treatments. Haematologica. 2008;93(3):372-380. doi:10.3324/haematol.12053
4. Barbui T, Vannucchi AM, Finazzi G, et al. A reappraisal of the benefit-risk profile of hydroxyurea in polycythemia vera: a propensity-matched study. Am J Hematol. 2017;92(11):1131-1136. doi:10.1002/ajh.24851