Comparative risk of major adverse cardiovascular events in polycythemia vera patients treated with ruxolitinib, hydroxyurea, or interferon: A real-world, propensity-matched cohort study

Abstract Background: Vascular complications are the primary cause of morbidity and mortality in patients with polycythemia vera (PV), yet consensus is lacking on the optimal cytoreductive therapy to reduce risk. Hydroxyurea (HU) is used commonly, though resistance or intolerance occurs in up to 25% of patients. Interferons (IFN) can lead to molecular remissions in select patients, while ruxolitinib (RUX), has consistently demonstrated hematocrit control and symptom reduction across studies. However, comparative cardiovascular safety among these agents remains unknown. We conducted a real-world, target trial emulation to assess the risk of major adverse cardiovascular events (MACE) in PV patients treated with RUX, HU, or IFN. Methods We identified adults (≥18 years) with PV initiating monotherapy with RUX, HU, or IFN between January 1, 2013, and December 31, 2023, using a federated EHR platform. Patients were required to have ≥3 prescriptions within the study period serving as a proxy for treatment continuity. Exclusion criteria included ECOG >3, Karnofsky <40, pregnancy/nursing, and any MI or stroke within 90 days prior to treatment initiation. Three pairwise analyses were conducted: RUX vs HU, RUX vs IFN, and HU vs IFN. Propensity score matching (1:1) was performed using baseline demographics, cardiovascular risk factors, comorbidities, and laboratory values. The primary outcome was 3-point MACE (MI, stroke, cardiac death). Secondary outcomes included all-cause mortality, venous thromboembolism (VTE), and hospital utilization. Cox proportional hazards models were used to estimate hazard ratios (HRs) with 95% confidence intervals (CIs). In all pairwise models, the second agent listed served as the reference group. Patients were censored at outcome, death, last follow-up, or 5 years Results A total of 10,508 patients met eligibility criteria. After 1:1 propensity score matching, 1,828 patients were included in the RUX vs HU analysis, 610 patients in the RUX vs IFN analysis, and 718 patients in the HU vs IFN analysis. Median follow-up was 53.4 months. In the RUX vs HU comparison, MACE incidence was 8.2% vs 6.5% (HR 1.364; p = 0.011). Kaplan-Meier curves demonstrated divergence in MACE-free survival beginning at 35 months. For RUX vs IFN, MACE rates were 5.7% vs 4.1% (HR 1.45; p = 0.156). In the HU vs IFN cohort, MACE occurred in 4.6% vs 3.5% of patients (HR 1.27; p = 0.257). RUX was associated with higher all-cause mortality compared to HU (HR 3.80; p<0.001) and IFN (HR 1.746; p<0.001). RUX was also associated with higher VTE compared to HU (HR 1.308; p=0.006), but no statistical difference was observed with IFN. IFN was associated with fewer hospitalizations and ICU admissions, whereas RUX was associated with greater all-cause hospitalizations and ICU admissions Conclusion In this large, real-world comparative study, ruxolitinib was associated with a higher risk of MACE, VTE, and mortality compared to hydroxyurea and showed no statistical difference in cardiovascular safety as compared to interferon. Interferon demonstrated the most favorable overall safety profile, including lower event rates and reduced healthcare utilization. However, given the smaller sample sizes, the analyses involving interferon may have been underpowered. These findings suggest that a patient's baseline cardiovascular risk profile should be a key factor in the individualized selection of cytoreductive therapy in PV.