Ruxolitinib and cardiovascular outcomes among patients with myelofibrosis

Abstract Introduction: Myelofibrosis (MF) is an advanced phenotype of Philadelphia chromosome-negative myeloproliferative neoplasm (MPN) characterized by extramedullary hematopoiesis (EMH). Moreover, patients with MF are at increased risk of cardiovascular disease (CVD), especially heart failure (HF) and pulmonary hypertension (PH), which occur more frequently compared with patients with essential thrombocythemia (ET) and polycythemia vera (PV). Potential mechanisms of increased HF and PH risk in MF are increased inflammatory milieu and EMH (particularly splenomegaly). Ruxolitinib, a JAK1/2 inhibitor, used in MF in order to improve symptoms and splenomegaly and its use in MPN is associated with reduced risk of thrombosis. However, whether JAK1/2 inhibition affects the risk of HF hospitalization (HFH), in patients with MF is unclear. Methods This was an analysis of a multicenter retrospective registry of MPN patients with ≥ 1 transthoracic echocardiogram (TTE) after diagnosis of MPN at Massachusetts General Hospital, New York University Langone Health, and University of Chicago from 2010 to 2024. Patients with MF at time of first TTE without known heart failure were included. Patients were divided by receipt of ruxolitinib at the time of first TTE and HFA-PEFF and H2FPEF scores were calculated. These scoring systems are validated in predicting HF outcomes and integrate risk factors for HF including age, body mass index (BMI), hypertension, atrial fibrillation, cardiac biomarkers and TTE findings, were calculated. Cardiovascular outcomes were incident HFH, thrombosis (arterial or venous), and all-cause death. To estimate the risk of outcomes between patients treated with versus without ruxolitinib at time of first TTE, multivariable Fine-Gray competing-risk regression models were performed. Model for HF hospitalization was adjusted for MIPSS70+, H2FPEF score, and spleen size and left ventricular end-diastolic dimension (LVEDD). Models for thrombosis and all-cause death also adjusted for prior atherosclerotic CVD (ASCVD) and aspirin use. Given time-to-event models only account for first event, multivariable negative binomial regression models were performed to estimate incidence rate ratio (IRR) of recurrent HFH and thrombotic events. Results Of the 144 patients included, 46 (31.9%) were treated with ruxolitinib at time of first TTE, 72.2% had primary MF, 40.3% were female, and 78.5% were White race. There was no significant difference in age, female sex, canonical driver mutation (JAK2, CALR, MPL), non-phenotypic driver mutation, and prior CVD or CV risk factors between groups. Patients treated with ruxolitinib were less likely to have primary MF (58.7% vs 78.6%, p = 0.034), had higher median MIPSS-70+ scores (3 vs 2, p = 0.006), have larger LVEDD (median 4.8 vs 4.5 cm, p = 0.033), but no significant difference in other TTE parameters and H2FPEF and HFA-PEFF scores. The median time from ruxolitinib initiation to first TTE was 7.3 months (IQR 2.0 – 22.2). After multivariable competing-risk regression, treatment with ruxolitinib was associated with decreased risk of HFH (8.7% vs 14.3%; adjusted SHR 0.29, 95% CI 0.10 – 0.90) but not thrombosis (28.3% vs 20.4%; adjusted SHR 1.90, 0.76 – 4.74) or all-cause death (45.7% vs 50.0%; adjusted HR 0.64, 95% CI 0.34 – 1.20). A total of 4 HFH occurred in 4 patients treated with ruxolitinib and 28 HFH occurred in 14 patients without. The median LVEF at time of HFH was 60% for both groups. After negative binomial regression ruxolitinib was associated with decreased incidence of recurrent HFH (adjusted IRR 0.14, 95% CI 0.03 – 0.81) but not thrombosis (adjusted IRR 1.18, 95% CI 0.55 – 2.54). Among patients managed with ruxolitinib, those who had HFH were treated with ruxolitinib for a numerically shorter amount of time prior to TTE compared with those who did not have HFH (median 6.4 vs 28.6 months, p = 0.093). Conclusions Among patients with MF without known HF, treatment with ruxolitinib was associated with decreased risk of HFH but not thrombosis or all-cause death. Our results suggest that JAK1/2 inhibition may reduce the risk of HFH. However, further studies are need to confirm these findings and explore the potential mechanism of JAK1/2 inhibition on HF risk in MPNs and other patient populations.