Impact of erythrocytosis on overall survival, thrombotic events, and risk of transformation to AML in patients with myelofibrosis.

e18637 Background: Erythrocytosis is an uncommon phenotype in myelofibrosis (MF), and its association with long-term survival and thrombotic risk remains poorly defined. We evaluated the longitudinal outcomes associated with erythrocytosis in a large real-world MF cohort. Methods: We performed a retrospective cohort study using a federated electronic health record network. Adult patients with MF were classified by the presence or absence of erythrocytosis. Patients with prior allogeneic hematopoietic stem cell transplantation were excluded. Propensity score matching (1:1) was conducted using demographics, cardiovascular comorbidities, and baseline laboratory variables. Outcomes included all-cause mortality, pulmonary embolism (PE), and venous thromboembolism (VTE), evaluated at 1, 3, and 5 years. Time-to-event analyses were performed using Kaplan–Meier methods with log-rank testing; hazard ratios (HRs) are reported with log-rank p values. Binary outcomes are summarized using odds ratios (ORs). Results: After matching, 4,357 patients were included in each cohort. Erythrocytosis was associated with significantly lower all-cause mortality at 1 year (4.9% vs 6.6%; HR 95% CI 0.61–0.87; log-rank p<0.001), with this survival advantage persisting at 3 years (10.0% vs 12.4%; HR 0.69–0.89; p<0.001) and 5 years (14.1% vs 16.4%; HR 0.75–0.93; p=0.001). In contrast, erythrocytosis was associated with higher thrombotic risk. PE incidence was increased at 1 year (1.1% vs 0.7%; HR 1.02–2.57; log-rank p=0.040), 3 years (2.1% vs 1.1%; HR 1.35–2.76; p<0.001), and 5 years (2.8% vs 1.4%; HR 1.42–2.66; p<0.001). Similarly, VTE rates were higher at 1 year (2.0% vs 1.3%; HR 1.07–2.12; p=0.020), 3 years (3.7% vs 2.6%; HR 1.10–1.81; p=0.007), and 5 years (5.3% vs 3.4%; HR 1.21–1.86; p<0.001). Conclusions: In a large propensity-matched real-world cohort, erythrocytosis in MF was associated with a durable survival advantage across 5 years of follow-up, despite a progressively increased risk of thromboembolic events. These findings suggest a biologically distinct MF phenotype and highlight the importance of individualized thrombotic risk assessment in patients with erythrocytosis.