P1062: A REAL-WORLD EVALUATION OF THE ASSOCIATION BETWEEN ELEVATED BLOOD COUNTS AND THROMBOTIC EVENTS IN POLYCYTHEMIA VERA: AN ANALYSIS OF DATA FROM THE REVEAL STUDY

Abstract

Background: Polycythemia vera (PV) is characterized by clonal hematopoiesis leading to elevated peripheral blood counts and an increased risk of thrombotic events (TEs). Advanced age and TE history form the conventional risk model used to determine TE risk/treatment strategy. Associations between TEs and elevated hematocrit (HCT) levels exist, but associations with white blood cell (WBC) or platelet (PLT) counts have not been assessed consistently. The large, real-world, prospective Observational Study of Pts with Polycythemia Vera in US Clinical Practices (REVEAL; NCT02252159) followed pts with PV treated in community or academic centers. Aims: This analysis evaluated associations between elevated blood counts and TEs in pts with PV using data from REVEAL. Methods: Eligible pts had ≥3 lab values (blood counts) post-enrollment; pts with a post-enrollment TE but no lab value <6 mo before that TE were excluded. The association between blood counts and TEs was assessed using a time-dependent covariate Cox proportional hazards model. Time to first post-enrollment TE was modeled with time censored at last known visit for pts with no TE. Each lab parameter was modeled with sex, age, disease duration, and TE history at enrollment as baseline covariates and treatment as a time-dependent covariate. Blood counts were included as binary time–dependent covariates using the following thresholds: HCT >45%, WBC >11×109/L, PLT >400×109/L. Linear interpolation was used to determine lab values between observed lab values. Alternative thresholds for WBC (<7, ≥7 to <8.5; ≥8.5 to <11, and ≥11×109/L, and >12×109/L with HCT controlled at ≤45%) and PLT counts (>600×109/L) were evaluated. Statistical significance was considered at P<0.05. Results: 2271/2510 pts were eligible (median age, 66 y [range, 22–95]; male, 54.1%). Median disease duration was 4.1 y (range, 0–56.3), 456 (20.1%) had TE history; 52.6% of pts received hydroxyurea. Of 106 pts who had TEs, 30 had arterial TEs (most commonly, transient ischemic attack [n=15]) and 76 had venous TEs (most commonly, deep vein thrombosis [n=37]). Elevated HCT levels (>45%, hazard ratio [HR]=1.84 [95% CI, 1.234–2.749], P=0.0028), WBC (>11×109/L, HR=2.35 [1.598–3.465], P<0.0001), and PLT counts (>400×109/L, HR=1.60 [1.088–2.359], P=0.0170) were each associated with increased TE risk (Table 1). WBC count ≥11×109/L is associated with the highest TE risk compared with WBC count <7×109/L (HR=2.61 [95% CI, 1.594–4.262], P<0.0001). Elevated WBC >12×109/L was significantly associated with increased risk of TE with HCT controlled at ≤45%. PLT count (>600×109/L) increased TE risk (HR=1.37 [95% CI, 0.763–2.468]) compared with PLT count ≤600×109/L, but this was not statistically significant (P>0.05). In all models, advanced age, female sex, and TE history were associated with increased TE risk. Image:Summary/Conclusion: This analysis of REVEAL, the largest real-world cohort of PV pts to date, demonstrated that elevated HCT levels (>45%), WBC (>11×109/L), and PLT counts (>400×109/L) were associated with increased TE risk. An association of elevated WBC >12×109/L with increased risk of TE was also observed when HCT was controlled, indicating that TE risk may be reduced by controlling WBC as well as HCT. These data support the need to incorporate blood count into risk stratification and treatment strategies for pts with PV in clinical practice and to move beyond the conventional risk model. Further studies to understand the causal relationship between elevated blood counts and TEs are warranted.