Polycythemia vera (PV) is a rare myeloproliferative neoplasm affecting 44 to 57 individuals per 100,000 in the United States.1,2 It is characterized by somatic mutations in the hematopoietic stem cell, resulting in hyperproliferation of mature myeloid lineage cells.2 Sustained erythrocytosis is a hallmark of PV, although many patients also have leukocytosis and thrombocytosis.2,3 These patients have increased inherent thrombotic risk with arterial events reported to occur at rates of 7 to 21/1000 person-years and venous thrombotic events at 5 to 20/1000 person-years.4-7 Thrombotic and cardiovascular events are leading causes of morbidity and mortality, resulting in a reduced overall survival of patients with PV compared with the general population.3,8-10
Blood Cell Counts and Thrombotic Events in PV
Treatment strategies for patients with PV mainly aim to prevent or manage thrombotic and bleeding complications through normalization of blood counts.11 Hematocrit (Hct) control has been reported to be associated with reduced thrombotic risk in patients with PV. This was shown and popularized by the prospective, randomized Cytoreductive Therapy in Polycythemia Vera (CYTO-PV) trial in which participants were randomized 1:1 to maintaining either a low (< 45%) or high (45%-50%) Hct for 5 years to examine the long-term effects of more- or less-intensive cytoreductive therapy.12 Patients in the low-Hct group were found to have a lower rate of death from cardiovascular events or major thrombosis (1.1/100 person-years in the low-Hct group vs 4.4 in the high-Hct group; hazard ratio [HR], 3.91; 95% confidence interval [CI], 1.45-10.53; P = .007). Likewise, cardiovascular events occurred at a lower rate in patients in the low-Hct group compared with the high-Hct group (4.4% vs 10.9% of patients, respectively; HR, 2.69; 95% CI, 1.19-6.12; P = .02).12
Leukocytosis has also been linked to elevated risk for vascular events as shown in several studies, including the real-world European Collaboration on Low-Dose Aspirin in PV (ECLAP) observational study and a post hoc subanalysis of the CYTO-PV study.13,14 In a multivariate, time-dependent analysis in ECLAP, patients with white blood cell (WBC) counts > 15 × 109/L had a significant increase in the risk of thrombosis compared with those who had lower WBC counts, with higher WBC count more strongly associated with arterial than venous thromboembolism.13 In CYTO-PV, a significant correlation between elevated WBC count (≥ 11 × 109/L vs reference level of < 7 × 109/L) and time-dependent risk of major thrombosis was shown (HR, 3.9; 95% CI, 1.24-12.3; P = .02).14 Likewise, WBC count ≥ 11 × 109/L was found to be a predictor of subsequent venous events in a separate single-center multivariate analysis of patients with PV.8
Although CYTO-PV remains one of the largest prospective landmark studies in PV demonstrating the impact of Hct control on thrombosis, it is worthwhile to note that the patients in the high-Hct group who received less frequent myelosuppressive therapy with hydroxyurea than the low-Hct group also had higher WBC counts.12,15 Work is needed to determine the relative effects of high Hct and high WBC counts on PV independent of each other.
The Veteran Population with PV
Two recently published retrospective analyses from Parasuraman and colleagues used data from the Veterans Health Administration (VHA), the largest integrated health care system in the US, with an aim to replicate findings from CYTO-PV in a real-world population.16,17 The 2 analyses focused independently on the effects of Hct control and WBC count on the risk of a thrombotic event in patients with PV.
In the first retrospective analysis, 213 patients with PV and no prior thrombosis were placed into groups based on whether Hct levels were consistently either < 45% or ≥ 45% throughout the study period.17 The mean follow-up time was 2.3 years, during which 44.1% of patients experienced a thrombotic event (Figure 1). Patients with Hct levels < 45% had a lower rate of thrombotic events compared to those with levels ≥ 45% (40.3% vs 54.2%, respectively; HR, 1.61; 95% CI, 1.03-2.51; P = .04). In a sensitivity analysis that included patients with pre-index thrombotic events (N = 342), similar results were noted (55.6% vs 76.9% between the < 45% and ≥ 45% groups, respectively; HR, 1.95; 95% CI, 1.46-2.61; P < .001).
In the second analysis, the authors investigated the relationship between WBC counts and thrombotic events.16 Evaluable patients (N = 1565) were grouped into 1 of 4 cohorts based on the last WBC measurement taken during the study period before a thrombotic event or through the end of follow-up: (1) WBC < 7.0 × 109/L, (2) 7.0 to 8.4 × 109/L, (3) 8.5 to < 11.0 × 109/L, or (4) ≥ 11.0 × 109/L. Mean follow-up time ranged from 3.6 to 4.5 years among WBC count cohorts, during which 24.9% of patients experienced a thrombotic event. Compared with the reference cohort (WBC < 7.0 × 109/L), a significant positive association between WBC counts and thrombotic event occurrence was observed among patients with WBC counts of 8.5 to < 11.0 × 109/L (HR, 1.47; 95% CI, 1.10-1.96; P < .01) and ≥ 11 × 109/L (HR, 1.87; 95% CI, 1.44-2.43; P < .001) (Figure 2).16 When including all patients in a sensitivity analysis regardless of whether they experienced thrombotic events before the index date (N = 1876), similar results were obtained (7.0-8.4 × 109/L group: HR, 1.22; 95% CI, 0.97-1.55; P = .0959; 8.5 - 11.0 × 109/L group: HR, 1.41; 95% CI, 1.10-1.81; P = .0062; ≥ 11.0 × 109/L group: HR, 1.53; 95% CI, 1.23-1.91; P < .001; compared with < 7.0 × 109/L reference group). Rates of phlebotomy and cytoreductive treatments were similar across groups.16
Some limitations to these studies are attributable to their retrospective design, reliance on health records, and the VHA population characteristics, which differ from the general population. For example, in this analysis, patients with PV in the VHA population had significantly increased risk of thrombotic events, even at a lower WBC count threshold (≥ 8.5 × 109/L) compared with those reported in CYTO-PV (≥ 11 × 109/L). Furthermore, approximately one-third of patients had elevated WBC levels, compared with 25.5% in the CYTO-PV study.14,16 This is most likely due to the unique nature of the VHA patient population, who are predominantly older adult men and generally have a higher comorbidity burden. A notable pre-index comorbidity burden was reported in the VHA population in the Hct analysis, even when compared to patients with PV in the general US population (Charlson Comorbidity Index score, 1.3 vs 0.8).6,17 Comorbid conditions such as hypertension, diabetes, and tobacco use, which are most common among the VHA population, are independently associated with higher risk of cardiovascular and thrombotic events.18,19 However, whether these higher levels of comorbidities affected the type of treatments they received was not elucidated, and the effectiveness of treatments to maintain target Hct levels was not addressed in the study.