Background: Recommendations for phlebotomy to treat erythrocytosis of Secondary Polycythemia (SP) are largely reactive and based on limited evidence acknowledging quality of life (QOL) metrics and thrombotic risk over Hct >54%. Unlike the large prospective polycythemia vera study (CYTO-PV), most data with erythrocytosis of SP derive from small retrospective studies and case series. We attempt to evaluate the association of pre-existing cardiovascular (CV) risk factors and Charlson Comorbidity Index (CCI) to determine outcomes with phlebotomy and create therapeutic cut-points to facilitate avoiding serious cardiovascular complications including DVT, pulmonary embolism, myocardial infarction, and cerebrovascular events.

Methods: A retrospective review of 143 patients with secondary polycythemia was conducted between December 2003 and March 2017. Patient demographics and outcomes after therapy were evaluated. CCI was used assess CV morbidity burden (low risk ≤4 vs high risk >4). The outcomes were defined as major events (DVT, PE, MI, CVA, PVD) and QOL symptoms or non-fatal events (headache, fatigue, and shortness of breath). The analysis included chi square or Fisher exact analyses where appropriate to identify associations. The protocol, and subsequent modifications, were reviewed and approved by the IRB at Texas Tech University Health Sciences Center.

Results: Our study evaluated 143 patients with a median follow-up of 46.8 months. The mean age of at diagnosis was 56.9 + 13.6y, and 71.3% were males. The most common etiologies were hypoxic Lung disease (COPD) 53.9%, obstructive sleep apnea (OSA) 33.6 % and testosterone use 6.9%. The mean Hct at diagnosis for the cohort was 52.3 + 3.3, the mean hematocrit after phlebotomy was 50.1+2.9. Achieving a Hct ≤ 52% during phlebotomy treatments was identified as low-risk group 40/143 (28%) vs Hct > 52% a high-risk group 103/143 (72%). Major CV events were recorded in 12.6% (13/103) patients in the low-hematocrit group and vs 37.5% (15/40) patients in the high-hematocrit group χ² (1, N=143) = 11.3, (p <0.001) Non-fatal/QOL events were recorded in 44.6% (46/103) of low Hct group and 62.5% (25/40) χ² (1, N=143) = 3.6, (p=0.055) of high Hct group. There was no association between QOL symptoms and major events χ² (1, N=143) =0.64, (p=0.423). A CCI >4 was associated with higher risk of major events in 50.0% (13/26) vs 12.8% (15/117) in the low CCI (≤4) χ² (1, N=143) =18.6, (p=0.0001). There was no association among non-fatal/QOL events and high 46.2% (12/26) or low CCI 50.4% (59/117) χ² (1, N=143) =0.155, (p=0.693). Use of ASA 325 mg daily was associated with decreased major events among aspirin users (6.6% vs 28.9%)χ² (1, N=143) =10.9, (p=0.001). There was no association between use of aspirin and non-fatal/QOL events. χ² (1, N=143) =0.56, (p=0.454).

Conclusion: Our cohort identifies a high-risk group with hematocrit of > 52 and a CCI > 4 that has significantly higher rates of serious CV events. ASA use was associated with decreased major events. The lack of relationship between QOL symptoms, Hct, major events, and CCI demonstrates a limitation in QOL symptoms to guide phlebotomy indication. Prospective studies are need to corroborate these findings.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution