Real-world analysis of polycythemia vera treatment reveals nonadherence to NCCN guidelines in a large proportion of patients Open Access

TO THE EDITOR:

Polycythemia vera (PV) is a myeloproliferative neoplasm characterized by clonal expansion of hematopoietic stem and progenitor cells and elevated red blood cell mass, with associated symptoms and risks of thrombosis and disease progression.¹ Treatment almost always involves low-dose aspirin and/or phlebotomy, but cytoreductive therapy is not always instituted. High-risk patients, that is, those aged ≥60 years and/or with prior thrombosis, and symptomatic low-risk patients require cytoreductive therapy to mitigate symptoms and/or lower the risk of disease-related complications.² Cytoreductive therapy has many potential benefits but may cause adverse events (AEs) that influence treatment decisions. In this real-world study, we evaluate the cytoreductive therapy choices and the AEs associated with the 3 established cytoreductive therapies for PV: hydroxyurea (HU), interferon alfa (IFN-α)–based therapies (mainly peginterferon alfa-2a and the newest approved ropeginterferon alfa-2b [ropeg]), and ruxolitinib (rux).

HU is an oral chemotherapy that inhibits ribonucleotide reductase and is commonly used as a first-line cytoreductive capable of controlling blood counts and reducing thrombotic risk. Its prolonged use can be complicated by cytopenias, mucocutaneous toxicities, and nonmelanoma skin cancers.³ Ropeg is a US Food and Drug Administration (FDA)–approved newer-generation IFN-α, and a first-line option that is preferred over HU for patients with low-risk PV needing cytoreduction, according to the 2025 National Comprehensive Cancer Network (NCCN).^2,4 It has a favorable in vivo pharmacokinetic-pharmacodynamic profile, allowing biweekly subcutaneous administrations or even monthly upon inducing durable clinical and molecular responses.^5-8 Unlike HU, IFN-α can selectively suppress neoplastic cells to inhibit disease progression and improve survival.^9-13 However, despite being the only cytoreductive not associated with immunosuppression, IFN-α treatment can be associated with heightened immune reactions such as influenza-like symptoms or, rarely, autoimmune manifestations.¹⁴ Rux, a Janus kinase 1/2 inhibitor, is indicated as a second-line agent.¹⁵ It can control hematocrit, symptoms, and splenomegaly, and induce molecular responses.¹⁶ It is associated with an increased risk of infections, nonmelanoma skin cancer, and cytopenias.¹⁷ 

Thrombotic events and disease progression remain the leading causes of morbidity and mortality in PV, underscoring the importance of effective cytoreduction, disease-modification, and cardiovascular risk-factor management.^1,18 There remain critical gaps in the initiation and choice of cytoreductive therapy among both academic and community centers, and concern that cytoreductive treatment in PV is neither optimized nor adherent to guidelines at the very least. The prospective observational REVEAL study identified a significant proportion of patients with “high-risk PV” who were not on cytoreductive therapy.¹⁸ Although it completed before the FDA approval of ropeg, the REVEAL study also identified <2% of patients were treated with IFN-α. The problem may stem from a lack of understanding of the comparative safety of cytoreductive therapies. Randomized trials in PV compared few therapies in defined patient populations. Real-world clinical data are essential to identify the long-term risk-benefit profiles, and help guide individualized therapy. To address this gap, we conducted a retrospective cohort study of patients receiving PV treatment at 30 cancer centers in the United States that contributed data to the IntegraConnect Database (www.integraconnect.com). It is a large database of deidentified claims and electronic health records from academic and community centers nationwide.

This study included 3731 patients diagnosed with PV after the World Health Organization 2016 criteria revision, who had received ≥180 days of cytoreductive or phlebotomy treatment without a diagnosis of myelofibrosis or acute myeloid leukemia (supplemental Figure 1). To ensure a valid comparison and to capture accurate line of therapy (LOT) information, ≥90 days of clinical records before the diagnosis of PV were required. Treatment-emergent AEs were assessed from the electronic health records system and practice management records, and classified into the following categories: anemia, cytopenia (leukopenia, neutropenia, thrombocytopenia, and anemia), cardiovascular, central nervous system/mood, gastrointestinal, respiratory, venous thromboembolism, other organ dysfunctions, and dermatological (supplemental Table 1). Baseline demographic and clinical characteristics were summarized and compared by PV-related treatment (Table 1). With phlebotomy as the “control” group, we then used a generalized estimating equation regression model with difference-in-difference design to calculate the adjusted odds ratio (OR) for each category of AEs of the cytoreductive treatments, while controlling for potential confounders, including age, sex, race, provider specialty, risk of thrombotic events (low-/high-risk), and line of treatment. Statistical modeling and analyses were done in SAS (version 9.4).

A total of 2089 patients (56%) were male, and 2871 (77%) had high-risk PV, with a mean age of 67.2 years (standard deviation [SD], 12.7; median, 69 years). Overall, 1824 patients (49%) were treated with HU, 41% (n = 1537) with phlebotomy only, 15% (n = 571) with rux, 1.1% (n = 42) with ropeg, and 0.9% (n = 34) with other IFN-α. Patients treated with other IFN-α (mean age, 54 years [SD, 15.8]) and ropeg (mean age, 62 years [SD, 14.3]) were younger compared with those on other treatments. The top 2 cytoreductive treatments used in first line were HU (62.5%) and rux (32.4%), whereas ropeg and other IFN-α were mostly used in second or third line. These findings show discrepancy with NCCN guidelines and FDA labels, because ropeg is FDA approved in the first-line setting, whereas rux is not FDA approved for first-line use in PV.^4,5,12 

As shown in Table 2, HU was associated with significantly increased odds of developing cytopenia compared with phlebotomy (OR, 1.72; 95% confidence interval [CI], 1.44-2.06). Patients receiving other IFN-α exhibited lower odds of new cardiac complications (OR, 0.38; 95% CI, 0.23-0.62) and respiratory conditions (OR, 0.14; 95% CI, 0.03-0.78) compared with those treated with phlebotomy alone. Ropeg recipients demonstrated lower odds of cardiac complications (OR, 0.25; 95% CI, 0.08-0.79) and venous thromboembolism (OR, 0.57; 95% CI, 0.42-0.78). For cardiac events, HU and rux had ORs of 0.89 and 0.97, respectively, which were not statistically significant from phlebotomy alone. Furthermore, ropeg demonstrated numerically lower odds of developing anemia.

Our results showed that the real-world cytoreductive therapy use in PV remains far below the proportions that might be expected if the FDA approval and NCCN guidelines for PV treatments were followed. First, the proportion of patients receiving phlebotomy only without cytoreductive was very high given most were considered high-risk patients. Among the high-risk patients in this study (n = 2871), 1003 (34.9%) did not receive cytoreductive therapy. Second, the proportion of patients receiving IFN-α was very low (2%), despite first-line designation, favorable evidence, and the >3 years of ropeg FDA approval. Most patients received HU (62.5%) over interferons (2%), despite the results of the phase 3 PROUD-PV/CONTINUATION-PV trials showing superior clinical benefits of ropeg over HU in the long term, including durable complete hematologic response, reduction of JAK2V617F allele burden, and event-free survival.¹² Third, rux is mainly as a second- or third-line agent but was used in 13.8% of patients with PV as first line, whereas the first-line agent ropeg was recorded in only 0.5%. This may be due to myriad factors, including lack of awareness of the NCCN guideline updates, reluctance to use IFN-α–based therapy due to conflation of historic AE profiles of older IFN-α formulations and non-PV indications (eg, solid tumors), and most importantly, a lack of access to care by hematology/oncologists with expertise in PV treatment. This study showed only 40.6% were under hematology/oncology specialty care (Table 1). Therefore, many patients with PV are not receiving optimal specialty care, as the real-world treatment patterns identify nonadherence to guidelines and approved FDA LOTs. Suboptimal PV treatment raises the risks of thromboembolic events and disease progression, and the associated morbidity and mortality.

We also found significant heterogeneity in treatment-emergent cytopenias between cytoreductive therapies. We observed a high frequency of cytopenias with HU, known for its myelosuppressive toxicity, and cytopenia without anemia with other IFN-α, whereas ropeg was not associated with significant cytopenia compared with phlebotomy alone.

Although this study underscores the nonadherence to NCCN guidelines and FDA-approved LOTs, it also identifies the safety profile of cytoreductive therapies compared with phlebotomy only in a real-world setting. In reference to thrombotic complication rates, ropeg was the only therapy that was associated with significantly lower rates of venous thrombosis vs phlebotomy. Both ropeg and other IFN-α displayed decreased odds of cardiac events during treatment compared with phlebotomy alone, whereas HU and rux did not. Although adjusted for age and other cardiovascular risk factors, this observation may remain confounded by the healthier profile of patients selected to receive IFN-α–based therapy. In this study, ropeg also showed the lowest odds of anemia, consistent with clinical trial results.¹⁹ Therapeutic phlebotomy without cytoreduction may therefore subject patients to unacceptable risks of thrombosis and iron deficiency-related symptoms (eg, fatigue, lethargy, cognitive impairment, and disease progression).^20-22 

Study imitations include the retrospective nature, reliance on administrative or clinical coding data, and variance of data standards in the real-world setting, which may introduce misclassification or underreporting, and relatively small numbers for IFN-α– and ropeg-treated patients. Second, our findings are potentially confounded by indication and LOT, as currently younger patients are more likely to receive interferons or ropeg, despite ropeg being approved for all adults with PV. Finally, there was a lack of data on treatment doses, patient compliance, and duration of therapy, and our real-world analyses may not capture all relevant AEs as specifically assessed in clinical trials.

In conclusion, our large-scale real-world analysis highlights significant divergence from the contemporary evidence-based landscape of cytoreductive therapies for PV treatment, FDA-approved LOTs, and NCCN guidelines. The findings emphasize that cytoreductive therapy in PV is underutilized and should be adherent to evidence-based practices. Furthermore, cytoreductive therapy should be tailored to the individual patient with treatment AE profiles considered, and carefully monitored, and cardiovascular risk mitigated.

Acknowledgments: The authors thank our colleagues for the discussions and help in editing the manuscript.

The study was supported by PharmaEssentia.

Contribution: G.A.-Z., A.M.H., J.J.S., A.Y., A.Q., H.-L.C., and R.A.M. were involved in the conceptualization, analysis, writing, and review and editing of the manuscript; H.-L.C. was involved in the statistical analysis of the data; and all authors reviewed and approved the final draft of the manuscript.

Conflict-of-interest disclosure: G.A.-Z. reports research support provided to the institution from PharmaEssentia. A.M.H. has received honoraria/consulting fees from GlaxoSmithKline (GSK), Cogent Biosciences, PharmaEssentia, Blueprint Medicines, CTI Biopharma (now Sobi), and Incyte; and declares research funding from Incyte, Cogent Biosciences, Ascentage Pharma, Blueprint Medicines, Syntrix Biosystems, Novartis, and PharmaEssentia. A.Y. reports consultancy for Incyte, CTI Pharma (Sobi), PharmaEssentia, Pfizer, Novartis, Servier, AbbVie, Karyopharm Therapeutics, GSK, Blueprint Medicine, Apellis, Gilead, Notable Labs, and Protagonist; and research funding from CTI Pharma (Sobi) and Stemline Therapeutics. A.Q. serves as the chief medical officer of PharmaEssentia. H.-L.C. is an employee of PharmaEssentia. R.A.M. reports consulting fees (honoraria) over past 3 years from AbbVie, Blueprint, Bristol Myers Squibb (BMS), CTI, Genentech, Geron, GSK, Incyte, Novartis, Sierra, Sierra Oncology, and Telios; and research support from AbbVie, Blueprint, BMS, CTI, Genentech, Incyte, MorphoSys, and Sierra. J.J.S. declares no competing financial interests.

Correspondence: Ghaith Abu-Zeinah, Division of Hematology and Medical Oncology, Weill Cornell Medicine, 1300 York Ave, Box 13, New York, NY 10065; email: gfa2001@med.cornell.edu.