Conventional interferon-α 2b versus hydroxyurea for newly-diagnosed patients with polycythemia vera in a real world setting: a retrospective study based on 286 patients from a single center

Polycythemia vera (PV) is a myeloproliferative neoplasm (MPN) characterized by clonal proliferation of multipotent bone marrow progenitors.¹ A clinical trial investigating the efficacy of pegylated interferon (IFN) for PV and essential thrombocytosis is ongoing in China. However, as only conventional IFN and hydroxyurea (HU) are covered by Chinese basic medical insurance, these cytoreductive agents are recommended as first-line treatment by the consensus of Chinese experts for the diagnosis and treatment of PV, including low-risk patients.²

As the difference in efficacy between conventional IFN and HU for newly-diagnosed PV is undefined, we retrospectively analyzed data of 286 newly-diagnosed PV patients who were treated at the Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences between June 1, 2007 and February 28, 2020. All patients received conventional IFN-a 2b or HU for at least 6 months. Patients were excluded if they changed groups. The flowchart for patient selection is shown in the Online Supplementary Figure S1A. Cases were diagnosed in accordance with the 2016 World Health Organization diagnostic definitions.³

Conventional IFN-a 2b was recommended first for young (age <60 years old) patients and older patients without contraindications. HU was usually recommended for other patients. In total, 82 and 204 patients received single-agent conventional IFN-a 2b (IFN cohort) and single- agent HU (HU cohort), respectively. Generally, the initial dose of conventional IFN-a 2b was 3×10⁶ IU three times per week; the initial dose of HU was 20 mg/kg/day. Treatment schedules were adjusted by monitoring peripheral blood counts with the target of hematocrit (HCT) <45%.

Quantitative measurements of the JAK2 V617F variant allele frequency (VAF) were performed by real-time polymerase chain reaction (PCR) as previously described.⁴ Hematologic and molecular responses were evaluated in accordance with the revised response criteria of the European LeukemiaNet (ELN) and International Working Group-Myeloproliferative Neoplasms Research and Treatment (IWG-MRT).⁵ Complete hematologic remission (CHR) was defined as HCT <45% without phlebotomy, white blood cell (WBC) <10×10⁹/L, and platelets ≤400×10⁹/L. Complete molecular response (CMR) was defined as indetectable JAK2 V617F mutation. Partial molecular response (PMR) applied only to patients with a JAK2 V617F VAF ≥20% before treatment and was defined as a ≥50% decrease in allele burden after treatment. ⁵ The clinical and laboratory features of subjects in the IFN and HU cohorts are displayed in Table 1. The median treatment duration in the IFN and HU cohorts were 51 months (interquartile range [IQR], 24–83 months) and 53 months (IQR, 31–84 months), respectively. The duration of exposure to IFN or HU for each patient is shown in the Online Supplementary Figure S1B and C.

Compared with the HU cohort, a higher proportion of patients in the IFN cohort achieved CHR (65% vs. 43%; P=0.001), control of HCT (72% vs. 43%; P=0.06), control of platelets (88% vs. 77%; P=0.04), and control of WBC (89% vs. 72%; P=0.002; Figure 1A) during follow-up.

A higher proportion of low-risk subjects who received IFN achieved CHR compared with those who received HU (64% vs. 32%; P=0.001; Figure 2A). Consistently, high-risk subjects who received IFN also had a higher CHR rate than those who received single-agent HU (68% vs. 47%; P=0.06; Figure 2C).

Interestingly, the median increase in mean corpuscular volume (MCV) from baseline, when patients achieved the best control of HCT, was 21.2 fL (IQR, 9.1–31.9 fL) in the HU cohort, which was much higher than in the IFN cohort (3.9 fL; IQR, −2.8–9.5; P<0.001). Because HCT equals the erythrocyte count multiplied by MCV, some patients in the HU cohort might have been phlebotomized with normal RBC.

The median duration from starting treatment to achieving CHR was 11 months (IQR, 7–23 months) in the IFN cohort, which was shorter than in the HU cohort (19 months; IQR, 10–47 months; P=0.001). Among the patients who achieved CHR, two (4%) and seven (7%) were lost during follow-up in the IFN and HU cohorts, respectively. Compared with the HU cohort, CHR rates in the IFN cohort were higher throughout the treatment duration and became significantly better after 3 years of continuous treatment (88% vs. 44%; P<0.001; Figure 1B), which was consistent with the results of the PROUD-PV and CONTINUATION-PV studies, which used pegylated IFN.⁶

Table 1.Clinical features of the interferon and hydroxyurea cohorts at baseline.

Figure 1.Comparison of hematologic and molecular responses between the interferon and hydroxyurea cohorts. (A) Hematologic responses, (B) complete hematologic remission (CHR) rates over time, (C) molecular responses, and (D) dynamics of JAK2 V617F variant allele frequencies (VAF) over time are compared between the interferon (IFN) and hydroxyuera (HU) cohorts. In (D) the horizontal lines indicate median values; bars represent minimum and maximum values; boxes represent values included between the 25% and 75% percentiles. (E) JAK2v V617F VAF waterfall plot in the IFN (n=22) and HU (n=31) cohorts; the y-axis indicates the absolute change of the JAK2 V617F VAF from baseline to the best molecular response; each bar represents a patient; dotted lines represent median changes of the JAK2 V617F VAF in each group. IFN: conventional (non-pegylated) interferon; HCT: hematocrit; PLT: platelet; WBC: white blood cell; PMR: partial molecular response; Pts: patients. *P<0.05; **P<0.01; ***P<0.001.

In addition to hematologic responses, the IFN cohort also showed better molecular responses than the HU cohort. In total, 31 and 40 patients had data regarding molecular responses in the IFN and HU cohorts, respectively. The median JAK2 V617F VAF at baseline were not significantly different between the IFN and HU cohorts (68% [IQR, 51–78%] vs. 62% [IQR, 40–70%]; P=0.21). Only one patient in the IFN cohort achieved CMR. The percentage of patients who obtained a JAK2 V617F VAF <10% was higher in the IFN (65%) than in the HU (33%) cohort (P=0.007; Figure 1C). Among patients with baseline JAK2 VAFs ≥20%, 95% (19/20) achieved PMR in the IFN and 59% (17/29) in the HU cohorts (P=0.007; Figure 1C). The median change in JAK2 V617F VAF from baseline to the best molecular response in the IFN and HU cohorts was −58% (IQR, −69% to −34%) and −30% (IQR, −51% to –0.4%) (P=0.001; Figure 1E). Finally, the JAK2 V617F VAF in the IFN cohort was significantly lower than in the HU cohort after 3 years of continuous treatment (Figure 1D).

Because the IFN cohort was younger than the HU cohort, we compared treatment responses between patients in the IFN and HU cohorts matched for age and sex. The baseline peripheral blood counts, JAK2 V617F allele burdens, follow-ups, and thrombosis risk stratifications were balanced between the two matched cohorts (Online Supplementary Figure S2A). The CHR rate (66% [44/67] vs. 34% [23/67]; P=0.001; Online Supplementary Figure S2B), control of HCT rate (73% [49/67] vs. 54% [36/67]; P=0.03; Online Supplementary Figure S2B), and PMR rate (95% [18/19] vs. 62% [8/13]; P=0.029; Online Supplementary Figure S2C) were significantly higher in the IFN cohort than in the HU cohort when matched for age and sex. When patients were stratified by thrombosisrisk, the CHR rate in the IFN cohort was higher than in the HU cohort for age- and sex-matched low-risk (63% [24/38] vs. 26% [9/35]; P=0.002) and high-risk (70% [19/27] vs. 45% [14/31]; P=0.067) patients.

Figure 2.Comparison of hematologic and molecular responses between the interferon and hydroxyurea cohorts stratified by thrombosis risk. Hematologic (A) and molecular (B) responses of low-risk patients. Hematologic (C) and molecular (D) responses of high-risk patients. IFN: interferon; HU: hydroxyurea; RBC: red blood cell; WBC: white blood cell; HCT: hematocrit; PLT: platelet; VAF: variant allele frequency; IQR: interquartile range; *JAK2 V617F VAF in JAK2 V617F-mutated patients; CHR: complete hematologic remission; PMR: partial molecular response. *P<0.05; **P<0.01; ***P<0.001.

In total, 14 of 82 subjects (17%) discontinued IFN treatment for the following reasons: normalized peripheral blood counts (n=8, 57%), adverse effects (n=2, 14%), disease progression (n=1, 7%), and unknown reasons (n=3, 21%). Fever was the most common adverse effect of IFN, which was reported in 23% (14/62) of patients, followed by bone pain in 11% (7/62) of patients.

Post-treatment thrombotic events occurred in two (2%) and six (3%) patients in the IFN and HU cohorts. The thrombosis rates were 0.5% (95% confidence interval [CI]: 0–1.1) patients per year for the IFN cohort and 0.7% (95% CI: 0.2–1.2) for the HU cohort. These rates were much lower than those published in a previous study (2.62%; 95% CI: 2.34–2.94 patients per year).⁷ In our study, the thrombosis rate in the low-risk cohort (95% CI: 0.6% [0.2–0.9]) was lower than that of low-risk PV patients treated by phlebotomy, as reported by Barbui et al. (95% CI: 2.0% [1.5-2.5]).⁸

The lower incidence of thrombosis in this study compared with previous studies might be related to racial differences in thromboembolism between Asian and Western populations. This is related to differences in genetic polymorphisms and environmental factors, such as obesity and healthcare facilities.^9,10 For instance, a study reported that Japanese patients with paroxysmal nocturnal hemoglobinuria (PNH) had a significantly lower incidence of thrombosis than American PNH patients.¹¹ Moreover, the ECLAP study and a matched study of 951 patients with PV reported a benefit-risk profile of HU therapy over phlebotomy with respect to the lower rate of arterial thrombosis.^12,13 Our findings suggested that early intervention with cytoreductive treatments for low-risk subjects rather than phlebotomy might also correlate with lower thrombosis rates.

Thrombosis-free survival rates were not significantly different between the IFN and HU cohorts (P=0.81), similar results were found when adjusted by age (P=0.73; Online Supplementary Figure S3A). There was no significant difference in overall survival (P=0.99; Online Supplementary Figure S3B) or myelofibrosis-free survival (P=0.98; Online Supplementary Figure S3C) between the IFN and HU cohorts when adjusted by age. A previous retrospective study of PV patients reported that IFN reduced the risk of mortality and transformation into myelofibrosis compared with HU or phlebotomy.¹⁴ The different conclusions that we report might be due to the relatively short follow-up in our study. Finally, there was no significant difference in thrombosis-free survival (P=0.40), overall survival (P=0.55), or myelofibrosis-free survival (P=0.26) between patients who achieved PMR or not.

A recent meta-analysis reported that CHR rates, thrombotic complications, and treatment discontinuations owing to adverse events were not significantly different between pegylated and conventional IFN.¹⁵ In our study, conventional IFN-a 2b was a good choice for PV, showing better efficacy than HU and acceptable tolerance.

In conclusion, this study found that the hematologic and molecular responses of newly-diagnosed PV to conventional IFN-a 2b were better than to HU. There are limitations to this study, such as it being a retrospective study from a single center with a short follow-up, a mixture of low- and high-risk patients, and only a few patients who were tested for molecular responses, which are all sources of potential bias.

Footnotes

• Received September 24, 2021
• Accepted December 24, 2021

Correspondence

*DL and ZX contributed equally as co-first authors

BING LI - libing@ihcams.ac.cn

RAAJIT K RAMPAL - rampalr@mskcc.org

ZHIJIAN XIAO - zjxiao@ihcams.ac.cn

Disclosures: RPG is a consultant to BeiGene Ltd., Fusion Pharma LLC, LaJollaNanoMedical Inc., Mingsight Pharmaceuticals Inc., and CStone Pharmaceuticals; is an advisor to Antegene Biotech LLC; medical director of FFF Enterprises Inc.; is a partner of AZAC Inc.; is a member of the Board of Directors of the Russian Foundation for Cancer Research Support; and member of the Scientific Advisory Board of StemRad Ltd. RKR has received grants and personal fees from Constellation, Incyte, Celgene/BMS, and Stemline, and personal fees from Promedior, CTI, Blueprint, Jazz Pharmaceuticals, Galecto, Pharmaessentia, AbbVie, and Novartis. All other authors declare no conflicts of interest.

Contributions: ZJX designed the study; DL and ZFX collected and interpreted the data and performed statistical analysis; PHZ and QS contributed to analysis of bone marrow histology; TJQ, SQQ, LJP, WYC, JQL, HJW, XJS, MJ, and QYG contributed to recruiting patients and collecting data; DL wrote the manuscript with contributions from ZJX, ZFX, BL, GH, RPG, RKR, ZXS, and HJH. All authors reviewed the manuscript during its preparation and approved the final version of the manuscript.

Funding

this study was supported in part by National Natural Science Funds (No. 81530008, 81870104, and 82070134), Tianjin Natural Science Funds (18JCZDJC34900, 16JCQNJC11400, and 19JCQNJC09400), and CAMS Initiative Fund for Medical Sciences (No. 2016-I2M-1-001, 2020-I2M-C&T-A-020, and 2020-I2M-C&T-B-090).