Prevalence and clinical outcomes of polycythemia vera and essential thrombocythemia with hydroxyurea resistance or intolerance

ABSTRACT Introduction To determine the prevalence, clinical outcomes, and factors associated with hydroxyurea (HU) resistance or intolerance among polycythemia vera (PV) and essential thrombocythemia (ET) patients. Methods This study was a retrospective cohort study including PV and ET patients diagnosed by WHO criteria and treated with HU between January 2000 and June 2020. Clinical features, laboratory data, and resistance or intolerance of HU were collected. The prevalence, clinical outcomes, and associated factors of HU resistance or intolerance were analyzed. Results There were 260 patients including 144 ET and 116 PV. The prevalence of HU resistance or intolerance was 11.9% (31 patients) which was more frequent in ET patients (14.6% vs. 8.6% in PV). Patients who had HU resistance or intolerance significantly increased the risk of bleeding events (HR 2.64; 95% CI 1.19–5.85, P = 0.017). The risk factors of HU resistance or intolerance were low baseline hemoglobin levels (HR 0.90; 95%CI 0.84–0.97, P = 0.01), age more than 60 years old (HR 3.98; 95% CI 2.08–7.62, P < 0.001) and splenomegaly (HR 2.08; 95% CI 1.03–4.21, P = 0.04). Conclusions The prevalence of HU resistance or intolerance in PV and ET patients was 11.9%. Patients with HU resistance or intolerance significantly increased the risk of bleeding complications.


Introduction
Polycythemia vera (PV) and essential thrombocythemia (ET) are myeloproliferative neoplasms (MPNs) characterized by proliferation of myeloid lineages due to driver gene mutations including Janus Kinase 2 (JAK2), Calreticulin (CALR), and Myeloproliferative Leukemia (MPL). PV is characterized by major proliferation of red blood cells whereas ET is a major proliferation of platelets [1,2]. Both PV and ET are gradually progressive diseases resulting in complications including thrombotic events, bleeding events, increasing the risk of developing myelofibrosis, and transformation to acute myeloid leukemia (AML) leading to the decreased survival time [1][2][3][4][5].
The major aim of the treatment of PV and ET is for decreasing the risk of thrombosis and bleeding. Hydroxyurea (HU) is a cytoreductive agent that is widely used as first-line therapy for high-risk patients who are more than 60 years old or have a history of thrombosis [3,6]. Moreover, HU is recommended in PV or ET patients who cannot tolerate frequent phlebotomy (for PV), progressive splenomegaly, platelet count higher than 1500 × 10 9 /L, or white blood cell (WBC) count higher than 15 × 10 9 /L [3]. However, some patients resistant or intolerant to HU defined by achieving one of the European LeukemiaNet (ELN) [7,8] and modified European Leukemia Net (ELNm) [9,10] (for PV) criteria. The causes of HU intolerance including hematological and non-hematological side effects such as leg ulcers or mucocutaneous toxicities [9,10]. The study from Belgium showed the prevalence of HU resistance or intolerance in PV was 20.7% and 39.6% according to ELN and ELNm criteria, respectively [10]. Besides, the resistance to HU in PV was associated with a higher rate of AML and myelofibrosis transformation as well as reduced survival [11]. Regarding HU resistance or intolerance in ET, the study from Spain revealed the prevalence of 19.9%, and HU resistance was also associated with decreased survival [12].
The management of HU resistance or intolerance including ruxolitinib, a JAK1/JAK2 inhibitor, showed the efficacy in PV patients [9] and was recommended as second-line therapy for PV patients with HU resistance or intolerance following the ELN guideline [3]. Other options in both PV and ET are anagrelide and interferon [3]. However, limited data about HU resistance or intolerance exists in Asia. As a result, the objectives of this study were to determine the prevalence, clinical outcomes, and factors associated with HU resistance or intolerance among PV and ET patients.

Materials and methods
This was a retrospective study. The patients diagnosed with PV or ET according to World Health Organization (WHO) [1,2] criteria and initially treated with HU as a single cytoreductive agent from 1 January 2000 to 30 June 2020 were enrolled. The patients who received other cytoreductive agents including anagrelide, ruxolitinib, and interferon in combination with HU were excluded.
Data regarding treatment with HU were analyzed including indication, duration, clinicohematologic response by ELN 2009 criteria [14], HU resistance or intolerance determined by ELN [7,8] (for PV and ET) and ELNm criteria [9,10] (for PV). Clinical outcomes included thrombosis, bleeding, AML transformation, myelofibrosis transformation, and death were recorded. Thrombosis was defined as the presence of blood clot in arterial or venous site documented by imaging studies. Bleeding was defined as any hemorrhagic event that occurred during the study period. Myelofibrosis transformation was determined according to WHO criteria [1,2] of post-PV and post-ET myelofibrosis. The diagnosis of AML transformation was made if the patients met the WHO criteria of AML [1,2] after the diagnosis of PV or ET.

Statistical analysis
The data including demographic data, clinical manifestations, laboratory findings, and prevalence of PV and ET patients with HU resistance or intolerance as well as clinical outcomes were analyzed by descriptive analysis.
The comparison of clinical manifestations and outcomes in patients with or without HU resistance or intolerance were analyzed by Chi-square or t-test for categorical data and non-categorical data respectively. Survival outcomes were measured through the Kaplan-Meier method. The factors associated with resistance or intolerance to HU including demographic data, disease, signs and symptoms, risk stratification, hematological parameters, and JAK2V617F mutation status were analyzed by univariate and multivariable analysis using logistic regression analysis with a p-value less than 0.05 meaning significant. All data were analyzed using the SPSS version 22.0 (SPSS Inc., IBM Corp, Armonk, NY, USA; 2013).
The clinical manifestations of PV and ET patients were mostly asymptomatic (33.0% in PV and 39.9% in ET). Thrombotic events were found in 20.7% and 11.8% of PV and ET patients, respectively. In contrast, the minority of patients presented with bleeding symptoms (6.9% of PV and 9.7% of ET). Splenomegaly was found in 18.1% and 11.1% of PV and ET patients, respectively.
Laboratory data showed that PV patients had mean Hb higher than ET patients (17.0 ± 3.9 g/dL and 12.1 ± 2.3 g/dL respectively).Twenty-eight PV patients (24.1%) had concurrent iron deficiency (23 patients) or thalassemia (5 patients) and resulted in Hb less than 16 g/dL but their bone marrow biopsies were compatible with PV. On the contrary, the mean platelet count in ET patients was higher than PV patients (1078 × 10 9 /L and 669 × 10 9 /L respectively). JAK2V617F mutation was detected in 90.8% and 60.9% of PV and ET patients, respectively. JAK2 exon 12 mutaion analysis was not available in our institute and the diagnosis of JAK2V617F-negative PV was based on bone marrow findings and serum erythropoietin level. CALR mutation was tested in a subgroup of JAK2V617F-negative ET patients and was found in 14 from 30 patients (46.7%). Only two patients (1 PV and 1 ET) were analyzed for MPL mutation and results were negative. All patients underwent bone marrow aspiration and biopsy for the diagnosis.
Assessment of risk stratification showed that 62.9% of PV and 53.5% of ET patients were stratified as high risk according to conventional risk stratification (age more than 60 years old or have a history of thrombosis). Similarly, most PV patients were classified as high risk according to the International Study of PV (57.8%). In ET patients, 42.4%, 47.2%, and 10.4% of them were classified as high, intermediate, and low risk, respectively according to IPSET survival. The majority of ET (61.5%) patients are stratified as high risk according to IPSET thrombosis.
Response to HU therapy, HU resistance or intolerance, and risk factors All patients received HU as first-line therapy according to national guideline. Median duration of HU treatment was 45 months (range 1-151 months). The majority of patients received HU at stable dose of 500 mg/day (101 patients, 38.8%) and 1000 mg/day (72 patients, 27.7%) with range from 250 to 3000 mg/day. Regarding the response of treatment with HU, approximately half of the patients achieved a complete response (50.9% of PV and 54.2% of ET).
The partial responses were achieved in 43.1% and 36.1% for PV and ET, respectively. (Table 1) Thirty-one from 260 patients with HU resistance or intolerance resulted in the prevalence of 11.9%. Patients who had HU resistance or intolerance included 8.6% of PV (10 from 116) and 14.6% of ET (21 from 144). For the PV group, two-thirds (6 patients) met the criteria of HU intolerance (3 hematologic side effects and 3 nonhematological toxicities), and one-third (4 patients) had HU resistance due to leukocytosis or thrombocytosis according to both ELN and modified ELN criteria. All of the ET groups met the criteria of HU intolerance with 12 patients (57.1%) met the criterion for anemia, two patients (9.5%) for leukopenia, and seven patients (33.3%) had non-hematological toxicities. (Table 2) The univariate analysis of risk factors of HU resistance or intolerance in PV and ET was shown in Table 3. The multivariable analysis of risk factors of HU resistance or intolerance in PV and ET were low baseline Hb levels (HR 0.90; 95%CI 0.84-0.97, P = 0.01), age > 60 years (HR 3.98; 95% CI 2.08-7.62, P < 0.001) and splenomegaly (HR 2.08; 95% CI 1.03-4.21, P = 0.04).

Clinical outcomes of PV and ET patients and risk factors
Regarding clinical outcomes (Table 4), the analysis about survival was not applied due to 107 patients (41.2%) were loss to follow-up and only 3 death (1 PV, 2 ET with 1 HU intolerance) patients were documented. The causes of death were sepsis in 2 ET patients and unknown in 1 PV patient. The multivariable analysis of risk factors of thrombosis in PV and ET revealed no significant risk factor. However multivariable analysis revealed that patients with HU resistance or intolerance increased risk of bleeding (HR 2.64; 95% CI 1.19-5.85, P = 0.017). Myelofibrosis and AML transformation was found in 1.7%and 2.1%, respectively. The multivariable analysis showed that patients with splenomegaly increased both myelofibrosis (HR 8.54; 95% CI 1.42-51.12, P = 0.019) and AML transformation (HR 17.49; 95% CI 2.45-124.82, P = 0.004).

Discussion
This study revealed the prevalence, clinical outcomes, and factors associated with HU resistance or intolerance among PV and ET patients in Asia. The prevalence of HU resistance or intolerance among PV and ET patients was 11.9%. PV patients with HU resistance or intolerance were found to only 8.6% which was lower than the study from Belgium [10] and Spain [15] that reported the prevalence of 20.7% and 15.4%, respectively. All studies showed a lower proportion of PV patients with HU resistance than HU intolerance (50% in this study, 29.4% in the Belgium study, and 34.7% in the Spanish study). Among ET patients, 14.6% had HU resistance or intolerance for which was comparable with the Spanish study [12] which found the prevalence of 19.9% (33 from 166 patients). Similarly, the majority of cases had HU intolerance and the proportion was higher than PV (100% in this study and 63.6% in Spanish study). In contrast, another study from Italy revealed only 1.7% of ET patients were withdrawn from HU due to side effects whereas 17% had no clinical-hematologic response after at least 12 months of HU treatment [16].
The causes of HU intolerance in this study were different from a study from European countries. This study showed that anemia was the most common criterion that led to HU intolerance in ET patients (57.1%)  whereas the Spanish study revealed leg ulcers and mucocutaneous toxicity were the most common reason (60.6%) [12]. The study of PV patients from Belgium [10] and Spain [11] showed leg ulcers and mucocutaneous manifestations occurring in 13.2% and 8.9% and accounted for 82.4% and 84.0% of HU intolerance, respectively. A large cohort of MPN patients from Italy also showed the prevalence of mucocutaneous toxicity of 5.4-8.3% including ulcers, precancerous lesions, and dermatitis [17,18]. These non-hematologic adverse events were reported by only 3.8% in this study and suggested the slightly lower prevalence of these toxicities in Asian countries. The multivariable analysis of risk factors of HU resistance or intolerance showed low baseline Hb level, age more than 60 years, and splenomegaly significantly associated with HU resistance or intolerance. Anemia is a common hematologic adverse event of HU with an incidence of 17-18.3% in randomized controlled trials in high-risk ET [19,20]. Since anemia (Hb < 10 g/ dL) was included in ELN criteria of HU intolerance in both PV [8] and ET [9], patients who had baseline lower Hb before treatment with HU might have a higher risk of developing HU intolerance. This selfreinforced association is one of the limitation in this study. Older patients (age more than 60 years old) had a risk of developing HU resistance or intolerance more than younger patients (HR 3.98). This finding was similar to a Spanish study in PV that revealed a more frequent and faster rate of HU resistance or intolerance as well as non-hematological toxicities in patients aged >60 years [15]. However, there were conflicting data in ET. Older age (more than 70 years) was reported as a factor associated with good clinicohematologic response to HU in ET patients but the association with HU intolerance was not reported [12]. Another study reported tolerance to cytotoxic therapy was comparable between ET patients aged 60-74 years and ≥ 75 years [21]. Finally, the patients who had splenomegaly increased the risk of HU resistance or intolerance (HR 2.08). A Spanish study in PV found that palpable spleen at diagnosis was associated with the development of cytopenias [11]. In addition, this group of patients might have more degrees of myeloproliferative and tended to be associated with poor clinico-hematologic response to HU in ET patients [12].
This study showed that bleeding symptoms significantly occurred more in patients with HU resistance or intolerance 2.64 times (P = 0.017) and supported data from previous studies that this group of patients had a higher risk of complications of MPN. These bleeding events might be explained by the presence acquired von-Willebrand disease due to thrombocytosis in patients with HU resistance, thrombocytopenia in cases of HU intolerance, as well as effects of antithrombotic therapies [1,3]. One study also demonstrated that response in platelet count (<400 × 10 9 /L) was associated with a lower incidence of bleeding [11]. Since testing for acquired von Willebrand syndrome was not done routinely in cases with extreme thrombocytosis in our institute, so the author could not clarify the cause of bleeding in this study.
Although survival data could not be analyzed in this study. HU resistance but not intolerance was previously reported as adverse factor for survival (HR 5.6; 95%CI, 2.7-11.9; P < 0.001) and transformation to AML or myelofibrosis (HR 6.8; 95% CI, 3.0-15.4; P < 0.001) in PV patients [11]. Similarly, a previous study showed ET patients with HU resistance but not intolerance had a higher risk of death (HR 6.2, 95% CI 2.3-16.7, P < 0.001) [12]. Both studies [11,12], as well as the current study, did not show an association between HU resistance or intolerance and risk of thrombosis.
The treatment guideline of PV with HU resistance or intolerance was changed [3] after ruxolitinib demonstrated efficacy in this setting in randomized-controlled trials [9,22,23]. However, the outcomes of ruxolitinib in ET patients with HU resistance or intolerance were not different from best-available therapy [24]. Pegylated interferon alfa-2a could also induce hematologic response as well as molecular response in Phase II study of both PV and ET with HU resistance or intolerance [25] and was recommended as an alternative treatment option [3]. Anagrelide in combination with low-dose hydroxyurea was also shown efficacy in platelet control in both ET and PV patients [26]. This study could not provide the data about the efficacy of these medications in HU resistance or intolerance due to a limited number of patients who could access novel agents.
The limitation of this study was it was a retrospective cohort study that could not provide complete data.The diagnosis of PV and ET were based on either 2008 WHO or 2016 WHO Classification depended on the time periods. As a result, some pre-fibrotic primary myelofibrosis patients might be classified as ET and were enrolled in this study. In addition, some patients were not evaluated for gene mutations so that this study could not precisely analyze the associations between gene mutations and HU resistance or intolerance. Only JAK2V617F mutation was analyzed but did not show a significant impact on HU resistance or intolerance. This study also did not show the impact of CALR mutation on HU resistance or intolerance in ET as the previous study [27] due to CALR mutation was tested only in subgroup of JAK2V617F-negative ET patients. The association between genomic characterization by next-generation sequencing and HU resistance in PV patients was evaluated in the MPN Spanish Group [28]. The data revealed patients with TP53 disruption/aneuploidy, spliceosome or chromatic aberrations, homozygous and heterozygous JAK2 mutation had a different risk of developing HU resistance. Moreover, genomic classification was associated with the risk of resistance to HU (HR 2.2; 95% CI 1.5-3.2, P < 0.0001). This aspect should be investigated for ET patients in the future.

Conclusions
The prevalence of HU resistance or intolerance among PV and ET patients was 11.9%. Patients with HU resistance or intolerance increased the risk of bleeding. Factors that influenced the occurrence of HU resistance or intolerance among PV and ET patients are low baseline Hb levels, age more than 60 years old, and splenomegaly.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
The author(s) reported there is no funding associated with the work featured in this article.