Significant association of cutaneous adverse events with hydroxyurea: results from a prospective non-interventional study in BCR-ABL1-negative myeloproliferative neoplasms (MPN) - on behalf of the German Study Group-MPN

Until today, hydroxyurea (HU) remains the most frequently used cytoreductive drug for long-term treatment of patients (pts) with classical BCR-ABL1-negative myeloproliferative neoplasms (MPN) [1]. Since HU is known to be associated with the occurrence of ulcers in a dose dependent manner, leg ulcers and mucocutaneous manifestations are defined criteria for HU intolerance by the European LeukemiaNet (ELN) [2–5]. However, the spectrum of cutaneous adverse events (CAE) caused by HU is much broader and ranges from skin dryness to malignant lesions like squamous cell carcinoma [6]. Since retrospective studies reporting HU associated CAE in 5–10% of pts may underestimate the frequency and the clinical relevance of such symptoms, we initiated a non-interventional trial focusing on prospective observation of CAE associated with different cytoreductive drugs used in routine MPN management [7–9]. We included pts with polycythemia vera (PV), essential thrombocythemia (ET), and primary or secondary myelofibrosis (MF) regularly presenting at our MPN outpatient clinics in Ulm and Minden (Germany), respectively. All pts included were diagnosed according to WHO criteria and treated with cytoreductive drugs according to ELN recommendations [10, 11]. Most frequently used agents were HU, anagrelide (ANA), conventional or pegylated interferon-alpha (IFN), and ruxolitinib (RUX). The institutional review board of each center approved the study. Prospective follow-up time was defined as the interval from giving informed consent to last visit. All data were collected in one database. Enrollment started at 09-Feb-2011 and date of last follow-up ( ́data cut-off ́) was 09-Jan-2018. In total, 172 MPN pts have been evaluated. Medical history was carried out by a set of specific questions at study entry and at every three-monthly visit. Possible relation between CAE and cytoreduction was assessed by the physician. One treatment course (TC) was defined as use of one cytoreductive treatment for at least one month. Combination therapy consisting of two drugs for longer than one month each was considered as two TC. The annual incidence of MPN-therapy associated CAE was calculated by dividing the number of events by the total number of patient-years. Differences in the proportions were estimated using chi square test. Skin alteration-free A full list of board members of the German Study Group-MPN (GSGMPN) and their affiliations appears in the Supplementary Information.


To the Editor:
Until today, hydroxyurea (HU) remains the most frequently used cytoreductive drug for long-term treatment of patients (pts) with classical BCR-ABL1-negative myeloproliferative neoplasms (MPN) [1]. Since HU is known to be associated with the occurrence of ulcers in a dose dependent manner, leg ulcers and mucocutaneous manifestations are defined criteria for HU intolerance by the European LeukemiaNet (ELN) [2][3][4][5]. However, the spectrum of cutaneous adverse events (CAE) caused by HU is much broader and ranges from skin dryness to malignant lesions like squamous cell carcinoma [6]. Since retrospective studies reporting HU associated CAE in 5-10% of pts may underestimate the frequency and the clinical relevance of such symptoms, we initiated a non-interventional trial focusing on prospective observation of CAE associated with different cytoreductive drugs used in routine MPN management [7][8][9].
We included pts with polycythemia vera (PV), essential thrombocythemia (ET), and primary or secondary myelofibrosis (MF) regularly presenting at our MPN outpatient clinics in Ulm and Minden (Germany), respectively. All pts included were diagnosed according to WHO criteria and treated with cytoreductive drugs according to ELN recommendations [10,11]. Most frequently used agents were HU, anagrelide (ANA), conventional or pegylated interferon-alpha (IFN), and ruxolitinib (RUX). The institutional review board of each center approved the study. Prospective follow-up time was defined as the interval from giving informed consent to last visit. All data were collected in one database. Enrollment started at 09-Feb-2011 and date of last follow-up (´data cut-off´) was 09-Jan-2018. In total, 172 MPN pts have been evaluated.
Medical history was carried out by a set of specific questions at study entry and at every three-monthly visit. Possible relation between CAE and cytoreduction was assessed by the physician. One treatment course (TC) was defined as use of one cytoreductive treatment for at least one month. Combination therapy consisting of two drugs for longer than one month each was considered as two TC. The annual incidence of MPN-therapy associated CAE was calculated by dividing the number of events by the total number of patient-years. Differences in the proportions were estimated using chi square test. Skin alteration-free times between treatment groups were compared via logrank test.
Regarding non-HU TC (i.e. RUX, ANA, and IFN), CAE under RUX occurred in 2/66 TC (3.0%) after a median time of 4.5 months. One pt developed a pityriasis and one pt had erythromelalgia under a daily RUX dosage of 10 mg. Neither of the two RUX associated CAE was leading to discontinuation of the drug. In contrast, no CAE occurred under ANA during the prospective study time, while CAE under IFN occurred in 3/24 (11%) TC after a median treatment time of 11 years (range, 3.8-17.9). Of these, two pts developed increased photosensitivity and one pt suffered from rash after subcutaneous application leading to IFN discontinuation.
Finally, we analyzed the CAE-free survival over time during the prospective part of the study. Compared by log-rank test, a significant difference between HU and non-HU TC regarding the occurrence of CAE was observed (p = 0.0076; Fig. 1d). This was also true for the complete study period including the retrospective observation time (p = 0.0036).
In summary, the major findings of this observational study were (i) the significant association of HU with the occurrence of CAE, (ii) an overall CAE incidence twice as high when data were evaluated prospectively compared to retrospective analysis, suggesting that CAE associated with HU may be underdiagnosed and (iii) a discontinuation rate of~20% for HU treatment due to CAE. However, it has to be taken into account that HU was the most frequent cytoreductive agent used in the pts accounting for approximately one half of all TC in both study parts.
In line with our retrospective observation part, Antonioli et al. reported a CAE rate of 4.9% under HU in a large retrospective study including more than 3400 MPN pts [9]. Here, CAE occurred after a cumulative HU dosage of about 1400 g and after a median treatment time of 60 month. Moreover, in one single-center study with prospective observation the incidence of HU associated CAE was even 60% after a median time of 3.6 years [12].
Among CAE associated with HU, ulcerous lesions (25%) and skin rashes (23%) occurred most frequently, while  Basal cell carcinoma (n = 3), squamous cell carcinoma (n = 1). c HU: skin rashes (n = 14), skin dryness (n = 8), oral stomatitis (n = 4), nail changes (n = 2), erythema of the foot, psoriasis plantaris pustolosa, and livid-colored leg (n = 1, each). precancerous lesions and local skin cancers accounted for 17% and 7% of events, respectively. To prevent the occurrence of CAE under HU, we strongly recommend to (i) inform pts about possible CAE before treatment start, (ii) examine pts thoroughly before and during therapy, and (iii) advise pts about adequate skin care such as UV light protection and regular dermatologic controls.
Taken together, our non-interventional study shows a significant association between HU treatment and the development of skin toxicity, in particular with regard to ulcerous and (pre-)malignant lesions. CAE occurred after a median treatment time of 4.5 years, with a cumulative median HU dosage of 1533 g. Consequently, physicians and pts need to be informed about possible side-effects of the drug on skin and mucosa to prevent that CAE become a limiting factor for the long-term use of HU in MPN management.