Anagrelide Is Effective in Treating Patients with Hydroxyurea-Resistant Thrombocytosis in Patients with Chronic Myeloid Leukemia

Home , Anagrelide

Leukemia (2005) 19, 39–43 & 2005 Nature Publishing Group All rights reserved 0887-6924/05 $30.00 www.nature.com/leu Anagrelide is effective in treating patients with hydroxyurea-resistant thrombocytosis in patients with chronic myeloid leukemia

RT Silver1

1The Leukemia and Myeloproliferative Disease Center, Division of Hematology-Oncology, Weill Medical College of Cornell University, New York, NY, USA

We report phase II trial results of the use of oral anagrelide median platelet count was approximately 2.0 million/ml; seven hydrochloride for treating 38 patients with hydroxyurea (HU)- patients had thrombohemorrhagic complications. All patients resistant thrombocytosis accompanying chronic myeloid leukemia (CML). Anagrelide’s efficacy was well established during responded to anagrelide. The median platelet count after a phase II study of more than 400 patients with one of the four treatment was 343 000/ml; thrombohemorrhagic complications myeloproliferative disorders: essential thrombocythemia, poly- disappeared or did not recur in all patients previously affected. cythemia, idiopathic myelofibrosis, and CML. In the last As part of a large phase II efficacy and safety trial of anagrelide subgroup, there were 114 CML patients with significant in treating patients with elevated platelet counts in the thrombocytosis treated with anagrelide. Out of these patients, myeloproliferative diseases, the Anagrelide Study Group studied 38 had symptoms of thrombosis or hemorrhage and had a group of 114 patients with CML and thrombocytosis thrombocytosis resistant to HU. They were then treated with 2,12 anagrelide at an initial dose of 2.0 mg/day, followed by (Figure 1). The majority of the patients had been given modifications based upon response and toxicity. In all, 71% anagrelide solely because the platelet count was elevated of these patients responded with platelet reductions of more according to study criteria. However, within this group, we than 50% in a median time of approximately 4 weeks. The subsequently identified a subset of 38 patients who had been response rate was not influenced by age, gender, or prior given HU prior to anagrelide specifically for the thrombohemor- thrombosis or hemorrhage. Importantly, the response rate to anagrelide in patients refractory to prior HU was essentially the rhagic complications accompanying thrombocytosis. same as that of the other 76 CML patients. Treatment with This report emphasizes the rare complications associated with anagrelide was well tolerated and without undue toxicity. thrombocytosis in CML and its treatment with anagrelide, Reduction of excessive platelet counts by anagrelide some- especially in patients not satisfactorily treated with HU that times occurring in CML may lead to the prevention of had been given for increased blood platelets. As in other thrombohemorrhagic complications occurring in this clinical myeloproliferative diseases, it suggests that aggressive platelet setting and is relevant even in those patients in whom imatinib mesylate is primary therapy. control aimed at reducing the number of excess blood platelets Leukemia (2005) 19, 39–43. doi:10.1038/sj.leu.2403556 with anagrelide is a treatment objective for CML patients. Published online 28 October 2004 Keywords: CML; thrombocytosis; anagrelide Materials and methods

The diagnosis of CML required the typical phenotypic clinical and hematologic picture and demonstration of the Philadelphia Introduction chromosome in the bone marrow in all cases. Molecular studies for rearrangement of the BCR–ABL oncogene and fluorescent in Chronic myeloid leukemia (CML) is uniquely distinguished from situ hybridization studies were not routinely performed in all the three other common myeloproliferative disorders, essential patients at the time of the initiation of this study. Other criteria thrombocythemia (ET), polycythemia vera (PV), and idiopathic included a white blood cell count of X60 000/ml, granulocytic myelofibrosis (IMF), by the Philadelphia chromosome and/or by hyperplasia of the marrow, and decreased or absent leukocyte the molecular rearrangement yielding the chimeric BCR-ABL alkaline phosphatase score. Patients were 18 years or older and 1,2 oncogene. Thrombocytosis is often seen in CML patients, but gave written informed consent. Children were treated as part of 2 thrombohemorrhagic complications are rarely reported. Re- a compassionate-use program. Women of childbearing age had duction in levels of the white blood cell and platelet counts in to have a negative pregnancy test. Failure of a prior therapy was patients treated with interferon has been implemented by using not required. HU and, less commonly, busulfan. Even in the imatinib era, HU The 114 patients with CML eligible to receive anagrelide were has been routinely used when thrombocytosis has been required to have a platelet count of at least 900 000/ml on two 2 excessive. separate occasions at least 1 month apart or a count of at least It is well known that increased platelet counts are associated 650 000/ml on two successive determinations and documenta- 3–5 with an increased risk of microvascular occlusion, thrombo- tion of symptoms attributable to thrombocytosis. 6–9 10 sis, and bleeding. In only one study in CML has anagrelide Not all of the 114 patients were given routine prophylactic been reported to control the thrombocytosis and thrombohe- aspirin. Detailed coagulation studies for hypercoagulable states 11 morrhagic complications. Of the 12 patients in this study, the were not systematically performed. During the analysis of this group, we noted a number of patients who had been treated Correspondence: Professor RT Silver, Division of Hematology- with HU specifically for platelet control because of thrombo- Oncology, New York Presbyterian Hospital-Weill Medical College hemorrhagic symptoms or signs. In these 38 patients, the use of of Cornell University, Box 581, 525 East 68th Street, New York, NY HU was not successful for this purpose; because of excessive 10021, USA; Fax: þ 1 212 746 8246; E-mail: [email protected] thrombocytosis, anagrelide had been substituted. We therefore Received 22 December 2003; accepted 13 September 2004; reviewed the data pertaining to this subgroup to determine the Published online 28 October 2004 nature of the complications and the usefulness of anagrelide in Anagrelide treats thrombocytosis RT Silver 40 577 Chronic myeloproliferative disorders

335 114 68 60 Essential thrombocythemia Chronic myeloid leukemia Polycythemia vera Undifferentiated disease

38 Hydroxyurea-resistant

19 19 Definitely refractory to hydroxyurea Probably refractory to hydroxyurea

Figure 1 Patients in the original Phase II study for thrombocythemic states of the Anagrelide Study Group.

Table 1 Grouping of 38 patients based on their refractory response Time to platelet response was computed as the time from day to hydroxyurea 1 of anagrelide therapy until a response was ﬁrst noted. Treatment time was determined from the ﬁrst dosing date to 1. Refractory to hydroxyurea (HU): 19 the end of therapy or the last follow-up of patients still receiving 1.1 Leukopenia with thrombocytosis (platelets treatment. Observations were censored at the time of the last X600 000/ml) after 3 months of HU therapy, or dosing date of the patients still being treated.12 Note that the end 1.2 Platelet count X900 000/ml despite a dose of at least point was neither hematologic nor cytogenetic response of CML, 2 g/day for at least 2 months, or 1.3 Platelet count X2 000 000/ml after at least 2 months but rather reduction of platelet counts and associated symptoms of HU at any dose and signs contemporaneous with administration of anagrelide.

2. Probably, but not definitely, refractory to HU: 19 X Platelet count 600 000/ml for at least 2 months with any Results dose of HU and documented symptoms and signs attributed to thrombocytosis During the phase II study, 114 patients with CML with elevated platelet counts were treated with anagrelide. For their leukemia, the patients had been previously treated with HU, interferon, this particular situation. In the majority of cases, HU had been busulfan, 32P, chlorambucil, thiotepa, cytosine arabinoside, given as a single agent; in the minority, it supplemented another nitrogen mustard, and/or alkeran as antileukemic therapy. drug used for treating the leukemia. Overall, more than 70% of the 114 patients in chronic phase For the purpose of analysis, this subset of HU-resistant disease had received HU. As they had thrombocytosis as patients was divided into two groups (Table 1). The first group defined in the major study, they were given anagrelide as part of (n ¼ 19) consisted of those whose thrombocytosis was consid- the single-arm, open-label trial, together with antileukemic ered refractory to HU, defined as one of the following three: (1) therapy. We identified 38 patients who had been specifically leukopenia with thrombocytosis (WBC p3000/ml; platelets given HU to treat thrombocytosis, prior to starting anagrelide. X600 000/ml) after 3 months of HU therapy; or (2) a platelet The use of anagrelide to treat the thrombocytosis after HU was count X900 000/ml despite a dose of at least 2 g of HU for at the result of an inability to control the platelet count, an adverse least 2 months; or (3) a platelet count X2 000 000/ml for at least event accompanying HU, or both. 2 months of HU treatment at any dose. The second group The entire group of 114 CML patients, including the subgroup (n ¼ 19) consisted of patients who were considered probably, of 38 HU-resistant patients, had the age, sex, and race but not definitely refractory to HU, defined as a platelet count distribution typical of a CML cohort (Table 2). For the 38 median of 600 000/ml for at least 2 months with any dose of HU patients, the median age was 54.5 years; only one patient was and documented symptoms and signs attributed to thrombocy- African American. More than 90% were Caucasian, 53% were tosis. For the most part, these patients had had a temporary women. The mean baseline body weight was 70 kg (range 41– suspension of other antileukemic therapy before institution of 128 kg). All 38 patients were in chronic phase of varying HU because of the relatively impressive thrombocytosis. duration; none had accelerated or blast phase phenotypes. The treatment program consisted of anagrelide 0.5 mg 4 times Of the 38 HU-resistant patients, 28 had a history of a day. Weekly adjustments were made to achieve and then thrombosis and 10 had a history of hemorrhage (Table 3). There maintain the platelet count at p600 000/ml and ideally between was no distinct linear correlation between the platelet count and 300 000 and 400 000/ml. The average dose was 2.0–2.5 mg/day, the development of thrombohemorrhagic symptoms in this but higher doses up to 7.2 mg/day were sometimes required.12 A relatively small sample, although the higher the platelet counts, response was defined as a decrease in platelet count to the more frequent the complications. p600 000/ml or 50% of the initial value before anagrelide. All In all, 27 patients (71%) of this subset met the criteria for values had to remain constant for at least 4 weeks. Treatment response to anagrelide. After treatment there were 27 respon- failure was defined as anything less than the aforementioned. ders, but 11 patients remained symptomatic (Table 3). Following

Leukemia Anagrelide treats thrombocytosis RT Silver 41 treatment, the mean platelet levels in responders and non- noted,12 the white blood cell count was not affected by responders were 250 0007360 400/ml (Table 4). In one-third of anagrelide during the observation period. the responders, the initial platelet count was reduced by 50%. At The symptoms of the group of patients with thrombosis 6–8 weeks, the median platelet count in two-thirds of the included transient ischemic attacks, myocardial infarction, responders was less than 600 000/ml. The median time to best erythromelalgia, deep venous thrombosis, and ischemia with response on both subgroups was 7.1 weeks. The time to or without cutaneous ulceration of the extremities. Ten patients response and response rate were the same in the thrombotic and had either transient ischemic attacks (9) or a myocardial hemorrhagic groups. The mean daily dose of anagrelide was 2– infarction (1). In the patient with myocardial infarction, it is 2.5 mg/day. Response was not correlated with either early- or not now known whether other cardiovascular risk factors existed late-stage chronic phase disease, sex, age, race, or splenome- or played a contributory role. The patient had serum cholesterol galy. The median platelet count prior to treatment with of 236 mg/dl. Lipoprotein values are not known. Transient anagrelide was 1 237 000/ml, the mean 1 354 000/ml (ran- ischemic attacks did not recur in seven of the nine patients, nor geE600 000–6 300 000/ml). The responders maintained their did the patient with the myocardial infarction suffer a recurrence counts for a median of 7 weeks and as long as 8 months; after the platelet counts were lowered to p600 000/ml, at least thereafter, the platelet counts in each patient were affected by during the observation period of the study. In two patients, change in censored status of CML to accelerated or blast phase symptoms persisted without change in platelet count. Additional disease (four patients), by alternative chemotherapy for CML (21 dose escalation of anagrelide was not documented. patients), marrow transplantation (eight patients), and by refusal Similarly, the symptoms of the eight patients with erythrome- of a physician to complete the paperwork (one patient). During lalgia completely abated upon reduction of the platelet count the anagrelide trial, the median white blood cell count remained (the use of low-dose aspirin was not documented). constant at about a value of 30 000/ml, but the median No associated malignancies were found in the patients with hemoglobin values fell from 12.0 to 9.5 g/dl. As previously deep venous thrombosis. Usual anticoagulation as well as platelet reduction was employed. Only one of four patients with ischemic ulcers had improvement with dermatologic care and Table 2 Demographic characteristics of 38 CML patients anagrelide as the platelet counts were reduced towards normal. The possible contribution of HU in the pathogenesis of the ulcers was appreciated,13 but a speciﬁc relation between ulcers, Characteristic Number of patients (N ¼ 38) and change in platelet count and HU was not established. Age (years) Hemorrhage took the form of ecchymoses, gastrointestinal p30 4 hemorrhage, or epistaxis (Table 3). In three of six patients, 31–40 4 ecchymoses diminished clinically with decrease in the platelet 41–50 4 51–60 12 X61 14 Mean (years) 58 Median (years) 54.5 Table 4 Platelet counts before and after anagrelide in HU-resistant Range (years) 12–90 population

Sex Baseline (/ml) On-study (/ml) Male 47 (%) Female 53 (%) Mean 1 354 000 250 000 Standard deviation 7918 000 7360 400 Race Minimum 526 000 21 000 African American 1 Maximum 6 264 000 2 435 000 Caucasian 37 No. of patients ¼ 38.

Table 3 Thrombohemorrhagic events in 38 HU-resistant patients before and after anagrelide treatment

Symptom Number of patients Symptomatic patients (no.) Response

Before anagrelide treatment After anagrelide treatment Patients (no.)

Thrombosis 28 28 5 23 TIA 9 9 2 7 Erythromelalgia 8 8 0 8 Deep venous thrombosis 6 6 0 6 Distal ischemia/ulcer 4 4 3 1 Myocardial infarction 1 1 0 1

Hemorrhage 10 10 6 4 Ecchymoses, easy bruising 6 3 3 3 Gastrointestinal bleeding 3 3 2 1 Epistaxis, persistent 1 1 1 0

Total 38 38 11 27 (71%) Median platelet count 38 1 237 000/ml o600 000/ml (in 2/3 of patients)

Leukemia Anagrelide treats thrombocytosis RT Silver 42 Table 5 Number of patients (N ¼ 38) with toxicity attributed to Limitations to this retrospective study also include the fact a anagrelide that contributory factors, such as hypercholesterolemia and/or hypercoagulable states, were not studied systematically, Side effects N % nor was the dosing of aspirin consistent. Furthermore, anagrelide was not unequivocally successful in all patients. Any 26 (69) No difference could be appreciated in this analysis between Gastrointestinal 14 (38) Headache, dizziness 11 (30) those responding to anagrelide or not. This retrospective review did not disclose any significant difference based upon Cardiopulmonary 10 (26) early or late stage of chronic disease, sex, WBC, or initial Palpitations 6 (16) platelet count. No patient was treated with anagrelide in Fluid retention 3 (8) accelerated or blast phase disease. Nevertheless, despite the Chest pain, dyspnea 1 (4) limitations of this study, amelioration of symptoms and signs Musculoskeletal aches 7 (18) of thrombosis or hemorrhage clearly coincided with the Requiring cessation of anagrelide 3 (7) decrease in platelet count towards normal after anagrelide therapy was begun. aSome patients had more than one type of toxic manifestation. Anagrelide appeared more effective in treating the thrombotic rather than the hemorrhagic complications of thrombocytosis of CML. Studies of acquired von Willebrand’s disease should be examined in future in similar situations in patients with CML count. The three patients with gastrointestinal bleeding had a with increased platelet counts and bleeding.14 thorough evaluation to exclude other causes of contribution to The response criteria stipulated at least 4 weeks of platelet bleeding; none was found. The bleeding was characterized by control, and for responding patients an attempt was made to melena and stopped within 3 days. In only one of three patients sustain counts below 600 000/ml. The median time to response was there a decrease in platelet count in temporal relation to of our CML patients of approximately 7 weeks was quite similar cessation of bleeding. The patient with epistaxis had packing of to the results of Fruchtman et al15 and those of Tsimberidou the nasal passages in addition to treatment with anagrelide. In et al.16 The patients of Tsimberidou et al were treated with both this patient, the platelet count was not affected by anagrelide anagrelide and imatinib mesylate. They suggested an indepen- and cessation of hemorrhage was attributed to local measures. dent mechanism of action of anagrelide on the thrombocytosis of CML.17 Thus, anagrelide may have a unique application for treating the thrombocytosis of CML not shared by other Toxicity attributed to anagrelide chemotherapeutic agents. Our response rate of 71% is also consistent with the series of Fruchtman et al.15 In a large group Approximately two-thirds of the patients had side effects of patients with myeloproliferative diseases, excluding essential secondary to anagrelide, and three required cessation of thrombocythemia, the total response rate of thrombocytosis to anagrelide therapy despite reduction in dose (Table 5). This anagrelide was 75% entirely consistent with the frequency of occurred in elderly patients with a history of cardiac disease in 71% reported here.15,16 Moreover, the median time to response whom anagrelide was discontinued because of palpitations. In was 72 days, and the mean dose of anagrelide used was general, symptoms were mild or moderate in severity. These are 2.15 mg/day, consistent with our observations. shown in Table 5 and included gastrointestinal symptoms, Response duration was measured from achievement of headaches, dizziness, palpitations, and fluid retention. Dividing response to therapeutic failure or reinstitution of drugs for the daily dose of anagrelide into multiple small ones amelio- treatment of the leukemia (which could affect the platelet rated side effects. count), whichever came first, or by other censored observations, such as marrow transplantation or by progression to accelerated or blast phase disease. For these reasons, duration of response to Discussion anagrelide was relatively short compared to studies in the other myeloproliferative diseases.12,13 Although more than half the patients with CML have platelet The introduction of imatinib mesylate in the treatment of counts above 1 million/ml, reported thrombotic or hemorrhagic chronic phase CML has resulted in a dramatic change in phenomena are infrequent.2 Nevertheless, the fact that these therapeutic results. HU, however, may be used in the initial events can occur in CML, as in other myeloproliferative phase of therapy for debulking, especially in patients with disorders, is worthy of emphasis. Elevated blood platelet counts significantly increased white blood cell counts.18 As experience in CML have been demonstrated to be responsive to anagrelide. with imatinib develops, it will be interesting to see whether such However, this is the first time that patients who have been thrombotic complications will also be noted in a larger series of refractory to HU have been shown to respond to anagrelide. The reported CML cases, especially those with high platelet counts. group of 114 CML patients was referred for anagrelide therapy To date, they have not. Nevertheless, combinations of imatinib because of abnormally high platelet counts; many of these with other agents are already being used, and in those patients patients did not have a history of thrombosis or hemorrhage. with refractory thrombocytosis anagrelide may be of consider- This subset of 38 patients resistant to HU with thrombohemor- able importance. rhagic symptoms or signs was studied retrospectively. Thus, the For patients with thrombocytosis as a manifestation of CML true figures for the prevalence and frequency of thrombotic with symptoms requiring prompt platelet reduction, anagrelide complications in CML still remain unknown. These 38 patients represents an important therapeutic option. It has a relatively had been given HU specifically for the thrombocytosis. Some high response rate, is active in patients in whom other platelet- had received other antileukemic agents both for the leukemia lowering agents may have failed, is not leukemogenic, has no and the thrombocytosis. In these 38 patients, HU was ineffective effect on white blood cells and relatively minor effects on red in controlling the thrombocytosis, not the leukemia. blood cells, is orally active, can be administered for relatively

Leukemia Anagrelide treats thrombocytosis RT Silver 43 long periods, and has a toxicity profile that is significantly 6 Barbui T, Cortelazzo S, Viero P, Bassan R, Dini E, Semeraro N. different from other agents. Thrombohaemorrhagic complications in 101 cases of myeloproliferative disorders: relationship to platelet number and function. Eur J Cancer Clin Oncol 1983; 19: 1593–1599. Conclusion 7 Fenaux P, Simon M, Caulier MT, Lai JL, Goudemand J, Bauters F. Clinical course of essential thrombocythemia in 147 cases. Cancer Based on the results of this open-label uncontrolled retro- 1990; 66: 549–556. 8 Cortelazzo S, Viero P, Finazzi G, D’Emilio A, Rodeghiero F, Barbui spective analysis, anagrelide effectively and rapidly reduces T. Incidence and risk factors for thrombotic complications in peripheral platelet counts in the majority of patients, with a historical cohort of 100 patients with essential thrombocythemia. thrombocytosis accompanying CML refractory to HU. Treat- J Clin Oncol 1990; 8: 556–562. ment with anagrelide is well tolerated without undue toxicity. 9 Wehmeier A, Daum I, Jamin H, Schneider W. Incidence and Reduction of significantly elevated platelet counts by anagrelide clinical risk factors for bleeding and thrombotic complications in in patients with CML may prevent the thrombohemorrhagic myeloproliferative disorders. A retrospective analysis of 260 patients. Ann Hematol 1991; 63: 101–106. complications that can occur in this disease. The role of 10 Colombi M, Radaelli F, Zocchi L, Maiolo AT. Thrombotic and anagrelide in the imatinib era remains to be determined. The hemorrhagic complications in essential thrombocythemia. A occurrence of thrombohemorrhagic complications in CML is retrospective study of 103 patients. Cancer 1991; 67: 2926–2930. relatively rare, but its possibility should be remembered in those 11 Trapp OM, Beykirch MK, Petrides PE. Anagrelide for treatment of patients with increased platelet counts. patients with chronic myelogenous leukemia and a high platelet count. Blood Cells Mol Dis 1998; 24: 9–13. 12 Anagrelide Study Group. Anagrelide, a therapy for thrombo- Acknowledgement cythemic states: experience in 577 patients. Am J Med 1992; 92: 69–76. 13 Best PJ, Daoud MS, Pittelkow MR, Petitt RM. Hydroxyurea- I thank Frieda Pearce, PhD, and Ms Carmen Dixon for their induced leg ulceration in 14 patients. Ann Intern Med 1998; 128: assistance in the preparation of this manuscript. This study was 29–32. supported in part by the United Leukemia Fund, Inc., and the 14 van Genderen PJ, Leenknegt H, Michiels JJ, Budde U. Acquired Cancer Research and Treatment Fund, Inc. von Willebrand disease in myeloproliferative disorders. Leuk Lymphoma 1996; 22 (Suppl 1): 79–82. 15 Fruchtman SM, Petitt RM, Gilbert HS, Fiddler G, Lyne A, References Anagrelide Study Group. Anagrelide: Analysis of long term safety and leukemogenic potential in myeloproliferative diseases (MPDs). 1 Silver RT. Chronic myeloid leukemia. Curr Opin Oncol 1992; 4: Blood 2002; 100: 70a (abstr.). 66–72. 16 Tsimberidou AM, Colburn DE, Welch MA, Cortes JE, 2 Silver RT. Chronic myeloid leukemia. Hematol Oncol Clin North Verstovsek S, O’Brien SM et al. Anagrelide and imatinib Am 2003; 17: 1159–1173. mesylate combination therapy in patients with chronic myelopro- 3 Ravandi-Kashani F, Schafer AI. Microvascular disturbances, liferative disorders. Cancer Chemother Pharmacol 2003; 52: thrombosis, and bleeding in thrombocythemia: current concepts 229–234. and perspectives. Semin Thromb Hemost 1997; 23: 479–488. 17 Fruchtman S, Pettit RM, Gilbert H. Anagrelide therapy significantly 4 Michiels JJ. Erythromelalgia and vascular complications in poly- reduces disease related symptoms in patients with myeloprolifera- cythemia vera. Semin Thromb Hemost 1997; 23: 441–454. tive disorders. Blood 2003; 102: 32a (abstr.). 5 van Genderen PJ, Michiels JJ. Erythromelalgia: a pathognomonic 18 O’Brien SG, Guilhot F, Larson RA, Gathmann I, Baccarani M, microvascular thrombotic complication in essential thrombocythe- Cervantes F et al. Imatinib compared with interferon and low-dose mia and polycythemia vera. Semin Thromb Hemost 1997; 23: cytarabine for newly diagnosed chronic-phase chronic myeloid 357–363. leukemia. N Engl J Med 2003; 348: 994–1004.

Leukemia