International Journal of Hematology

, Volume 91, Issue 3, pp 516–521

Emergence of chronic myelogenous leukemia during treatment for essential thrombocythemia


  • Shinsuke Mizutani
    • Division of Hematology and Oncology, Department of MedicineKyoto Prefectural University of Medicine
    • Division of Hematology and Oncology, Department of MedicineKyoto Prefectural University of Medicine
  • Daisuke Shimizu
    • Division of Hematology and Oncology, Department of MedicineKyoto Prefectural University of Medicine
    • Department of MedicineFukuchiyama City Hospital
  • Shigeo Horiike
    • Division of Hematology and Oncology, Department of MedicineKyoto Prefectural University of Medicine
  • Masafumi Taniwaki
    • Division of Hematology and Oncology, Department of MedicineKyoto Prefectural University of Medicine
Case Report

DOI: 10.1007/s12185-010-0502-3

Cite this article as:
Mizutani, S., Kuroda, J., Shimizu, D. et al. Int J Hematol (2010) 91: 516. doi:10.1007/s12185-010-0502-3


A 72-year-old male patient was initially diagnosed with essential thrombocythemia (ET), a Philadelphia chromosome-negative (Ph1) chronic myeloproliferative disorder (CMPD), and was treated with hydroxyurea (HU). After 9 years of diagnosis of ET, his peripheral leukocytes gradually increased, while his platelet count showed a decrease. Bone marrow analysis disclosed Ph-positive chronic myelogenous leukemia (CML) in the chronic phase. Administration of imatinib mesylate (IM), a Bcr–Abl tyrosine kinase inhibitor (TKI), induced complete hematologic response in a month, but was discontinued after 4 months because of Grade 3 pleural effusion (PE). The treatment was switched to nilotinib which successfully induced a complete cytogenetic response (CCyR) after 5 months of TKI therapy and resolved the PE. Despite CCyR, however, ET recurred. Since then, the patient has been treated for 8 months with a combination of nilotinib and HU which has successfully controlled both CML and ET. This report includes a review of the characteristics of 15 reported cases with co-occurrence of CML and Bcr–Abl-negative CMPDs, including ours. Although rare, care needs to be taken since, despite the often similar clinical features of the two diseases, they require completely different treatments.


Chronic myelogenous leukemiaEssential thrombocythemiaChronic myeloproliferative disorderTyrosine kinase inhibitor

1 Introduction

Chronic myeloproliferative diseases (CMPDs) are classified into two major categories, Philadelphia chromosome (Ph1)-positive (Ph1+) chronic myelogenous leukemia (CML) and Ph1-negative (Ph1) disorders, such as polycythemia vera (PV), essential thrombocythemia (ET) and idiopathic myelofibrosis (IMF). Although approximately 1–11% of the latter disorders transform into acute myelogenous leukemia (AML) in their late clinical phase regardless of JAK2V617F mutational status or leukocyte counts at diagnosis [15], the coexistence of or more CMPDs is extremely rare and only 14 cases with CML and Bcr–Abl-negative (Bcr–Abl) CMPDs have been reported previously [618]. Here, we report a case with CML emerging after a 9-year history of ET, who was successfully treated with a combination of Bcr–Abl tyrosine kinase inhibitors (TKIs) and hydroxyurea (HU). We also review previous cases with Bcr–Abl CMPDs co-occurring with CML.

2 Case report

A 72-year-old man was referred to our hospital because of thrombocythemia in 1999. The platelet count in his peripheral blood (PB) was 1.23 × 1012/L, while leukocyte count and hemoglobin (Hb) concentration were normal, 8.3 × 109/L and 12.2 g/dL, respectively. The patient was diagnosed with ET because bone marrow (BM) analysis revealed the increased number of morphologically abnormal megakaryocytes (Mgks) with hypersegmented nuclei, the presence of a large cytoplasm-producing abundant platelets and the absence of Ph1 (Table 1; Fig. 1a). The patient remained untreated for 5 years because of his refusal to undergo therapy, but in 2004 he suffered acute myocardial infarction (AMI) accompanied by an increase in platelet count to 1.23 × 1012/L. BM study findings were again compatible with a diagnosis of ET (Table 1). AMI was successfully treated with coronary arterial intervention and the administration of HU and aspirin was initiated to prevent thrombotic events. HU dosage was adjusted to maintain the platelet count at around 500.0 × 109/L. In September 2008, the leukocyte count in the PB gradually increased to 1.23 × 109/L with 67.0% of mature granulocytes, 4.0% of metamyelocytes, 14.0% of myelocytes, 1.0% of eosinophils and 2.0% of basophils, but no myeloblasts. The platelet count was 392.0 × 109/L, the neutrophil alkaline phosphatase (NAP)-positive rate 20.5% and the NAP score 40.5. BM study showed a markedly elevated BM nucleated cell count of 880.0 × 109/L without an increase in myeloblasts. The number of Mgks was 222 × 106/L and they were comparatively small (Table 1; Fig. 1b, c). Karyotype analysis showed the presence of Ph1 chromosome, 46, XY, t(9;22)(q34;q22) in 19 of the 20 metaphase spreads, and the remaining 1 showed double Ph1, 47, XY, t(9;22)(q34;q22), der(22)t(9;22)(q34;q22). A fluorescence in situ hybridization (FISH) study revealed that 97% of BM interphase cells were positive for Ph1, and RT–PCR of PB showed the presence of major bcrabl transcripts. The patient was diagnosed with CML in the chronic phase (CP), and treatment with imatinib mesylate (IM) was started while HU was discontinued. The patient attained complete hematologic remission 1 month after IM initiation. However, the maximally tolerable IM dose for our patient was 200 mg per day because of Grade 3 peripheral fluid retention and pleural effusion which required the administration of as much as 80 mg/day of furosemide, so that the treatment was switched to 800 mg per day of nilotinib, which resolved the pleural effusion and peripheral edema 4 months after the initiation of IM treatment. FISH examination of PB showed that, together with the clearance of Ph1+ clones, the platelet count had gradually increased again. The count reached the maximal platelet count of 923.0 × 109/L after 5 months of treatment with Bcr–Abl TKIs. At this time point, the leukocyte count in the PB was 3.82 × 109/L, and Hb concentration was 8.7 g/dL. His serum iron was 17 μg/dL and ferritin was 26 mg/L. A BM study showed that the Ph1 clones had disappeared and his PB showed a reduction in bcrabl transcripts from >10000 copies/μg RNA to 438 copies/μg RNA, indicating that the patient had attained complete cytogenetic response (CCyR). In contrast, his BM showed a marked increase in Mgks to 1.17 × 109/L, and their morphology had reverted to that typical of ET. The large masses of platelet debris are also frequently observed in BM smear (Table 1; Fig. 1d, e). No JAK2V617F mutation was identified at this time. The patient had no splenomegaly. On the basis of these findings, we diagnosed the disease status as recurrence of ET with CML-CP during CCyR, and re-started HU therapy of 500 mg every 3 days, in combination with 800 mg per day of nilotinib, to suppress both Ph1 clones and platelet count. At the time of writing, this treatment has resulted in the maintenance of CCyR and suppression of the platelet count to around 50.0 × 109/L for 8 months.
Table 1

Bone marrow analysis findings


1999 March

2004 July

2008 October

2009 May


ET diagnosis


CML diagnosis

ET recurrence CML in CCyR

ANC (×109/L)





M/E ratio





Blast (%)





MgK (×106/L)





Ph1 + metaphase in BM










AMI acute myocardial infarction, ANC all nucleated cell count, ET essential thrombocythemia, CCyR complete cytogenetic response, CML chronic myelogenous leukemia, MgKs megakaryocytes, NE not examined
Fig. 1

Megakaryocytes (Mgks) in BM [Wright–Giemsa staining, ×1,000]. a Mgk at the initial diagnosis of essential thrombocythemia (ET) contain hypersegmented nuclei with abundant cytoplasm-producing platelets. Myeloid hyperplasia and relatively small Mgks at the diagnosis of chronic myelogenous leukemia [high magnification (b) and low magnification (c)]. Mgks and platelets aggregation (arrows) at the recurrence of ET following TKI treatment [high magnification (d) and low magnification (e)]

3 Discussion

The co-emergence of CML and Bcr–Abl CMPDs is extremely rare with only 14 previous cases reported [618], so that there is scant information on clinical outcomes and treatment strategies for this occurrence. For this reason, we review here CML accompanied by CMPDs in terms of both clinical features and treatments.

Of the 15 cases of co-occurrence of CML and Bcr–Abl CMPD, including ours, CMPD preceded CML in 11 cases, and in 3 they were diagnosed simultaneously at the initial diagnosis. Even in a unique case reported by Krämer et al. in which the diagnosis of myelofibrosis was attained after 3 months treatment with IM for CML, JAK2V617F mutation was retrospectively identified in leukocytes at the diagnosis of CML, suggesting the presence of CMPD clones before the diagnosis of CML [15]. Accordingly, in none of the cases did Bcr–Abl CMPD develop after CML. The type of CMPD preceding CML was ET in five cases, MF in six and PV in four each. The median age of the onset of CML in those patients was 64.0 ± 13.2 years which is somewhat older than usual for CML. No recurrent chromosomal abnormalities were observed in ten cases examined for such abnormalities when CMPD was diagnosed, and JAK2V617F mutation was identified in six of the ten patients examined for this mutation. Eleven had been treated with HU and one with 32P before the emergence of CML. Generally, 32P exposure may increase the risk for of secondary AML, while HU does not [1], but the effect of these treatments on secondary leukemogenesis, such as in CML, remains unclear. Finally, the time until the development of CML from the diagnosis of other CMPDs in the 11 cases of non-simultaneous occurrence varied widely from 2 to 18 years (mean ± standard deviation 9.0 ± 5.0 years). These findings indicate that patients affected by CML following CMPD are unlikely feature any particular clinical background which may help in predicting the emergence of CML during the treatment for CMPD.

As for treatment, IM was found to be effective for CML accompanied by CMPD. For the 13 patients for whom data are available, IM treatment achieved major molecular response (MMR) in 5, the 2-log reduction in bcrabl transcripts in 1, CCyR in 4, and partial cytogenetic response (PCyR) in 3 patients. However, the outcome for TKI treatment of the preceding Bcr–Abl CMPDs was not successful in all cases. Of the 11 cases for whom data are available, CMPDs were controlled successfully with TKI treatment in 5 cases, but recurred in 6. In our case, thrombocythemia recurred after the switch to nilotinib treatment when the patient had achieved CCyR for CML. These findings suggest two hypotheses to account for the co-emergence of CML and Bcr–Abl CMPDs. According to the first hypothesis, the off-target effect of IM, such as on c-KIT, may suppress the expansion of tumorigenic Bcr–Abl clones in some cases including ours, while nilotinib, with an off-target effect that is less than that of IM, allows for the expansion of Bcr–Abl clones and the subsequent thrombocythemia as seen in our case. The other hypothesis is that long-term treatment with TKIs causes the suppression of Ph1+ clones, which then creates space for the re-expansion of Bcr–Abl CMPD clones in BM. The two mechanisms may co-exist, and whether CMPD clones re-expand or not following TKI treatment may depend on the balance of the off-target kinetics of TKI treatment and the proliferation potency of CPMD clones, which may differ from case to case. When Bcr–Abl CMPDs recur as in our case, the additional treatment with TKI may thus be required. Indeed, the addition of HU to TKI was successful in two cases those experience the re-emergence of CMPD following TKI treatment, including ours [14].

Finally, one question that remains unanswered is whether CML originates from the same clone as does Bcr–Abl CMPD by acquiring Ph1 as an additional aberration, or the two diseases originate from different clones independently. Considering that JAK2V617F is normally unidentifiable in CML [19], and CMPDs reoccurred after the successful suppression of CML clones in about half of the cases investigated (Table 2), we favor the latter hypothesis, namely, that the original clones for CML and Ph1 CMPDs are different in most cases with rare exceptions [15]. In any case, Ph1+ CML clones may have a clonal advantage over Bcr–Abl CMPDs in BM in terms of proliferation that would explain why CML always occurred after CMPDs, but was never before and why CMPDs reoccur following the suppression of CML clones by TKI.
Table 2

List of cases of chronic myelogenous leukemia co-occurring with Philadelphia 1 chromosome-negative chronic myeloproliferative disorders

M male, F female, CMPDs chronic myeloproliferative disorders, ET essential thrombocythemia, IMF idiopathic myelofibrosis, PV polycythemia vera, del deletion, HU hydroxyurea, Ph1 Philadelphia chromosome, INF-α interferon-α, UR-PBSCT unrelated peripheral blood stem cell transplantation, IM imatinib mesylate, CCyR complete cytogenetic response, PCyR partial cytogenetic response, ND no data available

In conclusion, this report is the first of a patient with ET accompanied by CML who was successfully treated with a combination of nilotinib and HU. We also reviewed other cases of CML co-occurring with Bcr–Abl CMPDs. Although the co-emergence of CML during treatment for Ph1 CMPDs is rare, care needs to be taken since, despite the often similar clinical features of the two diseases, they require completely different treatments.


We are grateful to Ms. M. Goto and Ms. R. Tanaka for their excellent technical supports. This work is partly supported by the grant from Japan Leukaemia Research Fund (to J.K.).

Copyright information

© The Japanese Society of Hematology 2010