Cytarabine and clofarabine after high-dose cytarabine in relapsed or refractory AML patients


  • Conflict of interest: Nothing to report


Clofarabine has been shown to be effective in AML patients, either as single agent or, mainly, in association with intermediate dose cytarabine. Based on these reports, we conducted a preliminary study combining clofarabine and intermediate dose cytarabine in AML patients who relapsed or failed to respond to at least two induction therapies. We treated 47 patients affected by relapsed/refractory AML with a regimen including clofarabine at 22.5 mg/m2 daily on days 1–5, followed after 3 hr by cytarabine at 1 g/m2 daily on days 1–5. Ten patients received a further consolidation cycle with clofarabine at 22.5 mg/m2 and cytarabine at 1 g/m2 day 1–4. Among the 47 patients, 24/47 (51%) achieved a complete remission, 5/47 (10.5%) a partial response, 10/47 (21%) had a resistant disease, and 6/47 (13%) died of complications during the aplastic phase. The most frequent nonhematologic adverse events were vomiting, diarrhea, transient liver toxicity, febrile neutropenia, and infections microbiologically documented. Among the 24 patients who obtained a CR 13 underwent allogeneic bone marrow transplantation. In 14 patients, complete remission duration was shorter than 12 months, whereas 10 patients experienced longer complete remission duration. These very preliminary results suggest that clofarabine-cytarabine regimen is effective in this particularly poor prognosis category of patients, representing a potential “bridge” toward bone marrow transplant procedures. Safety data were consistent with previously reported salvage therapies. Further studies and a longer follow up are warranted. Am. J. Hematol., 2012. © 2012 Wiley Periodicals, Inc.


Since the introduction of nucleoside analogues in the 1960s, therapeutic strategies for acute myeloid leukemia (AML) have focused on cytosine arabinoside, typically administered in association with an anthracycline such as idarubicin or daunorubicin [1]. Current therapies in adult AML are able to induce a complete remission in about 70% of patients, but most of them relapse and succumb to the disease or to associated complications.

Over the past years, a series of deoxyadenosine analogues have been synthesized, with the aim to improve drug efficacy and minimize the extramedullary effects, in particular renal and neurologic toxicities, associated to the use of high doses of cytarabine. Several of these compounds, such as fludarabine, cladribine, nelarabine, or gemcitabine have been introduced into the clinical practice for hematological malignancies [2–4].

Clofarabine (2-chloro-2′-fluoro-deoxy-9-β-D-arabinofuranosyladenine) is a second-generation nucleoside analog, synthetized as a hybrid molecule to combine the most favorable pharmacokinetic properties of fludarabine and cladribine, avoiding in the meantime the dose-limiting neurotoxicity of cytarabine. Halogenation of the 2 position of adenine and substitution of a fluorine group at the C-2′ position of the arabinofuranosyl moiety represent the essential chemical modifications of clofarabine, and are responsible for the main characteristics of this molecule: (1) high resistance to phosphorolytic cleavage by bacterial purine nucleoside phosphorylase; (2) potent inhibition of DNA synthesis; (3) prolonged retention of clofarabine triphosphate in leukemic blasts [2].

Based on similarity with fludarabine and cladribine, preliminary experiences were conducted in lymphoproliferative disorders. Subsequently preclinical studies in vitro and in murine models showed effects also in acute leukemias, and in the late 1990s phase 1 and 2 studies in acute leukemia were begun, with encouraging results in heavily pretreated patients with a tolerable toxicity profile [5, 6].

A phase II study, in pediatric patients with refractory or relapsed ALL, treated with clofarabine 52 mg/m2 daily for 5 days showed an overall response rate of 30% and based on these results clofarabine was approved by FDA for the treatment of this category of patients [7].

As single agent in AML clofarabine at 40 mg/m2 daily for 5 days was able to determine a response in 55% of patients (42% CR; 13% PR) [6]. Preclinical studies showed a strong rationale of combination of clofarabine with other purine analogues, especially cytarabine, since pretreatment with clofarabine leads to increase in intracellular ara-C triphosphate levels [8]. At the MD Anderson Cancer Center, several trials have been conducted with clofarabine in association with cytarabine and or idarubicin with very encouraging results even in primary refractory or relapsed AML [9–13].

On the basis of these reports, we focused our interest on AML patients relapsed or failed to therapies including HDAC; considering the heavily pretreated category of patients selected, we decided to explore effectiveness of clofarabine at lower dosage than previously reported [9–13], and we fixed the amount at 22.5 mg/m2.

Design and Methods

Study group

The study was conducted on 47 patients with refractory or relapsed AML referred to our institutions between 2008 and 2010. Refractory patients were unresponsive to at least two different induction regimens including HDAC, relapsed patients were required to be refractory to reinduction therapies with HDAC. All patients provided informed consent according to institutional guidelines. Patients were included in the analysis if they had an ECOG performance status ≤2, and adequate hepatorenal function (serum creatinine ≤2; total bilirubin ≤2 mg/dL; SGOT or SGPT ≤ upper limit of normal), absence of serious heart diseases, uncontrolled infection or any other severe concurrent disease.

Treatment (doses and schedule)

Clofarabine was administered at the dosage of 22.5 mg/m2 as a 1-hr intravenous infusion daily for 5 consecutive days, followed after 3 hr by cytarabine at 1 g/m2 as a 3-hr constant-rate intravenous infusion once daily for 5 consecutive days, as reported in previous experiences [10, 12].

Patients achieving CR or PR were slated to receive a further consolidation cycle with clofarabine at 22.5 mg/m2 and cytarabine at 1 g/m2 on days 1 through 4. Before starting the treatment, patients underwent a general evaluation including history, physical examination, complete blood counts with differentials and platelet count, a complete chemistry survey, and marrow aspiration or only peripheral blood analysis (if the diagnosis of relapse could be made unequivocally on this specimen).

Supportive measures for optimal medical care were provided as established by the treating physician and patient's medical need. Anti-infective prophylaxis was administered at the onset of neutropenia and included levofloxacin (500 mg/day, orally), and posaconazole 200 mg twice/day. Prophylaxis with acyclovir (400 mg orally, twice/day) was introduced after HHV6 reactivation documentation. Hematopoietic growth factors were administered in patients with febrile neutropenia or prolonged leucopenia.

Response criteria

A CR was defined as a normocellular marrow with ≤5% blasts and peripheral counts with noncirculating blast cells, a neutrophil count of more than or equal to 1 × 109/L and platelet counts more than or equal to 100 × 109/L. A CRp had similar criteria, but with an incomplete recovery of platelets. A partial response (PR) consisted of a blood recovery as for CR, but with persistence of 5–25% marrow blasts.

Statistical analysis

This was a single arm study with tumor response as the primary efficacy endpoint. Efficacy endpoints were CR, PR, and overall response rates (CR plus PR) duration of remission and survival. Results are presented for all treated patients. Kaplan-Meier methodology was used to describe time-to-event outcomes.


Study population

A total of 47 patients were considered. Their characteristics are depicted in Table I. The median age was 50.5 years (range 21–71 years); 24 patients were males and 23 females. Cytogenetic analysis at diagnosis showed a normal karyotype in 22 patients, and complex aberrations in 15. Two patients were t(8;21) positive. In eight cases, cytogenetic analysis was not available (not performed or absence of growth; Table I). Twenty patients were refractory AML, 14 in second or third relapse, 13 in first relapse after receiving preceding regimens including HDAC. Median CRD1 duration was 16.2 months (range 2–80 months; Table I).

Table I. Patient Characteristics
  1. CRD1: duration of first complete remission.

 Age (years); median (range)50.5 (21–71)
 Sex (M/F)24/23
 Cytogenetics, no.
Disease Status, no.
 First relapse13
 Second or third relapse14
 Primary refractory20
 CRD1, mo, median (range)16.2 (2–80)

Response and outcome

Among all patients, 24 (51%) achieved a CR. Another 5 patients (10.5%) had a PR, for an overall response of 60.5%. Ten patients (21%) were resistant and we observed six deaths (13%) during the induction phase. Among the cases obtaining a CR, seven patients had been experienced a CRD1 longer than 12 months, eight had been shown a CRD1 shorter than 12 months, and nine cases were primary refractory. Fourteen patients (30%) relapsed within 12 months, 10 (21%) showed a longer CR duration. Characteristics of responding patients (CR) are depicted in Table II.

Table II. Characteristics of Responding Patients (CR)
Patient no.Age/sexKaryotypeNPM1/FLT3 statusPrior therapyCRD1, mosResponse Duration, mosNo. of courseAllo BMT Y/N
  1. CRD1: duration of first complete remission (0 indicates primary refractory to induction therapy); n.a.: not available; HDAC: high-dose cytarabine; FLA: fludarabine, high-dose cytarabine; ABMT: autologous bone marrow transplant; alloBMT: allogeneic bone marrow transplant; ICE: idarubicin, cytarabine, and etoposide.

465/MMonosomy 7neg/negHDAC + ABMT7.861N
1249/FDiploidneg/negHDAC + ABMT12.142Y
1443/FDiploidn.a./negHDAC + alloBMT24.2132N
1533/FDiploidneg/negHDAC + ABMT1222Y
2154/Ft(8;21)neg/negHDAC + alloBMT851N
2244//FDiploidpos/negHDAC + alloBMT8101Y
2563/FAbsence of growthneg/negFLA123,51Y
2651/FTrisomy 11neg/negHDAC032Y
2759/Fn.a.n.a/n.aHDAC + alloBMT8092N
3163/MComplexneg/negICE + HDAC011N
3564/Ft(2;13)neg/negHDAC + ABMT851N
3728/FDiploidneg/posHDAC + alloBMT071N
3821/MDiploidn.a/n.aHDAC, MEC0191Y
3923/MComplexneg/negICE + AMBT + MEC6211Y
4056/FDiploidneg/negICE + HDAC14131N
4260/MDiploidneg/posICE + HDAC0212Y
4343/Mt(11;17)n.a./negICE + HDAC0182Y
4434/MDiploidneg/posHDAC + ABMT10172Y

Among the 24 patients, 14 received only one course, mainly because of complications during the first cycle or because of prompt availability of bone marrow transplant procedures. Only 10 patients received a consolidation course. Neither refractory nor partial responder patients received a further cycle. Thirteen patients underwent allogeneic bone marrow transplantation. Transplant was not an option in the other 11 responding patients because of lack of donor, older age, or early relapse.

Median overall survival of all treated patients was 197 days (range 4–1531 days; Fig. 1), with a median overall survival of responding patients of 478 days (range 58–1531 days; Fig. 2). Median remission duration in patients in CR was 235.6 days (range 30–1531 days). Median overall survival of transplanted patients was 445 days (range 30–1531 days; Fig. 3), with a median CR duration of 317.6 days (range 30–1531 days).

Figure 1.

Overall survival of all population censored at the time of the last follow up. [Color figure can be viewed in the online issue, which is available at]

Figure 2.

Overall survival of responders (CR) censored at the time of the last follow up. [Color figure can be viewed in the online issue, which is available at]

Figure 3.

Overall survival of transplanted patients censored at the time of the last follow up. [Color figure can be viewed in the online issue, which is available at]

Side effects

The principal side effects observed with clofarabine/cytarabine regimen were gastrointestinal; nausea and vomiting were common, as diarrhea and mucositis (Table III). Skin flushing were typically noticed during chemotherapy infusion, and resolved a few days after stopping the treatment. Two patients experienced a Grade 4 skin rash; in one case, it was associated with a HHV6 reactivation, but cutaneous manifestations were often concomitant to drug administration and were persistent even after reduction of the viral load. Transient liver alterations were common; seven patients (15%) showed elevation of transaminases and mild hyperbilirubinemia, three patients experienced Grade 3 toxicity, and among them one patient, a 63-year-old woman, previously highly treated, at the third relapse after bone marrow transplantation, evolved into multiorgan failure (MOF).

Table III. Nonhematologic Side Effects
No. (%)
Side effectGrade 1–2Grade 3 or higher
  1. MOF: multiorgan failure.

Nausea/vomiting21 (45)
Diarrhea18 (38)
Mucositis4 (8.5)2 (4.25)
Skin rash15 (32)2 (4.25)
Liver disfuction7 (15)3 (6)
MOF1 (2)

Notably, no neurological manifestations of toxicity were documented, even in patients heavily treated with HDAC regimens, where a neurological toxicity due to a cumulative dose of cytarabine might be expected.

Comparing with other salvage strategy, we did not observe a significant delay in neutrophil recovery, with a median time of neutrophil count below 1 × 109/L of 24 days (range 13–42). Unexpectedly, a patient (#15, female, 33-year-old, first relapse after HDAC and ABMT) experienced irreversible aplasia after the consolidation course, complicated by an unusual HHV6 reactivation and extensive skin rash. She underwent cord blood transplant; still aplastic, and died in day + 20 for myelosuppression-associated complications (pneumonia) without any sign of engraftment.

Febrile episodes were common during the aplastic phase (Table IV). Fever of unknown origin occurred in 20 (47%) patients. In 16 patients, an infection was documented; among them, 13 (28%) showed a bacteremia, 1 (2%) patient developed a pulmonary aspergillosis, and 2 (4%) patients presented a HHV6 reactivation. Four patients died of septic shock (2 Pseudomonas aeruginosa, 1 Streptococcus pneumoniae, and 1 Klebsella pneumoniae) during the myelosuppression phase. Based on these observations, we introduced antiviral prophylaxis with acyclovir and we monitored patients weekly with quantitative PCR for CMV, EBV, and HHV6, without documenting any further viral reactivation.

Table IV. Hematologic Side Effects
  1. Other indicates a HHV6 reactivation.

Median time to recovery for patients in CR
 Neutrophil count at or above 1 × 109/L d, mean (range)24 (13–42)
 Febrile EpisodesNo. (%)
 Fever of unknown origin22 (47)
 Bacterial/sepsis13 (28)
 Fungal1 (2)


Despite the high rates of complete remission observed in the AML setting, the median duration of CR is <1 year, and only about a 40% of young patients are cured [14]. Moreover, CR rates and OS are completely unsatisfactory in patients with unfavorable features at the diagnosis, such as complex karyotype, older age, or FLT3 ITD positivity [15–19]. Salvage chemotherapy is able to recover about a 50% of relapsed or primary refractory patients, but their prognosis remains poor. Allogeneic bone marrow transplant can at least determine a longer disease control in these categories of patients, but the majority of them relapses and succumbs of their leukemia [20]. Based on these considerations, it is clear that the introduction of new and more potent drugs in the treatment of AML represents a major need.

First phase I studies with clofarabine in acute leukemias demonstrated an activity and tolerability of the drug, and a schedule of 40 mg/m2 IV daily was indicated as the recommended dose for phase II studies [5]. In a phase II study conducted at MD Anderson in relapsed-refractory AML, MDS and BP-CML patients received clofarabine 40 mg/m2 IV daily for five consecutive days, with an overall response rate in AML of 55% [6].

Preclinical data showed that clofarabine is able to modulate ara-CTP accumulation and increase antileukemic activity of cytarabine [8]. For these reasons, a Phase I/II study was conducted combining clofarabine with cytarabine, with a overall response rate in AML of 40% [9, 11, 13]. Encouraging results were reported also in the phase I study conducted by Faderl et al., where clofarabine was associated to Idarubicin alone or with cytarabine [12].

Due to the acceptable safety profile, which emerged in all these investigations, several studies explored the use of clofarabine in elderly AML patients as frontline treatment, providing evidence of a significant activity especially in association with cytarabine, with no major increase in toxicity [21–23].

In this study, we focused our interest on primary refractory/relapsed AML, a category with a very poor prognosis where standard approaches are still unavailable and where effective treatments are an urgent medical need. Prognosis of our cases was particularly poor, since all patients were treated after receiving cytarabine at high dosage; in fact, primary refractory AML patients had not responded to at least two induction attempts, including high doses of cytarabine, while relapsed AML patients received clofarabine-cytarabine regimen after failure of reinduction with HDAC or in the case of early relapse after induction-consolidation strategy including HDAC. We decided to explore the efficacy of a clofarabine-cytarabine combination using clofarabine at a lower dosage than in previous experiences [11–13], and we fixed the amount at 22.5 mg/m2, reported as the maximum-tolerated dose in association with idarubicin and cytarabine [11]. Treatment schedule was the same used by Faderl et al. [11–13]. Clofarabine was administered as a 1-hr intravenous infusion followed ∼3 hr later by cytarabine as a 3-hr intravenous infusion, over five days.

An overall response rate of 61.5% (51% CR and 10.5% PR) was observed, with a median overall survival in responding patients of 478 days. Therapy was substantially well tolerated and both drugs could be administered at full dosage. The major toxicities observed were gastrointestinal, including transient liver test abnormalities, nausea and vomiting, diarrhea, and mucositis, but mainly at Grade 1–2. Skin rashes were frequent during the infusion, but were transient and responsive to steroid therapy at low dosage. Interestingly, even if patients had received cumulative cytarabine dose, we did not detected neurological side effects, even in older patients.

Infectious complications were comparable to other salvage chemotherapies, except an uncommon HHV6 reactivation documented in two patients, and related most likely to the relevant immunosuppressive activity of clofarabine. In particular, occurrence of HHV6 reactivation in a patient during a consolidation course and the concomitant irreversible aplasia was an event relatively unexpected. Based on these observations, we would suggest carefully monitoring patients for herpes virus reactivation after treatment with clofarabine; antiviral prophylaxis is strongly recommended. Six patients died during induction, resulting in an overall induction mortality of 13%, rather acceptable in this very heavily treated class of patients.

Incorporation of bone marrow transplant for these high-risk patients is certainly a strategy that could improve their outcome. We were able to transplant in complete remission, 13 patients that represents a considerable number if we consider the poor-prognosis patients we were managing. In this subgroup of cases, therapy with clofarabine and cytarabine definitely represented a “bridge” to transplant; and in this setting, we experienced longer remission durations.

These very preliminary clofarabine data are encouraging but still unsatisfactory, both in relapsed and in primary refractory AML patients. A possible strategy might be the use of clofarabine in an early setting, for example immediately at relapse, independent of previous treatment with HDAC, or directly in the refractory setting, after failure of standard “3 + 7” induction therapy.

Certainly, detection of new molecular markers to identify rapidly subgroups of very high-risk patients at diagnosis is mandatory. Moreover, a different approach could be represented for prompt determination of parameters, such as blast clearance and WT1 transcript reduction [24], simple to monitor during the induction treatment, and able to identify cases not responding to the standard treatment or at high risk for relapse, that could potentially benefit from an immediate intensification of therapy.

Furthermore, combinations with different and active drugs could be explored; preliminary results indicate at least the feasibility of the combinations of clofarabine with idarubicin and cytarabine [12] or with alkylating agents in adults with refractory acute leukemias [25].

In summary, our data confirmed a significant activity of clofarabine-cytarabine combination in relapsed-refractory AML patients, but long-term follow up remains still unsatisfactory. Further efforts should be directed to individuate alternative strategies to optimize the use of this promising drug in the AML setting.