Long-term safety and efficacy of imatinib mesylate (Gleevec®) in elderly patients with chronic phase chronic myelogenous leukemia: Results of the AFR04 study


  • Conflict of interest: Dr Philippe Rousselot received research funding from Novartis and Bristol-Myers Squibb.

  • This study was presented in part at the 2007 Annual Meeting of the American Society of Hematology (Atlanta, December 2007).


Data from registries suggest that the median age of chronic myelogenous leukemia (CML) patients is 10–15 years older than that of clinical trials. We conducted a prospective phase II study to evaluate imatinib mesylate (IM, 400 mg daily) in newly diagnosed chronic phase CML (CP-CML) patients. Patients aged 70 years and over diagnosed with CP-CML within 12 months were eligible. Thirty patients were enrolled from April 2002 to October 2004. Median age was 74.8 years (range, 70–90). Male/female ratio was 1.72. At inclusion, comorbidities were reported in all but one patient, Cumulative Illness Rating Scale for Geriatrics comorbidity mean index was 1.47, and 66% of patients had cardiovascular disease. The median daily IM dose was 392 mg (range, 256–445). IM was interrupted in patients with severe comorbidities. Treatment discontinuation was observed in 36.6% of patients. Cumulative incidence of complete cytogenetic response was 71.4 and 78.5% at 12 and 24 months, respectively. A high level of sustained responses was observed in patients with mild or moderate comorbidities. Seven-year estimated overall survival was 80.8% (95% CI: 59.0–91.7). Two-thirds of the patients were still on long-term therapy at cut-off, and no patients had died from progression. This trial was registered at http://www.clinicaltrials.gov as # NCT00219765. Am. J. Hematol. 2013. © 2012 Wiley Periodicals, Inc.


Chronic myelogenous leukemia (CML) is a myeloproliferative disorder characterized by the t(9;22)(q34;q11) translocation generating a der(22) chromosome commonly referred as the Philadelphia chromosome. This translocation results in a BCR-ABL fusion gene coding for a fusion protein with enhanced nonregulated tyrosine kinase activity in the cytoplasm of leukemia cells [1]. Median age reported in clinical trials with CML patients range from 50 to 55 years. However, studies from registries indicate that the observed median age in patients with CML at diagnosis is closer to 60–65 years [2, 3]. With a difference in median age between registries and clinical trials of 10–20 years, elderly CML patients are underrepresented in prospective clinical trials.

Imatinib mesylate (IM, Gleevec®, Novartis Pharma) is a rationally designed tyrosine kinase inhibitor (TKI) and has demonstrated powerful activity against leukemic cells in vitro and in vivo [4, 5]. Based on the results of the large phase III IRIS trial, IM was registered as first-line therapy in chronic phase CML (CP-CML) [6]. However, the median age reported for patients included in the pivotal IRIS trial was 50 years, with a range from 18 to 70 years. We thus decided to conduct a prospective study to evaluate the IM treatment in newly diagnosed CP-CML patients aged 70 years old and over.

Patients and Methods

This phase II nonrandomized multicenter trial was designed to evaluate the efficacy and tolerability of IM in elderly patients. Male and female patients aged 70 years or older with newly diagnosed CP-CML were eligible. Patients had to have been diagnosed within 12 months of inclusion, with no prior therapy other than hydroxyurea. Other inclusion criteria included serum AST and ALT inferior than three times the institutional upper limit of the normal range (ULN), serum bilirubin less than 1.5 times ULN, and creatinine clearance more than 60 ml/min. Patients were excluded if they had previously been treated with interferon or imatinib or had another uncontrolled malignancy or an Eastern Cooperative Oncology Group (ECOG) performance status more than 2. Patients with Philadelphia chromosome negative CML, CML in accelerated phase, or in blast crisis were also ineligible. Minimental state exam (MMSE) was determined for all patients before inclusion. The MMSE is a scored serie of questions and tests. Scores of 27 or above out of 30 are considered normal [7]. To minimize noncompliance, it was recommended that patients with an MMSE score less than 25 were not included. All patients were scored according to the Manual of Guidelines for Scoring the Cumulative Illness Rating Scale for Geriatrics (CIRS-G) using the following descriptions for a given level of severity: 0 (no problem), 1 (current mild problem or past significant problem), 2 (moderate disability or morbidity/requires “first line” therapy), 3 (severe/constant significant disability/“uncontrollable” chronic problems), and 4 (extremely severe/immediate treatment required/end organ failure/severe impairment in function) [8]. A total CIRS-G score was calculated for each patients. The mean CIRS-G index was calculated by summing each comorbidity score and dividing by the number of comorbidities. Both the MMSE and CIRS-G scores were calculated for all patients in collaboration with geriatric specialists.

IM was administered at 400 mg daily. Dose modifications wereallowed according to the criteria of the IRIS study for IM [6]. Cytogenetic response was assessed for each patient every 6 months for 2 years and every year thereafter. RTQ-PCR analysis for BCR-ABL quantification was performed in reference laboratories every 3 months and results expressed as a percentage of BCR-ABL/ABL ratio [9]. Efficacy endpoints were the rate of complete hematologic response (CHR) at 6, 9, and 12 months, the rate of complete cytogenetic response (CCyR) at 6 and 12 months, the level of molecular residual disease in patients in CCyR after 6 and 12 months of treatment, the rate of major molecular response (MMR: BCR-ABL = 0.1%), and the rate of complete molecular response (CMR: BCR-ABL = 0.0032%) referred to hereafter as molecular response with 4.5 log reduction (MR 4.5), event-free survival (EFS; including no response according to ELN criteria, disease progression, death from any cause, loss of CHR, loss of major cytogenetic response, and IM discontinuation) and overall survival (OS; calculated from the time of the start of IM treatment to the date of death). Survival was censored at the last recorded contact or evaluation for patients alive at the time of analysis. Duration of response was censored at the last examination date for patients with an ongoing response or patients who discontinued treatment for reasons other than adverse events, progression, or death. Patients who withdrew from treatment before a confirmed response were counted as nonresponders. Response and event rates were reported based on an intent-to-treat analysis. OS and EFS were estimated according to the Kaplan–Meier method. Tolerance was assessed by recording all adverse events observed on therapy. Safety results are reported for all patients who received at least one dose of IM.



Thirty patients with newly diagnosed CP-CML were enrolled in nine French centers from April 2002 to October 2004. Patient characteristics are described in Table I. Median age was 74.8 years (range, 70–90). A higher proportion of males were included (male/female ratio of 1.72). Patients were in good general condition with an ECOG PS of 0 or 1 in 20 patients (67%), 1 in nine patients (30%), and 2 in one patient (3%). Median MMSE score was 29, which is considered normal (range, 25–30). All but one of the 30 patients had at least one mild or moderate comorbidity. The median CIRS-G score was 6 (range, 0–12). Two patients had level 3 categories (cardiopathy in one, cardiopathy, and nephropathy in the other). The mean CIRS-G index value was 1.47 (range, 0–2). Of note, 66% of patients had cardiovascular diseases at inclusion, which were scored from 1 to 2 other than one patient who presented with a score of 3. Thirty-three percent of patients were classified as high-risk CML, 37% as intermediate risk, and 30% as low risk according to the Sokal score stratification [10]. Three patients (10%) presented with a variant translocation at diagnosis involving chromosomes 9 and 22, while other translocations included chromosome 8 and 15 (one case), chromosome 6 (one case), and chromosome 19 (one case). The interval between diagnosis and imatinib initiation was 51 days (range, 13–312 days), and median follow-up was 5.75 years.

Table I. Patient Characteristics
  • ECOG, Eastern Cooperative Oncology Group.

  • a

    One patient had a score of 0 and 66% of patients had cardiovascular disease.

Number of patients30
Median age in years (range)74.8 (70–90)
Sex ratio (M/F)1.72
ECOG PS, N (%)
 020 (67%)
 19 (30%)
 21 (3%)
Sokal score, N (%)
 Low9 (30%)
 Intermediate11 (37%)
 High10 (33%)
Median MMSE score (range)29 (25–30)
Comorbidity CIRS-Ga
 Median score (range)6 (0–12)
 Mean severity index1.47
Median follow-up (years)5.75
Table II. Main Adverse Events
 Patients (%)
  • a

    Thirty seven percent of all patients received erythropoietin.

  • b

    One patient received G-CSF.

  • c

    No grade 3 to 4 event.

  • d

    Grade 4 in one patient.

Hematological adverse events
 Anemia <10 g/dL9 (30%)a
 All grade neutropenia6 (20%)b
 All grade thrombocytopenia7 (23)%
Nonhematological adverse events
 Grade 1–2 edemas14 (47%)c
 All grade cramps5 (17)%
 All grade gastrointestinald9 (30%)

Treatment administration

The median daily dose of IM was 392 mg (range, 256–445), as calculated from the patients' records during the first year of therapy. Imatinib trough levels were assessed in nine patients, and the median through value was 682 ng/ml (range, 408–2,297). A transient or permanent daily dose reduction was recorded in 52% of patients and 44% experienced transient imatinib interruptions. Most patients who interrupted IM resumed at 300 mg/day. Treatment discontinuation was observed in 37% of patients, the most common reason being intolerance, reported in six patients, including the three patients with the highest comorbidity score. Three of these patients were switched to dasatinib and one to nilotinib. Other reasons for discontinuation were development of another cancer, cytogenetic failure, hematological relapse, death not related to disease progression or therapy, and discontinuation of imatinib due to CMR (all one patient each).

Safety and tolerability

The most frequently reported hematological toxicity was anemia, which was present in approximately one-third of the patients (nine patients with grades 2 and 3), eight of whom received rHU-EPO at a dose ranging from 10,000 UI per week to 10,000 UI per month, with complete resolution of anemia in all treated patients. Grades 2 and 3 transient neutropenia were observed in six patients, and only one patient received G-CSF (Table II).

The most commonly reported nonhematological toxicities were grades 1 and 2 fluid retention (14 patients). Two patients presented cardiac failure (one after 126 and the other after 365 days of IM therapy), both of whom had a CIRS-G cardiac category at level 3 at inclusion. With the exception of these two patients, all other patients responded to diuretics, and IM was maintained. Left-ventricular ejection fraction was assessed in five patients with fluid retention and was found to be within the normal range. NT-proBNP or BNP levels were not systematically assessed during the study. Gastrointestinal symptoms, mainly diarrhea, were also frequent (nine patients) and were severe (grade 3) in one patient. Other reported toxicities were rash (one case of grade 3), asthenia not in the context of anemia (three patients), and conjunctivitis (three patients) (Table II).


One patient was not evaluable for response having withdrawn consent before evaluation. Twenty-eight of the 29 evaluable patients (97%) were in CHR after 3 months of IM therapy, and all patients reached CHR at month 6. A summary of the cumulative incidence of cytogenetic and molecular responses is shown in Fig. 1. The rate of CCyR was 71.4 and 78.5% at 12 and 24 months, respectively. The cumulative incidence of MMR was 25, 53.5, and 66.6% at months 12, 24, and 72, respectively. The rate of MMR at months 12 and 24 was 25.0 and 51.7%, respectively. The cumulative incidence of MR 4.5 increased from 6.7% by month 24 up to 44% by month 72. The rate of undetectable residual disease was 27.6% at month 72. EFS at 7 years was 53.4% (95% CI: 29.8–72.3) (Fig. 2). OS at 7 years was 80.8% (95% CI: 59.0–91.7). No patients died from CML progression to accelerated or blast phase, one patient died while on study, and three others died after imatinib discontinuation.

Figure 1.

Cumulative incidence of cytogenetic and molecular responses over 6 years.

Figure 2.

Event-free survival (EFS; A) and overall survival (OS; B). EFS included no response according to ELN criteria, death, progression to accelerated phase or blast crisis, and loss of CCyR and treatment discontinuation. No patient died from CML progression to accelerated phase or blast crisis. The 95% confidence intervals are indicated as dashed lines.


Age is a well-recognized adverse prognostic factor for CML and is included in the calculation of the Sokal score [10]. As expected, the prognostic value of age in CML is not independent of the therapy used. Before the era of interferon, age was an independent prognostic factor for survival. Interferon, however, resulted in significantly worse side effects in older patients, and their poor prognosis may be due, at least in part, to poor treatment tolerance and inadequate treatment delivery, as suggested in one study showing a worse outcome for elderly patients treated with interferon compared to young patients [11]. On the other hand, in a study performed by the German CML group, survival of patients treated with interferon was identical irrespective of age [12]. Interferon doses used to achieve response were lower in older patients, and adverse events were similar in both age groups. With the use of IM, older age appears to have lost much of its prognostic relevance, suggesting that the poor prognosis previously observed with older age was related to treatment-associated factors rather than to the biology of CML in older patients [13].

We report here a prospective study with long-term follow-up of newly diagnosed CP-CML patients aged 70 years and over treated with IM as first-line therapy. The cumulative response rates compare very favorably with the data from the IRIS study with a similar follow-up [14]. The cumulative rate of CCyR at 12 months of 71.4% in our study is similar to the rate of 69% reported in the IRIS study. The cumulative rates of MMR at 72 months were 66.6% in our study and 65% in the IRIS study, respectively [14]. Seven-year OS was 80.8% (95% CI: 59.0–91.7) here after a median follow-up of 5.75 years, and remarkably no patients died from disease progression to accelerated phase or to blast crisis. Using the criteria reported for the Hammersmith Hospital cohort, EFS at 5 years was 66.2% (53.4% at 7 years) in our study, while it was 62.7% at 5 years in the Hammersmith Hospital study [15]. A recent retrospective survey from the GIMEMA group evaluated patients from three CML trials who were over the age of 65 and treated with IM as frontline therapy. Results were similar to those reported here [16, 17]. The median age of these 115 patients was 71 years (vs. 74.8 years in our study). The cumulative MMR rate was 89.7% at 72 months compared to 66.6% in the present prospective study. OS was 78% in the GIMEMA survey compared to 80.8% in the present study with a follow-up of nearly 6 years.

Comorbidities represent an underestimated parameter in prospective or retrospective studies evaluating the value of TKIs in elderly patients. To our knowledge, our study is one of the first including an assessment with a widely accepted comorbidity scoring system at inclusion. We suggest that a high-CIRS-G score value does not preclude the use of imatinib if the level of severity for a given comorbidity is below the value of 3. Physicians should pay close attention to the development of fluid retention and anemia as these are easily preventable with the prescription of diuretics and rHu-EPO. Two second-generation TKIs, nilotinib and dasatinib, were recently evaluated in the first-line setting, and both resulted in faster and better responses compared to IM [18, 19]. However, very few elderly patients were enrolled in these studies. Furthermore, cardiac comorbidities (present in 66% of our patients) were listed as criteria for noninclusion in both ENESTnd and DASISION trials. Cardiac comorbidities are also accepted warnings in the use of nilotinib. Moreover, being aged over 65 years is associated with a relative risk of up to 8.2% for developing pleural effusion, a serious limitation for the use of dasatinib in first-line treatment at the recommended 100 mg daily dose [20]. We thus recommend that IM remains the TKI of choice in elderly patients with newly diagnosed CML in the frontline setting.

Life expectancy at 75 years is around 12 years [21], and the 10-year survival probability of CML patients in MCR was estimated to be ∼65% with imatinib treatment patients in 2005 [22]. Thus, current life expectancy with CML may be longer than age-adjusted life expectancy in elderly patients aged 75 years or more. This suggests that the objective of the treatment of elderly CML patients may remain survival with a well-tolerated TKI in the long term, as imatinib in our study, even in cases of associated comorbidities.


The authors are indebted to the French CML group (Fi-LMC) for scientific and financial support, to Novartis Pharma, France for providing imatinib, and they thank Sarah MacKenzie for editorial assistance with this manuscript.