Intolerance to tyrosine kinase inhibitors in chronic myeloid leukemia

Definitions and clinical implications



Tyrosine kinase inhibitor (TKI) treatment targeting breakpoint cluster region-Abelson murine leukemia virus, the cause of chronic myeloid leukemia (CML), has revolutionized therapy for patients with this disease. The majority of patients with CML maintain favorable responses with long-term imatinib therapy; however, the availability of the second-generation TKIs nilotinib and dasatinib limits the need for patients intolerant to imatinib to continue with therapy. Unfortunately, there is currently no standard definition of intolerance to imatinib. Common Toxicity Criteria for grading adverse events, designed to identify acute toxicities, are often used to determine intolerance. However, because CML therapies are long-term, patient quality of life may provide a better measure of true intolerance. Several general methods of quantifying patient quality of life are in use for patients with CML, and a CML-specific variant of the M. D. Anderson Symptom Inventory is in development. An appropriate and consistent definition of intolerance will provide clinicians with an algorithm for managing their patients with severe or chronic adverse events during treatment with imatinib. As more long-term data become available for newer TKIs, the definition of intolerance in the context of CML treatment will continue to evolve to maximize the likelihood of durable responses and superior quality of life for patients. Cancer 2011. © 2010 American Cancer Society.

Despite the advances in patient outcomes since the approval of imatinib (Gleevec/Glivec; Novartis Pharmaceuticals Corporation, East Hanover, NJ) as first-line therapy for chronic myeloid leukemia (CML) and the favorable and tolerable safety profile of this drug, a substantial minority of patients do not benefit from therapy because of intolerance or resistance. In a study conducted at the Hammersmith Hospital in London, for example, 25% of patients had discontinued imatinib after 5 years because of unsatisfactory response or toxicity.1 Acceptance by patients and clinicians of chronic adverse events (AEs) is inevitably related to available treatment options. As treatment options have expanded and alternatives to imatinib have become available, the opportunity to avoid chronic toxicity attributed to imatinib has emerged. Approval of the second-generation tyrosine kinase inhibitors (TKIs) nilotinib (Tasigna; Novartis Pharmaceuticals Corporation) and dasatinib (Sprycel; Bristol-Myers Squibb Company, Princeton, NJ) increases treatment options and presents the opportunity to avoid the chronic toxicity paradigm, although it is important to note that these agents themselves have unique toxicities, and the long-term toxicity profile of these novel agents is largely unknown. Paramount to the argument to minimize chronic toxicity is the finding that prompt identification and appropriate management of AEs may improve adherence to CML therapy and thus result in better long-term outcomes and quality of life (QOL) for patients. The aim of this review is to examine current definitions of intolerance to imatinib and to discuss their implications on management strategies for patients with CML.

Defining Imatinib Intolerance

Clinical trials

There are subtle differences in the criteria used to define intolerance in clinical trials (Table 1). In the IRIS (International Randomized Interferon versus STI571) trial, intolerance to imatinib was defined as the recurrence of a nonhematologic toxicity of at least grade 3 despite appropriate dose reductions and optimal symptomatic management.2 In trials using nilotinib as a second-line therapy, imatinib-intolerant patients were classified as those with symptoms of intolerance who had never achieved a major cytogenetic response on imatinib therapy. Symptoms of intolerance were defined as any nonhematologic toxicity of grade 3 or higher severity, or of grade 2 or higher severity lasting >1 month or recurring >3× despite dose reduction and optimal supportive care or any grade 4 hematologic toxicity persisting for >7 days.3 In the phase 2 registration trial of dasatinib as a second-line therapy, intolerance was defined as any nonhematologic toxicity of at least grade 3 or a hematologic toxicity of grade 4 that persisted for >7 days in patients treated with any dose of imatinib.4 Response to imatinib was not incorporated into the definition of intolerance in this trial.

Table 1. Definitions of Imatinib Intolerance in Clinical Trials
TrialTrial IDDefinition of Intolerance
IRIS trial2NCT00006343Recurrence of nonhematologic toxicity of at least grade 3 despite appropriate dose reductions and optimal symptomatic management.
Nilotinib in imatinib-resistant or -intolerant patients3NCT00109707Patients with symptoms of intolerance who had never achieved a major cytogenetic response.
Any grade 2 nonhematologic toxicity lasting >1 month or recurring >3× despite supportive care and maximum dose reduction.
Any grade 3 or higher nonhematologic toxicity.
Any grade 4 hematologic toxicity lasting >7 days.
Dasatinib in imatinib-resistant or -intolerant patients4CA180013Occurrence of at least a grade 3 nonhematologic or grade 4 hematologic toxicity lasting >7 days during treatment with imatinib at any dose.

Clinical practice

There is currently no universally accepted definition of imatinib intolerance in clinical practice, with centers and individual practitioners using different definitions based on rules for dose interruptions/reductions in clinical trials, management recommendations, or their own previous experiences and their patients' expectations. Although the National Comprehensive Cancer Network (NCCN) publishes guidance on recognizing and managing imatinib toxicity,5 and the European LeukemiaNet6 provides recommendations on the use of second-line agents for patients with intolerance to imatinib, neither elaborates on the definition of intolerance. Furthermore, few reports exist about the rates of imatinib intolerance in clinical practice, undoubtedly in part because of the lack of an established definition of intolerance. The UNIC (Unmet Needs in CML) study is a retrospective chart review of patients with CML conducted in 8 European countries. A recent report from the French cohort of the UNIC study classified 197 (31.4%) of 654 patients with CML in chronic phase (CML-CP) as intolerant to imatinib, based on toxicity leading to a change or discontinuation of imatinib use as reported in the medical chart.7

The definitions of intolerance from clinical studies, and guidance from the NCCN,5 use the Common Toxicity Criteria (CTC) to grade the severity of AEs. CTC criteria are generally designed to identify acute toxicity of a particular agent, and the value/measurement is often based on the severity and level of intervention needed. Because targeted therapies for CML involve long-term chronic therapy rather than a typical cyclic and finite chemotherapy program, the applicability of the CTC to assist in defining intolerance to these compounds has been called into question. For example, although a grade 2 or even grade 1 AE (eg, diarrhea or muscle cramps) may be acceptable when experienced over the course of a short-term treatment, it may become unacceptable when it becomes permanent. In addition, individual patients have different thresholds and levels of acceptability for different grades of AEs; therefore, this type of assessment may not be an appropriate measure for all patients.

Focusing on patient QOL may be more appropriate for the definition of intolerance. Direct measurement of QOL parameters may be useful to the clinician when deciding on whether to continue with 1 line of therapy despite AEs or to switch to an alternative therapy and for clarification of the tradeoffs between different management strategies.8 Validated instruments that have been used for QOL measurement in CML include the Functional Assessment of Cancer Therapy–Biologic Response Modifiers9 and the Euro Quality of Life–5 Dimensions (Fig. 1),10-12 although these tools are not specific to patients with CML.

Figure 1.

Quality of life (QOL) measures as applied to patients receiving imatinib therapy are shown. (Top) The Functional Assessment of Cancer Therapy–Biologic Response Modifiers (FACT-BRM)9 is a method used to assess patient QOL and is comprised of 40 questions distributed over 5 subcategories. Patients were asked about their physical well-being, functional well-being, social and family well-being, and emotional well-being and were also asked treatment-specific questions. Responses were measured on a 5-point scale ranging from 0 (not at all) to 4 (very much). Responses from the physical, functional, and treatment-specific subcategories were used to calculate the Trial Outcome Index (TOI; primary endpoint). The TOI is comprised of 27 questions with a possible score ranging from 0 to 108. Social and family well-being (7 items; range, 0-28) and emotional well-being (7 items; range, 0-24) were calculated as secondary endpoints. (Bottom) The Euro Quality of Life–5 Dimensions (EQ-5D)10 uses 2 measures of patient health. Patients were asked qualitative questions about 5 dimensions of their health and to rate their overall health that day compared with that over the past 12 months. They were then asked, on a scale of 0 to 100, to rate how they felt that day. The results ranged from 100 to −100, with scores <0 indicating a feeling of “worse than death.”

In the IRIS trial, 315 (58%) patients in the imatinib arm and 198 (38%) patients in the interferon-α (IFN-α) arm completed the Functional Assessment of Cancer Therapy–Biologic Response Modifiers and Euro Quality of Life–5 Dimensions questionnaires at baseline and at 8 additional time points during the first 12 months of treatment.10 The primary endpoint of the QOL analysis was the Trial Outcome Index, a measure of physical well-being and function. The secondary endpoints were social and family well-being (SFWB) and emotional well-being (EWB), and crossover was accounted for in the analysis. These QOL tools were meaningful to demonstrate differences between the imatinib and IFN-α arms. Overall, 41% of patients randomized to first-line imatinib in the IRIS trial had a clinically relevant improvement in the Trial Outcome Index from baseline to Month 12, compared with 10% of patients randomized to first-line IFN-α (P<.001).10 More specifically, patients randomized to imatinib reported better daily functioning and well-being, less fatigue, fewer emotional and cognitive problems, and fewer AEs. Of note, the rate of treatment discontinuation was significantly less on the imatinib arm. Patients receiving imatinib also had more favorable scores of SFWB and EWB. Data from the Euro Quality of Life–5 Dimensions index showed that patients on imatinib treatment were more likely to report health equivalent to perfect health (P<.001). Taken together, these data indicate that QOL assessments can be a reliable indicator of physical well-being in patients with CML and can be used to assess the relative impact of different therapies.10 However, no data on the long-term effects of imatinib on QOL are available.

Results from randomized controlled trials suggest that treatment decisions influenced by QOL considerations may be beneficial in some patients.8 In some therapeutic areas, symptom-specific rating scales have demonstrated value as tools for assessing the effects of an intervention on treatment-related symptoms. In a retrospective analysis of 106 lung cancer patients treated in 3 clinical trials, 15% of patients reported fatigue on the Lung Cancer Symptom Scale, but the fatigue was measured as grade 0 according to the CTC.13 In another recent study of 163 lung cancer patients receiving chemotherapy, patients and clinicians independently reported CTC AEs and Karnofsky Performance Status.14 In general, patients reported symptoms earlier and more frequently than clinicians. Clinician, but not patient, report of fatigue, nausea, constipation, and performance status was significantly associated with death and emergency room admissions. Patient-reported symptoms, however, had greater correlation with the Euro Quality of Life–5 Dimensions and a global health state questionnaire than clinician-reported symptoms. From these studies, it appears that clinicians may be better at recognizing AEs with potentially serious consequences, whereas patients may be better at assessing more subtle changes that affect their overall QOL. This finding may be due in part to the finding that the QOL is a more direct measure from patients, whereas the CTC are greatly influenced by the interpretation of the clinician. Input from both parties may be needed to adequately assess patient well-being. These findings suggest that symptom-specific evaluation of patients may identify AEs that are relevant to patient QOL but are not identified with the CTC criteria.

To best facilitate the application of QOL measurements as a management tool for patients with CML, a specific tool for QOL assessment with long-term TKI therapy is needed. One such tool, a CML-specific variant of the M. D. Anderson Symptom Inventory, is currently in development. Preliminary findings suggest that this tool is a reliable means of detecting numerous symptoms of CML and can differentiate between levels of severity of illness.15 This assessment tool is still in development and being refined to more accurately capture CML- and treatment-specific symptoms and awaits prospective validation.

Imatinib Intolerance: A New Definition

Although there is no standardized definition of imatinib intolerance, there are 3 groups of patients who may be considered intolerant to imatinib: patients who are truly intolerant to imatinib at a dose that would be considered acceptable for adequate efficacy (usually at least 300 mg/d), those who are resistant or do not have an adequate response to standard dose (400 mg/d) imatinib and are intolerant to dose escalation, and those who experience persistent low-grade AEs that do not prevent them from continuing therapy with imatinib but that may adversely impact function and QOL. In each case, the management of AEs using the recommended guidelines should be attempted before classifying a patient as truly intolerant (Table 2).

Table 2. Suggested Management of Adverse Events on Imatinib Therapy
Adverse EventEvent Management
  1. ANC indicates absolute neutrophil count; FDA, Food and Drug Administration; GI, gastrointestinal.

 NeutropeniaDose interruption is not recommended for grade 1-2 toxicity.
Grade 3-4 toxicity defined as ANC <1 × 109/L.
Dose interruption until ANC ≥1-1.5 × 109/L, then resume at 400 mg/d if recovery within 2 weeks.
In the event of recurrent grade 3-4 toxicity or delayed recovery, dose reduction to 300 mg.
Consider growth factors for patients with recurrent or persistent grade 3-4 toxicity.
 ThrombocytopeniaDose interruption is not recommended for grade 1-2 toxicity.
Grade 3-4 toxicity is defined as a platelet count <50 × 109/L.
Dose interruption until platelet count is ≥50-75 × 109/L, then resume at 400 mg/d if recovery within 2 weeks.
In the event of recurrent grade 3-4 toxicity or delayed recovery, dose reduction to 300 mg.
 AnemiaDose interruption is not indicated for toxicity of any grade.
Dose reductions may help in cases with chronic anemia.
Erythropoietin has shown efficacy, but is not recommended by the US FDA.
 Gastrointestinal toxicityAdministration with food is recommended to reduce upper GI tract toxicities.
Some patients may benefit from antiemetics.
Supportive care (eg, loperamide) for management of diarrhea.
 Edema/fluid retentionGrade 1-2 toxicities do not require dose interruption.
Diuretics and supportive care recommended for symptomatic patients.
Dose interruption, reduction, or discontinuation may be required in some patients.
 Muscle crampsAdministration of a calcium supplement and high fluid intake.
Quinine or tonic water.
 RashAdministration of topical or systemic steroids is effective in the majority of cases.
In severe cases, dose interruption, dose reductions, and occasionally treatment discontinuation may be required.
 Hepatic toxicityWith grade 2-3 toxicities, dose interruption until toxicity is grade 1, followed by resumption of therapy at a reduced dose.
Alternate therapies should be considered for grade 4 toxicity.

Of significant importance is the finding that the advent of alternative therapeutic strategies for CML has affected the perception of intolerance. Before TKI therapy, patients experienced high rates of toxicity, mainly in comparison with IFN-α16; however, given the lack of alternative therapies, this toxicity was managed or tolerated by patients. The introduction of the first TKI, imatinib, resulted in a treatment with a dramatically improved toxicity profile. Grade 3-4 AEs were uncommon, and persistent grade 2 or lower AEs were considered acceptable, particularly when the recent experience was with a toxic agent such as IFN-α. Furthermore, no adequate alternatives were initially available for most patients, even when considered intolerant. The subsequent approval of the second-line TKIs, nilotinib and dasatinib, provided alternative agents to overcome chronic toxicity encountered on imatinib, thus rendering the compensated low-grade toxicities less acceptable. The long-term impact on CML and QOL outcomes of switching therapy from imatinib because of mild or moderate AEs is currently unknown, but is being investigated in ongoing clinical studies.

Symptomatic Management of AEs on TKI Therapy

The identification of patients intolerant to imatinib, however defined, is desirable for several reasons, including that it may allow for management of the intolerance, thereby increasing the likelihood of staying on therapy and achieving optimal long-term outcomes for the patient. Patients with intolerance to imatinib may be treated with symptom-directed supportive care to reduce symptoms, can undergo dose reductions of imatinib, can be switched from imatinib to an approved second-generation TKI, or can potentially even consider allogeneic stem cell transplantation or other investigational options.

Management of a treatment-related AE using a symptomatic therapy and/or dose reduction should first be attempted to determine whether the intolerance could be managed without discontinuing therapy. Most AEs occur early with imatinib therapy, and de novo toxicity rates diminish over time. In the IRIS trial, AEs predominantly occurred in Years 1 to 3; rates of newly occurring grade 3-4 AEs were <2% after 4 years.17 In the phase 2 registration and expanded access trials of nilotinib after imatinib failure, the pattern of AEs over time was similar, with the most common AEs occurring within the first 2 months of therapy.18, 19 In the phase 2 registration trial for dasatinib, the rates of treatment-related AEs were only slightly higher at a median follow-up of 15 months than in the initial cohort at a minimum follow-up of 8 months; however, AEs generally occurred early in treatment in this study as well.4, 20

Some specific measures can be taken to manage some of the AEs most commonly observed while on therapy with TKI, many of which are outlined by the NCCN5 and summarized in Table 2. With regard to myelosuppression, grade 3-4 neutropenia or thrombocytopenia may be managed with transient dose interruption and resumption of imatinib at 400 mg/d. If severe events recur, dose interruption followed by resumption at a reduced dose of 300 mg/d may be appropriate, preferably in the presence of active and sustained disease remission. Anemia of any grade is typically managed via dose reduction, although sustained chronic anemia may be more problematic and could lead to long-term use of reduced imatinib doses that may not be appropriate in the absence of sustained disease remission. Nonhematologic events, the most common being gastrointestinal symptoms, fluid retention, cramps, and rash, are preferably managed via palliative care rather than reduction or interruption of the dose. However, in some cases of persistent low-grade fluid retention, severe rash, or hepatotoxicity, permanent discontinuation of imatinib may be required.

The approved initial treatment with imatinib is 400 mg/d.21 For patients in need of dose reductions to manage toxicity, doses <300 mg once daily should not be routinely used, as this was the threshold at which significant response occurred in the phase 1 study of imatinib.22 Lower doses have been associated with decreased probability of response whenever they have been tested.

Toxicity on Nilotinib Therapy

Nilotinib is approved throughout the world for the treatment of patients with Philadelphia chromosome-positive (Ph+) CML-CP and CML in accelerated phase (AP) who are resistant or intolerant to prior therapy, including imatinib. Nilotinib was designed as a more potent analogue of imatinib with improved affinity for breakpoint cluster region-Abelson murine leukemia (BCR-ABL) and a structure that makes it less susceptible to point mutations.23 Nilotinib is now approved in the United States for patients with newly diagnosed CML-CP.24-27

Given that the target and structure of nilotinib is shared with imatinib, the possibility of cross-intolerance has been investigated to determine whether AEs causing imatinib intolerance and treatment discontinuation would recur with nilotinib.28 A total of 122 patients (95 with CML-CP and 27 with CML-AP) were enrolled in the phase 2 nilotinib registration trial after imatinib intolerance.28 Only 1 (1%) of 75 patients with nonhematologic imatinib intolerance experienced a recurrence of similar grade 3-4 AEs during nilotinib therapy, and only 4 patients experienced grade 3-4 or persistent grade 2 AEs with nilotinib. No dose reduction or discontinuation was required for management of the AEs. Whereas 17 (43%) of 40 patients with hematologic intolerance to imatinib experienced recurrence of the same grade 3-4 AE, only 7 (18%) of these patients discontinued nilotinib for this reason. In all of these patients, thrombocytopenia was the reason for discontinuation of nilotinib. Thus, cross-intolerance to nilotinib is uncommon in imatinib- intolerant CML-CP and CML-AP patients, particularly for nonhematologic AEs.

In imatinib-resistant or -intolerant patients, rates of nilotinib discontinuation because of drug-related AEs were low.28 The most commonly observed (>10% all grades) drug-related nonhematologic AEs observed in clinical trials included rash, pruritus, nausea, vomiting, fatigue, and headache, most of which were 1 or 2.18, 29 Some patients complain about pruritus on the scalp while taking nilotinib. Myelosuppression was the most common reason for treatment discontinuation in the phase 2 registration trial of nilotinib, necessary in 8% of imatinib-resistant or -intolerant patients.18 Biochemical abnormalities were generally mild and easily managed.18 These included hyperglycemia, hypophosphatemia, and elevation of bilirubin (mostly among patients with Gilbert syndrome), lipase and amylase (rarely associated with clinical symptoms), and transaminases.

Management of these AEs was similar to that described for imatinib. Transient treatment interruption and possible dose reduction is recommended for patients with grade 3-4 elevation of transaminases or bilirubin. The majority of laboratory abnormalities and myelosuppression occurred in the first 3 months of nilotinib treatment; rates decreased thereafter.18

After a report from the Jefferson Medical College suggested that imatinib was associated with congestive heart failure in 10 patients,30 the potential cardiotoxicity of TKIs has been extensively studied. A review of 1276 patients treated with imatinib at The University of Texas M. D. Anderson Cancer Center demonstrated that such events are rare on imatinib, occurring in 1.7% of patients, most of whom had pre-existing cardiac conditions, with an incidence similar to that seen in the general population.31 Nilotinib has the potential to cause QTc prolongation, as do other TKIs, and the product information for this drug contains a box warning about this condition. Before the start of therapy, patients should undergo electrocardiography to investigate pre-existing prolonged QTc, and electrolyte (eg, potassium and magnesium) abnormalities should be corrected and monitored throughout the course of therapy. Fortunately, few patients experience significant prolongations of the QTc interval, and only rarely does this condition have clinical consequences.

In the frontline setting, nilotinib has been shown to be well tolerated.24-27 In fact, as is the case with all TKIs, the rate of AEs is lower among patients treated in the frontline setting than those treated after failure with other therapies. ENESTnd (Evaluating Nilotinib Efficacy and Safety in Clinical Trials–Newly Diagnosed Patients) has provided the first randomized comparison of a second-generation TKI (in this case 300 or 400 mg twice daily nilotinib) versus the current standard of care (400 mg/d imatinib).24 At 12 months, nilotinib was shown to be superior to imatinib as a first-line treatment for CML (higher rates of major molecular response and complete cytogenetic response, deeper levels of molecular response, and significantly lower rates of progression to accelerated phase/blast crisis [BC]). The rate of discontinuation because of AEs was similar across arms, but somewhat lower for nilotinib at 300 mg/d than for nilotinib 400 mg/d or imatinib (7%, 11%, and 9%, respectively). In this large, randomized phase 3 trial, which required prospective cardiac monitoring, the incidence of clinically relevant QT prolongation was uncommon, with no patient in either arm having QT prolongation >500 ms, and <1% of patients in each nilotinib arm experiencing QT prolongation from baseline of >60 ms.

Similarly, high response rates and favorable toxicity have been observed in single-arm prospective studies of nilotinib in newly diagnosed patients.25-27 The most common AE with nilotinib in this setting was, again, myelosuppression; however, it was usually seen in the first 2 to 3 months of therapy, it was transient, it most frequently did not require dose adjustments, and it was usually grade 1-2. For example, in the single-arm study conducted at The University of Texas M. D. Anderson Cancer Center, anemia, thrombocytopenia, and neutropenia were observed in 49%, 43%, and 52% of patients, respectively; however, these AEs were grade 3-4 in only 5%, 11%, and 12% of patients, respectively.27 In general, AEs occurred at higher rates in the first 3 months of nilotinib therapy and decreased over time.26

Toxicity During Dasatinib Therapy

Dasatinib is a highly potent inhibitor of BCR-ABL, which also targets Src family kinases. Dasatinib is currently approved throughout the world for the treatment of patients with CML-CP, -AP, or -BC (and Ph+ acute lymphoblastic leukemia) who are resistant or intolerant to prior therapy, including imatinib. Despite the finding that dasatinib is structurally distinct from imatinib, investigation into cross-intolerance in dasatinib-treated patients previously intolerant to imatinib was warranted, given its potential use in patients with imatinib intolerance as well as its shared target, BCR-ABL, and other off-target effects. A total of 271 patients with imatinib-intolerant CML-CP were eligible for analysis.32 Imatinib intolerance was related to hematologic toxicity in 46 patients and nonhematologic toxicity in 225 patients. Cross-intolerance was uncommon, particularly in regard to nonhematologic toxicity. Only 9 (4%) patients experienced the same grade 3-4 nonhematologic AEs on dasatinib as on imatinib, and only 2 (1%) patients discontinued dasatinib as a consequence. Recurrence of the same grade 3-4 hematologic toxicity that led to imatinib intolerance was observed in 86% of patients, but only 6 (13%) patients discontinued dasatinib because of the same hematologic toxicity that had resulted in discontinuation from imatinib therapy. This analysis reflects the use of the initially approved dasatinib dose of 70 mg twice daily. However, a randomized trial investigating different doses and treatment schedules has demonstrated that 100 mg once daily is associated with fewer AEs (particularly myelosuppression), while maintaining efficacy.33 Thus, it is possible that even fewer patients would experience cross-intolerance at this dosing schedule.

In the treatment of patients who have experienced imatinib resistance or intolerance, dasatinib at 70 mg twice daily was associated with the occurrence of generally low-grade (1-2) nonhematologic AEs, the most frequent of which were fluid retention (including pleural and pericardial effusions), dyspnea, infection, headache, diarrhea, fatigue, and abdominal and musculoskeletal pain.20, 34 However, grade 3-4 myelosuppression was reported in nearly 50% of patients.35 In an effort to decrease toxicity, a phase 3, open-label, dose-optimization study was performed in which patients were randomized 1:1:1:1 to dasatinib 100 mg once daily, 50 mg twice daily, 140 mg once daily, or 70 mg twice daily.33 Treatment with 100 mg/d resulted in lower rates of discontinuation because of drug toxicity (11%) than treatment with 70 mg twice daily (21%). The most common nonhematologic AEs were nausea, headache, diarrhea, and fatigue. Treatment at 100 mg daily resulted in significantly reduced rates of grade 3-4 nonhematologic and hematologic AEs. Grade 3-4 pleural effusions decreased from 6% or more in earlier trials20 to 2% with 100 mg daily.33 Currently, 100 mg once daily is the standard dose of dasatinib for CML-CP.36

Some AEs that have been particularly noticed among patients treated with dasatinib include pleural effusion and bleeding. All grade pleural effusion has been reported in 10% to 36% of patients,20, 37-39 and more frequently among patients treated in the advanced stages. Many (54%-100%) pleural effusions observed are grade 1-2 in nature. However, it is important to note that grade 2 pleural effusions are symptomatic and require interventions such as diuretics or up to 2 thoracenteses. Grade 3-4 pleural effusions occur in 5% to 15% of patients and may require treatment interruption, dose reduction, permanent discontinuation, the use of interventions such as diuretics or corticosteroids, and >2 thoracenteses, chest tube draining procedures, or chemical or surgical pleurodesis.40 Bleeding episodes can occur in 20% to 25% of patients treated with dasatinib, but are grade 3-4 in only 10% of cases.41 Patients who develop bleeding while on dasatinib should discontinue therapy immediately. Bleeding frequently occurs in the setting of thrombocytopenia, but in approximately 30% of cases, bleeding occurs with adequate platelet counts.41 Recently, it has been reported that dasatinib may inhibit platelet aggregation.42 Thus, concomitant administration of drugs that inhibit platelet function, such as nonsteroidal anti-inflammatory agents, is not recommended.

Dasatinib is also currently being investigated as a frontline agent in patients with newly diagnosed CML. In a phase 2 study of dasatinib (100 mg once daily or 50 mg twice daily) in patients with newly diagnosed early CML-CP,43 rates of grade 3-4 nonhematologic AEs were low and included fatigue (6%), joint and muscle pain (6%), peripheral neuropathy (5%), dyspnea (5%), and memory impairment (5%). Most events were transient and manageable. Pleural effusions occurred in 13% of patients, with only 2% being grade 3-4 in severity. The majority of patients experienced some degree of myelosuppression on dasatinib (anemia [81%], thrombocytopenia [69%], neutropenia [63%]); grade 3-4 events occurred in 6%, 10%, and 21% of patients, respectively.43

The DASISION trial is a randomized comparison of dasatinib 100 mg once daily versus the current standard of care (400 mg/d imatinib).39 At 12 months, dasatinib was shown to be superior to imatinib as a first-line treatment for CML (higher rates of major molecular response and complete cytogenetic response, and numerically lower rates of progression to AP/BC). Pleural effusions occurred in 10% of patients treated with dasatinib, and all were grade 1 (2%) or grade 2 (8%). According to the published protocol, prospective baseline and 6-month chest x-rays were mandated, which may have decreased the severity of pleural effusions observed in the first 12 months of this study compared with prior studies. Overall, the rate of discontinuation because of AEs was slightly higher for dasatinib (5%) versus imatinib (4%). The incidence of clinically relevant QT prolongation was uncommon, with only 1 patient in either arm having QT prolongation >500 ms, and 5% of patients in the dasatinib and imatinib arms experiencing QT prolongation from baseline of >60 ms.


Although imatinib is generally well tolerated, and the majority of AEs are of mild or moderate intensity, a substantial minority of patients experience AEs that require dose interruptions or withdrawal from first-line therapy in addition to the commonly recognized group of patients for whom imatinib is not an effective long-term therapy. This niche has benefitted from the development of 2 approved alternatives and may benefit from further drug development.

Chronic low-grade toxicities are emerging as a common concern for patients receiving therapy for several years, possibly indefinitely. Patients receiving nilotinib and dasatinib appear to have lower rates of AEs, particularly those of higher grade. The follow-up of patients receiving these agents is shorter, which thus prevents a full assessment of the prevalence of chronic, low-grade events. Optimal symptomatic management of modest imatinib-related toxicities remains the preferred option before classifying a patient as imatinib-intolerant and to switching therapy.

A consistent definition and algorithm for clinicians to follow with respect to intolerance may improve identification of patients who are truly intolerant to imatinib, ensuring that patients are on the most appropriate treatment. Definitions of imatinib intolerance in clinical trials are based on the CTC criteria for defining the severity of AEs, which may not be appropriate for the long-term management of CML in clinical practice, because they may underestimate the impact of low-grade toxicities when experienced over a lifetime. Disease and symptom-specific assessment tools and QOL assessments may provide valuable insight into the impact of AEs on a patient's daily life and hence support treatment decisions, particularly in the area of mild chronic toxicity deemed unacceptable when treatment is indefinite—a paradigm somewhat unique to TKI use in CML. The definition of intolerance is inevitably relative to the alternative treatment options available and, therefore, continues to evolve with the emergence of new data and novel therapies. Thus, our tolerance for intolerance in practice has shifted in relation to available alternatives. Reassuringly, cross-intolerance with imatinib is minimal for imatinib-intolerant patients switching to a second-generation TKI such as nilotinib or dasatinib, although there may be some sharing of AEs, particularly hematologic, in a minority of patients. Trials of second-generation TKIs in previously untreated patients with newly diagnosed CML are ongoing, and the data obtained to date indicate a favorable safety and tolerability profile. The availability of new treatment options has accordingly further changed the landscape of CML treatment and brings the issue of tolerability appropriately more to the forefront. Given the availability of alternate therapies and the consequent evolving definition of intolerance to imatinib, more investigation into the role of QOL in managing patients with CML and continued focus on long-term safety data with newer agents is therefore needed to complete the picture with optimization of tolerability/toxicity to complement the strides made in the improvement of response and survival.


We thank Erinn Goldman, PhD (Articulate Science) for medical editorial assistance with this article.


J.P.-I. has served as a consultant/advisor to Novartis and Bristol-Myers Squibb, has received honoraria from Novartis, and has received research funding from Novartis and Bristol-Myers Squibb. J.C. has received research funding from Novartis, Wyeth, and Bristol-Myers Squibb. M.J.M. has received research funding from Novartis and Bristol-Myers Squibb. Financial support for medical editorial assistance was provided by Novartis Pharmaceuticals.