Prof. Philippe Travade, head of the Haematology Department of Clermont-Ferrand University Hospital died before submission of this paper. The authors would like to honor his memory and his invaluable role in the French Cooperative Group for CLL.
Department of Haematology, University Hospital, Clermont-Ferrand
Bruno Cazin, Service des Maladies du Sang, Hôpital Huriez, rue Michel Polonovski, CHU Lille, 59037 Lille Cedex, France. E-mail: email@example.com
A multicentre single-arm study testing the efficacy and toxicity of the oral combination of fludarabine and cyclophosphamide (FC) over 5 d in 75 patients with untreated B cell-chronic lymphocytic leukaemia. Oral FC demonstrated high efficacy with overall (OR) and complete response (CR) rates of 80% and 53%, respectively. Out of the 30 CR patients studied for Minimal Residual Disease (MRD) using 4-colour flow-cytometry and the 22 using Clonospecific polymerase chain reaction, 22 (66%) and 16 (68%), respectively, were MRD negative. Median survival and median treatment-free interval had not been reached at 7 years of follow-up. Median progression-free survival (PFS) was 5 years. Toxicity was acceptable, with 52% and 16% of National Cancer Institute grade 3/4 neutropenia and infections, respectively. Gastrointestinal toxicity was mild. Oral FC demonstrated a high efficacy and an acceptable safety profile and may be considered as the standard first line treatment in chronic lymphocytic leukaemia.
The introduction of purine nucleoside analogues for treatment of chronic lymphocytic leukaemia (CLL) led to a higher remission rate and a longer progression-free survival (PFS) compared to a regimen containing alkylating agents (Rai et al, 2000). Although intravenous (i.v.) formulation of fludarabine is possibly the most active single agent in CLL (Keating et al, 1993; Steurer et al, 2006), the high response rate observed had no impact on survival in randomized studies in first line of therapy, mainly because of treatment cross-over (Leporrier et al, 2001).
An oral formulation, comprising an immediate-release tablet containing 10 mg fludarabine phosphate has been developed recently. Pharmacokinetic study of a single dose in patients with B-cell CLL (B-CLL) and non-Hodgkin lymphoma showed that the 24-h area under the concentration-time curve was similar to that following i.v. administration, after dose adjustment. The bioavailability was approximately 60% and remained unaffected by food intake (Oscier et al, 2001). A single oral daily dose of 40 mg/m2 might therefore provide systemic exposure equivalent to 25 mg/m2 i.v. The efficacy of oral fludarabine did not differ from its i.v. formulation and its safety profile was similar (Boogaerts et al, 2001; Rossi et al, 2004).
Fludarabine inhibits the repair of DNA damage caused by agents such as mitoxantrone and cyclophosphamide. A synergistic effect has been demonstrated between fludarabine and cyclophosphamide (Bellosillo et al, 1999). Intravenous combination of fludarabine with cyclophosphamide (FC) has demonstrated interesting results in CLL (O’Brien et al, 2001). Two recent randomized trials confirmed that FC was the new front line gold standard in CLL (Catovsky et al, 2006,Eichhorst et al, 2006). Cyclophosphamide is currently used in a 5-d schedule in the ChOP (cyclophosphamide, doxorubicin, prednisone, vincristine) regimen, which has given significant results (Leporrier et al, 2001). This 5-d schedule was implemented in our oral FC regimen for convenience, and in order to limit the number of tablets to be taken daily.
Despite the introduction of several new therapeutic strategies in CLL over the last three decades, no improvement in survival has been demonstrated in randomized clinical trials. Experience suggests that overall survival does not improve until the majority of patients achieve a complete response (CR). The objective of this study was to assess the efficacy and safety of oral FC as a first line treatment in CLL and demonstrate that the increase in CR rate will translate to longer survival.
Patients and methods
Between October 1999 and February 2001, a multicentre, single-arm phase II clinical trial of the combination of oral FC was performed in patients with previously untreated CLL (Cheson et al, 1996). Patients were required to be less than 66 years old, Binet’s stage B or C, with a performance status <3 and life expectancy >3 months. Good end-organ function (kidney and hepatic) was required. Human immunodeficiency virus (HIV)-positive patients, patient with previous history or concomitant malignancies, autoimmune thrombocytopenia or autoimmune hemolytic anemia were not eligible. All participants signed an informed consent form before they entered the study. The study protocol was approved by the institutional review board of Lille University Hospital.
Cyclophosphamide 200 mg/m2 per day and fludarabine (FLUDARA® 10 mg, Bayer Healthcare, Bayer-Schering Pharma, Berlin, Germany) 30 mg/m2 per day were both administered orally on days 1 to 5, repeated every 28 d for six cycles. Courses were delayed in case of haematological toxicity as described below. Herpes/Zoster Virus prophylaxis with valaciclovir and pneumocystis carinii prophylaxis with trimethoprim/sulfamethoxazole were recommended during treatment and for at least two additional months after the last dose of chemotherapy. Therapeutic support with antibiotics, antiemetics or growth factors could be used at the discretion of the physician.
Assessment of safety and dosage adjustment
Toxicity was graded according to the National Cancer Institute (NCI) Working Group for CLL protocols criteria (Cheson et al, 1996). Doses of both fludarabine and cyclophosphamide were reduced for patients with a calculated creatinine clearance between 0·5 and 1·17 ml/s, according to the Cockroft formula. Courses were delayed if the neutrophil count was <1 × 109/l or the platelet count was <50 × 109/l. The delay could not exceed two weeks. If toxicity ≥ grade 3 was observed at this date, the doses of fludarabine and cyclophosphamide were reduced, to 25 mg/m2 per day and 150 mg/m2 per day, respectively, ×5 d, and patients were withdrawn in case of grade 4 haematological toxicity.
Staging and assessment of efficacy
Response was assessed according to NCI Working Group criteria (Cheson et al, 1996) at 2 months after the last cycle. After completion of therapy, patients were re-evaluated at 3-month intervals until month 6, then every 6 months. All patient files were reviewed by two independent experts for inclusion and response criteria. In case of persistent cytopenia, reassessment of the response was done within the 6 months following the last course. Patients withdrawn before cycle 3 were considered as failures.
Minimal residual disease assessment
In case of CR, minimal residual disease (MRD) evaluation was performed in the peripheral blood, using flow cytometry and polymerase chain reaction (PCR) evaluation. Blood flow cytometry analysis was carried out by the Laboratory of Hematology of Pitié Salpêtrière Hospital in Paris in order to detect residual CLL cells in the blood as described (Maloum et al, 2002). Briefly, we used a 4-color flow cytometry with anti-CD19, -CD20, -CD5 and -CD79b antibodies. The sensitivity of this technique is 10−4. MRD was also assessed by IGHV gene PCR analysis of the CDR3 region with a sensitivity of 10−2. In the cases in which CDR3 amplification was found to be polyclonal, this region was sequenced as described (Magnac et al, 1999) with a sensitivity of 10−6.
The study was designed as a single-arm, open-label trial with group sequential design based on the triangular test (Bellissant et al, 1990). The purpose was to compare the observed CR rate in our study to the reference rate of 20% observed with fludarabine i.v. alone (Keating et al, 1993). The design allowed for analysis every 15 patients to detect a 15% absolute increase in the CR rate from 20 to 35% with a 5% one-sided type I error and 90% study power. The recruitment was stopped following the second intermediate analysis, because an increased response rate with the oral FC was obtained with 76 patients already enrolled. Overall survival (OS), calculated from inclusion to death, progression-free survival (PFS), calculated from inclusion to time of disease progression or death, and treatment-free survival (TFS) from end of therapy to time of second-line treatment or death, were computed using Kaplan-Meier estimates and reported as median (range).
Seventy-six patients were recruited in 15 French centres. One had autoimmune hemolytic anemia before the beginning of the treatment and was excluded from the study. Seventy-five patients were evaluable for response and safety.
The median age at the time of inclusion was 57 years (37–66 years), comparable to other studies (O’Brien et al, 2001; Eichhorst et al, 2006; Flinn et al, 2007). Sixty-two patients (83%) were male. Fifty-nine patients (79%) were at Binet stage B and 16 (21%) at stage C. An average of 5·3 cycles (range 1–6) was administrated per patient. Mutational status of IGHV (Hamblin et al, 1999) was studied on frozen blood samples in 42 patients, 35/40 CR patients and 7/20 PR patients, but not in non-responders. Sixty-four % of the patients tested were unmutated for IGHV, 25 CR and 2 PR, respectively.
The OR rate reached 80%, with 40 patients (53%) in CR (Table I). Median survival (range 0–8 years) had not been reached at 7 years (84 months) of follow up (Fig 1A). The 7-year probability of survival was 71% and 36% in responders and non-responders, respectively (P = 0·0027, Fig 1B). The median PFS was 5 years (range 1–8) (Fig 1C). The median TFS was not reached (range 1–7 years) (not shown).
Table I. Response after six courses (n = 75 patients).
*One patient had Richter syndrome.
†Includes two deaths, one unexplained early after cycle 1, one in neutropenia after cycle 5. Five patients were considered as failures: one because of the lack of follow-up at time of evaluation after six cycles, with prolonged survival 82 months+, four were withdrawn early before three cycles mainly because of hematological toxicity: one with prolonged neutropenia who received further i.v. FC with good response and prolonged survival 95 months+, one with prolonged pancytopenia, and two immune cytopenias. All had important tumor reduction at the time of therapy discontination and could be considered in PR except for cytopenia.
Nodular partial response
Oral FC was a highly effective and well tolerated regimen, although quality of life was not prospectively studied (Table II). Ninety-one percent of the cycles were administrated without delay and 97% were given at the full dose. Toxicity included myelosuppression and infections, with frequent neutropenia, but was acceptable with 52% and 16% of NCI grade 3/4 neutropenia and infections, respectively. Only one patient was withdrawn because of grade 4 neutropenia and two early deaths might be related to infection in neutropenia but were not documented. Few opportunistic infections were observed during this study, within the limits of 6 months of assessment for adverse events. Infections may have been minimized by the use of prophylaxis. Compliance to the oral formulation was excellent with acceptable gastrointestinal side effects, generally limited to grade 1–2 severity.
Table II. Grade 3 and 4 toxicity (corresponding percentage of the 401 courses).
Infections included one aspergillosis, one hepatitis B reactivation and one Pneumocystis carinii pneumopathy (PCP). Four courses were delayed and six treatments discontinued because of infections.
*Only one patient discontinued the treatment due to gastrointestinal toxicity after cycle 5.
†Seven patients developed hemolytic anemia of which five during therapy resulting in treatment discontinuation. Three cases were associated with thrombocytopenia. All responded to steroids. Three of them relapsed early of their autoimmune cytopenia and resolved definitively after splenectomy.
Seven patients (9, 3%) developed hemolytic anemia. Five of these cases were considered to be autoimmune with positive direct antibody test. Five cases developed during therapy, resulting in treatment discontinuation. Three cases were associated with thrombocytopenia. All the patients responded to steroids. Three of them relapsed early of their autoimmune cytopenia and resolved definitively after splenectomy. Moderate tumor lysis syndrome was only observed in two patients.
Study of MRD
Thirty patients and 22 patients in CR were evaluated for blood MRD using 4-colour flow-cytometry (immunophenotypic remission) and using clonospecific PCR, respectively. Twenty (66%) and 16 (68%) were MRD negative using these two techniques, respectively.
Our data confirmed that the combination of fludarabine and cyclophosphamide is a highly effective and well tolerated regimen. Compliance to the oral formulation was excellent with acceptable gastrointestinal side effects, however, the impact of vomiting on drug absorption and the risk to compromise serum levels lead us to recommend an antiemetic prophylaxis.
The advantages of the oral formulation include convenience, for both patients and health care professionals and avoidance of hospitalization, thereby reducing need for transportation and overall costs, which ultimately lead to better quality of life.
The observed toxicity was mainly hematological with frequent neutropenia and a relatively high occurrence of autoimmune cytopenia. This relatively high occurrence of immune cytopenia has been described with variable incidence, from 5% (Dearden et al, 2007) to 6·5% (Borthakur et al, 2007). In this study, infections might have been minimized by the use of prophylactic trimethoprim/sulfamethoxazole and valaciclovir. Although 97% of patients in our study received the full dose and frequent grade 3 and 4 neutropenia was observed, few bacterial and fungal infections were encountered within the limits of 6 months of assessment for adverse events (Eichhorst et al, 2007). Infectious complications were less frequent than after the combination of fludarabine and chlorambucil (Morrison et al, 2001).
Despite the fact that response was assessed with stringent criteria following NCI guidelines (Cheson et al, 1996), including bone marrow biopsy and computed tomography scan, the response rate and CR rate of this oral combination was remarkable, and is one of the highest published in first line therapy with chemotherapy alone. These good results may be due to the treatment dose, corresponding to 30 m2/d of i.v. fludarabine over 3 d and 300 m2/d of i.v. cyclophosphamide as initially published by O’Brien et al (2001), who reported a 35% CR rate. Doses and schedules of fludarabine and cyclophosphamide have a broad spectrum in published studies and may influence the CR rate (Catovsky et al, 2006; Eichhorst et al, 2006; Flinn et al, 2007). The 5-d schedule, which is currently used in the ChOP regimen for cyclophosphamide, may ensure a longer exposure for both drugs when administered orally.
The median PFS was 5 years (60 months) and was comparable to that observed by the German CLL Study Group (GCLLSG) after FC (48 months) (Eichhorst et al, 2006) and superior to that reported by Flinn et al (2007) (31·6 months). The treatment-free interval (median not reached at 44 months) was higher than in the GCLLSG study (37 months) (Eichhorst et al, 2006). We believe that dosage and duration of treatment is of critical importance in order to obtain long duration of survival and greater outcome, and that our 5-d regimen, together with a high percentage of patients able to pursue FC treatment at full dose might have impacted on the survival and CR rates in our study.
Evaluation of MRD demonstrated a substantial reduction of the tumor mass, which represents the goal of an optimized front line treatment today. Indeed, two-thirds of evaluated CR patients achieved phenotypic remission with a high rate of molecular response detected by clonospecific PCR.
Despite a high remission rate and long disease-free survival, chemotherapy alone cannot cure CLL and almost all patients will experience a relapse. Complementary treatments aiming at eradicating MRD could lead to improved outcomes. Combination or consolidation with monoclonal antibodies are promising for acheiving higher response rates, eradication of MRD and prolonged DFS. The combination of FC with rituximab (FCR) has shown promising results (Keating et al, 2002) and is currently being assessed in a randomized international trial. Concurrent treatment seems better than subsequent use (Byrd et al, 2006). Alemtuzumab (Campath-1H) is effective for the eradication of MRD after good initial response (Montillo et al, 2006; Schweighofer et al, 2006). There is controversy, however, regarding the increased risk of infection in patients treated with alemtuzumab (Wendtner et al, 2004; Lin et al, 2007). Different doses and schedules are currently being studied. The combination of fludarabine with alemtuzumab has demonstrated a good tolerance and an appealing efficacy (Elter et al, 2005). A randomized trial comparing oral FC either with rituximab or subcutaneous alemtuzumab is ongoing in first line CLL, coordinated by the Groupe Ouest Est d’Etude des Leucémies et Autres Maladies du Sang (GOELAMS) in collaboration with the French cooperative group for CLL.
The 5-d oral fludarabine and cyclophosphamide regimen demonstrated a very high efficacy and an acceptable safety profile in younger CLL patients, with an extended disease-free survival and treatment-free interval. The good results of this 5-d combination therapy with oral fludarabine and cyclophosphamide indicate that this regimen should be included among standard therapies in the front line treatment of CLL.
Dominique BRON and Jean François ROSSI who reviewed inclusion and final response criteria, to Grégoire LEMAIRE and Eric GAUTHIER of Schering SA.
Conflict of interest
Financial support by Bayer Healthcare, Bayer-Schering Pharma France.