Abstract
Bendamustine is an established treatment option in chronic lymphocytic leukemia (CLL) and frequently used in Austria and Italy. Therefore, we analyzed 100 unselected, consecutive patients with CLL (treatment-naïve and relapsed/refractory) receiving bendamustine in a real-life setting. Most patients were treated with bendamustine in combination with rituximab (BR). However, bendamustine monotherapy was additionally evaluated. Patients treated with BR had a significantly higher overall response rate of 76% (complete response=22%) when compared to those treated solely with bendamustine (overall response rate=50%; complete response=13%). Overall survival (OS) and progression -ree survival (PFS) were significantly lower in the bendamustine-treated group (OS=14.3 months; PFS=8.3 months) compared to the BR group (OS=42.7; PFS=22.5 months; both p<0.001). In multivariate analysis, patients with a good cytogenetic risk and those receiving BR had a significantly better OS. Grade 3/4 hematological complications were seen in 32% of the patients. Hence, bendamustine, especially in combination with rituximab, is an effective therapy with manageable toxicity for non-selected patients with CLL including those pre-treated with fludarabine and the elderly.
Chronic lymphocytic leukemia (CLL) is the most common type of adult leukemia in industrialized countries (1, 2). Despite the advent of new drugs (3, 4), it still remains an incurable disease, except for the few patients who are able to undergo allogeneic stem-cell transplantation (5-8). Only recently, bendamustine, first synthesized in the early 1960s, was to become an efficient treatment for hematological malignancies and it was approved for rituximab-refractory indolent lymphoma (9), CLL (10) and multiple myeloma (11).
Up until now, the use of bendamustine in CLL was mainly investigated in clinical trials in combination with rituximab (12). Current guidelines recommend a bendamustine-based therapy in patients with Binet stage C or Binet stage A and B with symptomatic disease and notably physically non-fit patients (11, 13, 14). Only a few real-life CLL populations, which were either treated with bendamustine monotherapy or in combination with rituximab, have been analyzed retrospectively (2, 13). Therefore, we herein present data on the toxicity and therapeutic efficacy in 100 unselected, consecutive patients with treatment-naïve or relapsed/refractory CLL treated with bendamustine-based therapy.
The objectives of the present study were to provide a descriptive analysis of a CLL population who received bendamustine-based therapy to evaluate patients' characteristics, including the spectrum of bendamustine-based therapy combinations used, and efficacy in terms of response and survival; as well as to determine toxicity in routine clinical use.
Patients and Methods
We retrospectively assessed 100 unselected, consecutive patients with CLL who received bendamustine either as monotherapy (n=24) or in combination with rituximab (n=76). The study was conducted according to the rules of the Medical University of Innsbruck Ethics Committee, as reported in other studies (15-17).
Progression-free survival according to treatment modality. a: The median PFS for patients treated with bendamustine and rituximab (BR) was 22.5 months (n=76) compared to 8.3 months for patients treated with bendamustine monotherapy (n=24; p<0.001). b: The median PFS for treatment-naïve patients on BR was 27.3 months (n=21) and for pre-treated patients was 17.7 months (n=55), while that of patients on bendamustine monotherapy was 8.3 months (n=24). When comparing all groups together the p-value was 0.001 (bendamustine vs. BR pre-treated [p=0.010], bendamustine vs. BR not pre-treated [p=0.001], BR pre-treated vs. BR not pre-treated [p=0.120]).
Data concerning the therapeutic regimen, efficacy, toxicity and follow-up were obtained directly from the clinical charts and from primary physicians. Data for 26 patients from a previous evaluation were updated and included in the study (18). The final data update was in September 2013. All analyses were performed using SPSS software version 20 (IBM Corp. Armonk, NY, USA).
Complete remission (CR), partial remission (PR), stable disease (SD) and progressive disease (PD) were classified according to the current International Work Group on Chronic Lymphocytic Leukemia guidelines for CLL (iWCLL) (7). Treatment-related adverse events were classified according to the Common Toxicity Criteria (CTC v.4) (19). Survival curves were plotted according to the Kaplan–Meier method. Progression-free survival (PFS) was defined as the time from the initiation of bendamustine-based therapy until disease progression (PD) or death, whichever occurred first. Patients, who were lost to follow-up without any sign of progression, or were alive at the final data update, were censored at the time of last observation. Overall survival (OS) was measured from start of bendamustine-based therapy. Univariate survival comparisons between categorical variables were evaluated with the log-rank test. All parameters with a p-value of 0.05 or less in univariate analyses were included in a Cox proportional hazard model. The p-value of 0.05 or less was considered as significant in two-sided tests.
A subset of 41 patients' assessments of co-morbidity according to the cumulative illness rating scale (CIRS) was available. The patient fitness groups according to CIRS were defined by the Alberta Health Services clinical practice guideline LYHE-007 (20).
Results
Patients' characteristics and mode of therapy. Forty-five patients (45%) were treated in Tyrol (Austria), and 31 (31%) and 24 (24%) in Bolzano and Messina (Italy), respectively. Bendamustine-based therapy was initiated between August 2007 and April 2013. Detailed patient characteristics are shown in Table I. The administered bendamustine dose ranged between 60 mg/m2 and 120 mg/m2 body surface area on day 1 and 2 every 4 weeks. Rituximab was usually administered on day one of each cycle, mostly at the recommended dose of 500 mg/m2. The median number of cycles administered was 5 in the BR group and 3.5 cycles in the bendamustine-treated group.
Bendamustine was administered as monotherapy (n=24, 24%) or in combination with rituximab (n=76, 76%). A total of 64 patients were male (64%). The median age was 73 years (range=41-88 years). The majority of the cases were treated in a higher therapy line (76%) and presented with Rai stages III and IV (27% and 40%, respectively), and were characterized by CIRS scores over 6 [n=32/41 evaluable patients (78%), median CIRS = 9].
Cytogenetic status, by fluorescence in situ hybridization (FISH) was available in 81 patients. According to the algorithm of Döhner et al. (21), 44% of patients were high-risk, harboring 11q deletion (n=15) or 17p deletion (n=20), meaning that patients with these cytogenetic aberrations usually respond very poorly to conventional chemotherapy and have inferior survival. Most of the patients had a WHO performance status of 0-1 (87%) and presented without B-symptoms (61%) at the beginning of the treatment. Forty-three patients with CLL (43%) had previously received fludarabine, of whom seven (16%) had fludarabine-refractory disease according to the criteria proposed by Keating et al. (22). Bendamustine alone or in association with rituximab was administered in first, second and higher line of therapy in 24 (24%), 27 (27%) and 49 (49%) patients, with a median of one (range=0-7) previous therapy.
Overall survival (OS) according to treatment modality. a: The median OS in patients treated with bendamustine and rituximab (BR) was 42.7 months (n=76), compared to 14.3 months for those treated with bendamustine monotherapy (n=24; p<0.001). b: The median OS for treatment-naïve patients on BR was not reached (n=21) and for pre-treated patients on BR was 35.8 months (n=55), while that for patients on bendamustine monotherapy was 14.3 months (n=24). When comparing all groups together the p-value was lower than 0.001 (bendamustine vs. BR pre-treated [p=0.001], bendamustine vs. BR not pre-treated [p<0.001], BR pre-treated vs. BR not pre-treated [p=0.050]).
Response to therapy. Sufficient data for response assessment information were available for 87 patients. In 13 patients, response was not evaluable due to premature death (n=6), therapy switch (n=1), premature discontinuation due to toxicity (n=2), or due to loss to follow-up (n=4) (as shown in Table II). A response according to iWCLL criteria was achieved in 70 patients (ORR=70%) with CR in 20 (20%) and PR in 50 (50%). SD and PD, under therapy or within 2 months after the last dose of bendamustine, was observed in six (6%) and 11 (11%) cases, respectively. For response analysis, we divided the cohort into two subgroups, the group who received bendamustine in combination with rituximab and those who received solely bendamustine. Interestingly, patients receiving BR were characterized by significantly higher response, with ORR achieved in 58 patients (76%) (CR in 17 [22%]) than patients receiving solely bendamustine, with ORR in 12 (50%) patients, (CR in 3 [13%]) (p=0.014).
In the BR group, the most important factors determining ORR were cytogenetic risk category (p=0.004), the Rai staging (p=0.042) and the cumulative dose of bendamustine (p=0.005). Furthermore, patients over the age of 75 years, as well as the ones with a CIRS score greater than 6 and those with GFR less than 60 ml/min did not have any disadvantages in ORR compared to younger patients, those with a lower CIRS score and those with higher GFR. The bendamustine-treated cohort showed similar results, but some significant results concerning β2-microglobulin levels and CIRS score have to be regarded carefully, due to their one-sided distribution (shown in Table III).
Progression-free survival. The overall median PFS was 17.7 months. For the whole cohort, PFS was influenced by clinical characteristics and therapy-related factors, namely the FISH karyotype risk-group (p<0.001), type of bendamustine therapy (p<0.001), quality of response (p<0.001) and cumulative dose of applied bendamustine (p=0.005). The cohort treated with BR regimen reached a median PFS of 22.5 months, which is significantly longer than the median PFS of the cohort treated with bendamustine monotherapy (8.3 months; p<0.001; Figure 1a). When analyzing therapy line-related sub-groups, treatment-naïve patients (n=21) in the BR cohort had a tendency for a longer median PFS (27.3 months) than pre-treated patients in this cohort (n=55; PFS=17.7 months; p=0.120). Due to a relatively small cohort treated with first-line bendamustine monotherapy (n=3), this subgroup analysis was not included in the calculation (Figure 1b). The whole cohort PFS analyses are shown in Table IV. The statistically significant analyses for the whole cohort were also performed for the two treatment-associated subgroups and are shown in Table V.
Overall survival. For the whole CLL cohort, the median OS from the start of bendamustine-based therapy was 31.8 months. OS was influenced by several parameters and most importantly by the Rai stage (p=0.007), FISH karyotype risk-group (p<0.001), type of bendamustine therapy (p<0.001) and quality of response (p<0.001). In detail, the median OS was not reached for the cohort with 13q deletion only and the cohort with 17p deletion had the shortest median OS by far (8.3 months). As expected, the classification into patients with good risk (13q deletion, normal karyotype, trisomy 12) and poor risk (11q deletion and 17p deletion) cytogenetics proved to be of major prognostic importance (median OS not reached vs. 14.3 months, p<0.001). Patients who reached CR after a bendamustine-based therapy had the longest median OS (median not reached), when compared to the other categories. Those with PR and SD had similar outcomes (OS=35.8 and 31.8 months, respectively), whereas PD was associated with a significantly shorter median OS (9.2 months; p<0.001). In general, the BR cohort (n=76) was characterized by a superior median OS compared to the bendamustine-treated cohort (n=24; OS=42.7 months vs. 14.3 months; p<0.001; Figure 2a). The median OS of the subgroup of treatment-naïve patients (n=24; median OS not reached) was significantly superior to that of the pre-treated patients (n=76; OS=29.0 months; p=0.027). Furthermore, when analyzing solely the BR group, consisting of 76 patients, the treatment-naïve subgroup (n=21) had a significantly longer median OS compared to pre-treated patients (n=55) (not reached vs. 35.8 months; p=0.050; Figure 2b). The complete analysis is shown in Table IV and V.
Patient characteristics of all patients with chronic lymphocytic leukemia (CLL) divided into bendamustine-rituximab therapy cohort and bendamustine-monotherapy cohort.
Overall response rates (ORR) for all patients treated with bendamustine-monotherapy or bendamustine-rituximab therapy.
Overall response rates (ORR) for significant factors in the total cohort analysis, comparing bendamustine-rituximab therapy cohort to bendamustine-monotherapy cohort.
Multivariate analysis was performed for the whole cohort, and in the first model, only pre-treatment baseline characteristics that showed significance in the univariate analysis were included (n=81). Parameters independently influencing OS were lower Rai stage (p=0.030) and good-risk cytogenetics (p<0.001) (not shown in detail). The second model of multivariate analysis included pre-treatment baseline characteristics which were significant in the first multivariate analysis and treatment-related parameters that were significant in the univariate analysis (n=81) (shown in Table VI). The independent parameters associated with longer median OS were lower Rai stage (p=0.020); good-risk cytogenetics (p<0.001); administration of BR (p=0.009); and response to treatment (p=0.026).
Toxicity. All observed adverse events occurring from the beginning of bendamustine-based therapy until two months after the last administration were classified according to the CTC criteria version 4. Hematological and non-hematological adverse advents are detailed in Table VII. Briefly, grade 3/4 hematological adverse events were leukopenia in 26/91 patients (29%), neutropenia in 26/90 (29%), anemia in 14/94 (15%), and thrombocytopenia in 17/95 patients (18%). Febrile neutropenia was observed in 4/70 cases (6%). Among non-hematological toxicities, severe infections of grades 3/4 were reported in 21 out of 70 cases (30%), two of which were fatal. Nine patients died during bendamustine therapy or within 2 months after the last dose due to sepsis (n=2), rapid CLL progression (n=2), or unknown causes (n=5). Secondary malignancies, during or after therapy, occurred in five cases (peritoneal cancer of unknown primary; melanoma; gastric cancer; rectal cancer; myelodysplasia). Of note, interstitial pulmonary infiltrates considered potentially therapy-associated occurred only in patients receiving the BR combination treatment (n=4, 6%). We created sub-groups to check for adverse events (grade 3 or 4 leukocytopenia, neutropenia, anemia, thrombocytopenia, and infection) by bendamustine-based therapy in different settings, defined by age, gender, prior treatment (none vs. any), prior fludarabine, fludarabine refractoriness, CIRS score (>6 vs. ≤6), GFR (≥70 vs. <70 ml/min), bulky disease and cytogenetic risk category. No significant results in these sub-group analyses (not shown in detail) were found. Furthermore, no significant difference in toxicity between the BR and bendamustine-treated cohorts was found in these subgroup analyses.
Progression free survival (PFS) and overall survival (OS) for all patients treated with bendamustine-monotherapy or bendamustine-rituximab therapy.
Progression free survival (PFS) and overall survival (OS) for significant factors total cohort analysis, comparing bendamustine-rituximab therapy cohort to the bendamustine-monotherapy cohort.
Discussion
CLL is a lymphoproliferative disease occurring mostly in patients with advanced age and therefore the treatment decisions are essentially influenced by the comorbidities of the patients (4, 6, 23, 24). We aimed to study the use of bendamustine in a real-life CLL patient cohort in Austria and Italy. The patient cohort analyzed is well-balanced and reflects a real-world CLL cohort, with patients mostly in higher-therapy lines and with comorbidities (median CIRS score=9, median GFR=63 ml/min). In total, 76% of the patients were treated with bendamustine in combination with rituximab; 24% of the patients were treated with bendamustine monotherapy.
Another aim of this study was to evaluate patient response and survival, especially comparing patients who received BR and those who received bendamustine monotherapy. The ORR was in line with those of other studies, not only in the overall cohort, but also in the two sub-groups of patients which received BR or solely bendamustine. Of note, patients undergoing treatment with BR achieved significantly higher ORR (76%) and CR (22%) rates than those receiving single-agent bendamustine (ORR and CR of 50% and 13%, respectively). PFS and OS were also significantly lower in the single-agent bendamustine-treated group, with OS of 14.3 months and PFS of 8.3 months compared to the BR group with OS of 42.7 and PFS of 22.5 months (both p<0.001) (Figure 1 and 2). The results suggest that bendamustine should no longer be used as monotherapy in CLL, but that it has significant value as a chemotherapy backbone when combined with rituximab in older, less-fit patients with CLL, especially when considering that no significant differences in treatment-related toxicities were found. However, when comparing these sub-groups, we should bear in mind that the bendamustine monotherapy cohort included more high-risk patients in terms of cytogenetics, Rai stage and median number of previous therapies (shown in Table I).
Multivariate analysis of overall survival (OS) of the total cohort treated with bendamustine-monotherapy or bendamustine-rituximab therapy.
Toxicities of all patients treated with bendamustine-monotherapy or bendamustine-rituximab therapy.
No prospective study comparing bendamustine monotherapy and the BR combination therapy has yet been conducted in CLL and probably never will be. However, recognizing the epidemiological imbalances mentioned above as well as the limitations of a retrospective study, the data presented here are in line with the improved outcomes observed in prospective first- (25, 26) and second-line trials (27). Hence, the addition of rituximab to bendamustine may have a similar positive impact on patient response in CLL as seen in other immunochemotherapy studies including purine analogs with/without cyclophosphamide as chemotherapy backbone. Other retrospective studies showed similar responses, whereas the prospective studies have a noticeable better response (ORR=98%, with CR=38%) (2, 13). Moreover, a relatively small sub-group of 21 patients received BR in the first-line setting and the treatment combination proved to be very effective in this setting (Figure 1b and 2b).
Interestingly, a CIRS score over 6, GFR less than 60 ml/min, and patient age greater than 75 years were not unfavorable in ORR, PFS, and OS. Although chemotherapy with fludarabine, cyclophosphamide and rituximab (FCR) is currently considered to be the most active therapeutic regimen in CLL (26), it is associated with considerable toxicity and is often too toxic for older patient populations. Therefore, we await the final results of the CLL-10 study comparing BR and FCR in fit patients with CLL (28).
The applied cumulative dose of bendamustine (>750 mg/m2) seems to be an important prognostic factor. In our study, dose application was no problem and in general bendamustine was well-tolerated (median number of applied cycles in the BR group was 5). The most common grade 3 and 4 toxicities were hematological, with leukopenia (29%) and neutropenia (29%) being the most frequent. Anemia (15%) and thrombocytopenia (18%) were less common therapy-associated toxicities. Non-hematological toxicities were infections (30%) followed by skin toxicities (7%) and nausea (2%). The observed toxicities are common in bendamustine treatment and are in line with those already reported (2, 12, 29). Interestingly, toxicities did not occur more often in older patients, those with fludarabine-refractory disease, or those with a high CIRS score, or low GFR.
In conclusion, we describe an unselected population of patients with CLL (elderly patients, higher lines of therapy and mostly fludarabine pre-treated, with impaired renal function) in which bendamustine demonstrates a favorable efficacy and toxicity profile in routine clinical use.
Acknowledgements
This study was supported by the Verein für Tumorforschung, Innsbruck. We thank all centers participating and documenting patients for the study.
Footnotes
Conflicts of Interest
MN, AP, MM, PM, FK, CW, and IV have no conflict of interest to disclose. MF received a grant for this observational study; MF, MS and GG received honoraria from Mundipharma Austria GmbH for lectures. MM received honoraria from Mundipharma for participation in an advisory board.
- Received April 17, 2015.
- Revision received May 19, 2015.
- Accepted May 22, 2015.
- Copyright© 2015 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved