Daunorubicin Continuous Infusion Induces More Toxicity Than

Home , Daunorubicin

Leukemia (2001) 15, 898–902  2001 Nature Publishing Group All rights reserved 0887-6924/01 $15.00 www.nature.com/leu Daunorubicin continuous infusion induces more toxicity than bolus infusion in acute lymphoblastic leukemia induction regimen: a randomized study M Hunault-Berger1, N Milpied2, M Bernard3, J-P Jouet4, M Delain5, B Desablens6, A Sadoun7, F Guilhot7, P Casassus8 and N Ifrah1

Division of Hematology of the University Hospital of 1Angers, 2Nantes, 3Rennes, 4Lille, 5Tours, 6Amiens, 7Poitiers and 8Bobigny, France

We report the first randomized study assessing the efficacy Patients and methods and safety of daunorubicin (DNR) continuous infusion (CI) compared to the more conventional 30-min infusion (i.v.) in newly diagnosed adult acute lymphoblastic leukemia (ALL). Seventy- Treatment seven patients were initially randomized to receive either a 24- h CI DNR (60 mg/m2 days 2–4) (40 patients) or bolus DNR at Seventy-seven consecutive previously untreated patients with the same dosage (37 patients) with vincristine (2 mg i.v. days a non-Burkitt ALL, between the ages of 15 and 65 years, from 1, 8, 15) and oral prednisone (60 mg/m2 days 1–15), without nine hospitals, entered this study from May 1986 to May hematopoietic growth factor support, as an induction regimen. 1990. Patients with either a clinical history or electrical signs The distribution of adverse prognostic factors was comparable of coronary insufficiency or heart failure were not included. in the two-induction arm. Acute toxicity was more important 2 in the CI arm. Gram negative infection (9 vs 1 gram negative During the induction treatment, DNR (60 mg/m /day on days septicemia, P = 0.01) and infection-related deaths (6 vs 1 2 to 4) was delivered according to the initial randomization, deaths, P = NS) occurred more frequently in the CI arm during either by CI or by rapid 30-minute infusion (i.v.). DNR was the induction treatment than in the i.v. arm, leading to the study given with vincristine (2 mg administered intravenously on interruption. Neutropenia but not thrombopenia duration was days 1, 8, 15) and oral prednisone (60 mg/m2 on days 1 to significantly longer in the CI arm than in the i.v. arm (18 days vs 14 days, P . 0.05 and 16 days vs 12 days, P . 0.05, 15). Supportive care was performed according to each center respectively). Despite a similar CR rate according to the policy. Antibiotics were maintained up to neutropenia recov- method of DNR administration (68% in the CI DNR arm vs 76% ery. Patients did not receive hematopoietic growth factors. in the i.v. arm after the first course), there was a trend toward The complete treatment program is summarized in Figure 1. higher freedom from relapse (FFR) after DNR CI (48% vs 28% Bone marrow evaluation was carried out on day 21. At this = in the i.v. arm at 5 years, P NS), suggesting that despite this time, patients with significant bone marrow blasts cells high toxicity, DNR CI may improve the CR quality and decrease further the residual disease. Leukemia (2001) 15, 898–902. received a second induction cycle (named ADA containing 2 2 × Keywords: acute lymphoblastic leukemia; daunorubicin; continu- DNR 60 mg/m i.v. day 1, cytarabine 30 mg/m i.v. 2 days ous infusion; toxicity 3–7, asparaginase 1000 U/kg days 8–12) with DNR CI or i.v. according to the initial randomization. Patients who did not enter CR after the second induction cycle were considered as Introduction treatment failure and were taken off the study. Patients in CR, after one or two induction regimens, received a two-course Despite improvement in chemotherapy and supportive care consolidation therapy including an ADA course 10 days after over the past two decades, the prognosis of adults with acute CR, followed by a second course consisting of high-dose lymphocytic leukemia (ALL) is still poor. Complete remission methotrexate, l-asparaginase, prednisone and vincristine rates of 60% to 85% are reached with induction regimens (MAPO) (Figure 1). Late intensification by the BEAC regimen 2 2 including anthracycline. However, the toxicity of these induc- (BCNU 200 mg/m p.o. day 1, cytarabine 300 mg/m CI days 2 tion regimens is high, with 10% to 20% mortality rate during 1–4, cyclophosphamide 1.5 g/m i.v. days 2–3, etoposide 300 2 induction as a result of infections, and does not preclude a mg/m CI days 2–3) without stem cell support ended this high relapse rate, leading to 20–40% 5-year disease-free sur- intensive program. Patients under age 45 years, with an HLA- vival, depending in part on patient selection.1–3 Continuous identical donor were proposed for allogeneic bone marrow infusion (CI) of cancer chemotherapeutic agents may have transplantation after the first consolidation phase. In one insti- enhanced therapeutic effects because of an improved pharma- tution, patients under age 50 years received an autologous cological availability while the reduced peak plasma concen- bone marrow transplant (ABMT) after MAPO course. Six tration level seems to reduce some of their side-effects.4,5 weekly prophylactic intrathecal injections were followed, Studies with solid tumors have shown that anthracycline con- after completion of the chemotherapy, by cranial irradiation of tinuous intravenous infusion could reduce its dose-limiting the entire brain and the spinal cord above the second cervical cardiotoxicity.6 Few data concerning CI in ALL treatment are vertebra for a total midline dose of 18 Gy, performed over a available. In order to improve the remission rate and/or to 2-week period. decrease the toxicity of the ALL induction regimen, we compared in a randomized multicentric study, DNR CI during induction chemotherapy with the more conventional rapid Statistical methods infusion. The primary end-point of this study was the CR rate after the induction regimen while the secondary end-points were the induction toxicity, overall survival (OS) and freedom from Correspondence: N Ifrah, Me´decine D, CHU, 49033 Angers cedex relapse (FFR) duration. Differences among response rates and 01, France; Fax: 33 2 41 35 45 82 incidence of certain characteristics among subgroups were Received 23 October 2000; accepted 23 February 2001 analyzed by the ξ2 test for qualitative parameters, and the Stu- Safety and efficacy of daunorubicin continuous infusion in ALL M Hunault-Berger et al 899

Figure 1 Complete treatment program. dent’s t-test for quantitative parameters. Survival time was measured from the time of diagnosis to the death or last follow-up. FFR duration was measured from the time of remission until documented relapse. Global event-free survival (EFS) was measured from day 1 of therapy until failure Table 1 Patients’ characteristics to achieve CR, death or relapse. Survival and remission curves were generated with Kaplan–Meier product limit method and Daunorubicin CI Daunorubicin i.v. = = compared by log rank test. (n 40) (n 37)

Sex (M/F) 21/19 25/12 Median age (years) 36 36 Results Age ,35 17 18 Bulk diseasea 14 10 Patients’ characteristics Splenomegaly 15 12 Fever 12 8 Forty patients were initially randomized to CI DNR arm and Documented infections 7 3 37 to i.v. DNR arm. The distribution of clinical and laboratory LDH (n = 61) .2N 19/30 19/31 characteristics in the two induction groups (Table 1) was com- Median WBC (109/l) 18 18 parable and not significantly different, especially for adverse WBC .30 × 109/l 11 12 Hemoglobin (g/dl) 10 10 prognostic factors such as Philadelphia chromosome, leuko- 9 cytosis, non-T cell leukemia lineage. Platelets (10 /l) 45 44 Immunophenotype (n = 70) T-ALL 10 9 Disease response and toxicity according to Calla 17 18 Other pre-B ALL 5 5 randomization Tdt — 1 Myeloid antigen 2 1 Induction results are summarized in Table 2. Sixty-six patients Null ALL 2 — achieved CR, 11 of them after an ADA course leading to an Karyotype (n = 41) overall CR rate of 86% (66/77). Six patients died before their Normal 4 6 remission status could be ascertained while five failed to achi- Abnormal 11 11 eve CR. Remission rate was unaffected by the method of DNR Ph1 4 5 administration. CR was achieved after the first course in 27 patients allocated to CI DNR (68%) and 28 patients (76%) in aDefined by a mediastinal mass or an adenomegaly .5 cm.

Leukemia Safety and efﬁcacy of daunorubicin continuous infusion in ALL M Hunault-Berger et al 900 Table 2 Induction results

Daunorubicin CI Daunorubicin i.v. P (n = 40) (n = 37)

CR (ﬁrst induction) 27 (68%) 28 (76%) CR of patients alive at day 21 27/35 (77%) 28/36 (77%) CR after ﬁrst ADA 6 5 Overall CR 33 (82%) 33 (89%) Resistance after ADA1 2 3 Deaths 6 (15%) 1 (2%) 0.06 Time to reach CR (days) 25 (13–40) 21 (17–38) Median thrombopenia (,50 x 109/l) (days) 16 (0–31) 12 (0–26) Median red blood cells transfused (units) 8 (0–21) 8 (0–19) Median platelets transfused (apheresis) 5 (0–15) 5 (0–16) Median neutropenia (,0.5 x 109/l) (days) 18 (0–35) 14 (0–30) 0.05 Median duration of fever (days) 2 (0–27) 1 (0–13) Documented bacteriemia 19 10 NS Staphylococcus 6 7 Streptococcus 2 (mitis + B) 1 (D) Candida 1 1 Corynebacterium 1 0 Gram-negative 9 1 0.01 Pulmonary aspergillosis 1 0 Undocumented pneumopathia 1 1 Induction-related deaths 6 1 NS Gram – septicemia 4 (days 12, 13a, 15, 18) 0 NS Gram + septicemia 1 (day 22) 0 Candidemia 1 (day 26a) 1 (day 25) Aspergillosis 1 (day 40) 0

aThis patient developed a gram-negative septicemia and subsequently a candidemia.

the other group. In addition, the CR rates were also similar equally distributed in the two randomization arms. Two (77% in both arms) when only patients alive at day 21 were patients, one with hemorrhage pancreatitis and one with acute taken into account (27 out of 35 patients treated by CI as com- renal insufficiency were withdrawn from the study. At the time pared to 28 out of 36 patients in the i.v. arm). Neutropenia of analysis, 21 patients remained in CR, 36 relapsed, and nine (,0.5 × 109/l) duration but not thrombopenia (,50 × 109/l) died in CR. With a mean follow-up of 47 ± 6 months (0–158), duration was significantly longer in the CI arm than in the i.v. overall survival is 32% at 47 months with EFS and FFR of 27% arm (18 vs 14 days, P , 0.05 and 16 vs 12 days, P . 0.05, and 38% at 44 months, respectively (20 patients at risk). EFS respectively). This could account for the severity of infections and survival were higher after allogenic BMT (66% and 71% observed during the induction treatment: 19 patients in the CI at 44 months respectively, 11 patients at risk) than after late arm and 10 patients in the i.v. arm experienced documented intensification (18 and 26% respectively) (P = 0.0071 and bacteriemia, with an excess of gram-negative infections in the 0.002). The EFS and survival were 37.5% with 43% FFR for CI arm (three Escherichia coli, four Klebsiella pneumoniae, the eight patients who underwent autologous BMT in first CR. two Enterobacter in the CI arm vs one Moraxella infection in OS and EFS were not influenced by the method of DNR the i.v. arm) (P = 0.01). Infection-related deaths tended to be administration. However, although not significant, FFR tended more frequent in the CI arm (four gram-negative septicemia to be higher after DNR CI (48% in the CI arm vs 28% in the followed for one patient by a Candida tropicalis septicemia, i.v. arm at 5 years, P = NS) (Figure 2). one cocci gram-positive septicemia and one invasive aspergillosis) (15%) than in the i.v. arm (one from candidemia) (2%). No difference in terms of platelets or red blood cells Discussion transfusion were observed. No patients’ experienced grade 3 or 4 mucositis or requiried i.v. treatment. While cardiac tox- The increase effectiveness in terms of complete remission of icity was assessed by weekly EKG, only one patient in the CI ALL induction regimens containing DNR has been demon- arm experienced chest pain without any EKG modification. strated.7 However the best DNR administration modality Echocardiograms were not routinely available. However, 10 remains to be established. Pharmacokinetic data seem to favor years later, none of the living patients have clinical cardiac CI both in terms of efficacy and toxicity.8 DNR CI produced a dysfunction. more rapid cytoreduction than DNR bolus in ALL9 and, when performed over 4 h, reduced the peak plasma levels of DNR compared to bolus infusion without significant change in ALL Survival analysis cells concentrations or difference in acute toxicity.10 The absence of DNR peak serum concentration when adminis- Out of the 66 patients alive in CR, only 53 entered the late tered as 24-h CI5 is considered as a major reason why the intensification phase (18 received HLA genoidentical allo- risk of side-effects such as cardiomyopathy is relatively low in genic BMT, eight autologous BMT and 27 BEAC regimen) due multiple myeloma treated by the VAD regimen.11 to early relapse or toxic deaths. Late intensification type was Despite these theoretical advantages, few randomized data

Leukemia Safety and efficacy of daunorubicin continuous infusion in ALL M Hunault-Berger et al 901 toxicity (0/18 patients after DNR CI vs 4/18 patients after bolus DNR).9 The safety and efficacy of other anthracycline CI in ALL have been reported. Epirubicin, as a 4-day CI course followed by a vincristine 4-day CI course, induced 52% of response in 23 refractory or relapsed acute leukemia including four ALL with minimal extrahematologic toxicity, particularly cardiac and gastrointestinal toxicity.14 Idarubicin (24 mg/m2/day days 1–2 CI) induced 49% response in 51 children with prognostically poor recurrence of ALL but long intervals of myelosuppresion and high rates of systemic infection.15 CI or bolus mitoxantrone (10 mg/m2 days 1–3) were equally effective and well tolerated in 39 patients with newly diagnosed ALL.16 Low doses of adriamycin in the VAD program (vincristine 0.4 mg/day days 1–4 CI, adriamycin 12 mg/m2/day days 1–4 CI and dexamethasone 40 mg/day on day 1 to 4, 9 to 12 and 17 to 20 followed at day 24 by a second VAD cycle plus 1 g/m2 of cyclophosphamide on day 1 of the second course) induced a CR rate of 84% (88/105) with three deaths from infection.17 However, in a subsequent publication among 268 ALL, the CR rate was lower (58%) among elderly patients with an increased mortality rate during induction (12%).18 In our randomized study, continuous infusion of DNR (60 mg/m2/day x 3 days) resulted in a higher gram-negative infection rate (9 vs 1, P = 0.001), more toxicity-related early deaths (15% vs 2%), especially from gram negative septicemia (4 vs 0) and longer neutropenia (18 vs 14 days, P = 0.05). Such an high rate of gram-negative sepsis has also been reported in relapse ALL with lower doses of DNR CI (60 mg/m2/day x 2 days).19 Other toxicities such as cardiotoxicity or thrombocy- topenia were not increased. The excess infection is not related to the presence of a central venous catheter, as all patients had such venous access. The increased incidence of gram- negative bacteriemia suggests that DNR continuous infusion increases bacterial translocation from the gastrointestinal tract although an excess of mucositis or gastrointestinal toxicity was not reported. Because of this high morbidity in the CI arm observed at a planned interim analysis, without any benefit in terms of CR, the study was prematurely interrupted. Neverthe- less, after a long follow-up, we observed a trend toward an improved FFR in the CI arm (48% vs 28%) suggesting that CI DNR might improve the quality of the remission and decrease the residual disease. We strongly believed that this should be reported despite the low statistical power due to the small number of randomized patients. Indeed, although data are still conflicting, supportive care such as the addition of hematopoietic growth factor20 or quinolone prophylaxis21 might decrease the high morbidity and mortality related to gram- negative sepsis. Therefore, additional studies using route of administration designed to improve the tolerability of the anthracyclines and increase the antitumor activity such as liposomal DNR22 should be done. Figure 2 (a) Freedom from relapse, (b) overall survival and (c) The results of our randomized study are disappointing as event free survival according to the daunorubicin method of adminis- DNR CI at this dosage not only did not improve the response tration. rate but also increased chemotherapy-related toxicity. How- ever, the trend toward a better FFR in the CI arm would sug- on acute leukemia are available. In AML, DNR CI, used at gest that the quality of the CR in term of residual disease might 45 or 30 mg/m2/day days 1–3, seems to be efficient without be improved. increased toxicity, in association with either cytosine arabino- side CI (100 mg/m2/day days 1–10) and 6-thioguanine (100 2 mg/m every 12 h orally for 8 days) or carboplatin (200 References mg/m2/day days 3–7) respectively in non-randomized stud- 12,13 ies. In ALL, among 44 children, reduction in leukemic cell 1 Hoelzer DF. Therapy of the newly diagnosed adult with acute lym- population is more rapid after 2 days DNR CI infusion (60 phoblastic leukemia. Hematol Oncol Clin North Am 1993; 7: mg/m2/day) than after bolus and has a lower acute cardiac 139–160.

Leukemia Safety and efficacy of daunorubicin continuous infusion in ALL M Hunault-Berger et al 902 2 Copelan EA, McGuire EA. The biology and treatment of acute lym- tinuous intravenous infusion is effective in the treatment of acute phoblastic leukemia in adults. Blood 1995; 85: 1151–1168. nonlymphocytic leukemia. Br J Haematol 1985; 61: 261–265. 3 Pui CH, Evans WE. Acute lymphoblastic leukemia. N Engl J Med 13 Archimbaud E, Anglaret B, Thomas X, Devaux Y, Ffrench M, Seb- 1998; 339: 605–615. ban C, Troncy J, Fiere D. Continuous-infusion daunorubicin and 4 Carlson R, Sikic B. Continuous infusion or bolus injection in can- carboplatin for high-risk acute myeloid leukemia in the elderly. cer chemotherapy. Ann Int Med 1983; 99: 823–833. Leukemia 1992; 6: 776–779. 5 Speth P, Linssen P, De Pauw B, De Witte T. Cellular pharmacoki- 14 Liso V, Specchia G, Pavone V, Capalbo S, Dione R. Continuous netics of daunomycin administered as continuous intravenous infusion chemotherapy with epirubicin and vincristine in relapsed infusion in the treatment of acute non-lymphocytic leukaemia. Br and refractory acute leukemia. Acta Haematol 1990; 83: 116–119. J Haematol 1986; 63: 603–605. 15 Neuendank A, Hartmann R, Buhrer C, Winterhalter B, Klumper E, 6 Garnick M, Weiss G, Steele G, Isreael M, Schade D, Sack M, Frei Veerman AJ, Henze G. Acute toxicity and effectiveness of idarub- E. Clinical evaluation of long-term, continuous infusion doxorubicin in childhood acute lymphoblastic leukemia. Eur J Haematol icin. Cancer Treat Rep 1983; 67: 133–142. 1997; 58: 326–332. 7 Gottlieb AJ, Weinberg V, Ellison RR, Henderson ES, Terebelo H, 16 Koc Y, Akpek G, Kansu E, Kars A, Tekuzman G, Baltali E, Guler N, Rafla S, Cuttner J, Silver RT, Carey RW, Levy RN, Hutchinson J, Barista I, Gullu I, Ozisik Y, Firat D. Bolus and continuous infusion Raich P, Cooper M, Wiernik P, Anderson J, Holland J. Efficacy of mitoxantrone in newly diagnosed adult acute lymphoblastic leu- daunorubicin in the therapy of adult acute lymphocytic leukemia: kemia: results of two consecutive phase II clinical studies. Cancer a prospective randomized trial by cancer and leukemia group B. Invest 1998; 16: 355–365. Blood 1984; 64: 267–274. 17 Kantarjian HM, Walters RS, Keating MJ, Smith TL, O’Brien S, Estey 8 Bassan R, Chiodini B, Zucchetti M, Lerede T, Cornelli PE, Cortel- EH, Huh YO, Spinolo J, Dicke K, Barlogie B, McCredie K, Freire- ich E. Results of the vincristine, doxorubicin, and dexamethasone azzo S, Barbui T. Short course infusional idarubicin plus intermit- regimen in adults with standard- and high-risk acute lymphocytic tent cytarabine and etoposide for refractory hematologic malig- leukemia. J Clin Oncol 1990; 8: 994–1004. nancies: clinical and preliminary pharmacological results. 18 Kantarjian HM, O’Brien S, Smith T, Estey EH, Beran M, Preti A, Haematologica 1998; 83: 27–33. Pierce S, Keating MJ. Acute lymphocytic leukaemia in the elderly: 9 Steinherz PG, Redner A, Steinherz L, Meyers P, Tan C, Heller G. characteristics and outcome with the vincristine-adriamycin- Development of a new intensive therapy for acute lymphoblastic dexamethasone (VAD) regimen. Br J Haematol 1994; 88: 94–100. leukemia in children at increased risk of early relapse. The Mem- 19 Morland BJ, Shaw PJ. Induction toxicity of a modified Memorial orial Sloan- Kettering-New York-II protocol. Cancer 1993; 72: Sloan-Kettering-New York II protocol in children with relapsed 3120–3130. acute lymphoblastic leukemia: a single institution study. Med Ped- 10 DeGregorio MW, Carrera CJ, Klock JC, Pegelow CH, Wilbur JR. iatr Oncol 1996; 27: 139–144. Cellular and plasma kinetics of daunorubicin given by two 20 Ottmann OG, Hoelzer D. Growth factors in the treatment of acute methods of administration in a patient with acute leukemia. Can- lymphoblastic leukemia. Leuk Res 1998; 22: 1171–1178. cer Treat Rep 1982; 66: 2085–2088. 21 Engels EA, Lau J, Barza M. Efficacy of quinolone prophylaxis in 11 Koskela K, Pelliniemi TT, Remes K. VAD regimen in the treatment neutropenic cancer patients: a meta-analysis. J Clin Oncol 1998; of resistant multiple myeloma: slow or fast infusion? Leuk Lym- 16: 1179–1187. phoma 1993; 10: 347–351. 22 Hortobagyi GN. Anthracyclines in the treatment of cancer. An 12 Lewis J, Meyers F, Tanaka L. Daunomycin administered by con- overview. Drugs 1997; 54: 1–7.

Leukemia